Caplin Point Laboratories makes generic medicines and sells them across 23 countries, with the bulk of its revenue coming from Latin America. The company was founded in Chennai in 1990 as a contract manufacturer, meaning it made medicines for other companies under their labels. Over the following three decades, it evolved into a business that owns its own brands, controls its own distribution infrastructure in foreign markets, and manufactures complex pharmaceutical products for both emerging and regulated markets.

Latin America and Francophone Africa

This is where Caplin built its franchise. Latin America accounts for 76% of revenue, Africa for 3%.

Caplin has established legal subsidiaries in individual countries across Central and South America, including Guatemala, Nicaragua, Honduras, El Salvador, Ecuador, Colombia, Chile, and Mexico. It holds inventory in-country, registers its own products with local drug regulators, and supplies pharmacies and hospitals through its own distribution relationships. The company has over 22,000 distribution touchpoints across the region.

Because Caplin controls the supply chain down to the pharmacy level, its customers in these markets pay quickly, often in advance. This produces a structurally negative working capital cycle, which is rare for a manufacturing business of this kind.

The product mix in these markets is 75% generic and 25% branded generic. A branded generic is a generic drug sold under a company’s own brand name, which allows for a price premium over unbranded alternatives. The branded generic share has grown from just 5% in FY2012, and that shift is what drives the gross margin profile. Across 36 therapeutic segments, the portfolio covers over 65% of the WHO essential drug list.

United States (through Caplin Steriles)

The US business contributes 21% of revenue and operates through Caplin Steriles Limited, a subsidiary with a USFDA-approved injectable manufacturing facility at Gummidipoondi, Chennai.

Injectables are more complex to manufacture than oral tablets. They must be sterile, and the facility must pass regular USFDA inspections to remain eligible to supply the US market. Caplin Steriles has cleared four consecutive inspections with clean outcomes.

The US business currently runs on two models: supplying other American pharmaceutical companies under their labels (B2B, 75% of CSL revenue) and selling under Caplin’s own label through Caplin Steriles USA Inc, its American front-end entity (B2C, 25%). As of FY26, Caplin holds 59 ANDA approvals, each representing regulatory clearance to sell a specific generic drug in the US. The front-end entity achieved profitability in its first year of operations.

Theme

Latin America: The Market Nobody Wanted

Let’s start with geography, because that is where the story begins.

Latin America has a combined population of roughly 660 million people. Per capita incomes across most of the region sit well below OECD levels. The average patient in Guatemala, Honduras, or Ecuador cannot pay Rs. 1,500 for a branded drug. But they still get diabetes, hypertension, and bacterial infections that need antibiotics and parasitic infections that need dewormers. Generic pharmaceuticals in these markets are the only realistic option.

The Latin American pharma market is approximately $75 billion today and is expected to reach $102 billion by 2030. Generics account for 62% of that market. The generics sub-segment grows faster than the overall market because governments across the region have spent the past two decades passing legislation that actively promotes generic adoption to reduce public healthcare spending. Brazil passed its landmark generic medicines law in 1999, Mexico has progressively mandated that pharmacists offer a generic substitute for any branded prescription, Colombia has similar substitution policies.

Also, Latin America is urbanizing rapidly. As populations shift from rural subsistence into urban wage employment, healthcare utilization through formal pharmacy channels rises. This is the patient who walks into a pharmacy, asks for a medicine by name, and buys it every month for years. Chronic disease medications for hypertension, diabetes, cardiovascular conditions, and respiratory illness are repeat-purchase, high-frequency products. This is the best possible kind of demand for a pharmaceutical company.

What makes Caplin’s position unusual is that it has been building this market presence since 1994. When Caplin entered Angola and then Central America, no significant Indian pharmaceutical company was paying attention to the region. The regulatory complexity was seen as too difficult, markets were seen as too small, not to mention the currency and political risk. Caplin went anyway, registered its products country by country, built relationships with distributors and pharmacies, put its own brand names on medicines, and compounded those relationships over three decades.

The result today is a business with product registrations across 23 countries, over 5,000 product licences across 36 therapeutic areas, distribution reach covering tens of thousands of pharmacy touch points, and brand recognition that took twenty years to build. A competitor entering LatAm today would face the same queue Caplin faced in 1995, except Caplin has already crossed the finish line and the competitor is still at the starting blocks.

The financial proof of this market position is Caplin’s working capital structure. In most generic pharma supply relationships, the supplier ships goods, extends credit, waits 60 to 90 days, and hopes collections come in. Caplin’s LatAm distributors pay in advance or on very short credit terms. This happens because Caplin controls the shelf. If you are a distributor in El Salvador and Caplin is your most reliable supplier of antibiotics, antifungals, and antihypertensives across twenty therapeutic categories, you do not make them wait for payment. The consequence is that Caplin’s LatAm business operates with a structurally negative working capital cycle. This is a rare thing in manufacturing and an almost unheard of thing in emerging market pharma exports.

The United States: The Right Door into a Complex Market

Most Indian pharmaceutical companies that tried to build a meaningful US business went through the oral generics market first (tablets, capsules, oral liquids). It seemed like the natural starting point because the manufacturing was simpler and the regulatory pathway was well understood. What happened was predictable in hindsight. When a branded drug patent expires and a generic tablet is approved, the first few manufacturers earn strong margins. Then more manufacturers get approved, prices collapse 80 to 90% within two to three years, and the entire segment becomes a commodity with thin margins and brutal price competition. Dozens of Indian companies have spent the past decade writing down goodwill from US generic acquisitions and watching their US margins compress quarter after quarter.

Caplin watched this happen and went through a different door entirely: Sterile injectable manufacturing.

Injectable drugs delivered intravenously or intramuscularly are technically far harder to produce than oral solids. The manufacturing facility must meet extremely stringent contamination control standards. Cleanrooms classified to ISO specifications, aseptic fill-finish lines, and overall a rigorous monitoring environment. The capital investment required runs into hundreds of crores and the technical expertise is genuinely scarce. This means that when a drug requiring an injectable form goes off-patent, far fewer manufacturers can produce a generic version compared to a tablet. Competition is thinner. Prices hold up better. Margins are structurally higher.

The US generic sterile injectable market is approximately $18 billion today, growing toward $43 billion by 2035 at a 9.2% annual growth rate. This growth is driven by a large and ongoing wave of branded drug patent expiries, many of which involve injectable formulations. Every patent expiry creates a new ANDA opportunity. Caplin Steriles has 59 approved ANDAs today, with 55 more in the development pipeline for filing over the next four years.

Within the injectable category, Caplin has made a specific product selection choice that is worth understanding. The company has deliberately prioritized pre-filled syringes, ophthalmic products, and Blow-Fill-Seal technology for its pipeline. These are the most technically demanding forms in the injectable category, and they have the fewest qualified manufacturers competing for market share.

There is also a structural demand dynamic that helps new qualified entrants. The US FDA maintains a public drug shortage list at any given time containing roughly 100 to 200 drugs, the majority of which are generic sterile injectables. Shortages occur when a manufacturer has a quality event and must suspend production, or when demand spikes faster than capacity can respond. The FDA has created expedited review pathways for ANDAs addressing drugs in shortage. For Caplin, a USFDA-compliant injectable manufacturer, these shortages represent a recurring, self-replenishing set of opportunities to enter segments where the competitive field has temporarily narrowed.

GLP-1 Generics Optionality

GLP-1 receptor agonists are a class of drugs that have gone from a diabetes treatment to a cultural phenomenon in the span of five years. Drugs like Ozempic and Wegovy have demonstrated dramatic weight loss results in clinical trials and have become among the best-selling pharmaceuticals in history. The US GLP-1 market reached $71 billion in 2024. The global market is projected to exceed $100 billion by 2030.

The branded versions cost upwards of $1,000 per month in the United States. In Latin America, they are essentially inaccessible for the vast majority of patients.

This is where the Caplin angle becomes interesting. Liraglutide, one of the first major GLP-1 drugs, had its primary US patents expire in late 2023, and generic versions entered the US market in late 2024. The timeline for semaglutide generics is longer and more complex due to Novo Nordisk’s patent portfolio, but the direction of travel is clear.

The reason this matters specifically for Caplin comes down to geography and manufacturing capability.

Latin America has some of the highest rates of diabetes and obesity in the world, and it’s disproportionate relative to their income levels. The patient population that would benefit most from GLP-1 drugs is concentrated precisely in the markets where Caplin has its deepest distribution infrastructure. A generic liraglutide or semaglutide at 1/10th the branded price, supplied by a company that already sits on pharmacy shelves across Guatemala, Honduras, El Salvador, Ecuador, and Colombia, is a fundamentally different commercial proposition than the same generic drug trying to build a distribution network from scratch.

On the manufacturing side, GLP-1 drugs are subcutaneous injectables delivered by pre-filled pen devices. Caplin’s injectable manufacturing capability is directly applicable. The company has stated in its Q4 FY26 earnings presentation that it awaits regulatory approvals from Central American countries for its own internally developed GLP-1 products, with plans to expand to other South American markets post-patent expiry in FY28.

Oncology: Institutional Channel Advantage

Cancer is among the fastest-growing disease burdens in Latin America. The World Health Organization projects that Latin America will see a 65% increase in cancer incidence by 2040, driven by aging populations, urbanization, and changing lifestyle patterns. Governments across the region are expanding public healthcare coverage for cancer treatment as political pressure from patient advocacy groups intensifies.

Oncology drugs are overwhelmingly dispensed through institutional channels, meaning hospitals, oncology clinics, and government health programmes, rather than retail pharmacies. This is exactly the channel that Caplin has been building in Latin America through its institutional sales program, which already accounts for 20% of its emerging markets revenue.

A company that already has established relationships with hospital procurement departments and government health ministries across 23 countries has a significant advantage when it moves into oncology.

Caplin’s oncology facility at SIDCO Kakkalur, near Chennai, has completed its first regulatory inspection successfully. The company has acquired several oncology ANDAs from third parties. A separate Oncology API facility at Thervoy SIPCOT is under construction, with first Drug Master File filings targeted for FY28. The backward integration into oncology APIs means Caplin would eventually control the full stack from the active ingredient through to the finished drug, which is the same cost structure advantage it has been building in general APIs for its existing portfolio.

Oncology is a long-duration build. The therapy area requires careful regulatory navigation, specialised manufacturing protocols, and deep clinical relationships.

Africa: The LatAm Story, 30 Years Earlier

Africa contributes approximately 3% of Caplin’s revenue today. That number invites dismissal, but let’s review a previous pattern.

In 1994, Caplin entered Angola with a Stock and Sale model when Latin America was widely considered too difficult, too fragmented, and too risky to justify serious investment. Thirty years later, Latin America is 76% of the company’s revenue and the foundation of its compounding machine. Africa today sits in a similar position to where LatAm was in 1994.

The African pharmaceutical market is approximately $27 billion, growing toward $37 billion by 2033. Roughly 60% of medicines consumed across the continent are imported, primarily from India and China. Local manufacturing capacity is grossly insufficient relative to the population and regulatory frameworks are fragmented on a country-by-country basis, which serves as a barrier to entry for any company unwilling to do the country-level regulatory work systematically over many years. Distribution infrastructure is underdeveloped and healthcare spending per capita is rising but from a very low base.

Every one of those characteristics describes Latin America in the mid-1990s. Caplin navigated that environment and built a near-monopoly position. The playbook is proven, and the question is whether management has the patience and institutional capability to replicate it in a new geography over a similar timeframe.

The African Continental Free Trade Agreement adds a potential structural catalyst. Signed by 54 of 55 African Union member states, it is progressively working to reduce intra-African trade barriers. For a pharmaceutical company that has invested in country-level regulatory approvals across multiple African markets, a future where a product registered in one country can be sold across a broader trade zone would multiply the value of those registrations significantly.

The Structural Shift in Generic Pharma: Complexity Wins

There is a broader industry transition underway that provides the backdrop for everything Caplin is building in the US.

The era of straightforward generic pharmaceuticals, simple tablets and capsules competing primarily on price, is largely over for Indian companies in developed markets. The returns from oral solid generics in the US have been competed away. Companies that built large businesses around generic tablet manufacturing in the 2000s and 2010s have spent the past several years watching margins compress and writing down acquisitions made at the peak of that cycle.

The next decade of opportunity in generic pharmaceuticals sits in technical complexity. These products require significant technical investment, take longer to develop and manufacture, and attract fewer qualified competitors. Caplin entered the US through injectables from the beginning, precisely because the oral generics path was already showing signs of deteriorating economics.

The China plus one dynamic accelerates this transition. COVID-19 exposed how concentrated pharmaceutical supply chains had become around Chinese API manufacturers. Approximately 70% of active pharmaceutical ingredients consumed globally come from China, so when Chinese API production was disrupted, the downstream effects on drug availability in hospitals across the United States were immediate and severe. Post-COVID, both US government procurement policy and private hospital procurement behaviour have shifted toward qualifying alternative suppliers, specifically non-Chinese sources, for critical pharmaceutical ingredients.

Caplin is building its own API manufacturing capability as a direct response to this environment. The Vizag API facility has received its manufacturing licence and completed validations for several APIs already. The company has completed research and development for over 90 APIs intended for backward integration into both its US injectable pipeline and its existing LatAm portfolio. As this backward integration completes, Caplin’s supply chain becomes simultaneously less expensive, more resilient, and better positioned to meet the sourcing standards that US hospital procurement is increasingly demanding.

The China 2.0 strategy adds another dimension to this. Rather than developing all new products from scratch, Caplin has established partnerships with Chinese pharmaceutical companies that already hold approved ANDAs or marketing authorizations in the US and EU. Caplin then files those same products in its LatAm markets, accessing a broader product range faster than its own R&D pipeline could supply, and paying for proven product development rather than uncertain early-stage research. The first biosimilar product filing in Central America came through this program.

Value

Caplin’s EBITDA margin in FY26 was 38%. The Indian pharma sector average sits between 18% and 22%. Branded domestic pharma companies, which sell under their own names to Indian doctors and pharmacies, typically run 22 to 26%. Caplin, a generic pharmaceutical exporter selling primarily into Latin America and the US, runs 38%.

Why the Margins Are This High

Most consumer-facing pharmaceutical companies spend 25 to 35% of revenue on selling, marketing, and distribution. Caplin’s LatAm business is institutional and wholesale-led. It sells to established distributors who already reach the pharmacies. There is no advertising spend, so the operating cost structure is fundamentally leaner than a consumer pharma model.

Vertical integration: Caplin manufactures approximately 60% of its products in-house and has progressively built its own API manufacturing capability. Gross margins in FY26 were 60.4%, expanding from approximately 50% in FY17.

The Negative Working Capital Model

Working capital is the money a business needs to bridge the gap between paying its suppliers and receiving payment from its customers. A company with 90-day receivables needs to finance three months of its own revenue. This costs money, either through bank borrowings or by tying up internal cash.

Caplin’s LatAm business runs with a structurally negative working capital cycle.

Estimated receivable days in its core markets are approximately 8 to 12 days. Distributors in Guatemala, Honduras, Ecuador, and Colombia pay Caplin within days of receiving goods, because Caplin is often the most reliable supplier of quality generic medicines across twenty therapeutic categories in their market.

This means Caplin’s LatAm operations are funded by their own customers. Cash comes in before it goes out. The business does not need bank facilities to manage its working capital. This is the financial proof of market power, and it happens automatically when a business controls the relationship.

ROCE Through a Capex Cycle

Return on Capital Employed has stayed above 25% for five consecutive years. The Indian pharma sector average is 12 to 18%.

The more revealing number is that this ROCE was maintained through the injectable facility construction cycle of FY20 to FY22, when Caplin was deploying approximately Rs. 650 crore into a new USFDA-compliant manufacturing plant at Gummidipoondi. Most companies see ROCE compress sharply during heavy capex because capital is being deployed before it earns revenue. Caplin’s ROCE floor during this period was approximately 25%. That tells you the underlying LatAm business was generating returns so far above 25% that the idle capital drag of a half-built facility could not pull the blended rate below the threshold.

The injectable facility is now substantially complete. As US injectable revenue scales through this already-depreciated asset base, each new rupee of US revenue flows at very high incremental ROCE. This is the margin expansion catalyst that has not yet appeared in the numbers.

The Balance Sheet

Finance costs in FY26 were Rs. 0.87 crore. On Rs. 650 crore of PAT, that is 0.13% of annual profit going to interest. Caplin has never borrowed meaningfully in 35 years of operations. The injectable facility, the US front-end, the oncology facility, and the API plants are all being built from what the business generates.

The zero-debt balance sheet eliminates constraints on management decision-making entirely. When Caplin acquired 10 ANDAs in January 2026 targeting a $473 million US market, it was funded quietly from cash with no equity dilution and no press release about raising capital.

Operating Leverage: The Compounding Effect

EBITDA margin was approximately 21% in FY14. It is 38.1% today. Revenue has grown roughly 12 times over that period.

There is more runway. The US injectable facility is a fixed cost base currently earning modest revenue. As the ANDA portfolio matures and US front-end revenue scales, each new rupee of US revenue uses infrastructure whose costs are already absorbed. The operating leverage from the US business has not yet appeared in the consolidated margins. When it does, the progression from 38% to something above it will be a function of scale rather than any change in the underlying business.

Growth

Latin America

The bear case on LatAm has always been saturation, but the evidence says otherwise. Caplin products occupy 41% of shelf space across independent pharmacies in Central America. The company’s stated target is 57%, achievable by adding a liquid manufacturing facility to register categories currently missing from the portfolio.

A business with 41% shelf space in its core markets, with a clear product-led path to 57%, is nowhere near saturation. It is mid-penetration in markets it has operated in for over twenty years.

Mexico and Chile are the newer engines. Mexico has 25 product registrations so far, with over 120 more to be filed in the next 18 months from the combined pipelines of Caplin Point, Caplin One Labs, and Caplin Steriles. The company has won 11 general and oncology products for tender supply in Mexico worth approximately $4 million over the next 24 months. More importantly, Caplin has purchased land to build its first local manufacturing facility in Mexico, which unlocks a 16% price advantage in government tenders once local production begins. In Chile, 135 product registrations are in place, $10 million of tender wins have been secured for the next 24 months, and management is actively evaluating two distribution company acquisitions to reach the private pharmacy market more directly.

Brand marketing is the next layer on top of distribution depth. Caplin is entering CNS therapy as a branded marketing play in Dominican Republic, Guatemala, and Nicaragua, with bioequivalence already completed for over 12 products and commercial launches planned for September 2026. This is the shift from selling generics through a wholesale channel to building prescription demand at the doctor and pharmacist level. The margin implication of that shift, as the branded generic share moves from 25% toward 35% of LatAm revenue, is significant.

United States

Caplin Steriles Limited generated Rs. 470 crore in FY26 revenue at a 30% EBITDA margin for the full year, expanding to 33% in Q4 alone. Caplin Steriles USA, the front-end commercial entity, completed its first full year of operations with $11 million in revenue and a 26.2% EBITDA margin.

Caplin went into the US through pre-filled syringes, ophthalmics, and IV bags rather than the commodity vial segments where margins erode quickly. The product quality is also visible in per-ANDA revenue: Vivek Partheeban noted on the concall that Caplin’s average revenue per ANDA is more than 50% above the industry average of $800,000 to $1 million.

There is a more important point buried in the breakdown. Of the US growth Caplin delivered in FY26, over 90% came from products that were already launched before the year began. The 10 ANDAs approved during FY26 and the 15 ANDAs acquired in January 2026 have mostly not yet been launched commercially. The largest ones are in the process of being launched over the coming weeks. Management guided 25% to 30% growth for the overall Caplin Steriles business in FY27, and separately set a target of doubling the Caplin Steriles USA own-label revenue from Rs. 100 crore to Rs. 200 crore. Both targets will be driven substantially by products already approved but not yet generating revenue.

The capacity picture makes the next three years even clearer. Today, Caplin Steriles operates with five injectable lines (three older lines and two new ones). The Phase III expansion will bring the total to 17 injectable lines within two to three years. The COL-II facility at Gummidipoondi is completing by December 2026 and will house five additional lines for injectables, ophthalmics, and Blow-Fill-Seal products. Caplin is simultaneously moving high-volume, lower-margin injectable products to compliant contract manufacturers, freeing existing in-house lines for higher-margin complex products.

The non-US regulated market piece is often overlooked entirely. Caplin Steriles has filed 54 products across Canada, EU, Australia, Mexico, Brazil, South Africa, Saudi Arabia, and UAE, of which 32 are approved. Management expects meaningful revenue from these markets in FY27. This is a third revenue stream within the US business.

What Hasn’t Started Yet

The oncology facility completed its first regulatory inspection successfully and has been capitalized post April 2026. The Thervoy oncology API facility is completing by Q3 FY27 with first Drug Master File filings planned for FY28. Oncology products will flow through distribution relationships with hospital channels that Caplin has spent twenty years building in LatAm.

The GLP-1 pipeline is the one that nobody is talking about. Caplin awaits regulatory approvals from Central American countries for internally developed GLP-1 products, with plans to expand to other South American markets post patent expiry in FY28. The combination of an injectable manufacturing capability, an existing distribution network reaching diabetic and obese patient populations across Latin America, and a generic GLP-1 at 1/10th the branded price creates an opportunity that has no precedent in Caplin’s history.

The China 2.0 strategy continues to build. Multiple partnerships with Chinese companies holding approved ANDAs or marketing authorizations in the US and EU have generated filings across Mexico, Colombia, and Chile, plus the first biosimilar product filing in Central America. This expands the product range at a fraction of the cost of in-house development and brings biologics into a company that has historically been a small-molecule generics manufacturer.

Valuation

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

Credit to Sathya Kalyanasundaram for recommending me to study this!

Explaining Sakar

Founded in 2004, Sakar Healthcare Ltd started as a contract manufacturer of oral liquid formulations. Two decades later, it has evolved into a vertically integrated, research-driven pharmaceutical company with a growing oncology franchise, EU-GMP certified manufacturing, and a presence in over 50 countries.

For most of its early life, Sakar stayed under the radar manufacturing quality generics at competitive prices and exporting them to emerging markets. But over the last few years, management has been quietly repositioning the business toward one of the highest-value, highest-barrier segments in the entire industry: oncology.

The business has 3 main segments:

1. Domestic CDMO

Sakar manufactures drugs on a contract basis for Indian pharma companies — Zydus, Cipla, Intas, Torrent, Glenmark, Abbott, Emcure, and others. This business is steady, recurring, and provides baseline capacity utilisation and cash flow. They do CDMO for both oncology and non-oncology.

They’ve got very strong and long-dated relationships with some of these companies:

2. Technology Transfer Projects

A company like Accord Healthcare (the European generic arm of Intas) wants to manufacture an oncology product at a site in India for supply into EU and UK markets. They approach Sakar, transfer the technology to Sakar's facility, and then apply to the European Medicines Agency for a site variation approval. They’re essentially telling the regulator that this product, previously manufactured elsewhere, will now also be made at Sakar's plant. Once EMA approves the variation, Sakar can supply the product to Accord for commercial distribution in Europe. Sakar has received 4 such site variation approvals as of the Q4 FY26 call (two for Accord, two for Torrent UK and Germany). The Accord portfolio alone is 9 oncology molecules.

3. Own Brand Exports

Here, Sakar develops a product entirely in-house, prepares an EU CTD (Common Technical Document, the standard regulatory submission format for Europe), completes bioequivalence studies using EMA or USFDA-approved contract research organizations, and either licenses the dossier to an international partner who registers it locally, or registers it directly and supplies under its own brand. This is the most margin-accretive route because Sakar captures more of the value chain. As of the Q4 FY26 call, 130+ dossiers have been submitted globally, 12 have received marketing authorizations, and oncology export revenue from this route was still essentially zero.

Facilities

Sakar operates from two main sites, both in Gujarat. The older Changodar facility near Ahmedabad handles the non-oncology business (cephalosporins, general injectables, oral liquids) and continues to generate the bulk of stable CDMO revenue. The newer Bavla facility is where the growth thesis sits. Bavla is EU-GMP approved and purpose-built for oncology, with integrated API synthesis and finished formulation capabilities under one roof. Critically, it has OEL Level 4 containment infrastructure.

OEL Level 4 is worth understanding. OEL stands for Occupational Exposure Limit, and Level 4 represents the most stringent containment tier in pharmaceutical manufacturing (the kind required for highly potent, highly toxic cytotoxic molecules like docetaxel or paclitaxel). Building a facility to this standard requires isolated negative-pressure processing suites, containment granulation and compression lines, specialised HVAC, and rigorous waste disposal systems. This is expensive to build and complex to maintain, which is precisely why it creates a meaningful barrier. Most Indian pharma manufacturers do not have this capability.

Bavla also houses flow chemistry infrastructure from Vapourtec in the UK, which enables continuous API synthesis as opposed to conventional batch processes. This is more efficient, produces lower volumes of solvent waste, and is better suited to complex reaction sequences. De-Dietrich glass reactors support additional API synthesis.

I’ve analyzed Sakar using the TVG framework:

Theme

Cancer is a growing epidemic, and this won’t change

The most fundamental tailwind is epidemiological. Cancer is a disease of DNA damage accumulating over time, driven primarily by aging, and secondarily by lifestyle. The global cancer burden is growing because the world’s population is aging.

In 2022, the International Agency for Research on Cancer estimated approximately 20 million new cancer cases globally. By 2040, that number is projected to reach 30 million annually. That is a 50% increase in new patients in less than two decades. The primary cancers driving this growth are lung, colorectal, breast, and prostate cancer. All age-related, all treated with chemotherapy and targeted therapy drugs that are either already generic or approaching genericization.

In India specifically, the ICMR data points to approximately 14.6 lakh new cancer cases in 2022, rising to an estimated 15.7 lakh by 2025 and potentially 20 lakh by the early 2030s. The growth here is driven by two forces simultaneously: epidemiological increase (urbanization, dietary shifts toward processed food, tobacco use, air pollution) and detection improvement (more diagnostic centers, more oncologists, better awareness). As diagnosis rates improve, the apparent caseload will continue to rise even if underlying biology did not change.

Every new cancer patient requires treatment. Every patient on chemotherapy is consuming formulated oncology drugs. Most of those drugs, particularly the cytotoxics that Sakar manufactures, have no therapeutic substitute. The price elasticity of oncology drug demand is essentially zero.

This is the demand floor under the entire thesis. The baseline demand for oncology formulations is growing at a high single-digit percentage annually and will continue to do so for structural demographic reasons.

The Oncology Patent Cliff

Here is the basic mechanic. A branded drug under patent is exclusively sold by the originator company at a high price. When the patent expires, generic manufacturers are legally permitted to make and sell the same molecule. Generics enter at prices 70-90% below the brand. The branded revenue collapses, but total volume consumed often expands because affordability increases. The generic market that forms is large and growing.

The oncology category is currently in the early stages of the most significant patent cliff in its history.

Lenalidomide (brand name Revlimid, marketed by Bristol-Myers Squibb) had global sales of approximately USD 12-13 billion at its peak. It is a treatment for multiple myeloma and certain lymphomas. Patent expiry commenced in key markets around 2022-2023. The generic lenalidomide market that is forming globally is one of the largest single-product generic opportunities in pharmaceutical history. Sakar has lenalidomide capsules across six different strengths (2.5mg, 5mg, 10mg, 15mg, 20mg, 25mg) in its oncology OSD dossier list.

Imatinib (Gleevec/Glivec, Novartis) already went off patent in most markets and is now a large generic market. Sakar manufactures Imatinib and this is one of the first products for which it received an EMA marketing authorization via the Accord technology transfer.

Abiraterone (Zytiga, Johnson and Johnson) for prostate cancer is a genericized, growing generic market. Sakar has Abiraterone Acetate API with DMF confirmation and Abiraterone tablets in the formulation list.

There are many more, but the pattern is a concentrated wave of blockbuster oncology drugs, many of which have annual branded sales of USD 2-5 billion each, are losing patent protection between now and 2030. The total branded oncology revenue subject to generic competition in this window is estimated at USD 80-120 billion globally. Every product in Sakar’s oncology pipeline maps onto this specific patent cliff window.

Complexity: Unlike a straightforward generic tablet where any manufacturer with WHO-GMP certification can compete, most of the next-generation oncology patent cliff products require specialized formulation technology. Enzalutamide is a BCS Class II molecule (poorly soluble) requiring amorphous solid dispersion technology (Hot Melt Extrusion) to achieve bioequivalence. Olaparib is similarly BCS Class II. Doxorubicin liposomal requires lipid nanoparticle manufacturing. These are not products where 50 manufacturers file simultaneously on day one of patent expiry. The pool of qualified manufacturers globally is perhaps 5-10 for each molecule.

Big pharma is exiting manufacturing

The traditional model was that large pharmaceutical companies like Pfizer, Novartis, Roche, and Johnson and Johnson owned their entire manufacturing chains. They synthesised APIs, manufactured formulations, packaged, and distributed. This was capital-intensive and operationally complex.

They now want to own intellectual property, manage regulatory relationships, and sell commercially without operating factories. The manufacturing function has been progressively outsourced.

Pfizer has been selling manufacturing plants for years. Novartis spun out its generics business as Sandoz. AstraZeneca sold several manufacturing facilities. The logic is that a pharmaceutical company’s return on capital from manufacturing a commodity injectable is 8-12%, while its return on capital from selling a branded drug is 30-50%. Capital is better deployed toward R&D and commercial operations than factory floors.

This created the Contract Development and Manufacturing Organisation (CDMO) industry. The global CDMO market was estimated at approximately USD 165 billion in 2022 and is projected to grow at 7-10% annually. But within this, oncology CDMO is growing faster, at 10-15% annually, for three reasons:

1. Cytotoxic manufacturing requires high containment infrastructure that is increasingly uneconomical for companies without scale. A high-potency OEB Level 4 facility requires tens of millions of dollars to build and specialized compliance overhead to maintain. For a company that manufactures 5 oncology products, this cost structure is devastating. For a CDMO that manufactures 50 oncology products across this infrastructure, the per-product cost is manageable.

2. The patent cliff described above is creating waves of new generic product launches. Every brand that goes off patent requires generics manufacturers to quickly establish manufacturing. Instead of building their own factories, they outsource to CDMOs with existing capacity and approvals.

3. The regulatory complexity of oncology manufacturing has increased. EU and US regulatory authorities have tightened containment requirements, validation requirements, and process analytical technology requirements for cytotoxic products. The cost of compliance is rising, making in-house manufacturing less attractive for all but the largest players.

For Sakar, the CDMO opportunity works on two levels. Domestically, every Indian pharma company with an oncology portfolio is a potential CDMO customer because they cannot all afford to build Bavla-equivalent facilities. Internationally, global generic companies like Accord (operating in 100+ countries with a European focus) and Torrent need manufacturing partners in India who can supply their European registered products at Indian cost structures. Sakar’s EU-GMP approved, high-containment facility is specifically the type of asset that this structural shift in big pharma creates demand for.

India’s domestic oncology market is at an inflection point

The Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (PM-JAY) launched in 2018 provides health insurance coverage of Rs 5 lakh per family per year to approximately 500 million people at the bottom of the income distribution. The scheme covers cancer treatment, including chemotherapy, radiation, and surgical oncology. Before this scheme, a significant proportion of India’s cancer patients either went untreated or delayed treatment until it was unaffordable.

When a patient who previously could not afford chemotherapy can now access it through insurance, that translates into incremental demand for oncology formulations. The utilization of PM-JAY for cancer treatment has been growing every year since the scheme’s launch. This is new demand creation, not market share redistribution.

India’s broader health insurance penetration is also rising. IRDAI data shows health insurance premiums growing at 20%+ annually. Employer-provided health insurance is expanding as formal employment grows. This creates a middle-income oncology patient segment that can access branded generic oncology products and is less price-sensitive than the PM-JAY segment.

The supply side is also expanding. The National Cancer Grid coordinates 250+ cancer care centers across India. The government has been establishing new AIIMS campuses in smaller cities and tier-2 towns. The number of oncologists in India, while still insufficient, has grown significantly over the past decade. More oncologists in more locations means more diagnoses and more treatment initiations. Each new oncologist treating patients in a tier-2 Indian city is creating demand for generic oncology drugs that need to be manufactured.

To put it simply:

🌍 Cancer burden rising globally and in India → more diagnoses, more treatment starts, more chemotherapy cycles → demand growth for oncology formulations

💊 Oncology patent cliff accelerating → blockbuster molecules becoming generic → affordability improves, volumes expand, and complex generic manufacturers get a large opportunity

🏭 Big pharma outsourcing manufacturing → cytotoxic compliance and containment costs rising → oncology CDMOs with approved high-containment facilities become more valuable

🇮🇳 India oncology access improving → PM-JAY, insurance penetration, cancer centers, and oncologist availability expanding → domestic oncology consumption moves from under-treatment to formal demand

Value

Return Ratios

The two most important return metrics for a capital-intensive business are Return on Equity (ROE) and Return on Capital Employed (ROCE).

10% ROE: For a specialty pharma company claiming structural margins and a premium positioning, 10% ROE is unimpressive. The Indian small-cap pharma sector median ROE for well-run companies is 15-25%. Companies like Divi’s Laboratories and Navin Fluorine have sustained 20-25% ROE over cycles. Sakar is well below this.

12% ROCE: This is similarly below the 18-25% that well-run Indian pharma manufacturers sustain.

The Bavla facility was built for Rs 800-1,000 crores of revenue and currently generates approximately Rs 96 crores of oncology revenue at 29% utilization. The return ratios are depressed because the denominator (capital employed) includes the full cost of an asset that is earning only a fraction of its potential. This is a common pattern in capital-intensive businesses before utilization reaches its inflection point.

Margin Profile

FY26 full-year gross margin was 51% (Rs 12,846 lakhs gross profit on Rs 25,174 lakhs revenue). EBITDA margin was 27.4%. EBIT margin was approximately 18%. PBT margin was 15.8%. PAT margin was 12.1%.

A comparison with FY25: gross margin 53.7%, EBITDA 28.0%, PAT 9.9%. So full-year gross margin actually compressed 2.7 percentage points year-on-year despite the oncology mix growing from 21% to 38% of revenue. This is counterintuitive and needs explanation.

The explanation lies in COGS dynamics. Revenue grew 42% from Rs 17,758 lakhs to Rs 25,174 lakhs. COGS grew 50% from Rs 8,226 lakhs to Rs 12,328 lakhs. This likely reflects one or more of three dynamics. First, as Sakar does more of its own-brand oncology manufacturing (buying APIs from third parties rather than having CDMO clients supply the API), the raw material cost is now in Sakar’s P&L rather than off it. Second, oncology APIs are expensive, molecules like lenalidomide, bortezomib, and docetaxel cost thousands of dollars per kilogram. As the oncology CDMO volume grew, the API cost embedded in COGS grew with it. Third, there may have been some API cost inflation in specific cytotoxic molecules during the year.

The more revealing data point is the quarterly gross margin trajectory: Q3 FY26 was 49%, Q4 FY26 was 60%. Nearly identical revenue (Rs 7,034 versus Rs 7,110 lakhs), but an 11 percentage point gross margin difference in one quarter. Management attributed this partly to inventory management improvements. A more complete explanation is that the product mix within Q4 was tilted toward higher-margin oncology products, possibly with some domestic oncology batches that carried lower input costs.

A 60% gross margin quarter, if sustained and accompanied by the operating leverage on fixed costs, implies an EBITDA margin capability well above the 27% full-year FY26 figure. Q4’s 37% EBITDA margin on Rs 71 crores of revenue suggests that at Rs 500 crores of revenue with a similar product mix (but more high-margin oncology exports in it), the blended EBITDA margin could sustainably be 30-35%. The question is whether Q4 is an outlier or a leading indicator of the direction.

Working Capital

Inventory days moved from approximately 90 days to 102 days. There are two possible readings. The optimistic reading: Sakar is pre-stocking expensive cytotoxic APIs in anticipation of export orders from the tech transfer partnerships commencing in FY27. Buying ahead at current prices and avoiding supply disruption risk is a sensible capital decision for complex oncology molecules. The cautious reading: finished goods are accumulating because expected export orders have not yet arrived. If this is finished goods inventory for EU markets awaiting serialisation and artwork approvals, there is working capital at risk if the export timeline slips further. This is a specific line item to track in the Q1 FY27 results.

Trade Receivables grew 63% increase against 42% revenue growth. Receivable days extended from approximately 63 days to 73 days. This is consistent with a revenue mix shift toward clients with longer payment cycles. European pharma companies and international distributors typically pay in 60-90 days, versus domestic Indian CDMO clients who often pay in 30-45 days. If exports were still negligible in FY26, the receivable days extension may reflect domestic oncology clients having somewhat longer credit terms than non-oncology CDMO. As exports grow in FY27, receivable days could extend further. Monitoring this metric is important because a stretch to 90+ days on a growing revenue base implies material working capital consumption.

Trade Payables is the most unusual balance sheet movement. Payable days based on COGS extended from approximately 94 days to approximately 153 days.

A 153-day payable cycle is long and warrants scrutiny. There are three readings. The first and most benign: as Sakar sources more expensive cytotoxic APIs for oncology manufacturing, it has negotiated extended payment terms with API suppliers. The second: Sakar’s large clients (Zydus, Torrent) are taking longer to pay, which is forcing Sakar to extend its own payables to manage liquidity. The third and most concerning: Sakar is stretching payables to fund working capital growth because the business does not generate enough cash to fund its own growth. The net cash position of Rs 25.96 lakhs (essentially zero) against Rs 5,154 lakhs of payables suggests the company is running tight on liquidity.

The payables jump is what actually makes the Cash Conversion Cycle look healthy — approximately 21 days in FY26 versus 59 days in FY25. But a CCC that is only short because payables are stretched is a fragile kind of working capital efficiency. If suppliers tighten terms, the CCC deteriorates sharply and the working capital funding gap widens.

Cash Conversion

Free Cash Flow (OCF - investing outflow) was Rs 2,091 lakhs in FY26 compared to Rs 332 lakhs in FY25. This is a significant improvement and reflects the completion of the Bavla capex. Investing outflow fell from Rs 3,071 lakhs to Rs 2,835 lakhs, and as the facility is now largely complete, FY27 capex should be substantially lower, perhaps Rs 1,500-2,000 lakhs of maintenance and minor expansion capex rather than facility construction capex.

The FCF improvement trajectory is the most encouraging signal. In FY22-FY24, the company was a net cash consumer as it funded Bavla construction. In FY25, it generated a negligible Rs 332 lakhs of FCF. In FY26, it generated Rs 2,091 lakhs. If FY27 capex falls to Rs 1,500 lakhs and OCF grows with revenue, FCF could see a significant improvement that begins to fund organic growth without requiring new debt or equity.

Growth

1. Domestic CDMO

As Sakar’s domestic CDMO customers grow their own businesses, Sakar grows with them. The Indian pharmaceutical market is expanding at 8 to 12% annually. The domestic oncology market is growing faster, at 12 to 18%, driven by rising cancer incidence, expanding health insurance coverage through Ayushman Bharat, and the rapid proliferation of oncology treatment infrastructure across tier-2 Indian cities. Sakar needs its existing customers to grow their volumes, which they are doing.

The numbers reflect this. Oncology revenue grew from approximately Rs 37 crores in FY25 to approximately Rs 95.7 crores in FY26, a 158% increase. In Q4 FY26 specifically, the domestic oncology quarterly run rate was Rs 31.46 crores, implying an annualized domestic oncology base of approximately Rs 125 crores. Non-oncology domestic revenue ran at approximately Rs 156 crores for the full year. Combined, the domestic business accounts for virtually all of Sakar’s current revenue.

The margin profile differs between the two. Non-oncology CDMO work runs at 45 to 55% gross margin. Oncology CDMO work runs materially higher, probably 60 to 70%, because the complexity premium and the scarcity of qualified manufacturers creates pricing power. This means even within the domestic business, the mix is shifting in a margin-positive direction as oncology grows faster than non-oncology.

The realistic growth expectation for this combined engine going forward is 10 to 12% annually.

2. Technology Transfer with Regulated Partners

Current pipeline as of Q4FY26:

• Accord-Intas: 9 molecules and 12 SKUs are under technology transfer. 2 products have received EMA site variation approval and are in the process of first commercial supply. The remaining 7 products have been submitted to EMA and are awaiting approval. Management confirmed on the call that these 7 are expected to come through “one by one, maybe one or two this quarter and the following quarter the rest.” The full Accord portfolio represents a potential Rs 50-100 crores of annual revenue for Sakar.

• Torrent (UK): 2 products have received site variation approvals. Commercial supply is starting from Q1 FY27.

• Torrent (Germany): included in the 5 tech transfer projects commercializing from Q1 FY27.

• A third UK company: 1 product also commercializing from Q1 FY27.

• Glenmark and Emcure: technology transfers ongoing at the Bavla plant, EMA submissions not yet made.

Total current EMA site variation approvals: 4 products. Each of these 4 is essentially ready for commercial supply. 7 more Accord products plus Torrent’s European portfolio plus Glenmark and Emcure is the pipeline for the next 18-24 months.

The critical distinction between this revenue category and generic export cold-start efforts is that the regulatory work is substantially done by the partner, not by Sakar. Sakar’s obligation is to execute technology transfer successfully and maintain compliance. The commercial risk (market distribution, pricing, competition in individual EU countries) sits with Accord.

Revenue conversion math: Management says the Accord portfolio alone is Rs 50-100 crores of potential annual revenue across 9 molecules. If 5 of 9 molecules are commercially active by Q3 FY27, and the annual potential is Rs 50 crores at a conservative baseline, that implies Rs 25-30 crores of Accord revenue in FY27 (partial year effect). Adding the rest (conservatively another Rs 10-15 crores) and the total tech transfer revenue in FY27 could be Rs 35-50 crores. This would be up from essentially zero in FY26.

The export revenue context: Currently, non-oncology exports from the Changodar facility run at approximately Rs 115-120 crores annualised (based on Q4 FY26’s Rs 29 crores quarterly figure). This is the established, semi-regulated market branded exports business covering Africa, MENA, CIS, APAC, and LatAm. The tech transfer oncology exports from Bavla are additive to this and carry materially higher margins because oncology products in European markets command European price levels.

3. Own Brand Exports

The economics are better than CDMO because Sakar owns the product development.

Current pipeline as of Q4FY26:

1. 55 oncology products developed in-house. 32 of these have dossiers ready for global launch. 21 of the 32 have been shared with partners. 11 have received marketing authorisation approvals. Separately, 11 bioequivalence studies have been completed from EMA or USFDA-approved CROs. Reference listed drugs (RLDs) are from EU, which is the standard required for EU regulatory submissions.

2. For non-oncology, there are 16 EU CTD dossiers for oral solids and 16 for injectables. 250 dossiers total have been shared globally across all product categories, 125 have been formally submitted for registration, and 12 have received marketing authorizations.

The two-year business plan targets 300+ dossier approvals. Currently at 12 MAs, this implies a target of obtaining approximately 290 additional approvals over two years, an aggressive target that probably reflects the pipeline of 125 submitted dossiers beginning to convert.

In the EU, an oncology generic with limited competition might sell at EUR 50-500 per pack depending on the molecule. A product like Lenalidomide (which Sakar has across 6 strengths) could command EUR 200-500 per 28-pack in certain European markets even as a generic. Volume may be modest (a few hundred thousand packs annually) but revenue per unit is high.

In Australia, TGA-registered generics command prices similar to EU levels. Sakar mentions Australia as a submission market. An Australian registration for a molecule like Imatinib or Abiraterone could generate AUD 2-5 million annually.

In Africa and MENA, the volume is higher but prices are lower. A registration across 10 African countries for a cephalosporin or oncology product might generate Rs 3-8 crores annually across the portfolio.

Revenue conversion math: Out of the 31 oncology dossiers currently submitted or in final stages, if 80% convert to registrations by Q3 FY27, that is 25 products. Each product needs a 120-150 day commercial lead time for first supply. First supplies from these 25 products would therefore begin Q3-Q4 FY27 in the best case. This is the source of management’s confidence in FY27 being a materially better export year than FY26.

The long-term potential of this pipeline is the most significant growth number in the story. If 300 dossiers are registered and each generates an average of Rs 2-4 crores annually, the portfolio value is Rs 600-1200 crores in revenue.

4. API Integration and Revenue

First, Sakar can sell its CEP-approved oncology APIs to third-party manufacturers globally. When a European or Indian pharma company wants to make a generic oncology product but does not synthesise the API themselves, they need a qualified supplier with CEP certification. Sakar’s APIs with CEP approval are saleable to this market. Second, when Sakar uses its own CEP-approved API in its formulation products for EU export, it captures the API margin within the vertically integrated supply chain rather than paying it to a third-party API supplier.

Current state: 22 oncology APIs developed in-house. Gefitinib and Capecitabine have CEP approvals. Two more CEP applications have been submitted and results are expected within a few months. Five more are planned for submission, bringing the potential total to 9 out of 22 APIs with CEP status within 12-18 months.

The economics: Oncology APIs are among the highest-value bulk pharmaceutical chemicals. Prices range from USD 500 to USD 5,000 per kilogram depending on the molecule’s synthesis complexity and patent status. Cytotoxic APIs like Docetaxel and Paclitaxel are complex to synthesise and command premium pricing. Even for molecules like Imatinib Mesylate (which is more genericized), the API is sold at USD 300-600 per kilogram in regulated markets with CEP certification.

For Sakar’s formulation business, the API integration benefit is a permanent gross margin improvement on every product that uses in-house API. If a finished formulation previously had an API cost of 40-50% of its selling price (which is typical for oncology generics where the API is expensive), and Sakar now makes that API itself, the cost falls to manufacturing cost + margin. This potentially reducing the effective COGS by 25-35% on integrated products.

Valuation

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

JNK India builds the heating equipment that sits at the core of India’s refineries, petrochemical plants, and fertilizer factories. Before any of those facilities can run, someone has to design and install the equipment that heats raw materials to the precise temperatures needed for chemical reactions to happen.

The company was founded in 2010 as a joint venture with JNK Heaters Co. Ltd., a KOSDAQ-listed South Korean company. The Korean parent brought the thermal engineering technology and the Indian entity built the execution capability.

JNK has 3 business segments:

Heating Equipment (~73%)

1. Process fired heaters: Think of them as large industrial furnaces where fuel burns around tubes carrying crude oil fractions or chemical feedstocks. The fluid inside gets heated to a precise temperature before moving to the next step in the refining or chemical process. Every refinery uses several of these.

2. Reformers: Inside a reformer, a chemical reaction happens inside the tubes rather than just heating. The most common use is hydrogen production, where natural gas and steam are fed into catalyst-packed tubes at around 800-900°C and broken down into hydrogen. Reformers are central to fertilizer plants and hydrogen production facilities. This is where JNK’s Korean technology partnership provides the most direct advantage.

3. Cracking furnaces: Used in petrochemical plants to produce ethylene and propylene (the raw materials for most plastics). Large hydrocarbon molecules are heated to around 800°C and broken apart into smaller, more useful ones. These are the most technically demanding products JNK makes.

Flares, Incinerators, and Process Plants

These are safety and environmental systems that burn off excess or hazardous gases at industrial facilities. Demand here is increasingly driven by stricter emission regulations rather than new plant construction, which makes it relatively steady.

Process plants covers technology-based EPC contracts where JNK executes complete plant packages rather than supplying individual equipment.

JNK Chemdist Technologies

In August 2025, JNK formed a joint venture with Chemdist, a technology company with patented processes in green hydrogen and sustainable chemicals. JNK holds 51%. In its first six months of operation, the JV contributed approximately 7% of group revenue. The near-term business is conventional process plant engineering for chemical and pharma clients, but the longer-term focus is a proprietary process that produces hydrogen as a byproduct of converting ethanol into ethyl acetate, a widely used industrial solvent. The pilot project is running in Pune. Commercial validation is still two years away, but the technology is genuinely differentiated from standard hydrogen production routes.

I’ve analyzed JNK India using the TVG framework:

Theme

India needs more refineries, petrochemical plants, and fertilizer factories

India is the world’s third-largest oil consumer, and demand is growing every year. To meet that demand, India needs more refining capacity. Every new refinery and every expansion of an existing one needs fired heaters, reformers, and related thermal equipment, which is exactly what JNK supplies.

Petrochemicals are the second leg. India currently imports a significant share of its plastic and chemical intermediate requirements from abroad. The government and large industrial groups have been pushing to build domestic petrochemical capacity to reduce that import bill. The building blocks of plastics are ethylene and propylene, and producing them requires cracking furnaces. JNK just won the largest cracking furnace order in its history, the BPCL Bina project worth Rs. 1,050 cr, which is evidence that this investment wave is arriving.

Fertilizers are the third leg. India is the world’s largest importer of urea, and reducing that dependence is a stated policy priority (the ongoing Middle East conflict has taught us many things). New urea plants require ammonia, and ammonia production requires reformers. JNK is already pre-qualified for domestic fertilizer reformer tenders and has received active inquiries. Bid submissions for at least one fertilizer reformer package are expected in Q1 FY27.

Taken together, the DRHP estimated that demand for heating equipment from Indian refineries, petrochemicals, and fertilizers would run at roughly Rs. 45,000 crore per year between FY24 and FY29, with 61% of that coming from petrochemicals, 37% from refineries, and the balance from fertilizers. JNK held approximately 27% market share in India by order booking at the time of its IPO. Even maintaining that share against a growing market produces meaningful order inflow growth.

The qualification barrier

JNK’s products are installed inside operating industrial facilities at very high temperatures and pressures. A fired heater that fails in a refinery can cause a fire. A poorly designed reformer can destroy the catalyst bed inside it and shut down an entire ammonia plant. A cracking furnace that does not perform correctly can coke up and halt ethylene production for weeks. Because of this, the customers who buy this equipment, and the engineering consultants who specify it on their behalf, do not invite every contractor to bid. They work from approved vendor lists built on track records.

In India, EIL (Engineers India Limited) acts as the engineering consultant for most large public sector refinery and petrochemical projects. Getting onto EIL’s approved list for a specific type of equipment requires a proven execution history. Once on the list, a company participates in every tender EIL manages for that equipment type.

Internationally, the dynamic is slightly different but structurally similar. Process licensors, the companies that own the technology behind reformers and crackers, maintain their own approved vendor lists. A licensor like Haldor Topsoe (which licensed the technology for Dangote’s fertilizer complex) will specify vendors it has worked with before. For the BPCL Bina cracking furnace, JNK needed JNK Global’s existing reference with the process licensor to get pre-qualified. Once Bina is complete, JNK India has its own independent cracking furnace reference in India. That unlocks the next cracker tender without needing to rely on the Korean parent’s credentials.

This is why JNK’s real competitors are actually a short list of technically qualified players: Thermax domestically, and Heurtey Petrochem Solutions internationally. In the Q4 FY26 concall, management made a revealing comment that Heurtey no longer quotes for EPC projects in frontier markets like Nigeria. If that is accurate, the most formidable international competitor has effectively stepped back from the playing field where Dangote Phase 2 will be contested. What remains is a smaller group of credible bidders, with JNK among the strongest given its Phase 1 execution record through JNK Global.

The Dangote Phase 2 opportunity

Nigeria’s Dangote refinery is one of the largest single-site refining complexes in the world. Phase 1 is complete and phase 2 involves doubling refining capacity and adding a major fertilizer complex. JNK Global executed the heater package for Phase 1 roughly a decade ago, when the contract was worth approximately $140 million. Management noted in the Q4 call that equipment prices have roughly doubled since then, putting the Phase 2 heater package in the $250-300 million range.

The fertilizer complex is a separate and additional opportunity. Dangote signed agreements for four fertilizer production streams licensed by Haldor Topsoe. Each stream needs a reformer package. Management estimated each reformer at $30-40 million, putting the total fertilizer reformer opportunity at $120-160 million across four units.

JNK has confirmed pre-qualification for both the heater and reformer packages. Management expects inquiries within one to two quarters from Q4 FY26 results and order finalization around Q3 FY27. The bids are being submitted in combination with JNK Global, with JNK India handling all engineering, procurement, and India-side execution. For the heater package specifically, a reasonable estimate of JNK India’s revenue share would be 60-70% of the total contract value, given the proportion of India-based engineering and supply work.

To put this in context, JNK India’s entire FY26 revenue was Rs. 838 crore. A Dangote heater package at current pricing could represent more than one full year of revenue.

The Middle East and Africa

Beyond Dangote, management quantified the export bid pipeline at roughly Rs. 4,000 crore in active bids as of the Q4 FY26 call, with expected finalization over two to three quarters. The Middle East is the largest component.

Gulf states are investing in downstream petrochemical and fertilizer capacity as part of their economic diversification strategies. Rather than exporting crude oil, they want to produce higher-value chemicals and fertilizers domestically. That investment creates demand for reformers and fired heaters. Management sees $200-300 million in Middle East opportunities over two to three years. A waste gas handling package worth Rs. 200-250 crore in the Middle East was described as in final commercial discussions at the time of the Q4 call.

JNK has also had proposals in Russia for roughly two years, stalled by geopolitical complications. Management said in Q4 that one project is now at an advanced stage of discussion, with finalization possible within a quarter or two.

In FY24 and FY25, a large wave of domestic Indian opportunities arrived simultaneously and absorbed JNK’s full execution capacity, so the export pipeline was deprioritized. As domestic projects move into execution phase and engineering bandwidth frees up, the export pipeline is being actively pursued again. Management expects international orders to scale substantially over the next two to three years, and the bid pipeline supports that expectation with specific numbers.

The energy transition creates demand rather than destroying it

Many investors assume that the global shift toward renewable energy means less investment in refineries and petrochemical plants over time. That assumption is broadly correct over a very long horizon but substantially wrong over the 5-10 year window for investment thesis today, particularly in India.

India’s liquid fuel demand is growing, and more than half of India’s energy consumption still comes from coal and oil. Fertilizer demand is tied to agriculture and food security, which is independent of energy transition dynamics. Global plastics demand, which drives petrochemical investment, continues to grow in developing economies even as recycling and substitution gain ground in developed ones. The plants being built today will operate for 20-30 years.

More interestingly, the energy transition is directly creating new demand for JNK’s core products. Hydrogen production, whether classified as grey, blue, or green, requires reformers. India’s National Green Hydrogen Mission targets significant scale-up of hydrogen capacity. Even the pathway to green hydrogen runs through grey and blue hydrogen in the near term, because electrolysis-based green hydrogen remains expensive and intermittent renewable supply creates operational challenges. Every hydrogen plant, regardless of its color classification, needs thermal processing equipment.

The JNK Chemdist JV is positioned precisely at this intersection. Chemdist has a patented process that converts ethanol into ethyl acetate, a widely used industrial solvent, with hydrogen produced as a byproduct. The economics are different from standard hydrogen routes because the ethyl acetate revenue partially covers the production cost, making the effective cost of the hydrogen lower. The pilot project at Hydrogen Valley Pune, currently being executed, will be the first commercial proof-of-concept. If it validates the process, this technology becomes licensable, and the addressable market is any industrial user who needs both hydrogen and a commodity chemical solvent.

Flares and incinerators add one more energy-transition-adjacent demand source. Stricter environmental regulations globally are forcing industrial facilities to upgrade waste gas handling systems that were built to older, more permissive standards. This demand is retrofit-driven rather than tied to new plant construction, which makes it more consistent and less cyclical than the rest of JNK’s order pipeline.

Putting it simply:

🛢️ India is building more refineries, petrochemical plants, and fertilizer factories → every single one needs JNK’s equipment to function

🌍 Large export markets, led by Dangote Phase 2 in Nigeria, are opening up simultaneously → potential for a single order worth more than a full year of revenue

🔒 Only a handful of companies globally are qualified to build this equipment → JNK is one of them, and that list does not grow easily

⚡ The energy transition is creating new demand, not destroying old demand → hydrogen plants, clean fuels, and emission compliance retrofits all need JNK’s products

🔄 Every project JNK completes strengthens its qualification record → which directly unlocks access to larger, more complex projects next time

Value

The first question any investor should ask about an EPC company is simple: does it actually make money, or does it just look like it does on the income statement while cash bleeds? JNK’s recent financial history is the story of a company transitioning from the second to the first.

The legacy project problem and why it is now behind them

When JNK scaled up rapidly after its IPO, it took on a set of larger, more complex projects than it had historically executed. Some of these older contracts, referred to consistently in concalls as “legacy projects,” were bid at thinner margins during a period when the company was establishing itself in new product categories and larger order sizes. Executing them took longer than expected, costs ran higher, and the revenue recognition methodology at the time (output-based, tied to completion milestones) created lumpy, difficult-to-interpret quarterly results.

EBITDA margins compressed to single digits in some quarters, operating cash flow was negative for three consecutive years, and investors who looked at the balance sheet saw debtor days running above 200.

The transition started in H2 FY25 when the company shifted to input-based revenue recognition, which ties revenue to costs incurred rather than milestones hit. This created more stable quarterly reporting. More importantly, by Q3 FY26, management confirmed that legacy projects were substantially complete.

The FY26 numbers

The quarterly trajectory within FY26 is more important. EBITDA margin was 7% in Q1, 12.1% in Q2, 14.3% in Q3, and 15.2% in Q4. This is the legacy project drag washing out in real time. By Q4, with legacy orders gone and new, better-structured projects dominating execution, the margin settled at 15.2%. Management guided 14-15% as the normalized run-rate going forward.

The working capital transformation

Cash conversion cycle fell from 162 days in FY25 to 57 days in H1 FY26.

Part of this improvement is the accounting methodology change and the roll-off of legacy PSU projects reduced unbilled revenue balances. But the more durable driver is customer mix, as PSU contracts in India are structured so that the bulk of payments come at the end of the project, tied to commissioning milestones that routinely slip. Private sector customers like Reliance, and export customers backed by Letters of Credit, pay differently. Advances are higher, milestone payments are more frequent, and collection timelines are shorter.

The BPCL Bina cracking furnace order is the critical test of whether this improvement holds. Management confirmed that essentially no revenue from Bina had been booked through Q4 FY26. As this Rs. 1,050 crore project moves into peak execution in FY27, the billing and collection structure will determine whether the 76-day cash conversion cycle was a temporary phenomenon.

Operating cash flow

Three consecutive years of negative operating cash flow is the fact that kept quality-conscious investors away from JNK. An investor asked about this directly in the Q4 call, and management’s response was that operating cash flow has improved significantly, is now positive in FY26, and the trend should continue as new projects with better payment terms dominate the book.

The asset-light model and what it means for ROCE

JNK’s manufacturing footprint is deliberately limited. The Mundra facility in Gujarat covers roughly 20,000 square meters and has a 5,000 MT fabrication capacity, primarily oriented toward export work. For domestic projects, JNK uses customer-approved third-party fabrication shops near the project site, which reduces logistics costs and avoids large fixed asset investments.

The real asset of this business is its engineering team. JNK employs approximately 450 people, of which 120 are engineers focused on thermal and mechanical design. This team, combined with the qualification record and the JNK Global technology access, is what creates the moat. It is accumulated technical credibility and the ability to design and execute safely at very high temperatures.

This asset-light structure is why ROCE of 19.1% is achievable at relatively modest absolute profit levels. Capital employed stays low because the company does not need to build factories or hold large raw material inventories to grow. Growth comes from winning more orders and hiring more engineers, both of which consume far less capital than traditional manufacturing scale-up.

The JNK Global relationship

For large contracts, bidders are required to post bank guarantees, typically 10% of the contract value, to give the customer assurance of performance. On a Rs. 1,000 crore contract, that means finding Rs. 100 crore of bank guarantee capacity. For an Indian company of JNK’s current size, arranging that from Indian banks is expensive and capacity-constrained.

JNK Global provides bank guarantees from Korea for large international projects and for certain domestic contracts where the scale exceeds JNK India’s comfortable BG capacity. This effectively allows JNK India to bid contracts that would otherwise strain its balance sheet. The cost is a 2% technical fee on independently won overseas orders, which is currently minimal given that most large orders are either domestic or routed through JNK Global. The benefit is access to a much larger addressable project universe than JNK India’s balance sheet alone would allow.

As JNK India’s own balance sheet strengthens through retained earnings over FY27 and FY28, this dependency naturally reduces. But in the near term, it helps JNK punch above its weight on large bids.

Growth

Current order book

As of March 31, 2026, JNK’s consolidated order book stood at Rs. 1,961 crore. Against FY26 revenue of Rs. 838 crore, that is a book-to-bill ratio of approximately 2.3x, meaning the company has more than two years of revenue already secured. This is the foundation of any near-term growth case.

The single most important item in that order book is the BPCL Bina cracking furnace.

BPCL Bina, formerly known as Bharat Oman Refinery, is expanding its refinery and adding a new petrochemical complex in Madhya Pradesh. As part of that expansion, BPCL needed a cracking furnace package, the equipment that will produce ethylene and propylene at the new facility. JNK Global won the overall contract, valued at approximately Rs. 2,600 crore. JNK India’s share of that contract, covering all Indian engineering, supply, and construction work, was booked at Rs. 1,050 crore into JNK India’s order book, with an additional Rs. 400-600 crore expected to be added in subsequent quarters as scope is finalized.

The revenue recognition timeline for Bina is what makes it the dominant growth driver for FY27. Management was explicit in the Q3 FY26 concall: no revenue from this project was booked until very late in FY26, with only 3-4% recognized in Q4. The peak revenue year is FY27, when management guided that 50-60% of the Rs. 1,050 crore already booked would be recognized. That is Rs. 525-630 crore of revenue from a single project in a single year, layered on top of everything else the company is executing.

New order pipelines

Management guided revenue growth of 25-30% in FY27, which on Rs. 838 crore of FY26 revenue implies Rs. 1,050-1,090 crore. That guidance is conservative in an important sense: it is achievable largely from the existing order book, without significant new order wins contributing to FY27 revenue. New orders won in Q1 FY27, management noted, would not begin generating meaningful revenue until Q3 at the earliest.

1. Dangote Phase 2

Dangote is doubling the capacity of what is already Africa’s largest refinery, and adding a fertilizer complex alongside it. The engineering consultant for the project is EIL, which means the bidding process follows a structured, qualification-based framework that favors established vendors over new entrants.

JNK Global executed the heater package for Dangote Phase 1 a decade ago, when the contract was approximately $140 million. Equipment prices have roughly doubled since then. The Phase 2 heater package is therefore estimated in the $250-300 million range by management. JNK India’s share, covering engineering and India-side supply and construction, would represent 60-70% of that value.

The fertilizer complex adds a second opportunity. Dangote signed agreements for four fertilizer production streams using Haldor Topsoe’s licensed technology. Each stream requires a reformer package, and management estimated each reformer at $30-40 million. Four units put the total fertilizer reformer opportunity at $120-160 million.

Order finalization for the heater package was guided around Q1 FY27, and for the fertilizer reformers around Q2-Q3 FY27. A confirmed Dangote order would represent a step-change in JNK’s order book, adding Rs. 1,500-2,000 crore in a single announcement.

2. Middle East

Alongside Dangote, management described a Middle East waste gas handling package worth Rs. 200-250 crore as being in final commercial discussions at the time of the Q4 call, with closure expected within weeks. A domestic fertilizer reformer package tender was in active preparation, with bid submission expected in Q1 FY27 and finalization in Q2-Q3.

The broader Middle East pipeline, covering fertilizer and petrochemical opportunities, was sized at $200-300 million over 2-3 years by management. Russia adds a smaller but real element, with one project at an advanced stage of discussion after two years of slow progress.

3. Domestic pipeline

On the domestic side, the large refinery projects that management referenced, BPCL Andhra, IOCL Paradip Phase 2, HPCL various expansions, are still six months to a year away from active tendering by management’s own estimate. The near-term impact of oil price weakness on PSU investment decisions is causing some delay. But these projects are not cancelled. When they arrive, they represent a substantial second wave of domestic demand layering on top of what JNK is already executing.

The petrochemical side is moving faster. Management noted in the Q3 FY26 concall that two to three additional cracker projects are at preparatory stage in India, with consultancy and licensing work already underway. Each of these represents a potential cracking furnace opportunity for JNK, and Bina’s successful execution is the reference that would determine JNK’s ability to win them independently.

Total order inflow and conversion rate

Management guided Rs. 1,300-1,500 crore in new order inflows as sufficient to support two years of growth at the current guidance level. The historical conversion rate on bids submitted is approximately 25-30%. Against a live bid pipeline that management described as approximately Rs. 4,000 crore, the math produces Rs. 1,000-1,200 crore of expected inflows under a normal hit rate, consistent with the guidance. Dangote winning would take that number considerably higher.

3. JNK Chemdist Technologies

The JV’s order book stood at Rs. 70 crore as of March 2026, with a bid pipeline of approximately Rs. 200 crore in chemical, pharma, and water-related process plant projects.

Management guided the JV to contribute 10-15% of consolidated revenue within two to three years. At FY27 guided revenue of Rs. 1,050-1,090 crore for the group, 10-15% implies Rs. 105-163 crore from the JV.

The near-term business is relatively straightforward: turnkey process plant engineering for chemical and pharmaceutical clients, leveraging Chemdist’s manufacturing capability in reactors, distillation columns, separators, and pressure vessels. This is a competitive market, and Chemdist’s existing order pipeline suggests genuine commercial traction.

The longer-term story sits in the technology. Chemdist holds patents on a process that converts ethanol into ethyl acetate, a widely used industrial solvent, with hydrogen produced as a byproduct. The economic logic is that ethyl acetate is the primary product with commercial value, which means the hydrogen is produced at a structurally lower effective cost than standard hydrogen production routes. This is a differentiated technology pathway, one that does not require expensive electrolysers or depend on continuous cheap renewable power. The pilot project, a commercial-scale installation at Hydrogen Valley Pune, is being executed in FY27. If it validates the process, the JV has a licensable technology with a pipeline of potential clients across pharma, paints, flexible packaging, and any industrial user who needs both hydrogen and a commodity solvent.

Management was measured in the Q4 call about the technology timeline, noting that commercial-scale proof-of-concept is approximately two years away and that early JV projects are being executed on competitive rather than technology-premium terms. What is already proven is that the JV can win and execute process plant contracts, which gives it a viable revenue base even if the hydrogen technology takes longer to commercialize than hoped.

The operating leverage that ties it all together

One growth dynamic that does not fit neatly into any of the three layers above but connects all of them is the operating leverage embedded in JNK’s current cost structure.

In FY26, employee costs grew 46.5% as the company hired ahead of its pipeline. The Mundra facility is underutilized, used primarily for export work that has not yet scaled. Engineering and project management bandwidth is at roughly 70% utilization by management’s own estimate. These are largely fixed costs in the near term.

When revenue scales from Rs. 838 crore toward Rs. 1,050-1,100 crore in FY27, those fixed costs get spread across a significantly larger revenue base. The incremental margin on that additional revenue is therefore higher than the average margin. EBITDA growing faster than revenue is not a projection of operational improvement. It is simple arithmetic from a fixed cost base meeting higher volumes.

If FY27 plays out at 25-30% revenue growth with stable or improving margins, the PAT impact is even more pronounced because finance costs and depreciation, both relatively fixed, grow slowly. Management has guided 14-15% EBITDA margins as the sustained run-rate. On Rs. 1,050-1,100 crore of revenue, that produces Rs. 147-165 crore of EBITDA and PAT in the range of Rs. 95-110 crore, compared to Rs. 64.8 crore in FY26. That is a 47-70% increase in PAT on 25-30% revenue growth.

Valuation

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

(Ridham Desai)

Frank Knight, writing in 1920, made an observation that has lost none of its force: all profit resides in uncertainty. The point of maximum profit is at the point of maximum uncertainty. When the future narrows into something knowable, the excess return disappears with it. Look at India in mid-2024, after the election results settled and the narrative became clear. Valuations re-rated, certainty arrived, and within two or three months the market had made its peak, as there was nothing left to price in.

Today the environment is the reverse, as nobody knows when the conflict ends. Supply chains are breaking down in ways that are still working through the system which blurs the macro across multiple fronts simultaneously. The current moment is not maximum uncertainty, but it is high uncertainty. The darkest point was probably late March into early April. Things are slightly less murky now, but the gap between the bull case and the bear case remains wide, and that gap is exactly where returns are manufactured.

What makes this moment unusual is the positioning data. On relative valuation, relative performance over the trailing twelve months, and FII positioning, India is at a lifetime low. In thirty-five years of coverage, Ridham has not seen levels like this. India has not traded this cheaply relative to global markets since it opened to foreign investors in 1993.

The market is not ignoring this. Indian retail investors bought in March while institutional capital was heading for the exit. Most commentary treated this as unsophisticated behavior, but the more accurate reading is the opposite. When you buy risk assets during periods of genuine uncertainty, the prospective result is usually good. Tops and bottoms are for fools, as the saying goes. The retail investor who bought the dip was pricing in the right thing at the right time.

II: What the Strait of Hormuz Is Actually Doing to India

(Ritesh Jain)

Markets are watching oil, but that’s the wrong variable. Crude accounts for 20-25% of what passes through the Strait of Hormuz. The rest is methanol, sulfur, helium, ammonia, LNG, and urea. None of these attracted headlines but are what holds global food production together.

Half the world’s population depends on nitrogen fertilizers to eat. 30% of all urea moves through the Strait. Virtually none has passed through in the past two months. The price signal arrived with a lag (a tanker that cleared the Strait sixty days ago only reached Europe recently), which is why the disruption felt abstract for so long. It no longer does. India is purchasing fertilizer at prices double what they were before the conflict began, Thailand has been unable to sow its crop this season because the inputs did not arrive, US wheat prices have risen to their highest level since 2024, with projections pointing to the lowest domestic harvest in fifty years.

The oil picture is quieter only because floating inventories have been drawn down aggressively to buffer the shock, which is finite. Exxon’s stock price today sits below where it was when the conflict began, which tells you how thoroughly the market has discounted the physical reality. This is an availability crisis rather than a price shock, which matters because availability crises are not solved by demand destruction alone.

The bear case for India is a sustained escalation that drives oil to $130, which causes real pain. But it doesn’t cause a balance of payments crisis: India’s oil intensity relative to GDP has roughly halved since 2008, when crude touched $140 and the economy absorbed it. The 1991 episode, when oil doubled in a single month and triggered a full BOP collapse, is a different order of magnitude from where India sits today. The current account pressure is manageable under most scenarios.

What changes the risk calculus is the El Niño overlay. Independent research points to an incoming super El Niño potentially stronger than the event of 1876 to 1878, which was the episode that caused the southwest monsoon to fail across multiple consecutive seasons, destroyed harvests across the subcontinent, and produced one of the worst famines in recorded history. That event unfolded in a world without nitrogen fertilizers, yet this one arrives on top of a fertilizer supply shock already in progress.

III: The Nominal GDP Problem

(Ritesh Jain / Ridham Desai)

From 2013 to 2023, India’s nominal GDP grew at roughly 12% per year on average, and that rate justified the valuation multiples the market assigned. After 2024, nominal GDP growth collapsed to 8.5-9%. The market cap-GDP ratio did not follow it down, so valuations stayed elevated while the growth engine that supported them slowed materially. You cannot award the same multiples to an economy growing at 9% as you would to one growing at 12%. That mismatch explains more about where foreign capital has gone.

The rebuttal: Nominal GDP recovery back toward 10-10.5% is plausible by next year, and at that level the valuation conversation changes. The government and the RBI both pivoted on policy in 2025, and the effect is already showing in corporate earnings (the reflation trade is on). The question is whether it arrives fast enough, and in large enough magnitude, to shift the attention of capital that is currently positioned elsewhere.

Underneath the nominal GDP debate sit structural facts that do not move with a single election cycle or a commodity shock:

1. India’s median population age is around 27 years. That demographic advantage compounds for the next 70 to 80 years and produces a growing domestic consumption market that no robot-driven economy can replicate (robots do not buy food or drive cars). China’s fertility rate is 0.9. Its domestic market is shrinking.

2. India has a capacity shortage across nearly every sector (infrastructure, energy, defense, fertilizers, basic manufacturing). China has the opposite problem because it overbuilt and must export excess capacity to the rest of the world to sustain its industrial base. India can invest internally for decades without exhausting the need. Investment creates earnings, and earnings drive stock markets over time.

3. Property rights: India’s saving to consumption ratio of roughly 20:80 reflects a population that feels secure in its ownership. People who are afraid their assets will be confiscated save compulsively and consume little. China’s excessive household savings rate is partly a response to exactly that insecurity. The US has gone the other direction, consuming future income it does not yet have. India sits in the correct position.

4. Respect for capital: Indian entrepreneurs do not deploy capital unless they can see a return on it of at least 15 percent. That discipline is what has made India one of the best-performing equity markets in the world in dollar terms over a 30-year horizon, in an economy that is a fraction of China’s size but has produced a far stronger stock market. Growth without capital discipline produces GDP. Growth with capital discipline produces wealth.

Section IV: AI Disruption, Delay, and the Application Layer

(Ridham Desai / Ritesh Jain)

The AI trade has cost India more in relative terms than any single macroeconomic event of the past year. The combined profit of SK Hynix and Samsung this year will be approximately 3x the total earnings of all Nifty 50 companies. That ratio captures the problem precisely: India has no memory chip industry and has no position in the hardware layer that is absorbing the bulk of AI capital expenditure right now. While Korea re-rated 50% in six weeks on the back of that trade, India moved 7-8%.

The near-term pain runs deeper than relative underperformance: The world’s blunt measure of AI adoption is headcount reduction. If a company fires people, the market concludes it is embedding AI into its processes. A significant share of the work being automated sits in India (lower-end coding, data processing, back-office functions that American firms had offshored). The instinct is to let those people go, which is already translating into hiring freezes and headcount declines across the IT services sector. Ridham Desai calls it a 4-6 quarter transition, and the market, having priced it in aggressively, has likely turned too pessimistic on the other side.

Ritesh Jain adds a data point that reframes the AI story: For the first time in 65 years, the cost of compute exceeds the cost of hiring a human being. Technology has historically been the cheaper input, which has inverted. The implications run in both directions, as it makes AI adoption more expensive than projected and creates a floor under human labor demand that most models have not accounted for. By 2027 and 2028, the application layer opportunity begins reflecting in earnings.

V: FIIs, Dollar Returns, and the Momentum Trap

(Ridham Desai / Kenneth Andrade)

FII selling in India over the past year has three distinct causes. The first cause was valuation, India was simply too expensive going into September 2024, and this has since has unwound. The second was a growth slowdown that made the valuation even harder to defend, which has also largely unwound with most companies now reporting numbers that reflect the policy pivot by both the government and the RBI. The third cause is the absence of an AI trade which has not gone away.

Let’s look at the mechanical dimension of FII selling. Domestic investors are putting roughly 40,000 crore rs into the market every month, and that capital has to find a seller on the other side. With corporate issuances not absorbing the demand, foreign institutions are functionally providing liquidity to domestic flows. It is partly a structural consequence of the volume of domestic buying.

The deeper issue is the composition of global institutional capital. Around 80% of it is momentum-driven. When momentum shifted away from India in September 2024 (tighter fiscal conditions, tighter monetary policy, China beginning to ease, the global AI trade gathering speed) that capital moved. It will return when the price action changes, or when something goes wrong in the markets currently absorbing that flow.

Kenneth Andrade says we are stuck in the middle. India has lost its momentum entirely. Its valuations, while meaningfully cheaper than a year ago, are not yet cheap enough to force the hand of deep-value capital. It is attractive without being compelling, which is the least useful combination for attracting flows.

The metric that argues most clearly for a floor is dollar returns. Indian equities have compounded at roughly 15% in dollar terms over the past five years and we are near the rolling low for that measure. Historically, that level marks the point from which markets recover from a dollar-return perspective, absent a further deterioration in the underlying. Kenneth’s positioning is to deploy half now and stagger the remainder through the first half of 2027. The worst is mostly in the price, and the upside requires patience and a catalyst that shifts relative momentum.

VI: The Rupee

(Ritesh Jain)

Every major conflict since 1990 has produced the same market response: capital flows into the dollar, US bond yields fall, and the rest of the world absorbs the volatility. The Iran war has produced the opposite. The dollar is receiving no meaningful bid, US thirty-year borrowing costs crossed 5% for the first time since 2007, UK yields are at five-year highs, German and Canadian yields are rising. The bond market is where this crisis is expressing itself in the West, and it’s the currency markets in emerging economies.

The rupee operates inside this new configuration under pressure. Ritesh Jain’s pre-war target was 92 against the dollar. The current range of 95 to 100 reflects the fertilizer import bill, sustained domestic gold buying, and the reality that oil prices have not been fully passed through to consumers. The question of how far the rupee falls is partly a question of how much gold Indians keep buying. If that continues, the depreciation pressure persists. If it slows, the exchange rate stabilizes around 95 to 96, which is also where Bernstein’s current target sits.

VII: Where to Invest

(Ritesh Jain / Kenneth Andrade)

The Nifty 50 benchmark is 50-55% services. When the Indian Prime Minister and the US President both say they want manufacturing, reshoring, and energy sovereignty, they are describing a world in which the benchmark is the wrong place to look. The investment thesis for India is not the index, but what the index does not yet contain.

Four themes have anchored Ritesh Jain’s India positioning for two years: electrification, defense and engineering services, tourism, and wealth management. Three of the four have delivered. Tourism was supply-constrained and did not deliver on its potential. The three that worked sit in sectors where the Indian government is actively deploying capital and where global supply chains are beginning to redirect work toward India.

The manufacturing angle is the medium-term frame. India’s manufacturing sector is currently 14% of GDP. Moving it to 16-17% over 5-7 years is achievable and would produce a set of winners that are not yet in any benchmark. Kenneth Andrade’s version of this is that the companies that win are not those in a particular sector but those that have built dominant domestic cash flow positions and are using that base to step out into global markets. A competitive cost structure and supply chain discipline matter more than sector designation. India’s depreciated currency makes the country one of the few large economies capable of exporting deflation to a world that is otherwise importing inflation.

Within that framework, the sector-level picture is differentiated. Automotive, defense, industrial engineering, pharmaceuticals, and solar all have structural tailwinds. India is already the second largest solar module producer globally, with significant capex committed for both domestic and export markets. Ceramics, commoditized steel, and agriculture sit further back in the formation but follow the same logic that India has capacity, the world needs an alternative source, and India’s share of global trade is still small enough that gaining share moves the needle domestically without displacing incumbents globally.

IT services remains the unsolved question. Kenneth Andrade has exited the sector pending clarity on how the AI transition redistributes work across the value chain. Valuations are cheaper than they have been in years, but cheaper is not the same as investable. The shape of the industry in five years looks genuinely different from what it was, and that uncertainty argues for patience over conviction. The revisit happens in six to twelve months when the transition pattern becomes clearer.

Two sectors warrant explicit caution. Consumer staples and discretionary are entering the same consolidation phase that capital goods endured between 2008 and 2020. Those companies will rework their models, invest their balance sheets, and eventually return. But the cycle suggests 5-8 years of grinding before the re-rating arrives. Financials face a different problem: in an inflationary environment, pricing power belongs to the companies manufacturing products, not the institutions financing them.

Something worth tracking is labor export. High-income countries and post-conflict reconstruction zones are running out of blue-collar workers, and India obviously has a surplus. The next five years will see a material increase in Indian tradespeople (plumbers, electricians, carpenters, drivers) moving to Japan, Germany, Canada, and eventually the economies that will need rebuilding when the current conflict ends. White-collar workers remain in India and blue-collar workers leave. The domestic scarcity that follows creates its own set of investment implications in wages, productivity, and the sectors that depend on that labor.

I would like to credit most of this work to my teachers Ritesh Jain, Ridham Desai, and Kenneth Andrade. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

Kross Limited is a Jamshedpur-based manufacturer of safety-critical automotive components. Kross manufactures load-bearing, suspension and driveline-related components for heavy trucks, trailers and tractors, ranging from trailer axle and suspension assemblies to axle shafts, flanges, gears and tractor parts. The company has been at this since 1991, supplies to OEMs including Tata Motors and Ashok Leyland, and listed on the NSE in September 2024 after a Rs 500 crore IPO.

The reason these products matter is baked into what they do. If a differential spider fails inside a truck’s axle, the vehicle cannot move. If an axle beam cracks under a loaded trailer, the trailer tips. These are safety-critical components, and that designation shapes every commercial relationship Kross has: OEMs take months to qualify new suppliers, run extensive testing before approvals, and are reluctant to switch even when cheaper alternatives exist.

There are two distinct businesses inside Kross:

Trailer Axles and Suspension (roughly 43% of revenue)

A trailer is the load-carrying unit that attaches behind a truck. Every trailer needs axle assemblies, which carry the trailer load and hold the wheel ends, and suspension assemblies, which absorb road shocks and stabilize the load. Kross designs and assembles these products with substantial in-house manufacturing of critical parts such as forgings, castings, machining and surface treatment.

This segment is relatively new. Kross entered trailer axles only in 2019, and has grown it quickly enough that it now represents nearly half the business. The organized domestic market has relatively few players with in-house trailer axle and suspension manufacturing capability; Kross’s edge is that it offers an integrated domestic alternative in a market historically served by a mix of domestic manufacturers, imports and assemblers.

In February 2026, Kross commissioned India’s first axle beam extrusion plant, producing seamless single-piece axle beams. The conventional Indian axle beam is welded together from four pieces of steel. The weld points are the weakest part of the structure. Kross’s seamless beam is extruded as a single piece, which removes those structural weak points and produces an axle that is stronger, lighter, and more reliable. According to the company, this gives Kross a domestic first-mover advantage in single-piece extruded axle beams.

Forged and Machined Components (roughly 57% of revenue)

This is the older and more established business. Kross makes a wide range of safety-critical forged parts for trucks and tractors: axle shafts, companion flanges, differential spiders, bevel gears, planet carriers, anti-roll bars, suspension linkages, and tractor-specific components like PTO shafts and hydraulic lift parts.

Kross is substantially backward integrated across key processes (die design, forging, casting, heat treatment, machining, testing and surface protection) which helps control quality and cost on safety-critical parts.

Tractor components currently represent around 9% of revenue. The company supplies tractor components to large OEM relationships including TAFE/Eicher and International Tractors/Sonalika, and has added a third large tractor OEM relationship. The tractor segment behaves differently from M&HCV: it is driven by monsoon cycles and rural income rather than freight rates, which gives Kross a partial cushion when the commercial vehicle industry goes through a down year.

FY26 revenue was Rs 673 crore, EBITDA was Rs 87.9 crore at a 13.1% margin, and PAT was Rs 55.2 crore. The full year was modest on growth at 8.5%, largely because the M&HCV industry went through a soft patch in H1 FY26. The Q4 exit was much stronger: revenue up 22% year-on-year and EBITDA margin expanding to 14.9%.

I’ve analyzed Kross using the TVG framework:

Theme

India’s commercial vehicle industry runs in long cycles. It peaks, overshoots, contracts, and then recovers. The current cycle matters for Kross because the industry went through a multi-year trough after FY19 and is now in the early stages of a genuine upcycle.

The structural driver behind this recovery is fleet age. The average truck on Indian roads is roughly 10 years old, and that is a fleet that has been sweating its assets and deferring replacement purchases through a period of uncertain freight economics. As freight rates have recovered and GST rate cuts have lowered the cost of buying new vehicles, fleet operators are beginning to refresh. Replacement demand of this kind tends to be durable rather than episodic, because the aging cohort does not retire all at once. Nomura estimates M&HCV industry volumes grew around 8% in FY26 and could grow a further 10% in FY27. Crisil is more conservative at 4-5% for FY27, but both point in the same direction.

For a component supplier like Kross, every new truck that leaves an OEM factory generates demand for axle shafts, flanges, gears, and suspension parts. The upcycle is a volume tailwind that flows directly into order throughput.

The more specific and more interesting tailwind, though, is in trailers.

A trailer is the unpowered unit that attaches to a truck. India’s logistics network runs on tractor-trailer combinations for long-haul freight, and this segment has been gaining market share. Trailer combinations accounted for roughly 9% of the M&HCV mix in FY21. By FY25, that share had grown to approximately 22%. However, this also means the tractor-trailer mix may not keep expanding at the same pace; some normalization is possible if rail or dedicated freight corridors capture parts of bulk freight. Better highways, faster turnaround times and route flexibility have supported tractor-trailer adoption in industrial cargo, although the exact share shift from rail should not be overstated. That shift in freight mix directly expands the addressable market for trailer axles.

India’s ongoing infrastructure build-out reinforces this. More expressways, port connectivity projects, and the development of logistics parks and multimodal hubs all raise freight throughput and put more trailers on the road. Higher utilization also accelerates the replacement cycle for axle and suspension components, which are wear items on heavy-duty trailers.

One concern worth addressing directly is the dedicated freight corridor. Investors occasionally raise this as a structural risk to trailer demand, and the question came up on the most recent concall. Management’s response was that road economics remain far better than rail for most freight categories. Management illustrated this with a Jamshedpur–Chennai example, saying road dispatch costs around Rs 3 per kg versus over Rs 18 per kg by train, after related costs. This should be treated as management’s route-specific example, not a universal rail-versus-road cost ratio. The broader point management made is that trailers carry an extremely wide range of goods, from coal and iron ore to cars, FMCG goods, and road-building materials, and that even if the freight corridor takes some volume, it is more likely to affect rigid body vehicles carrying smaller goods than the tractor-trailer segment specifically.

The tractor segment adds a second, steadier dimension to the theme. India’s tractor market is growing from a base of around USD 9.4 billion in FY25 toward an estimated USD 16.8 billion by 2034, driven by agricultural mechanization, rural credit access, and an expanding role for tractors in construction and infrastructure work. Kross supplies safety-critical components to several of India’s major tractor manufacturers. The tractor cycle is driven by monsoon and rural income rather than freight rates, which means it does not move in lockstep with M&HCV. That partial independence provides a buffer when the commercial vehicle industry softens.

The final layer of the theme is exports. India is becoming a credible alternative manufacturing base for auto components as global OEMs diversify supply chains away from China. Indian forging and precision machining capabilities have matured over the past decade, and cost competitiveness remains strong. Kross has export business with Sweden-based Leax Falun AB across multiple product segments, has developed and submitted samples to a Japanese OEM, and has additionally secured orders from a European Tier-1 customer across two product families. Exports contributed 4% of FY26 revenue. The absolute number is still small, but the customer wins are specific and the direction is clear.

Value

Auto component manufacturers often look cheap on the surface and expensive once you understand the working capital cycle. The real test is whether the business generates cash, keeps its balance sheet clean, and earns a return on the capital it deploys. Kross’s picture on these dimensions is mixed, and worth going through carefully.

Start with the balance sheet, because the IPO changed it significantly. Before listing in September 2024, Kross carried debt at a debt-to-equity ratio of 0.8x. The IPO proceeds cleared that overhang, and by end of FY25 the company had Rs 82.8 crore of cash on the books against minimal debt. By end of FY26, the picture has shifted again. Cash has come down to Rs 4.4 crore as 100% of IPO proceeds have now been deployed across the stated objects, including capex, debt repayment and working capital. Long-term borrowings have risen from Rs 5.3 crore to Rs 29.1 crore, reflecting the partial debt funding of the seamless tube facility. The company remains lightly leveraged in absolute terms, but it is no longer sitting on a cash surplus. The meaningful improvement is that finance costs fell from Rs 12.3 crore in FY25 to Rs 8.1 crore in FY26, a 34% reduction from post-IPO debt repayment, and that flows directly to the bottom line.

Return ratios for FY26 come in at 16.4% ROCE and 12% ROE. These are respectable for the segment but not outstanding. The context is that FY26 was a heavy deployment year. The Rs 25 crore axle beam extrusion plant and the Rs 167 crore seamless tube facility are both being commissioned without contributing revenue yet. The more relevant question is what ROCE looks like once those assets are running at reasonable utilization, which begins to answer itself from FY27 onward as the axle beam plant ramps and the tube facility moves toward commissioning.

EBITDA margin has improved from 9.9% in FY22 to around 13% in FY24–FY26, with Q4 FY26 reaching 14.9%. Management has guided for 14 to 15% as the range to expect in the quarters ahead. The structural reason these margins hold is backward integration across key processes (die design, forging, casting, heat treatment, machining, testing and surface protection). This means Kross captures value across the production chain rather than ceding upstream economics to a sub-supplier. One number that does not get enough attention: gross margins expanded from 43% in FY25 to 45.7% in FY26, a 276 basis point improvement. That suggests better mix, procurement and integration economics, although part of the movement can still be influenced by steel-price cycles and product mix.

In OEM components, the moat is largely supplier qualification, process capability and switching cost. Once Kross is qualified as a supplier for a specific safety-critical component at an OEM like Tata Motors or Ashok Leyland, displacing it requires the OEM to run a new supplier through an expensive and time-consuming qualification process, validate dimensional tolerances, conduct fatigue and failure testing, and get sign-off from its engineering teams. OEMs avoid this unless they have strong financial motivation to switch. That stickiness is real and has supported Kross’s long-standing relationships with its major customers. The first product Kross sold to Tata Motors was in 1997, Ashok Leyland followed in 2006. These relationships are sticky, but pricing and volumes are still commercially negotiated, especially when raw-material costs move.

In trailer axles, the moat is more about cost, product reliability, field service, dealer and fabricator reach and brand acceptance. Extrusion gives Kross a first-mover product and cost advantage, but the moat is execution-led rather than IP-led: cost, quality, utilization, service network and customer acceptance will decide the outcome. A seamless single-piece axle beam is structurally superior to the welded four-piece construction that has been the domestic standard: it is stronger at the weld points, lighter, and more dimensionally consistent. According to the company, this is India’s first axle-beam extrusion plant, giving Kross a domestic first-mover advantage. But replicating this process takes capital and time, not a patent licence.

On segment margins, the concall provided a clear hierarchy worth understanding. Exports carry the best margins of any product line. The axle and suspension business and tipping jacks sit below that. The components business is next. And the tractor segment is, by management’s own description, the most price-competitive of all. This matters because the tractor wallet share target of 15% is a volume and customer diversification story, not a margin expansion story. The margin drivers are the extrusion line reaching utilization, more exports, and eventually the seamless tube facility reducing input costs.

Now the yellow flag. Working capital days have risen from 87 to 125 days over the past year. For a company of Kross’s scale, a 38-day deterioration in working capital represents a meaningful cash drag. A Rs 105 crore working capital build in FY25 was large enough to turn operating cash flow negative despite healthy reported profits. The company was earning on paper but not collecting fast enough.

The second risk worth naming is raw-material pass-through. The two parts of Kross's business handle this differently, and the difference matters. In OEM components, price increases from steel or consumable cost moves are passed through retrospectively, typically with a one-quarter lag. The settlement happens, but it is delayed. In the trailer axle and suspension business, there is no retrospective mechanism. Kross has to take price hikes in the market immediately, and how far it can go is constrained by the market leader. Management said on the concall that Kross is the number two player in trailer axles and cannot price materially above the number one. In Q4 FY26, the company took roughly a 5% price hike in trailer axles, partly to cover LPG and consumable cost increases from the Middle East conflict. Whether that was sufficient will show up in Q1 FY27 margins. The broader point is that gross margin can be volatile even in quarters where volume is strong, and the extrusion thesis should be underwritten as a cost-saving and market-share story, not a pricing-premium story.

The likely explanation is a combination of factors. The trailer channel, which now represents a large part of the business, collects through dealers and fabricators rather than directly from OEMs. Those customers pay more slowly than large OEMs on formal purchase orders. On top of that, the M&HCV slowdown in H1 FY26 would have caused inventory to build as production plans held but actual offtake from OEMs softened. The Q4 FY26 recovery is encouraging. But working capital is a variable that needs to be tracked over the next two or three quarters before treating the deterioration as fully resolved.

Growth

The growth story for Kross has two parts. The first is recovery: getting back to the run rate the business should be at after a difficult H1 FY26. The second is expansion: new capacity, new products, and new geographies that add revenue on top of the core business.

On recovery, the trajectory is already visible. The M&HCV industry went through a soft patch in the first half of FY26, driven by the GST transition and a temporary demand pause from fleet operators, and Kross felt that directly. Revenue in Q2 FY26 declined 6% year-on-year. But Q3 recovered, and Q4 FY26 came in at Rs 225 crore, up 22% year-on-year with EBITDA margins expanding to 14.9%. The exit rate matters more than the full-year average, and the exit rate is strong. With the M&HCV industry now in an upcycle supported by replacement demand and improving fleet economics, the base business should grow steadily through FY27 without needing any of the new capacity to contribute.

The expansion levers are where the more interesting growth comes from.

Seamless Axle Beams

The axle beam extrusion plant was commissioned in February 2026 and began commercial sales in May 2026. The plant brings total axle beam capacity from 5,000 to 7,500 units per month. In strong months the old capacity could become a constraint, and management has indicated it is currently ramping to meet demand and the new line provides the headroom needed to grow volume from here.

The extruded single-piece axle beam is a superior technical product, but near-term economics are more likely to come from cost savings, utilization and market-share gain than premium pricing. Management has guided that EBITDA margins on the extrusion line are expected to improve significantly above 50% utilization, which sets a clear internal milestone to track. Kross appears to have a domestic first-mover position in extruded single-piece axle beams: if customer acceptance is strong, this can become a share-gain lever before competitors replicate or import equivalent capability.

The most underappreciated optionality from the extrusion plant is TAG axles. Trailer axles are sold primarily through fabricators, but TAG axles are fitted directly onto trucks at the OEM level, which means a different, more sticky customer relationship and a longer validation cycle. Kross's own strategy slide notes that the axle-beam extrusion capability may enable entry into TAG axles. This is a FY27-FY28 call option rather than an immediate revenue driver, but it represents a meaningful expansion of addressable market if the extrusion line demonstrates product quality at scale.

Seamless Tube Facility

This is the largest and longest-dated investment on the growth roadmap. Kross is putting Rs 167 crore into a facility that will produce seamless steel tubes in-house. Currently, steel tubes are sourced externally and processed into axle beams. The seamless tube facility can capture some upstream economics, but its bigger strategic purpose is supply security, reduced import dependence and input control for axle beams, tipping jack cylinders and other tube-based components.

Revenue contributions are not expected until FY28. The Rs 167 crore project will depress asset turns and ROCE during the commissioning period; the payoff depends on timely FY28 ramp-up, captive consumption and external tube spreads.

Tipping Jacks

In December 2025, Kross launched tipping jacks for trailers. A tipping jack is the hydraulic mechanism on a tipper trailer that lifts the cargo body to discharge its load. The product is a natural adjacency to the axle business: same customer base, same sales channel, same trailer ecosystem. Kross sold 75 to 80 kits in Q4 FY26, is targeting 150 to 200 in Q1 FY27, and expects to reach 300 units per month by end of Q1 and 500 by Q3 FY27. Total installed capacity is 800 kits per month, and the total OEM market for tipping jacks is estimated at roughly 7,000 units per month. Kross is starting from zero share in a market it did not previously address, and management has guided for margins of approximately 15% on this product line.

Tractor Segment Share

Kross currently derives around 9% of revenue from tractor components and is targeting 15% over the next two years. Management was specific about how this gets done: two routes rather than one. The first is a new OEM relationship (the third large tractor customer Kross has added) where component development is now complete and revenue will begin from Q1 FY27. The second is product range expansion with the two existing OEM relationships of 15 to 20 years each, moving deeper into forgings, shaft components, and castings beyond the current product set. The 15% target is a two-year ambition. One honest note from management: the tractor segment is the most price-competitive of all the product lines Kross sells. Growing its share improves customer and segment diversification, but it will not expand group margins.

Exports

Export revenue was around 4% of FY26 total, and the company is targeting 8% over the next two years. Kross has export business with Sweden-based Leax Falun AB across multiple product segments, has developed and submitted samples to a Japanese OEM, and has additionally secured orders from a European Tier-1 customer across two product families. These are specific customers with specific programs. The ramp from 4% to 8% is achievable if these execute on schedule, though export revenue has a longer gestation from qualification to volume production than domestic OEM business.

High-Pressure Moulding Line

The high-pressure moulding line, targeted for September 2026, should double casting capacity and support trailer axle volumes. Margin improvement will depend on utilization, rejection rates and captive use.

Pulling these together: the near-term growth is driven by M&HCV recovery and the initial ramp of seamless axle beam sales. The medium-term growth is driven by the tipping jack ramp, tractor wallet share expansion, and the export push. The longer-dated growth, beginning in FY28, comes from the seamless tube facility reaching production and contributing both input cost savings and a new revenue stream.

Valuation

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

Here are three companies I think are worth looking at:

ARM Holdings ($ARM)

Business Overview

ARM does not actually make chips. It designs the blueprint for how chips should work, and then licenses that blueprint to companies who actually manufacture them. The blueprint is called an instruction set architecture, or ISA. Think of it as the language a chip speaks: the set of rules that defines how software communicates with hardware. Every company that wants to build a chip using ARM’s ISA pays ARM for the privilege.

The entire business is intellectual property. No factories, no silicon, no physical product.

Revenue has two streams:

1. Licensing fees are paid upfront. A company like Apple or Qualcomm pays ARM to access the architecture and then customizes it for their own chip design. The more sophisticated the license (e.g., access to Compute Subsystems rather than just the base ISA), the higher the fee.

2. Royalties are ongoing. Every time a chip built on ARM’s architecture ships inside a product, ARM collects a small cut per unit. Multiply that across billions of devices annually and the royalty stream becomes the dominant revenue engine.

How does it compare to x86 architecture (Intel & AMD)?

Intel’s x86 architecture trades raw single-thread performance for a heavy legacy compatibility burden. Every new Intel chip must be backward-compatible with decades of software written for x86, which constrains design choices and drives up power consumption. Intel is losing data center share and has been consistently outcompeted by AMD on server CPUs.

AMD improved the x86 playbook significantly through chiplet design: splitting one large chip into multiple smaller dies, which reduces cost and improves manufacturing yields. AMD EPYC is currently the preferred host CPU in AI servers, offering more memory bandwidth and PCIe lanes than Intel for GPU-heavy workloads. But AMD still carries the same x86 legacy constraints as Intel at the architectural level.

ARM competes on a different axis: efficiency over raw speed, and licensing flexibility that allows customers to customize deeply. Apple’s silicon performance is the clearest proof point, as its ARM-based M-series chips outperform x86 equivalents in performance-per-watt by a significant margin. The licensing model also means ARM has no manufacturing risk, no yield pressure, and no capacity constraints.

Theme

The core theme is that ARM sits at the foundation of almost every meaningful computing trend happening right now, and increasingly in AI specifically.

The AI capex buildout is primarily a GPU story at the training layer, and Nvidia dominates there. But there are two places where ARM’s architecture is increasingly central to AI infrastructure.

The first is custom silicon. Every major hyperscaler is now designing its own AI chips rather than buying off-the-shelf. Google’s TPU8s replace x86 with ARM. Amazon’s Graviton and Trainium are ARM-based. Microsoft’s Cobalt is ARM-based. Apple’s M-series chips are ARM-based. NVIDIA is using ARM CPUs for AGI workloads. Meta is working with ARM for AI data center infrastructure. ARM does not compete with any of these companies as it collects a royalty on all of them. As custom AI chip design proliferates across the industry, ARM’s royalty base widens automatically.

The second is the edge. AI is currently centralized (models are trained and run in large data centers). But inference is moving toward local devices: phones, laptops, cars, industrial sensors. Every single one of those devices uses ARM chips. When AI moves from centralized training to distributed inference on actual devices, ARM’s installed base of ~250 billion chips becomes the delivery infrastructure.

There is also a longer-dated optionality layer: AGI. Current AI is narrow, it is very good at specific, bounded tasks, all of which are GPU-dominated workloads. AGI would involve more diverse and unpredictable tasks, longer context, and more complex real-world interactions. That type of computing depends more heavily on CPUs than GPUs. ARM would be a direct beneficiary. This is speculative and uncertain, but it is worth noting as upside optionality. The Q4 commentary already flagged $2bn in new revenue visibility for FY27-28 just for AGI CPUs, which suggests this is beginning to show up in actual demand signals.

ARM’s Moat: The 3 lock-ins

1. Switching costs: Around 80% of engineers globally already develop software using ARM’s platform. Retraining an entire engineering base on a different architecture is expensive and slow.

2. Compute Subsystems: ARM has moved beyond licensing the base ISA toward offering full system blueprints and moved to complete reference designs for chip subsystems that customers can adopt directly or customize. This reduces development time significantly for chip designers and increases the value ARM adds per engagement. It also makes the relationship stickier, because customers are now co-developing on ARM’s platform rather than just licensing a specification.

3. Product fit: Power consumption has become one of the central constraints in AI data center design. ARM’s architecture was built around efficiency over raw speed, which means its natural product positioning aligns with where the industry’s pain point is right now.

Growth Triggers

The clearest near-term growth lever is the ARMv9 transition. ARMv9 carries approximately 2x the royalty rate of ARMv8, and sost devices in the field are still running v8 chips. As the installed base naturally refreshes and new chips roll out on v9, ARM collects higher royalties on the same volume without needing to win a single new customer.

The second lever is Compute Subsystems. As more customers adopt full system blueprints rather than just the base license, the revenue per chip design engagement rises. This is what drove recent data points: Armv9 and Compute Subsystem growth together enabled higher fees per chip.

The third lever is data center penetration. ARM-based server chips (Graviton, Cobalt, and others) are taking meaningful share from x86 in cloud infrastructure. Google replacing x86 with ARM in its TPU8 line is a significant signal. Meta exploring ARM for AI data center workloads is another.

Celestica ($CLS)

Business Overview

Celestica builds the physical hardware that makes AI data centers work. It designs, manufactures, and assembles the networking switches, servers, storage systems, and rack infrastructure that hyperscalers need to run large-scale AI workloads. It does this on behalf of companies like Google, Microsoft, Amazon, and Meta, who find it more efficient to outsource complex hardware manufacturing than to do it themselves.

The business has two segments:

1. Communications and Cloud Solutions (CCS) is now 80% of revenue. This is where all the AI-driven growth is happening. Within CCS sits the Hardware Platform Solutions (HPS) division, which is the strategically critical part. HPS is where Celestica goes beyond simple assembly and actually co-designs the hardware platforms with its customers. This is the business Celestica is building toward.

2. Advanced Technology Solutions (ATS) is the remaining 20%. This covers aerospace, defense, medical devices, and industrial equipment. Slower-growth markets but higher quality and stickier than consumer electronics. ATS is currently flat year on year and undergoing a deliberate reshaping toward higher-margin lines.

Theme

The structural tailwind is straightforward: hyperscalers are spending at a scale with no historical precedent on AI infrastructure. Industry capex from the top five (Microsoft, Amazon, Google, Meta, and Oracle) is estimated to exceed $600bn in 2026, up 36% YoY. Every dollar of that spending needs physical hardware: networking switches to move data between chips, servers to house the compute, storage systems, and the increasingly complex rack infrastructure that integrates all of it together.

Celestica sits in the middle of this supply chain, since they are the backbone of how AI data centers function at scale. And critically, the company has secured its position through binding, non-cancellable, non-returnable (NCNR) contracts with hyperscalers extending into 2028. This contract structure transfers demand risk from Celestica to its customers. Celestica uses this visibility to lock in component supply agreements years in advance, which also explains why lead times on some materials have now extended beyond twelve months.

The deeper theme is a technology transition in networking. Data centers are moving from 800G Ethernet switches to 1.6T (1.6 terabits per second) switches, and then eventually to 3.2T. Celestica holds ten 1.6T programs. Its DS6000 series 1.6T switches were announced as available in late April 2026, with mass production expected to ramp in the second half of FY26 across two hyperscaler customers. This is a direct revenue catalyst in the near term.

Beyond that sits co-packaged optics (CPO). Traditional switches use pluggable optical modules that connect separately. CPO integrates the optics directly into the switch chip package, which dramatically reduces power consumption and latency, both of which are critical constraints as AI data centers scale. Celestica has secured a contract to mass-produce the first major commercial-scale 1.6T CPO Ethernet switch, using Broadcom’s Tomahawk 6 Davisson module, with mass production begining 2027. There are very few companies in the world capable of doing this at scale.

Value

The central value argument is that the market is still pricing Celestica as an electronics manufacturing services (EMS) company when it is increasingly behaving like an original design manufacturer (ODM).

An EMS company assembles hardware to a customer’s specification. It owns no design IP, competes primarily on cost and reliability, and earns thin margins. Peers like Flex, Jabil, and Sanmina operate at non-GAAP adjusted operating margins of roughly 5-6.5%. An ODM co-designs the product, owns the IP, and earns a premium on its engineering contribution. Celestica’s Q1 FY26 adjusted operating margin was 8%, a record, and management is targeting 10%.

The mechanism driving that margin expansion is HPS. As HPS grows as a share of total revenue, it pulls the blended margin higher.

The moat here is the technical challenge of CPO manufacturing, solving thermal management and signal integrity at the package level is a genuine barrier. Once Celestica proves 1.6T CPO with a hyperscaler customer, the path to 3.2T CPO becomes the logical next contract award with that same customer, because switching ODM suppliers mid-generation is too risky for the hyperscaler.

Growth Triggers

1. Current product cycle: 800G switches are already generating strong revenue. The transition to 1.6T switches through the DS6000 series begins contributing in H2 FY26, with two hyperscaler programs going into mass production. This is already in the order book and has direct visibility.

2. Program pipeline ramping into FY27: The 1.6T CPO contract goes into mass production in 2027. The AMD Helios rack-scale AI platform (where Celestica co-develops a complete AI compute rack, integrating GPUs, networking, and cooling as a single system) begins ramping in FY27. The Google TPU co-design program scales through 2026 and into 2027. Separately, machine intelligence compute programs with another hyperscaler also begin ramping in FY27.

3. Margin shift: As HPS mix rises from 42% toward 50% of revenue, and as ATS returns to mid-single-digit growth with margins holding at 6%, operating leverage compounds.

Cipher Digital ($CIFR)

This company used to be a bitcoin miner who realized their core asset (large amounts of cheap, grid-connected power) was worth far more to AI hyperscalers than to a mining operation. It is now in the middle of that transition.

Cipher Digital is a landlord. It builds data center infrastructure and rents it to hyperscalers on 10 to 15-year leases. Its tenants are AWS and Google and the revenue model looks closer to infrastructure real estate than technology.

Business Overview

Cipher Digital is best understood as a specialized real estate developer for AI infrastructure. It builds large-scale data center campuses on power-advantaged land in Texas, then leases them to hyperscalers on 10 to 15-year contracts at near-100% NOI margins. The tenants pay the rent; the project-level debt amortizes itself from the lease cash flows; the holding company collects the residual.

The company currently has three energized sites with 807MW of total capacity. 600MW is allocated to HPC under long-term leases. The remaining 207MW at Odessa continues operating as bitcoin mining under a fixed-price PPA through July 2027, serving as a short-term cash source while the transition completes.

The two flagship contracts:

Black Pearl (Texas, 300MW) is under a 15-year lease with AWS at a $5.5bn total contract value. Amazon provides a full corporate guarantee on all base rent and operating expenses, producing a project-level NOI margin approaching 100%. Rent commencement is targeted for October 2026.

Barber Lake (Texas, 300MW) is under a 10-year lease with Fluidstack, with Google providing a financial backstop for up to $1.73bn of Fluidstack’s obligations. This is effectively an investment-grade Google guarantee. NOI margin is 86%. Energization is also targeted for October 2026. Financed by $1.73bn in non-recourse bonds at 7.125%, again amortizing from lease cash flows.

Combined, these two sites generate an average of $669mn in annual NOI over the base lease term, rising to $754mn by 2035.

A third contract was announced in March 2026 with an undisclosed investment-grade hyperscaler at the Ulysses, Ohio site (200MW). This is Cipher’s first development outside Texas and signals the pipeline is actively converting.

Theme

The core theme is a concentrated angle around power scarcity. West Texas is a structurally advantaged market for AI infrastructure: ERCOT electricity rates between $0.027 and $0.031 per kWh are among the lowest in the United States. Extending a high-capacity transmission line in this geography takes years of permitting and bureaucracy, and Cipher has already done that work across multiple sites.

The hyperscaler demand dynamic strongly favors Cipher’s negotiating position. Both AWS and Google committed to decade-plus contracts before Cipher had built the data centers. That sequencing inverts the normal infrastructure development risk. Management has explicitly flagged that lease pricing is moving in their favor as the supply-demand gap widens, suggesting future contracts in the 3.4GW pipeline may be struck at better economics than the existing ones.

Value

The value case for Cipher is anchored in a simple NAV framework rather than earnings multiples, because the current income statement (still predominantly bitcoin mining revenue declining toward zero) tells you almost nothing about the business it is becoming.

At $669mn of average annual NOI from the two contracted sites, applying a 16x infrastructure multiple (third-party research) implies an enterprise value of roughly $10.7bn for just those two assets. Subtract $3.73bn in project-level debt (non-recourse), add back cash and bitcoin holdings, and the implied equity value materially exceeds the current market cap. And this is before attributing any value to the 3.4GW development pipeline.

The structural cost advantage deserves specific attention. Cipher’s Black Pearl facility operated at 13.9 joules per terahash efficiency during its mining phase, better than IREN’s 15 J/TH, which is considered industry best practice. This operational efficiency is the proof point that Cipher’s engineering team can meet hyperscaler standards. AWS would not have committed $5.5bn over 15 years to an operator it did not believe could deliver.

The non-recourse debt structure is an important value point. The $3.73bn in bonds is underwritten against AWS and Google credit quality, not Cipher’s balance sheet. If the projects fail, the bondholders look to the hyperscaler tenants, not the holding company. This structure significantly limits downside risk at the holding company level.

Growth Triggers

Growth in the near term is simply the two flagship sites completing construction and commencing rent collection. October 2026 is the target for both Black Pearl and Barber Lake. Once rent begins flowing, the financial profile transforms to a cash-generating infrastructure business with locked-in revenue through 2036 and beyond.

The medium-term growth is the 3.4GW pipeline converting site by site into contracted, energized assets. The pipeline breaks into three layers. The near-term layer includes Stingray (100MW, Q4 2026), Reveille (70MW), and Ulysses (200MW, Q4 2027 in Ohio). The 2028 layer includes McLennan (500MW), Colchis (1GW with a direct AEP grid connection already negotiated), and Mikeska (500MW). Additionally, Barber Lake extension (500MW) and Milsing (500MW) near Houston round out the longer-dated optionality.

Management’s comment on future lease pricing is worth taking seriously: each successive site contract is likely to be struck at better economics than the current ones, as hyperscaler demand continues to outpace available power-ready supply.

Disclaimer: This is not investment advice. I am not a registered analyst. Please do your own research before making any investment decisions.

The closure of the Strait of Hormuz was always the risk scenario that analysts modeled and policymakers quietly feared. It has now materialized, and the response from Washington has revealed more about the limits of American power than it has about its reach.

The initial U.S. position was straightforward: bomb Iran, force the Strait open, and demonstrate that no regional actor could hold global energy flows hostage. That logic collapsed on contact with reality. Tactically, Iran absorbed significant damage. Strategically, it was winning. The bar for Iran was never symmetrical. They did not need to defeat the United States military, they only needed to survive and keep Hormuz closed.

What followed was a sequence of improvisation dressed as strategy. Treasury Secretary Bessent, on March 22nd, stated that 50 days of pressure would buy 50 years of stability. 23 days later, with nothing resolved, the timeline was quietly extended.

The blockade of the blockade followed. After Iran began charging vessels transit fees to pass through Hormuz, the U.S. Navy moved to interdict that arrangement and the move was presented as a show of strength. However, a more careful reading suggests it was optics management. What credible sources on the ground were reporting was that Iran had already won the strategic exchange, and the blockade of the blockade was an attempt to obscure that outcome rather than reverse it. A sanctioned two-million-barrel VLCC passed through with its transponder on and was not touched. If the blockade were what it claimed to be, that tanker does not pass unmolested.

The real forcing function was rare earths. Interceptor missiles require Chinese rare earth inputs, and the U.S. and Israel were burning through interceptors faster than Iran was expending offensive capability. When Bessent remarked that China was not being a good partner in wartime and was choking off exports of certain goods, the diplomatic translation was not difficult. The ceasefire and the opening of talks did not follow from military superiority, but from a supply chain constraint that Washington could not publicly acknowledge without admitting the deeper vulnerability.

II: Supply Chain Math

The most useful mental model for what a prolonged Hormuz closure does to the global economy is COVID.

Travel and tourism account for roughly 10% of global GDP. A 30-40% contraction in that sector alone translates to a 3-4% hit to global output, which puts us at borderline flat global growth before any second-order effects are counted. Then that shock hits a levered system, so private credit tightens, loan losses rise, sovereign receipts fall as activity contracts.

Let’s take a look at oil replacement. The proposition is that the United States, as a major producer, can compensate for Hormuz disruption by redirecting domestic supply. But the numbers do not support it: the U.S. is a net importer of crude. It is a net exporter only when refined petroleum products are included in the accounting, and some of that export-import activity involves Canadian bitumen processing that inflates the gross figures. The idea that American barrels can be redirected at scale to replace what flows through Hormuz is not exactly an energy policy.

There is also the question of reversibility. Even in a scenario where a ceasefire holds and hostilities formally end, the assumption that supply chains snap back to their pre-conflict state is wrong. Physical infrastructure has been damaged. Refineries, processing facilities, and energy nodes that were struck do not restart on a timeline measured in days. The system that existed on February 27th is not the system that will exist when the guns go quiet.

What is striking in this environment is actually the absence of volatility, as markets have remained remarkably composed, and that composure is not entirely organic. Policymakers understand that if the bond market begins to price the actual risk embedded in this situation, the United States loses negotiating leverage and may be forced toward a settlement on terms it would prefer not to accept. Calm markets are a strategic asset right now, meaning the complacency being observed is partly manufactured.

The supply chain pressure is accumulating. Taiwan imports 97% of its energy and draws roughly 40% of its grid power from Middle Eastern natural gas, with TSMC holding only a few weeks worth of LNG on site. 30% of the world’s helium supply comes from Qatar, and helium is what cools the magnets in the machines that manufacture every advanced chip on earth. When that supply stops, so does chip production. Chip fabs are still running because they were built to absorb brief interruptions, but stockpiles that take months to rebuild can be exhausted in days.

III: China has America by the Short Hairs

Understanding why China holds the leverage requires tracing a sequence of decisions that stretch back decades.

The Chimerica cycle, as it came to be known, was simple: China entered the WTO, absorbed American manufacturing, produced goods at scale, exported them westward, received dollars in return, and recycled those dollars into U.S. capital markets. Financial conditions in America loosened, consumption rose, more goods were imported from China, and the cycle repeated. For roughly two decades, this arrangement produced the appearance of mutual benefit while quietly hollowing out the industrial base on one side and building it up on the other.

The inflection point was 2008. When the American financial system imploded, the response was entirely consistent with how Washington had always prescribed crisis management to others, except that the prescription was reversed. For decades, the IMF recipe applied to Southeast Asia, Latin America, and Russia involved austerity, currency devaluation, political restructuring, and the dismantling of overleveraged institutions. When the crisis arrived in America, the response was quantitative easing, institutional preservation, and the effective devaluation of the bonds that China, Russia, and others were holding.

By late 2013, China had quietly stopped accumulating U.S. Treasuries. What they did instead was a reallocation of surplus dollars into physical assets: copper mines, nickel deposits, port infrastructure, and strategic footholds across Africa, Latin America, and Southeast Asia. The Port of Piraeus, investments in the Panama Canal, resource extraction deals structured as loans with asset collateral. China had looked at what the Americans considered real value and decided to own it.

Made in China 2025, announced in 2015, was the industrial complement to this financial repositioning. The goal was dominance in the industries that would define the next half-century, and the execution over the following decade was serious enough that by 2021, the same Western commentators who had dismissed it as overreach were acknowledging that China had largely delivered. The remaining gaps, advanced semiconductors and a handful of other areas, are meaningful but narrowing, and they exist against a backdrop of near-total Chinese dominance in the material inputs that every advanced technology requires regardless of where it is assembled.

That dominance is the source of the leverage that now constrains American options at every turn. Interceptor missiles require rare earth inputs that China controls. Pharmaceutical supply chains run through Chinese precursor chemistry to a degree that most Western populations have never been asked to confront directly. The engineering depth differential alone, illustrated by the comparison between Tesla’s total global headcount and BYD’s engineering workforce, captures something about the scale mismatch that cannot be closed in a budget cycle or an election cycle.

IV: The Dollar’s Quiet Demotion

The process began earlier than most accounts acknowledge. When Iran was removed from the SWIFT messaging system in 2012, the immediate effect was the intended one: Iran was isolated, transactions became difficult, and pressure mounted. If the U.S. could hyperinflate another country’s economy overnight by flipping a switch on the messaging layer of global finance, then the only rational posture was to build a second set of rails before that switch was pointed at you. By 2015, China had CIPS operational, an integrated platform combining messaging, clearing, and settlement. Volumes through that system have since grown and the network has continued to deepen.

Alongside CIPS, China established offshore yuan clearing banks in every major gold hub in the world (London, Zurich, Dubai, Singapore, Hong Kong) effectively building the settlement infrastructure for a parallel monetary system before most Western observers had registered that one was being constructed. The e-yuan was tested first in oil markets and gold markets. Those are the two markets that matter most if the objective is to create a credible alternative to dollar settlement for commodity flows, and the sequencing revealed exactly what the priority was.

Gold has already supplanted U.S. Treasuries as the largest reserve asset held by global central banks on an adjusted basis, a development that received brief coverage and was then largely set aside by mainstream financial commentary. At $6000 per ounce, gold exceeds Treasuries even on a gross basis. Central banks do not restructure their reserve portfolios casually or quickly, which means the accumulation visible in the data reflects decisions made over years and reflects a considered view about the long-run reliability of dollar-denominated assets as a store of value.

Take a look at U.S. trade data. For four of the last five months, the single largest export from the United States by value has been non-monetary gold, surpassing aircraft, pharmaceutical preparations, and automotive products. That gold is flowing to China, Hong Kong, Switzerland, and Gulf countries, and what leaves Switzerland subsequently follows the same eastward trajectory. The mechanism Luke Gromen describes (sell dollars, buy gold, deliver gold to China, receive yuan, use yuan to pay for rare earths or components) may sound like a thought experiment, but the export data suggests it is operational to some degree already. The comparison to 1940, when the United States demanded gold or dollars from Britain in exchange for arms regardless of sterling’s reserve history, is instructive.

V: Gold in Inflation and Deflation

The case for gold in an inflationary environment is intuitive enough that it doesn’t require elaboration. When nominal claims are diluted through money creation, an asset with finite issuance and no counterparty absorbs the purchasing power that paper instruments surrender. The more interesting and less examined question is what gold does when the shock runs in the opposite direction.

The base case, given the Hormuz closure and the supply chain dynamics described above, is not a clean deflationary depression but a stagflationary depression in which economic activity contracts while input costs remain elevated or rise further. Oil and gas shortages produce a simultaneous compression of real incomes and an increase in the cost of producing everything that requires energy as an input, which is functionally everything. That combination is historically difficult for levered systems to navigate and impossible for policymakers to address with conventional tools.

But the deflationary scenario deserves engagement on its own terms, because it is the one that gold skeptics most frequently deploy as a counterargument. The claim is that in a true deflation, cash and cash equivalents outperform everything, gold included, because the purchasing power of money rises as prices fall. That logic holds in a world where sovereign debt is not a concern, but not in the world that currently exists. U.S. true interest expense (entitlements + interest payments) already stands at roughly 104% of receipts. In a deflationary environment, receipts fall faster than interest expense, because rate adjustments operate on a lag while entitlement spending rises counter-cyclically as unemployment and economic stress increase. A government that cannot cover its interest and entitlement obligations from tax receipts in an environment of falling prices faces a binary choice between nominal default and monetary expansion, and the historical record on which path is chosen is unambiguous. They always print.

This is precisely what makes gold durable across both scenarios rather than just one. In inflation, it protects against the dilution of nominal claims. In deflation, it protects against the credit risk that attaches to every piece of sovereign debt when the fiscal arithmetic stops working.

The swap lines extended to Gulf countries and Argentina fit within this framework as a rearguard action to preserve dollar relevance at the margin. The Gulf lines are designed to prevent reserve managers from selling Treasuries to raise dollars, keeping them inside the dollar system rather than accelerating their migration toward alternatives. The Argentina line is designed to prevent further drift toward Chinese financial infrastructure, given that Argentina had already refinanced an IMF loan in yuan. These interventions carry an assumption that Hormuz reopens on a short timeline and that the pressure on the system remains manageable. If that assumption fails, and the evidence from the ceasefire negotiations suggests it is not a safe one, then Bessent’s trade of Treasury stability today for slightly higher inflation tomorrow becomes a trade of Treasury stability today for a significantly more disorderly situation at a horizon that markets are not yet pricing.

VI: Why I think Equity Markets will stay intact

I recently ran a useful exercise in looking at capital gains tax receipts as a share of total U.S. government revenue.

Over the past three decades, the share of capital gains in total U.S. receipts has reached the 5% range on four occasions. The first was in the late 1990s preceding the dot-com collapse, second was in the mid-2000s ahead of the global financial crisis. The third was 2017-18 which was followed by average gains, and the fourth was in 2021, which preceded a relatively sideways market. The pattern is not law, but it is consistent enough to warrant attention, and we are currently sitting at approximately 5% again.

The tempting conclusion is that this is simply another instance of the same cycle, and that there’s a correction due. But let’s look at another variable, the debt-to-GDP. In first two occasions, debt-to-GDP was running at 50-70%, pre-QE, in an environment where the fiscal position still retained meaningful room to absorb a shock through conventional means. The next two occasions occurred at approximately 100-110% debt-to-GDP, and the correction was largely a sideways and rotational market. So my conclusion is that, post-QE, the Fed hasn’t let markets fall steeply at concerning levels.

The current environment sits at roughly 130% debt-to-GDP with capital gains at the same 5% threshold that preceded each of the prior episodes. A 40-50% percent drawdown in U.S. equities, which is the range that some analysts have begun to discuss seriously, would produce a fiscal shock of a magnitude that the system has not been asked to absorb in the post-QE era.

My view is simple: the Fed has to print, as the alternative is a sovereign debt crisis in the world’s largest economy.

Keeping this in mind, I don’t see American equities going through significant corrections any time soon. The consequences that follow would be too severe. The recent movement tells me that the market is behaving quite irrationally and is ignoring all macro issues the western world faces. Instead of looking at the market as a whole, there are good opportunities in specific bottoms-up stock picking.

Welcoming any pushback or feedback.

I would like to credit most of this work to my teachers Luke Gromen, Ritesh Jain, Alex Divinsky, and Bob Murphy. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

I covered Gravita India recently. Pondy Oxides is a similar company, but operates at a much smaller scale. I consider it as a Gravita in the making (it will take time), and believe that more money is made in the first derivative, which is in the rate of change, not the thing itself.

Pondy Oxides is in the metallic and non-ferrous recycling business. It’s one of India’s largest secondary lead manufacturers. The word "secondary" is important here. Primary lead comes from mining. Secondary lead comes from recycling used material, mainly spent lead-acid batteries. POCL takes scrap, melts it down, refines it, and sells the output as metal or alloys. It does this across four materials: lead, copper, aluminium, and plastic.

Business Segments

1. Lead and Lead Alloys (~90%+ of revenue)

This is the original and dominant business. POCL carries out smelting of lead battery scrap to produce secondary lead metal, which is further transformed into pure lead and specific lead alloys. The alloys they make include calcium, antimony, tin, silver, and babbit alloys, each used in different battery chemistries and industrial applications. The primary end market is lead-acid batteries, which go into automobiles, inverters, and UPS systems.

The economics here are commodity-like. You buy scrap at a discount to LME lead prices, process it, and sell refined metal or alloys at or above LME. The spread between input cost and output realization, after processing costs, is the margin.

2. Copper (the fastest-growing vertical)

The copper segment is rapidly emerging as POCL’s most ambitious growth engine, reflecting a strategic shift toward diversification from lead. While lead still contributes over 90% of revenue, management has explicitly outlined copper as a core vertical that could rival or even surpass lead in terms of revenue over a 7-8 year horizon.

As of FY25, POCL’s copper processing capacity stood at 6,000 MTPA and revenue contribution was at Rs. 55 crore. However, H1 FY26 alone generated Rs. 172 crore in revenue, and management is guiding for a full year at Rs. 400 crore. Capacity is expected to reach 12,000 MTPA by end of FY26 and scale to 24,000 MTPA by FY27, with a planned capital investment of Rs. 100-110 crore.

3. Plastics (a byproduct turned business)

The raw material for this business primarily comes from POCL’s own battery recycling process, where about 7% of each battery comprises plastic. Instead of disposing of it, POCL processes this plastic in-house through washing, grinding, and extrusion to make polypropylene granules. It also procures additional plastic scrap domestically and internationally.

What makes this segment especially promising is the shift from basic recycled granules to value-added products, where additives and fillers are blended with base polymers to create high-performance plastics. POCL ultimately plans to raise monthly production capacity from roughly 1,000 MT to 2,000-3,000 MT in phases.

4. Aluminium (currently paused)

The aluminium segment was strategically launched to diversify beyond lead, copper, and plastics. While it initially showed promise, it has since entered a deliberate pause phase as part of a broader business recalibration. Unlike the lead division, the aluminium business was launched with a commoditized product mix centered around standard die-cast alloys like ADC12 and LM series, which suffer from intense price-based competition, low entry barriers, and critically, the absence of an effective hedging mechanism.

I’ve analyzed Pondy Oxides using the TVG framework:

Theme

The core theme for POCL is organized recycling as infrastructure. This is a more precise framing than simply “recycling company”, as what POCL actually does is sit at a structurally advantaged position in supply chains that cannot function without secondary metal feedstocks.

The secondary lead cycle is structurally unavoidable. Lead-acid batteries are not going away. They power ICE vehicles, e-rickshaws, inverters, UPS systems, and telecom towers. Every battery that gets sold today becomes a recycling feedstock 3-5 years from now. This is a replacement cycle that is mechanically guaranteed by the installed base of vehicles and backup power systems. India’s battery demand grows with vehicle penetration, and vehicle penetration in India still has a long runway. Management cited 5-6% incremental domestic demand growth per year, with 2-3% internationally.

The more important point is what happens to that scrap. Secondary lead accounts for the majority of global lead supply. In India, a large portion of battery recycling historically happened in the unorganized sector (small smelters with no environmental controls, no traceability, no compliance). The regulatory shift underway is systematically dismantling this. The Battery Waste Management Rules (BWMR) and Extended Producer Responsibility (EPR) framework are the two key instruments. BWMR mandates that battery producers take responsibility for end-of-life collection and recycling. EPR creates a credit mechanism where formal recyclers generate tradeable credits that battery manufacturers must buy to meet compliance. Management confirmed in Q3 that BWMR enforcement has materially improved domestic scrap availability and reduced leakage to the unorganized sector. This directly expands POCL’s addressable base because scrap that previously went to informal smelters now flows to compliant formal players.

POCL the first Indian smelter listed on the London Metal Exchange. When a battery manufacturer needs to meet EPR compliance, they need to buy credits from LME-registered, formally certified recyclers. That is a very short list. The regulatory tailwind is already showing up in procurement metrics. Management noted that domestic procurement of lead scrap improved meaningfully as the formal collection ecosystem tightened.

The EPR credit business is worth watching separately. POCL has been accumulating EPR credits but has not sold any yet because pricing has not matured. Management said they are waiting for a sustainable price to develop. Once the government enforces compliance more stringently from April 2026, demand for these credits should firm up. This is currently an unquantified upside that sits off the P&L entirely.

Copper is a structural diversification into a much larger addressable market. Secondary copper recycling in India is deeply fragmented and mostly informal. POCL is applying the same playbook it used in lead: build scale, get LME registration, and position as the formal organized player in a market that regulatory trends will eventually consolidate toward. Management said explicitly that they see copper potentially exceeding lead revenues over a 7-8 year horizon. Global copper demand is structurally growing driven by electrification, EVs, grid infrastructure, and data centers. India has limited primary copper mining capacity. Secondary copper from scrap is the only realistic way to fill the gap between demand growth and primary supply constraints. POCL is sourcing scrap globally (US, South America, Australia) and processing it domestically. The India-EU trade deal management flagged in the Q3 call matters here too: zero duty on Indian metal exports to Europe opens a large, quality-conscious market that POCL is already approved with several customers in, though business has been opportunistic rather than structural. Post-FTA, that changes.

The regulatory moat is deeper than it looks. Running a formal secondary metal recycling plant in India requires pollution control clearances, hazardous material handling licenses, EPR registrations, LME certification, and ongoing environmental compliance. This is a genuine barrier that the unorganized sector cannot cross. POCL has built this compliance infrastructure over 30 years. As regulation tightens, the moat widens because new entrants face the same compliance burden from day one, but POCL already has it.

The Mundra option is a medium-term theme lever. POCL owns 123 acres at Mundra, which is a port city in Gujarat with deep connections to Middle Eastern and European trade routes. Management confirmed plans to develop this in the second half of CY2027 once copper expansion is absorbed. Mundra gives POCL better freight economics to Europe and the Middle East, access to import-heavy scrap flows through a major port, and an opportunity to expand lead and copper capacity at a greenfield site with modern infrastructure. This is not in current numbers at all and represents a distinct second chapter of growth that only begins after FY27.

Lithium-ion is a long-duration optionality play. POCL has a stake in ACE Green Recycling, which is doing R&D on lithium-ion recycling across all battery chemistries. Management was measured about this, commercial entry is targeted for 2027, and the R&D is not complete. They flagged two real constraints: feedstock availability in India is still thin because EV penetration is early, and the technology for lithium-ion recycling is evolving rapidly. Management explicitly said they expect feedstock to become meaningful only around 2028. The rational position is to not price this into the thesis yet, but to note it as optionality. If POCL executes its R&D through ACE Green and positions itself before the EV scrap wave arrives, it has the infrastructure, regulatory relationships, and operational experience to be the dominant formal recycler in lithium-ion as well.

Value

The value case for POCL rests on a few interconnected things: the quality of the margin structure, the hedging model, balance sheet health, and the EBITDA per ton trajectory.

The margin model is operationally driven, not commodity price driven. This is the most important thing to understand about POCL’s economics. Revenue moves with LME metal prices because scrap procurement costs and finished metal selling prices both track LME. Management confirmed they run a full hedge book, commodity risk hedged through a direct LME broker account, forex risk hedged through banks, and natural hedges maintained where imports and exports offset. The result is that EBITDA per ton is structurally insulated from metal price movements. Management confirmed sustainable lead EBITDA per ton of Rs. 15,000-17,500 irrespective of price movements. The Q2 peak of ~Rs. 20,000 per ton reflected a better value-added product mix and operational efficiency gains.

What drives the actual EBITDA per ton are three things management explicitly quantified in the Q2 call: value-added product mix (responsible for ~Rs. 6 per kg improvement), operational efficiencies (responsible for ~1.5% margin improvement), and fixed cost absorption from higher volumes (the balance). None of these are commodity-dependent.

The hedge model has a wrinkle worth understanding. When metal prices move sharply upward in a short period (as copper did in Q3) mark-to-market provisions on open hedge positions show up as P&L charges. POCL reported a Rs. 7.28 crore MTM charge in Q3 for exactly this reason. This is not a cash loss, it is an accounting timing difference that reverses when the hedged position converts to sales. Management confirmed the Q2 MTM charge was only Rs. 2 crore, and the Q3 spike was because copper prices moved almost vertically.

Value-added product mix is the margin lever management controls. About 70% of lead revenue in H1 FY26 came from value-added products, which are specific alloys made to OEM specifications. The remaining 30% is commodity refined lead. The alloy business is higher margin, stickier, and less price-sensitive because OEMs are buying performance specifications rather than just the metal. POCL makes over 100 different alloy formulations for various customers. Q3 saw the value-add share drop to 55% as new capacity ramped up with some volume going to non-OEM channels initially, but management explicitly flagged this as temporary and said Q4 value-add share will recover as OEM contracts secured for the new capacity begin.

For copper, the current business is plain recycling: buy scrap, process to refined copper cathode or rod, sell. EBITDA per ton on copper recycling is around Rs. 34,000-35,000 per ton, which sounds high in absolute terms but is a lower percentage margin than lead because copper prices are much higher. The forward integration into value-added copper products (flat rolled products, extruded products, and eventually foils and coils) is where margins improve substantially. This is scheduled to begin in H1 FY27 as the product-wise expansion at TKD commences.

The balance sheet is genuinely clean. POCL achieved zero net debt and a net cash position of Rs. 71 crore as of Q2 FY26. Cash on books as of December 2025 (Q3) was Rs. 35 crore, lower because capex spending accelerated in H2. Working capital days are at 47 days as of Q3, with receivables at just 15 days. This is a sharp improvement from where the company was historically. Management said inventory should normalize at Rs. 230-250 crore by year-end. The clean balance sheet matters because it means growth capex is being funded from internal accruals. The Rs. 100-110 crore copper expansion program across FY26-27 is being funded this way. Free cash flow is guided to exceed Rs. 100 crore for FY26.

The one balance sheet item worth tracking is a ~Rs. 75-80 crore IGST refund receivable from the government. This arises because POCL procures domestically, pays IGST, and then exports the finished product with duty payment, making them eligible for the IGST refund.

Competitive positioning is structured around formalization advantages, not process differentiation. POCL does not have a proprietary smelting technology that competitors cannot replicate. What it has is: LME registration (quality signal that takes years to achieve), regulatory compliance infrastructure (takes years and capital), long OEM relationships (stickiness), scale in procurement (better sourcing economics), and a hedging capability that most unorganized and smaller formal players cannot match. These together create a positioning moat that is durable but not impenetrable if a large well-capitalized entrant chose to build it from scratch. The more relevant competitive risk is within the formal organized sector, where management acknowledged other players are also expanding lead capacity. Management’s response was essentially that demand growth is sufficient to absorb multiple formal players expanding, and that they are differentiating on value-added alloys rather than competing purely on commodity volume.

Growth

Lead: volume-led in the near term, mix improvement sustaining margins.

The most near-term visible growth driver is simply utilization of new capacity. POCL’s lead capacity went from 132,000 MTPA in FY25 to 168,000 MTPA after Phase 1 (36,000 tons commissioned Q1 FY26), and then to 204,000 MTPA after Phase 2 (further 36,000 tons commissioned December 2025). Total capacity expansion is over 50% in one year. Management guided 70,000 tons of lead sales in H2 FY26, against ~50,000 tons in H1. The Phase 2 capacity is expected to ramp to 70-80% utilization through FY27.

What is important is that contracts for the new Phase 1 capacity are already secured with OEMs. So the volume growth in Q4 and FY27 is contracted volume that was waiting for capacity to exist.

The value-add share trajectory matters for margin quality. At 70% value-add in H1, dropping to 55% in Q3 as volumes ramped, management guided it returns to higher levels in Q4 as OEM contracts kick in. The medium-term target is to maintain 70%+ value-add share even as volumes grow, which management said requires being selective about which orders to take.

Separately, the modernization of older plant capacity is underway in parallel. This should improve efficiency and potentially allow better product mix from legacy capacity over FY27-28.

Copper: the high-conviction growth story.

This is where the investment thesis gets interesting. The copper segment went from essentially a small diversification experiment to a Rs. 400 crore annualized revenue business in FY26. The full-year guidance for copper revenue was Rs. 400 crore for FY26, having done Rs. 172 crore in just H1. H2 was expected to deliver the balance.

The next phase is a 4x capacity expansion from 6,000 MTPA to 24,000 MTPA by end FY27, with Rs. 100-110 crore total capex. The expansion is structured product-wise, not phase-wise, meaning different product lines come online as capital is deployed, rather than large lump capacity additions. First value-added copper products are guided for H1 FY27.

Management’s revenue ambition from this Rs. 100-110 crore capex is Rs. 900-1,000 crore in copper revenue as the capacity matures and value-added products come online. Given copper’s per-ton pricing, 24,000 MTPA at blended realization with a mix of recycled and value-added products can plausibly generate that range. The margin profile improves as the product mix shifts from plain recycled copper (current EBITDA ~Rs. 35,000/ton) to fabricated value-added products (margins not guided yet but management confirmed they will be higher).

One caution from the transcripts: copper volumes in Q3 were 1,235 metric tons, which is below the H1 pace. Management attributed this to the sharp vertical rise in copper prices disrupting buyer purchasing patterns temporarily as buyers pause when input costs jump 40-45% because their own working capital requirements spike. Management said this restabilization happens in 1-2 quarters. This is consistent with how commodity markets behave and is not a thesis-breaking observation, but it does mean the copper revenue ramp is not linear.

The LME registration ambition for copper is also worth noting. POCL is already LME registered for lead. Management said they will pursue LME registration for copper as well, product by product. Getting there would meaningfully improve copper pricing and reduce counterparty risk, mirroring the competitive advantage they built in lead.

Plastics: small but interesting.

The plastics business is now fully shifted to the TKD facility and has been merged with the parent via amalgamation of the POCL Future Tech subsidiary. This consolidation improves cost efficiency and cash management. Current capacity is ~9,000 MTPA with production around 800-1,000 MT per quarter. The near-term product shift is from basic PP granules to compounded materials (ABS, nylon blends, and filled polymers) which carry materially higher margins.

Management gave modest revenue guidance of Rs. 80-100 crore for plastics as a segment. What makes it structurally interesting is that the raw material is essentially a byproduct of battery recycling. Every battery POCL processes yields ~7% plastic by weight. As lead volumes scale toward 200,000+ MTPA, the internal plastic feedstock generation also scales, reducing external procurement and improving unit economics. EPR credits are also being accumulated in plastics as that regulatory framework matures.

Mundra: the FY28+ growth option.

123 acres at Mundra represents the next chapter after FY27 copper expansions complete. Management’s stated intent is to add lead and copper capacity there alongside potential new verticals. The port proximity creates real freight economics advantages for European and Middle Eastern exports. Post India-EU FTA, this becomes even more valuable. No capex has been committed yet, but the land is owned and the strategic rationale is clear. This is an option, not a commitment, and should be treated as such.

Company-level guidance.

Management has guided for 20%+ volume growth and 20%+ revenue CAGR. EBITDA margins are guided at 7-8% sustainable, with aspiration to push above 8% as value-added mix improves. ROCE above 20%. These are ambitious but supported by the capacity additions now fully commissioned and the copper ramp underway. The 9M FY26 numbers (revenue up 33%, EBITDA up 96%, PAT up 114%) are tracking well ahead of this guidance, though the base effects from margin expansion make the percentage growth look dramatic.

Valuation

Assumptions:

1. Multiples - Conservatively below 1Y median

2. Margins - Derived from EBITDA per ton calculations above.

Risks

1. The copper expansion is betting on a product mix that doesn't exist yet - POCL has never manufactured fabricated copper products before. The customers, quality certifications, product development cycles, and operational know-how for making copper sheets or rods for industrial buyers are meaningfully different from what they do today. Management hasn’t disclosed which products, which customers, or what margins to expect because the products aren’t in the market yet. The first rollout is guided for H1 FY27.

   
   The Rs. 100-110 crore capex has already been committed. If the product launch is delayed, or if the products don’t achieve the margin premium management is assuming, the copper segment could generate substantially lower returns than the narrative suggests. Right now, investors are paying for the value-added copper story, not the recycling-only copper story.

2. POCL doesn't fully control the scrap sourcing - Everything POCL does starts with getting scrap at the right price. For lead, ~70-86% comes from imports. For copper, 100% is imported. This means POCL is permanently exposed to global scrap availability, shipping costs, currency movements, and trade policy.

   The practical risk is this: when metal prices spike sharply (copper in Q3 FY26) the economics of the whole chain get temporarily disrupted. Processing margins compress because input costs and output realizations don’t reprice simultaneously. Management said it takes 1-2 quarters to normalize.

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

Explaining Gravita

Lead recycling is the core, making up roughly 87% of volumes. The raw material is used lead-acid batteries from cars, telecom towers, inverters, and industrial backup systems. These batteries come in through Gravita’s global procurement network, get broken down, smelted, and refined into pure lead, lead alloys, lead oxide, red lead, and lead sheets. The customers on the other end are battery manufacturers: Amara Raja, Exide, Clarios, Luminous.

Aluminium is the second vertical, at around 9% of volumes. The process is similar: source aluminium scrap, melt it, produce alloys for die-casting and automotive applications. This business is currently running at a fraction of its potential in India, for reasons I’ll get into soon.

Plastic recycling contributes about 4% of volumes. Gravita sources plastic scrap and converts it into granules (polypropylene, HDPE, polycarbonate) that go back into industrial manufacturing. This segment also generates EPR certificates, which are sold to companies that need to demonstrate recycling compliance under Indian regulations.

Rubber recycling is newer. Gravita acquired a tyre recycling company in Romania and is commissioning an India plant at Mundra that should be operational in Q1 FY27. The outputs are tyre pyrolysis oil, carbon black, and crumb rubber.

Beyond these four verticals, Gravita also has a Turnkey Solutions division that designs and builds recycling plants for third parties globally. It has delivered over 70 such projects across countries including Qatar, UAE, Saudi Arabia, Poland, and Chile. This generates modest revenue but builds technical credibility and extends geographic relationships.

I’ve analyzed Gravita using the TVG framework:

Theme

The Global Picture

Secondary lead is the dominant form of lead supply globally. Recycled lead meets more than half of total world demand, and lead-acid batteries have a recycling rate above 95%, the highest of any consumer product anywhere. When investors think about lead, they often think about mining, but the actual supply chain runs almost entirely through recyclers like Gravita.

The demand base for lead-acid batteries is actually more resilient than most people expect. The global lead-acid battery market was valued at over $100 billion in 2025 and is projected to grow at around 3-5% annually through 2035. The main driver is not EVs or novel technologies, it is the steady expansion of the installed base of vehicles, telecom towers, data centers, and UPS systems across emerging markets. Every vehicle, regardless of powertrain, carries a 12V lead-acid auxiliary battery. Every telecom tower in Africa and South Asia runs on lead-acid backup power. Data center UPS systems rely on lead-acid batteries for their reliability profile. None of this is going away.

The lead acid battery recycling market specifically is projected to grow from around $17.6 billion in 2025 to $23.9 billion by 2030, at a CAGR of 6.3%. Asia Pacific accounts for roughly 65% of global volumes, driven by China and India. Gravita is explicitly named in industry research alongside Ecobat, Glencore, Clarios, and Johnson Controls as one of the top five players globally by recycled lead market share.

The regulatory direction globally is moving in favor of large formal recyclers. The EU Batteries Regulation (2023/1542) mandates 85% recycled-lead content in industrial batteries by 2031 and sets a 95% recovery target. Regulations like these require traceability, so manufacturers need to demonstrate their lead came from a certified recycling chain. That is not something an informal smelter in West Africa can provide, so it rewards exactly the kind of vertically integrated, globally networked operator that Gravita has built.

Recycled lead also has a structural cost advantage over primary production. The energy required to recycle lead is 40-60% lower than primary smelting. As carbon disclosure requirements tighten and ESG mandates become more commercial, battery manufacturers will increasingly preference sourcing from certified recyclers with documented sustainability records. Gravita has an AA- credit rating from ICRA and India Ratings, ILA-registered plants, and ISO certifications across its facilities.

The India Picture

India’s recycling rate across materials sits at around 20%. In developed markets, it is 60-80%. That gap is part of the thesis.

For lead specifically, the informal sector handled 80% of battery recycling as recently as FY16. That number is now around 25%, and the shift is being driven by three interlocking regulatory mechanisms:

1. The first is the Battery Waste Management Rules (BWMR 2022). Battery producers are legally required to collect spent batteries in escalating proportions of their annual output, ramping to 70-90% by FY25-26, and hand them exclusively to registered recyclers. Scrap that previously disappeared into informal channels now has to flow through certified players.

2. The second is the Extended Producer Responsibility framework. Compliant recyclers generate EPR certificates, which battery manufacturers must purchase to prove they’ve met their collection obligations. This creates a direct financial link between formal recyclers and their own customers. Battery companies buying recycled lead from Gravita are simultaneously getting a raw material and a compliance instrument. Management confirmed in concalls that major OEMs are now specifically choosing to source from recyclers who can issue EPR certificates.

3. The third is the GST Reverse Charge Mechanism. When a buyer purchases scrap from an unregistered informal operator, the GST liability falls on the buyer. This destroys the tax-evasion cost advantage that informal recyclers relied on for decades. Informal operators who didn’t charge GST could undercut formal prices.

The combined effect is measurable. Gravita’s domestic battery scrap procurement grew 35% year-on-year in Q2 FY26, and the domestic-to-import ratio for their Indian plants shifted from 36% domestic / 64% imported a year ago, to 52% domestic / 48% imported today.

The numbers on the total addressable market tell the same story. Gravita’s own estimates put the formal Indian lead recycling market at around Rs. 4,800 crore in FY25 (40% of a Rs. 12,000 crore total). By FY26, the formal segment is expected to account for 75% of a Rs. 13,875 crore market, over Rs. 10,400 crore. Even if the pace is slower than management expects, the direction is clear and the incumbents with scale, procurement networks, and certifications absorb the majority of that incremental flow.

The rubber story is following the same regulatory arc, running about five years behind lead. EPR rules for waste tyres were introduced in 2022, with targets starting at 35% of production and reaching 100% from FY24-25 onwards. India generates roughly 395,000 tonnes of waste tyres per year and growing. Around 90% of tyre recycling was in the informal sector before these rules came in. The crackdown on batch pyrolysis plants by CPCB and new certification requirements for continuous pyrolysis are pushing informal operators out. Gravita is entering this vertical now, before the formalization curve has played out, which is the right timing.

The broader pattern is worth stepping back to appreciate. Lead, rubber, plastic, and eventually lithium-ion batteries are all going through the same transition but at different stages. Gravita is sequencing its entry into each vertical to arrive early, build the procurement network, and then scale as regulatory enforcement tightens. It is applying the same playbook it ran in lead, now in rubber, and planting flags in lithium-ion for 2027-2030 when the first wave of EV batteries reaches end-of-life.

Value

Recycling companies often trade at a discount because the business model looks risky on paper. You are buying a commodity, processing it, and selling a commodity, so margins should be thin and commodity price swings should eat into earnings. Gravita has structurally addressed all three of these concerns, which is why its financial profile looks different from what you might expect.

The hedging mechanism is the most important feature of this business

Lead prices on the LME move constantly. A recycler that buys battery scrap today and sells refined lead three weeks later is exposed to that price movement, and Gravita has been fully hedged against this since June 2019. The moment they buy scrap, they simultaneously sell an equivalent amount of lead on the LME, locking in the processing spread. The company earns on the difference between scrap cost and refined product realization, and takes no view on where lead prices go.

LME prices moved from $1,947/tonne to $2,283 to $2,046 across FY21 to FY25. EBITDA margins over the same period rose from 7.2% to 10.4% in a straight line, irrespective of where lead traded.

The absence of an equivalent hedging mechanism for aluminium is the biggest structural gap in the business today. ADC12 aluminium alloy is not listed on MCX, which means Gravita cannot hedge Indian aluminium purchases and sales the way it can for lead. The India aluminium plant is running at roughly 5% utilization as a result. MCX approval has been pending for over a year, the board has approved it, the documentation is complete, but the actual launch of the contract keeps slipping. Management has said it will happen ‘any time now’ in each of Q1, Q2, and Q3 FY26 concalls, and when it eventually happens, it’ll unlock meaningful incremental capacity.

The financial track record is strong and improving

Five-year revenue CAGR is 23%, PAT CAGR is 57%. In 9M FY26, volumes grew 5%, revenue grew 9%, EBITDA grew 15%, and PAT grew 32%. Revenue growth was deliberately held back as management made a consistent choice to divert overseas semi-processed material into India for final refining when the India-LME price arbitrage was favorable, capturing higher margin at the cost of reported consolidated volumes.

Track EBITDA and PAT rather than volumes because the volume optics obscure the economics.

EBITDA per tonne for lead in Q3 FY26 was Rs. 23,035 against official guidance of Rs. 19,000-20,000. That is a consistent beat, driven by two things: the India-Africa price arbitrage, and the increasing share of value-added products. Value-added products (lead oxide, red lead, customized alloys, lead sheets) were at 46% of volumes in Q3 and the target is 50% by FY28. Management quantified the margin impact: each percentage point increase in value-added share adds roughly 2.5-3% to gross margin.

ROIC has been 25-31% over the last five years, against a stated minimum threshold of 25% for any new project. Every new vertical and incremental capex allocation has to clear that hurdle before capital is deployed. The company actually deferred paper and steel recycling explicitly because they hadn’t yet met the ROIC test.

The balance sheet is net-debt-free. The total capex plan to reach 700,000 tonnes of capacity by FY28 is Rs. 1,225 crore, which has already been reduced by Rs. 300 crore by opting for Brownfield expansion at Mundra and Phagi instead of Greenfield projects. The company targets a maximum 3-year payback and asset turns above 8x on new projects.

The competitive position in lead is genuinely strong. The barriers to entry are concrete: import licenses for scrap into India are performance-based and not available to new entrants without a track record; OEM quality approvals from major battery manufacturers take years to obtain; and Gravita’s procurement network of 33 owned yards and 1,900+ touchpoints across 34 countries took three decades to build.

Growth

The immediate lever: capacity coming online now

Gravita’s India lead capacity was constrained through most of FY26 because new plants at Mundra and Phagi were awaiting ‘Consent to Operate’ clearances from Gujarat state authorities. 125,000 tonnes of new lead capacity is being commissioned through Q4 FY26 and Q1 FY27. This is a 37% increase in installed India lead capacity. Management had pre-built inventory in anticipation and said utilization would reach 60-70% within one to two quarters of commissioning.

The math on FY27 volumes is straightforward. Current capacity runs at roughly 65% utilization. New capacity adds 37% to the base. Even at modest utilization on incremental capacity, volume growth of 25-30% in lead alone is achievable. Add rubber and aluminum as contributing verticals, and management’s 25%+ volume CAGR guidance becomes more credible than the FY26 numbers suggest.

Domestic scrap formalization as a structural tailwind

Only about 35% of Indian battery scrap currently flows through organized channels. Management’s expectation is that this reaches 90% within two to three years. That is a roughly 2.5x increase in the pool of formal-sector scrap accessible to certified recyclers, without any change in total battery consumption. Gravita, with 200+ domestic customers across 20 states and OEM relationships with every major Indian battery manufacturer, is positioned to absorb a disproportionate share of that flow.

Rubber recycling beginning to contribute.

The Romania plant generated Rs. 3.5 crore in Q3 FY26 and the Mundra India plant is targeting commissioning in Q1 FY27 with commercial revenue from Q2 FY27. Management has guided Rs. 70-80 crore in rubber revenue for FY27 with unit economics cited at Rs. 7-8 per kg EBITDA at minimum, which translates to roughly 30% EBITDA margin, superior to lead.

The rubber business benefits from the same regulatory-driven formalization happening in lead. Gravita also has global scrap sourcing infrastructure that can supply cheaper rubber scrap from overseas locations, improving procurement economics versus domestic-only operators.

Aluminium

The India aluminium plant is essentially idle because of the MCX hedging issue. Sustainable EBITDA for aluminium is guided at Rs. 12-14 per kg once hedging is in place. The business already generates Rs. 14-15 per kg in overseas operations. The moment MCX lists an ADC12 contract, Gravita can start buying Indian scrap, hedge it, and run the plant. Full utilization of existing capacity would add a meaningful new revenue line that is currently contributing almost nothing.

Valuation

Assumptions:

• Topline - Assumed 35% growth for FY26 and 30% for FY27.

• EBITDA - Although the copper business has lower margins, management has guided that they will ensure that, over time, the business is not margin dilutive.

• EV / EBITDA - I’ve been very conservative here, since I’m assuming that the current market sentiment is to continue till FY28.

Risks

• Aluminium hedging - The India aluminium business cannot be properly run without MCX hedging. Management has been expecting the ADC12 contract to launch every quarter since Q1 FY26. If it doesn’t arrive in FY27, aluminium growth in India gets pushed out and the non-lead mix target slips.

• Capacity timelines have already slipped once - The 125,000 tonne India lead expansion was expected to be commissioned in Q3 FY26. It has slid into Q4 FY26 and Q1 FY27, held up by state regulatory clearances. The Dominican Republic expansion, supposed to receive approvals in Q3 FY26, has also been quieter than expected.

• Rubber recycling is early and unproven in India - Management has guided 30% EBITDA margins for rubber, but no India rubber revenue has been generated yet. The Romania plant is a proof of concept at small scale.

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

India: Bear Market Math and What Comes Next

Excluding March 2020, March 2026 was the worst month for the Nifty 50 since October 2008, an 11.5% drawdown in the country’s top fifty companies. Although some of it has been reversed recently, we’re still in a bear market.

Three out of four small and micro cap stocks are currently sitting significantly below their peaks. Large and midcap median drawdowns are running around 30%.

History provides some context. Peak-to-peak cycles in Indian equities have averaged roughly six years. The bear phases within those cycles have lasted 14, 31, and 26 months respectively. The current phase began in September 2024. We are now approximately 19–20 months in. That places us well inside historical bear territory, but not obviously beyond it.

The Job in a Bear Market Is Simple

The bull cycles that bookend bear phases in India have been extraordinary. The post-COVID rally produced nearly 5x returns. The 2013–18 NDA cycle delivered 265%. Post-GST recovery: 226%. The asymmetry is striking, and it is the reason bear market discipline matters so much. Rather than focusing on generating returns, the job is to survive, avoid permanent impairment, and not fall too far behind the index, because the cycle that follows will do the heavy lifting.

At current valuations, the setup for medium-term investors is reasonably compelling. The Nifty 50 is trading around 20.5x earnings. Historical return data by entry PE tells a clear story: at 20x, the one-year median return has been approximately 25%, with an 88% win rate. At a five-year horizon, the win rate goes to 100%, and this holds even if the entry threshold is pushed to 25x.

The Geopolitical Overhang

The specific event that drove much of March’s weakness is now in the open. For the situation to deteriorate meaningfully from here, a second escalation would be required, and on current evidence, that appetite does not appear to exist (thank the bond market!). Containing the damage and managing the optics is the more likely path.

The more important signal came in February, before the event fully materialized. That month saw 18–20% selling in IT stocks, and yet FII flows were net positive, the currency held, and the index stayed positive. The March weakness is therefore largely attributable to the conflict itself, not to a broader deterioration in India’s fundamental picture. Once the dust settles, there will be residual effects in specific sectors (think hotels, airlines, oil refiners that will carry the scars longer than the broader market). Their numbers will disappoint and they will not snap back to pre-event valuations quickly. But for the index overall, a modest allowance on fiscal math does not meaningfully change the return calculus, the markets can absorb that.

Let’s invert this: Can the conflict extend into something structural? India runs $40 billion in annual remittances from the Middle East. Its oil supply is heavily linked to the region. Travel, tourism, and trade flows are meaningful in ways the Russia-Ukraine situation simply was not for India. A prolonged conflict would carry real economic consequence. The base case, however, is resolution. Iran-US back channels were not dramatically far apart before the situation escalated. And the domestic political calendar in the US (midterms approaching, inflation sensitivity) creates natural pressure toward de-escalation.

The FII Problem

Foreign institutional selling in India has attracted significant attention. FIIs have been reducing exposure across emerging markets broadly, and Korea and Taiwan have absorbed more than double India’s selling in absolute terms. The problem is not India-specific, it’s more about return math.

The equation that used to attract foreign capital looked something like this: 15% INR returns, no capital gains tax, minus 3–3.5% currency drag, net roughly 11% in dollar terms. That was a compelling offer. The same equation today looks quite different: 12% nominal INR returns, subtract approximately 2% for taxes, leaving 10% in rupee terms, then subtract 5% on currency, netting roughly 5% in dollars. Against that backdrop, incremental reallocation into India requires a strong conviction call rather than a routine carry trade.

The corrective that would actually move the needle: In 2013, India raised $30 billion through NRI deposits to stabilize the currency under pressure. Today the capacity exists to raise $75 billion. Even at a 2% interest premium on three-year deposits, the carry cost is roughly $1.5 billion annually, a manageable price for currency appreciation of 4–5%, lower import valuations on oil, and restored external confidence.

Where Capital Is Concentrating

One visible exception to the broad FII retreat has been the capital goods sector. The thesis is straightforward enough that multiple allocators arrived at it independently: the world needs power. Data centers, private capex, industrial policy, and electrification are all converging on the same input constraint. Investors who acted on this early discovered, after the fact, that they had company.

So, beneath the surface of macro disruption, the operating economy is not as impaired as the index suggests. Quarterly business updates coming through now are strong across sectors: auto sales, consumer names, gold, retail. The public was spending through March’s turbulence, even as markets fell.

The Framework That Survives Cycles

The clearest way to think about portfolio construction across regimes is through two benchmarks applied in sequence. In bull markets, the goal is to deliver small cap index returns and capturing the outsized gains that accrue to risk during expansion. In bear markets, the goal shifts to beating the Nifty 50, protecting capital, and not falling too far behind. If a portfolio can achieve both with a predominantly small cap orientation, it has produced genuine alpha.

We are still in the phase where the second benchmark applies. That is not a reason for pessimism. The tide being out is how you find out who built something real.

I would like to credit most of this work to my teachers Ishmohit Arora, Siddhant Bhandari, and Samir Arora. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

KSH International makes the wire that goes inside electrical machines. Not the cables that run between buildings, but the precision-engineered copper conductors that get wound into coils inside transformers, motors, and generators. Without these coils, no transformer can change voltage, no motor can spin, and no generator can produce electricity.

It operates four manufacturing facilities near Mumbai and Pune with combined capacity of 43,000 MTPA. It is the largest exporter of magnet winding wires from India and the second largest domestic manufacturer by installed capacity.

Every product falls into one of two categories: specialized winding wires at 75% of revenue, and standard winding wires at 25%.

Specialized Winding Wires

The flagship product is the continuously transposed conductor, or CTC. This is a precision assembly of multiple flat copper strands. For any large power transformer above a certain size and voltage rating, CTC is essentially the only engineering answer.

KSH is India’s largest CTC manufacturer and the only Indian company currently approved by PGCIL to supply CTC for HVDC transformer windings. HVDC is the transmission technology used to move large amounts of power over very long distances, the backbone of India’s renewable energy evacuation infrastructure. Every HVDC transformer built for the Indian grid needs PGCIL-certified components, and KSH is currently the only domestic supplier that qualifies.

The other specialized products are paper-insulated rectangular conductors (PICC), which go into large oil-filled grid transformers, and rectangular enamelled wires, which go into dry-type transformers, traction motors, and industrial equipment. All of KSH’s exports go to transmission and distribution customers, the transformer-building divisions of ABB, Siemens, Hitachi Energy, Toshiba, GE Vernova, and similar OEMs across 24 countries.

Standard Winding Wires

Round enamelled wires are the commodity end of the portfolio. A copper or aluminium wire coated with thin enamel insulation, used in motors, compressors, home appliances, and EV traction motors. Many suppliers can make these, competition is price-driven, and margins are lower. KSH participates here for volume and to serve the full product needs of customers who buy both specialized and standard wires from the same supplier.

The interesting development within standard wires is PEEK-coated round wire, used in 800-volt traction motors in premium electric vehicles. These motors run at temperatures standard enamel cannot survive. KSH is developing this product under an exclusive licence with Germany’s HPW Metallwerk. Margins on PEEK wires are significantly above the standard wire average. Two-wheeler EV volumes are already being supplied and four-wheeler volumes are 12 to 18 months from being material.

How the Business Model Works

KSH sells exclusively to OEM manufacturers. Every order is make-to-order. When a customer places an order, KSH books the copper from its supplier at the same price on the same day. The invoice to the customer carries that same copper price. Copper movements pass straight through to revenue with no impact on profit per ton. The company earns its margin entirely on the fabrication value-add.

This means reported EBITDA margins as a percentage of revenue will move with copper prices and are not the right lens for this business. EBITDA per ton is what matters. That figure was Rs 66,044 in the nine months ended December 2025, up 32% from Rs 50,133 in the same period last year.

I’ve analyzed KSH International using the TVG framework:

Theme

The Global Picture

The world is running short of transformers. It is showing up in utility earnings calls, national security reports, and infrastructure project delays across the United States and Europe.

The US imports approximately 80% of its transformer MVA capacity. Lead times for large power transformers have stretched to two to four years. Wood Mackenzie estimated a 30% supply shortfall for power transformer units in 2025. Since 2019, demand for generator step-up transformers has grown 274% and demand for substation power transformers is up 116%. Two data centers in Silicon Valley were built and ready to operate but could not turn on because the transformers needed to connect them to the grid had not arrived.

Several forces are converging to create this shortage simultaneously. AI and data center buildout is the most recent addition. Data center firms requested 40.2 GW of new power connections from Dominion Energy Virginia alone in early 2025, up from 21.4 GW just six months earlier. Each of those gigawatts requires transformers at multiple points in the supply chain. The global data center transformer market is projected to grow from $7.8 billion in 2024 to $11.2 billion by 2030.

Renewable energy integration is the larger structural force. Every solar park and wind farm needs step-up transformers to connect generation to the grid. Every long-distance transmission corridor needs substation transformers at each end. The IEA estimates that meeting global climate goals requires adding or refurbishing over 80 million circuit kilometers of grid by 2040. All of it is transformer-intensive.

Grid modernization adds a third layer. More than half of the roughly 60 to 80 million distribution transformers in service in the US are 33 years old or older, approaching or exceeding their designed service life. They need to be replaced regardless of new demand growth.

This creates a sustained multi-year demand environment for transformer OEMs, which flows directly upstream to winding wire suppliers. When ABB, Hitachi Energy, Siemens, and Toshiba are all simultaneously expanding transformer manufacturing capacity and reporting double-digit order growth in their grid divisions, their demand for CTC and PICC increases proportionally.

HVDC: The Most Consequential Sub-Theme

HVDC transmission is the technology that enables very large amounts of renewable power to be moved over very long distances with minimal losses. It is not optional infrastructure for countries with ambitious renewable energy targets.

India’s large renewable energy hubs in Gujarat and Rajasthan are hundreds of kilometers from the cities that need the power. Saudi Arabia, Germany, and other countries face similar geography. The only viable transmission solution at that scale is HVDC.

HVDC transformers are among the most technically demanding power transformers built. The winding wire requirements are correspondingly stringent. Every HVDC transformer built in India for the transmission grid requires locally approved components with local content. That regulatory structure, combined with KSH’s sole qualified status, creates a near-captive demand relationship for as long as HVDC transformer manufacturing happens in India.

KSH began supplying against 37 HVDC transformer orders in Q3FY26, 11 from BHEL and 26 from another existing customer. At current volumes, HVDC is a small revenue contributor. The significance is not the current revenue but the pipeline it points toward. As India’s renewable evacuation infrastructure scales, HVDC orders will grow, and KSH is the only domestic supplier positioned to capture that growth.

The China-Plus-One Export Opportunity

Western transformer OEMs have historically sourced winding wire components from a mix of domestic and global suppliers. Chinese suppliers have been a meaningful part of that mix. The tariff and geopolitical environment has changed that calculation substantially.

China faces a 34% tariff on the fabrication component of winding wire exports to the US. India, following recent trade developments, is expected to land at 18 to 25%. In markets where the duty rate is equalized, KSH management states plainly that they are cost competitive with Chinese manufacturers, citing Japan as a live example where both compete.

The US transformer manufacturing base is expanding aggressively, with companies like Hyosung, Virginia Transformer, and others investing in new domestic capacity specifically to address the supply shortage. All of these facilities need winding wire. With Chinese suppliers at a structural tariff disadvantage and Indian suppliers improving in relative competitiveness, KSH is positioned to grow its US presence from approximately 9% of annual revenue today.

India: A Multi-Decade Infrastructure Cycle

The National Electricity Plan targets 191,000 circuit kilometers of new transmission lines and 1,270 GVA of new transformation capacity by 2032, with total transmission sector investment expected at Rs 9.16 trillion. Power generation capacity is expected to double by March 2032. Non-fossil power capacity is targeted to grow from 217 GW to 613 GW between FY24 and FY32.

Every element of this expansion requires transformers at multiple voltage levels, and transformers require CTC, PICC, and high-spec winding wire. The large transformer segment, which is where KSH’s specialised products are used, is expected to grow at 11.5% CAGR versus 8.6% for small and medium transformers. KSH’s product mix is aligned with the faster-growing end of this market.

Railways - Traction transformers for electric locomotives and metro rail systems use CTC windings. India is expanding metro networks across multiple major cities with a planned doubling of operational metro lines over the next four to five years. KSH carries RDSO approval for CTC used in locomotive traction transformers, which means it participates in every new locomotive order placed by Indian Railways for electric traction.

Data centers - India’s data center capacity is projected to grow from 1.4 GW to 9 GW by 2030. This trajectory creates direct transformer demand at the grid connection level and dry-type transformer demand within the data center facilities themselves. KSH lists data centers as an end-use industry in its investor materials, and management referenced AI data center power demand as a structural tailwind in their most recent earnings call.

EVs - India’s EV penetration is projected at 10 to 12% by FY26 and 30 to 35% by FY30. Two-wheeler EV volumes are already flowing through KSH’s standard wire lines. Four-wheeler volumes will follow as domestic production scales. The PEEK wire development positions KSH for the premium end of this market once 800-volt architectures become more common in Indian manufacturing.

Value

The business model has three features that collectively make the quality of earnings meaningfully better than the reported margins suggest.

The copper pass-through structure eliminates commodity price risk from the equation entirely. Reported EBITDA margins of 6 to 7% look thin, but they compress and expand mechanically with copper prices while actual profitability per unit of output stays constant. EBITDA per ton is the right measure. At Rs 66,044 per ton in 9MFY26, up 32% from Rs 50,133 in 9MFY25, the underlying earnings quality is improving substantially.

The certification and approval stack creates high switching costs on both sides of the relationship. KSH does not lose customers easily because re-qualifying a new supplier with PGCIL, NTPC, NPCIL, or a global OEM like Siemens takes years. Over 94% of revenue comes from repeat customers. Five of the top ten customers have been buying from KSH for more than a decade.

The balance sheet has improved dramatically following the IPO. KSH repaid Rs 226 crore of long and short-term debt in December 2025, including the full Supa Phase 1 term loan. The interest cost that was dragging PAT in Q3FY26 includes Rs 2.7 crore of non-recurring Supa loan interest that is now gone, and a Rs 1.6 crore one-time labour code implementation charge. Stripping those out, Q3FY26 underlying PAT was ahead of what the reported number suggests.

One area for improvement is working capital days, currently 75 to 80 days versus peers closer to 50 to 60 days. The gap is primarily the creditor side. KSH currently pays copper suppliers in approximately 5 days, while peers like Precision Wires take 50 to 60 day credit terms. The CFO has committed to shifting toward credit-based copper procurement, and this improvement was already visible in Q3FY26 with creditor days improving by 4 to 5 days sequentially.

The product mix is a continuous tailwind for per-ton economics. Specialized wires are approximately three times more profitable per ton than standard wires, directly from management. With specialised at 75% of revenue and growing, with HVDC volumes building, and with PEEK wire development underway, the mix is structurally moving in the right direction.

Growth

This is primarily a volume story.

KSH sold 23,345 MT in FY25. Full year FY26 guidance is 28,000 to 29,000 MT. On the 43,445 MT base that exists today, management targets 80 to 85% utilization in FY27, implying roughly 35,000 to 37,000 MT. Phase 2 of Supa adds another 15,600 MT by end of Q4FY27, bringing total installed capacity to 59,045 MT. At 80 to 85% utilization of the full base, addressable annual volume is 47,000 to 50,000 MT. That is more than double FY25 volumes within approximately two years of the current date.

HVDC is the growth trigger that would improve both volume and per-ton economics. The pipeline of Indian HVDC projects is large and growing, and KSH has no domestic competition. As this pipeline converts to orders over the next two to three years, HVDC volumes at KSH could grow substantially. Management has acknowledged that a meaningful shift in HVDC order flow would move the per-ton economics above the current guided range, given it is the highest value-add product in the portfolio.

Export re-acceleration is another lever. The Supa capacity constraint had been the primary limit on export volume growth. With that constraint now removed, export growth has already accelerated to 37% year on year in Q3FY26. The improving US tariff environment, from a 54% fabrication duty to an expected 18 to 25%, structurally improves KSH’s competitiveness in its largest export market. US revenue was 9% of total in FY25. Growing that meaningfully as US transformer OEMs expand their own manufacturing capacity is a credible near-term growth driver.

EV and PEEK wires represent optionality that is not yet in the numbers. Two-wheeler volumes are live today. Four-wheeler volumes are 12 to 18 months away from being material. The PEEK wire development under the HPW Metallwerk licence positions KSH for the premium end of the EV motor market where margins are meaningfully higher than standard round wire. The Indian EV market is projected to grow at a 66% CAGR from 2022 to 2029. Even a small share of winding wire content across that volume growth creates a significant demand layer over time.

Valuation

Assumptions:

• EBITDA: I derived FY28 figures using EBITDA/ton. Management has guided Rs 66,000 per ton as a sustainable level. With the new Supa facility ramping and total installed capacity reaching 59,045 metric tons by end of FY27, the company is targeting 80 to 85% utilisation on that expanded base. At approximately 47,000 metric tons of volume in FY28 and Rs 60,000 per ton, a modest discount to the current run rate the business generates roughly Rs 280 crore of EBITDA.

• EV/EBITDA: Assumed the current multiples will sustain.

Risks

• The moat might not be durable

  The way I see it, the core moat is the approval stack, but these approvals are not a permanent barrier. Any well-capitalized Indian manufacturer who decides to go through the qualification process, albeit tedious, can eventually get there (or maybe already in the process). So my question is, what stops a well-funded new entrant from deciding to pursue the same certifications?

• The HVDC timeline is uncertain

  KSH’s sole qualified status for HVDC transformer CTC is the most defensible part of the investment case. The problem is that HVDC project timelines in India have historically slipped. If HVDC order flow is slower than the renewable energy buildout suggests, the per-ton economics improvement that investors are implicitly pricing in will take longer to materialise.

• Capacity utilisation at Supa needs to keep ramping

  The entire valuation case assumes Supa reaches 80% utilisation on the 43,445 MT base and that Phase 2 follows on schedule by end of Q4FY27. Supa hit 50% utilisation in its first quarter, which is a good start. If utilisation ramps more slowly than expected, the operating leverage story gets pushed out and the Rs 280 crore FY28 EBITDA assumption becomes optimistic.

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

Senores makes generic drugs. Generic drugs are copies of branded drugs whose patents have expired. The active ingredient is the same, the clinical effect is the same, but the price is lower. To sell a generic in the US, a company needs FDA approval through a filing called an ANDA, or Abbreviated New Drug Application. Getting that approval is the hard part. Once you have it, you have the legal right to manufacture and sell that drug.

Most large Indian generics companies, your Sun Pharmas and Dr. Reddy’s of the world, chase the biggest markets: drugs with billions of dollars in annual sales where even a small market share is meaningful. Senores has deliberately gone the other way. It focuses on drugs with total addressable markets in the $5-200 million range in the US. These are products large enough to be worthwhile but small enough that the giants do not prioritize them. In that space, Senores often competes against two or three players rather than twenty, which means better pricing and better margins.

The business has four segments:

1. Regulated Markets (65% revenue in 9MFY26)

Within regulated markets, Senores runs two businesses side by side.

The first is its own ANDA portfolio. Senores identifies a drug it wants to make, develops the formulation in its R&D labs, files for FDA approval, and once approved, sells the drug through a marketing partner. Senores manufactures, the partner sells, and revenue is split through a combination of an upfront licensing fee, a transfer price on the product, and a share of profits. As of December 2025, Senores has 46 approved ANDAs covering 137 different strengths, of which 18 are currently being sold. The remaining 28 are approved and waiting to be launched. Partners include names like Alkem, Dr. Reddy’s, Sun Pharma, Lannett, and Jubilant Cadista.

The second business within regulated markets is CDMO and CMO. The key asset enabling this is the Atlanta, Georgia manufacturing facility, which is USFDA approved, DEA licensed for controlled substances, and BAA compliant for US government supply. Most Indian pharma companies manufacture in India, which disqualifies them from US government contracts under the Buy American Act. Several large Indian companies, including Jubilant, Alkem, and Wockhardt, closed their US plants in recent years because of the cost. Senores kept its Atlanta plant open and has been picking up the CDMO work that those departing companies left behind. Currently 16 CDMO products are commercial with 16 more in the pipeline. The current revenue split between own ANDAs and CDMO is roughly 55:45.

The regulated markets segment runs at a 40% EBITDA margin. Own-product ANDAs carry 4-6 percentage points higher gross margins than CDMO work, which is why management is targeting a shift to 65:35 in favour of own products by FY27.

2. Emerging Markets (21% revenue in 9MFY26)

This segment operates through a subsidiary called Ratnatris Pharmaceuticals, out of a WHO-GMP approved facility in Chhatral, Gujarat.

The logic here is different from the US business. In Africa, Southeast Asia, Latin America, and Central Asia, the pharmaceutical market is growing rapidly as incomes rise, healthcare infrastructure builds out, and disease burden from both chronic and infectious conditions increases. These markets need affordable drugs, and many of them need the kind of specialty and complex products that local manufacturers cannot make.

The emerging markets portfolio includes drugs like Apixaban, a blood thinner used for stroke prevention; Sacubitril plus Valsartan, a heart failure treatment; Sugammadex, used to reverse muscle relaxants during surgery; and Ferric Carboxymaltose, an iron deficiency treatment given by infusion. These are specialty molecules that require real formulation capability and regulatory navigation to bring to market affordably.

The portfolio currently has 450 registered products across 40-plus countries, with another 858 products under registration. The top markets include the Philippines, Peru, Ghana, Nigeria, Myanmar, Guatemala, and Kenya. Revenue is generated through four models: selling through local distributors, partner-to-partner bulk supply, CDMO arrangements for regional companies, and Senores’ own branded products in select markets.

3. India Branded Generics

This is the domestic hospital business. The segment supplies critical care injectables directly to hospitals across India. 60 products have been launched. This segment also gives the company a domestic revenue base that is insulated from currency movements and US regulatory risk.

4. API and Others (7% revenue in 9MFY26)

The smallest part of the business. Senores entered the API, or active pharmaceutical ingredient, segment through an acquisition called Ratnagene Lifesciences, and has now commenced manufacturing at a greenfield API plant in Gujarat.

Currently 17 APIs are commercialized, sold primarily to the domestic market and SAARC countries. Management has been clear that API is a strategic play for backward integration rather than a standalone revenue driver.

I’ve used the TVG (Theme, Value, Growth) framework to analyze the business:

Theme

The cleanest way to understand the tailwind here is to look at one number: the US saw 114, 160, and 156 new drug shortages in 2021, 2022, and 2023 respectively. These happen because a large number of generic drugs are no longer profitable enough for manufacturers to produce reliably. Margins got squeezed over years of price erosion, and companies rationalized their portfolios away from the difficult or low-volume drugs.

The FDA created a mechanism to fix this called Competitive Generic Therapy, or CGT. When the FDA determines there is inadequate competition for a particular drug, it grants the first approved filer 180 days of marketing exclusivity. That policy is what Senores has systematically built its pipeline around.

Senores hunts for drugs that are likely to receive CGT designation, files first, gets approved, and captures that exclusivity window before the market opens up. The company claims a 40% CGT hit rate on its approvals versus a roughly 29% industry average, with 75% of those approvals securing the actual exclusivity period. Of the pipeline ANDAs still in development, 37 strengths are flagged as potential CGT candidates.

The second layer of the theme is what happens when Indian pharmaceutical companies shut down their US manufacturing plants. Jubilant closed its US facility in 2024. Alkem closed in 2023. Wockhardt and Aurobindo did the same in 2022. The reason is simple: manufacturing in the US costs 55-60% more than manufacturing in India. For most companies running commoditized generics, that math does not work.

But for certain products, US-based manufacturing is not optional. US government contracts under the Buy American Act require domestic manufacturing. The DEA requires licensed US facilities to handle controlled substances. Senores kept its Atlanta plant open precisely because it serves these high-value, protected segments. When the larger players walked away, Senores became the outsourcing partner for some of those same companies. That is how the CDMO business was built.

The third layer is simpler: the US pharma market is growing at 7.3% annually and will cross $1 trillion by 2028. Generic adoption is increasing as healthcare costs rise and payers push physicians toward lower-cost options.

Value

A year ago, the Senores balance sheet had an obvious problem. Operating cash flow was negative. The company was growing fast, deploying IPO proceeds, building the CDMO business, and absorbing working capital simultaneously. The FY25 operating cash flow was around negative Rs 46 crore.

That number has been turning around. For 9MFY26, operating cash flow was positive at Rs 51 crore. For Q3FY26 alone, it was Rs 19 crore. It is still a business in transition rather than a fully matured cash generator, but the direction is clear.

Working capital has improved sharply. Trade receivable days fell from 228 in FY24 to 108 in FY25. Inventory days came down from 69 to 43. The net working capital cycle currently sits at around 90-94 days. Part of why this is manageable is the way the company manufactures: Senores produces largely against confirmed purchase orders from partners, meaning finished goods inventory rarely exceeds 15 days. There has been no inventory write-off in the company’s history.

The margin structure is worth understanding because it varies meaningfully by segment. The regulated markets business runs at 40% EBITDA margin. Within that, own-product ANDAs have 4-6% higher gross margins than CDMO work. As the mix shifts toward own products over FY27, there is a straightforward path to margin improvement within the segment without needing topline to accelerate.

The emerging markets segment has been the transformation story of the past year. EBITDA margin in that business went from roughly 1% in Q3FY25 to 13% in Q3FY26, a 1,200 basis point improvement in twelve months. Management’s target for FY27 is 18-22% EBITDA margin in emerging markets, driven by a better product mix and more direct-export dollar revenues replacing lower-margin domestic distribution. If that plays out, the blended company margin could move meaningfully higher.

The blended EBITDA margin for the company is already at 29% for 9MFY26 against 25.6% in the same period last year. Management guided around 30% for the full year.

On capital intensity: the business is not heavy. The major capital deployments have been acquiring FDA-approved facilities rather than building greenfield. Apnar Pharma, acquired in January 2026 with 75% stake for an enterprise value of about Rs 91 crore, came with USFDA, MHRA, and Health Canada approvals, 5 existing ANDAs, and significant expansion capacity. That is a lot of infrastructure for a modest price. The annual maintenance and growth capex going forward is guided at Rs 50-100 crore, modest given the revenue base.

The one point that deserves a flag here is the capital position. The company raised Rs 500 crore in the IPO, of which Rs 362 crore had been deployed by December 2025 and Rs 138 crore remained. But the Chairman acknowledged in the Q3 concall that an additional Rs 75-100 crore may be needed over the next 12 months for ANDA acquisitions and working capital. The Apnar acquisition also came with Rs 76 crore of debt assumed. A small promoter share pledge was done during the quarter to consolidate some borrowings.

Growth

The growth story at Senores comes down to one simple idea: today’s revenue is generated by 18 commercialized ANDA products. Over the next two years, that number expands to 46+ products.

The portfolio grew 4x in a year, driven partly by organic R&D and partly by acquisitions from Dr. Reddy’s, Breckenridge, and now Apnar. Of those 46 ANDAs, 18 are currently generating revenue. The other 28 are approved by the FDA and ready to sell, but have not yet been launched. These 28 cover 102 different product strengths, and Senores has identified a total addressable market of $550 million-plus for that unlaunched pool. Management expects 3-5 launches per quarter over the next 6-8 quarters.

Another 22 ANDAs with 52 strengths are still in development, which keeps the pipeline running beyond the current unlaunched pool. This means the launch cadence has fuel for at least three years from today’s position.

The Apnar acquisition adds a separate, parallel track. Three of Apnar’s five ANDAs were scheduled for launch in Q4FY26, and management guided Rs 120-150 crore of revenue from the Apnar facility specifically in FY27. The facility also opens UK and Canada markets, since Apnar holds both MHRA and Health Canada approvals. And it enables a cost arbitrage: low-margin products that currently must be manufactured in Atlanta at high cost can shift to the Apnar facility in Gujarat, freeing up Atlanta capacity for higher-margin business.

The controlled substance angle is a separate growth layer. Currently around 15-20% of total revenue comes from controlled substances, which carry higher margins than standard generics because of the DEA licensing requirements. Management said two more controlled substance product launches are planned in FY27.

In emerging markets, the registered product portfolio grew from 237 in Q3FY25 to 450 in Q3FY26, with another 858 products currently under registration. In Q3FY26 alone, 56 new products were approved. Those approvals feed commercialization over the following quarters, and management guided Rs 170-180 crore of emerging markets revenue in FY27 against a current run rate of around Rs 130 crore annualized. A PIC/S approval for the Chhatral facility, when obtained, would also unlock access to Vietnam, South Africa, and other mid-tier markets.

For FY26 overall, management guided 50% topline growth and 100% PAT growth. With 9MFY26 total income already up 65% year-on-year and PAT up 110%, they are tracking ahead on both metrics. Q4 is typically the strongest quarter. The FY26 PAT is expected to land around Rs 100-115 crore. For FY27, management stopped short of a specific number but said 25%-plus growth is the minimum they are working toward, with upside from Apnar and the launch pipeline.

Valuation

Assumptions:

• Topline: 650 cr for FY26, then 25% growth in FY27 and 20% growth in FY28

• NPM: Minimal growth to be conservative

• P/E: Lower end of recent average

Risks

1. ANDA launch execution risk: The 28 approved-but-unlaunched ANDAs look compelling on paper, but each launch requires finding and finalizing a marketing partner, setting up manufacturing batches, bidding for government contracts where applicable, and managing the timing of market entry against any remaining branded competition. If launch cadence slips from the guided 3-5 per quarter, revenue recognition will be delayed.

2. Emerging markets margin credibility: The jump from 1% to 13% EBITDA margin in emerging markets in one year is impressive, but it needs to be sustained and expanded to the 18-22% target. Much of the improvement was driven by specific product commercializations that happened to land in Q3. The management acknowledged that the prior quarter had some delayed commercializations. If the product mix shifts or certain registrations take longer to monetize than expected, the margin trajectory could flatten.

3. Ongoing capital requirements and acquisition integration: The company has been growing through acquisitions, Havix in Atlanta, Dr. Reddy’s and Breckenridge ANDAs, Ratnatris for emerging markets, and now Apnar. Each acquisition brings integration work, debt, and execution risk. The admission that Rs 75-100 crore of additional capital may be needed over the next twelve months suggests the balance sheet is being stretched even as cash flows improve. Investors should watch whether the operating cash flow improvement holds up as Apnar comes online and more launches are executed, or whether working capital absorbs the gains again.

Disclaimer: This is not investment advice. I am not a SEBI-registered analyst. Please do your own research before making any investment decisions.

When missiles struck Iran on a Friday night, the instinctive assumption was straightforward: war breaks out, gold goes up. That is not what happened, Gold sold off sharply.

For a year and a half, the dominant macro position had been short US assets, long everything else. Gold was one of the primary vehicles for that trade, and it had performed exceptionally well. When the war started and markets initially concluded, within the first 48 hours, that the United States had decisively won, the entire trade snapped into reverse. People do not sell their losers first, they sell what has gains. Gold had gains, and it went.

There was also a specific physical dynamic at work. The conflict caused widespread flight cancellations from Middle East to Switzerland, meaning dealers could not ship out finished inventory. At the same time, insurance limits on physical gold holdings were maxed out with no additional cover available. To avoid accumulating further exposure they could not insure or move, wholesalers stopped buying back gold from traders and investors entirely. The market effectively seized. A handful of flights have since resumed, and liquidity has partially returned.

Gold also dropped as expectations of the Fed not cutting rates in the near future became clear, given energy-related cost inflation. Also, the expected yield spike for UST10Y that ‘supposedly’ would give a higher risk-free return and strengthen the dollar.

II. The Gold Thesis Has Never Been Stronger

Set aside the war premium entirely and look at the valuation. The cleanest way to frame gold’s price is to ask: what percentage of outstanding foreign-held US Treasury bonds is collateralized by official US gold reserves at market price? In 1980, that figure was 135%. In 1989 (the previous all-time low) it was 20%. The long-run average sits between 40 and 60%. Today, even after the substantial rally of the past two years, that figure is 13 to 14%. Gold would need to rally 50 to 60% from current levels just to return to the worst point ever recorded. On this measure, it remains the cheapest asset on the board.

III. What Actually Happened: The US Miscalculation

The United States went into this conflict expecting Venezuela. Iran is not Venezuela.

The initial market reaction told the story clearly. In the first two days, oil spiked and then collapsed, down $5-6. Markets were pricing a short, decisive US victory. Buy American assets, sell everything else. The narrative from Washington was that Iran’s nuclear program was obliterated and the operation was successful. Then the second and third derivative information began to surface, and the picture changed.

Credible reports emerged that US strategic radar infrastructure across Gulf bases had been heavily damaged or destroyed. More significantly, Joint Chiefs Chairman General Caine was quoted across major outlets essentially distancing himself from the operation, the kind of language that tells you that things had not gone as planned. Secretary Rubio’s language shifted quietly within a week: the stated objective moved from regime change to eliminating missile capability and sinking the navy.

The most consequential development was not the strikes themselves but what the Strait of Hormuz closure revealed about the state of military power. The United States Navy (the most dominant in history) would not enter the Strait. Iranian missiles and drones, costing a fraction of a carrier battle group, had turned a strategic chokepoint into a kill zone. The Mahan doctrine (control the naval chokepoints and you control the world) has been inverted. Chokepoints no longer favor the navy, they favor the missile.

For three to four hundred years, the powers that controlled blue-water navies controlled global trade (think Portugal, Spain, Britain, America). What played out in the Strait of Hormuz suggests that era is over. The United States could not open the chokepoint it has implicitly guaranteed for decades.

The outcome, at best, is a strategic draw. Iran absorbed enormous damage. Its navy was significantly degraded. But it closed the Strait, damaged US infrastructure, forced oil to $100 and beyond, and extracted a ceasefire on terms that left its core deterrent, the will to use missiles, intact.

IV. The Impact

Macro Impact

Oil is the only honest signal in this conflict. Ignore the official narratives from all sides and watch the price of crude. When it rises, Iran is winning. When it falls, the US is winning. For most of the conflict’s duration, it has been rising.

20 million barrels of oil pass through the Strait of Hormuz daily, out of roughly one 100 million barrels supplied globally. That is twenty percent of world supply. Oil does not price on the average barrel, it prices on the marginal one. Removing twenty percent of supply does not raise prices by twenty percent. It reprices the entire one hundred million barrels at the margin. The consequences cascade across every commodity and supply chain that touches energy, which is to say nearly all of them. Urea plants in India and the Gulf were already filing force majeures. Fertilizer disruptions and food prices follow. These second and third derivative effects are barely being discussed.

Treasury markets have registered the stress clearly. The mechanism is straightforward: the world holds Treasuries and needs oil. If oil is disrupted, the world sells Treasuries to bid up oil and food. The oil-importing creditor nations (China, Japan, South Korea, Europe) face a binary: hold their dollar reserves, or get the energy their economies require. Oil-exporting creditors face a different version of the same problem: they can hold Treasuries, or they can buy the weapons they now realize they need. They cannot do both unless the Federal Reserve prints the difference, which is not a scenario that supports holding bonds either.

The dollar’s position is more complicated than it appears. More US oil exports, a plausible outcome as buyers redirect away from the Gulf, would in theory support the dollar. But every time the dollar gets too strong, the Treasury market dysfunctions. Foreigners have borrowed $134 trillion in US dollars. When the dollar rises sharply, they get squeezed on those borrowings and sell what they can and what they can sell, first, is $9.5 trillion in Treasury bonds.

Country Impact

For Gulf states, this conflict has broken something that cannot easily be repaired. The petrodollar arrangement was elegant in its simplicity: Gulf sovereigns would provide the US with security access and recycle their surplus dollars into US financial assets, in exchange for American protection. But when the moment arrived, the protection was not there in the way it was promised. These governments understand what this means for their sovereign wealth funds, which will increasingly need to fund their own defense rather than American capital markets.

India is the most acutely exposed of the major economies. Every $10 increase in oil adds approximately $15 billion to India’s current account deficit. The two-year windfall from discounted Russian crude was largely spent on subsidies rather than saved as fiscal buffer. That room is now gone. Rerouting shipping around the Strait turns a two-to-four-week voyage into an eight-week voyage, requiring roughly double the number of ships. There are not enough ships in the world to absorb that displacement. The RBI faces a genuine dilemma: tighten to defend the rupee and hurt growth, or let the currency absorb the pressure and accept higher inflation. India’s starting inflation of around 1.5% provides some leeway. The longer the Strait remains closed, the faster that leeway is consumed.

South Korea and Japan source 50-60% of their oil through the Gulf. Both are already exhibiting the stress: Korean equities sold off sharply in the early days of the conflict, and Japan’s position is compounded by its pre-existing monetary trap.

V. The Real Winners: Russia and China

Russia won the moment the conflict began. Russian crude became more globally acceptable and more valuable overnight. The most telling signal came from Scott Bessent, who announced a temporary sanctions waiver on Russian oil (even as the Washington Post was reporting that Russia was actively helping Iran target American assets!). That is a government acknowledging, in real time, that it cannot afford to lose access to Russian supply. Who has the leverage in commodities? The answer was made visible for anyone watching.

China’s response will be characteristically indirect. It will condemn the strikes, call for dialogue, and say nothing further for several weeks. Then, quietly, rare earth flows will slow. Shipments of critical materials needed for interceptor missiles and advanced defense systems will encounter administrative delays. Treasury and trade officials will describe severe supply constraints. That is how you will know China has responded. The US cannot produce the critical materials for Tomahawk missiles domestically. The idea that China will supply those materials so the US can fire them into one of China’s largest oil suppliers is not a serious proposition.

The broader strategic picture is this: Iran, Russia, and China form a commodity and manufacturing bloc that the US-led financial system cannot easily coerce. Take Iran’s oil offline and you need Russia running full. Go after Russia and you need Iran’s cooperation. Pick a fight with either and you blow up the oil market. Blow up the oil market and you blow up the Treasury market. Blow up the Treasury market and you blow up the government’s ability to fund itself.

VI. Investment Implications

Gold

The short-term technical flush changes nothing about the fundamental position.

What the conflict has accelerated is the pace of de-dollarization at the reserve asset level. Foreign governments that watched the US freeze Russia’s reserves in 2022 already had reason to diversify. Governments that have now watched the US assassinate a sitting head of state and demonstrate that Treasuries and dollar-denominated assets can all be weaponized or frozen have even more reason. Gold is the only major reserve asset with no counterparty, no maturity, and no dependency on any government’s credibility.

The one area of nuance is gold and silver miners. Roughly 25% of their input costs are energy-related. Sustained high oil prices pressure margins at the producer level even as the metal price itself holds.

Equity Themes

The most important equity reorientation coming out of this conflict is from the new economy to the old one. Manufacturing, supply chain infrastructure, inventory management, industrial capacity, these are where capital will flow as the world reprices the cost of just-in-time global trade. The conflict has demonstrated what happens when the physical supply chains that underpin the paper financial system are disrupted.

Defense is the obvious beneficiary, and the obvious trap. Defense spending will increase. The US government has already requested an additional $200 billion. Trump has summoned defense company heads and demanded they move faster. But they cannot move faster than China allows them to. The critical materials for interceptor missiles, advanced radar systems, and precision munitions flow through Chinese supply chains. Defense stocks price the spending but do not price the supply constraint. That gap is worth understanding before assuming the trade is straightforward.

Energy self-sufficiency is another interesting theme. There are effectively two countries in the world that sit in this category: the United States and Russia. America’s domestic production make it structurally insulated from what is happening to Korea, Japan, and India right now. US energy producers benefit not just from higher prices but from the redirecting of global demand toward supply that does not pass through a contested chokepoint.

I would like to credit most of this work to my teachers Ritesh Jain, Luke Gromen, David Lin, Peter McCormack, Danny Knowles, and Neeraj Bajpai. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

VA Tech Wabag operates across the water value chain, but the simplest way to understand the company is this: it builds and runs the infrastructure that turns unusable water into usable water. That can mean cleaning sewage, producing drinking water, recycling industrial wastewater, or removing salt from seawater. They work across the full lifecycle: design, procurement, construction, installation, commissioning, and long-term O&M. It also executes projects under formats like EPC, DBO, O&M, BOOT, and HAM, depending on client requirements.

1. Municipal Water & Wastewater - This is the largest part of Wabag’s business. Here, the company works with governments and utilities to build and operate drinking water and sewage treatment infrastructure for cities. Municipal projects made up around 82% of the order book in 9M FY26, showing how central this segment still is.

2. Desalination - Desalination is one of Wabag’s most important and specialized businesses. This involves taking seawater or brackish water and converting it into freshwater. Think of it as manufacturing fresh water out of salt water, especially relevant in water-stressed coastal regions and the Middle East.

3. Industrial Water & Wastewater - Another major leg is industrial water. Factories often need highly purified water for production, and they also generate wastewater that must be treated before discharge or reuse. Wabag provides solutions such as effluent treatment, ultra-pure water, recycle/reuse, and ZLD systems across industries like oil & gas, power, steel, fertilizer, and food & beverages. This business is becoming more interesting as the company enters newer areas like solar manufacturing, semiconductors, green hydrogen, and Bio-CNG.

4. O&M - Wabag also operates plants. That matters because O&M revenue is more recurring and less lumpy than pure EPC. In 9M FY26, O&M contributed 18% of revenue, but it already makes up 36% of the closing order book, which suggests the future mix is becoming structurally better.

I’ve used the TVG (Theme, Value, Growth) framework to analyze the business:

Theme

Water is a structural theme: Cities need drinking water systems, sewage needs to be treated before it is discharged, and industries need process water to run factories. In coastal and water-stressed regions, desalination is often a necessity. That is what makes VA Tech Wabag interesting.

The Indian side of this theme is straightforward. The country still has a large wastewater treatment and reuse gap, and policy thinking is increasingly shifting from simply disposing sewage to treating it as a resource that can be reused for industrial and non-potable applications. NITI Aayog’s work on urban wastewater and treated wastewater reuse makes that clear. This is exactly the type of backdrop that supports Wabag’s core municipal business in drinking water, sewage treatment, and reuse projects. Management also indicated that tendering momentum in India continues across both municipal and industrial water, supported by EPC, DBO, and selective PPP opportunities.

The international theme may be even more powerful. In the Middle East, desalination and wastewater reuse are strategic priorities because water scarcity is structural. Saudi Arabia continues to push reuse and long-term water security under its national planning framework, and the country remains one of the most important desalination markets globally. Wabag is already seeing this on the ground, and management called the Middle East its next key growth engine and said the region presents sustained opportunities in desalination, wastewater treatment, and reuse. That is backed by real projects already in hand, such as Yanbu and Al Haer in Saudi Arabia.

The more interesting shift, though, is that the water story is widening beyond cities. New industries are becoming water-intensive in ways many investors do not fully appreciate: Semiconductor manufacturing requires ultra-pure water, solar cell manufacturing needs ultra-pure water, effluent treatment, and ZLD systems, and green hydrogen projects require purified water and, in many regions, associated desalination or reuse solutions. The IEA’s hydrogen work shows electrolysis capacity has been scaling rapidly, while semiconductor industry materials show how central water and reclaim systems are to plant operations. Wabag is already positioning itself for this layer of demand. The company has won the RenewSys order for Ultra-Pure Water, ETP and ZLD, and management said it is in active discussions around semiconductors, solar manufacturing, hydrogen, and related applications.

So the broader theme here is that the world is moving toward a much more complex water economy, one where cities need better sewage and reuse systems, coastal regions need desalination, and new industries need highly specialized treatment infrastructure. That creates a much wider opportunity set than the market may assume.

Value

Water EPC companies often trade at a discount for good reason. Order books can look impressive on paper, but collections are weak, working capital gets trapped, debt builds up, and execution quality varies. Wabag’s recent numbers suggest it is increasingly moving away from that stereotype. In 9M FY26, revenue grew 18.3%, EBITDA grew 19.9%, PAT grew 23.7%, EBITDA margin stayed at 13.7%, and the company reported net cash of ₹891 crore, or about ₹1,006.5 crore excluding HAM debt. It has also remained net cash positive for 12 consecutive quarters.

Gross cash stood at around ₹1,080 crore, interest income was higher than interest expense, and management said free cash generation for 9M FY26 was about ₹300 crore. Net working capital days improved further to 101 days, continuing a multi-year trend of tighter execution and better collections.

The more important point, though, is that order book quality matters more than order book size. Wabag’s closing order book stood at ₹16,342 crore, but the real signal is in the composition. The current revenue mix is still roughly 82% EPC and 18% O&M, while the closing order book is 64% EPC and 36% O&M. Also, international orders are increasingly contributing to the order mix. The backlog is structurally better than the current P&L mix.

That matters because O&M revenue is more recurring and less lumpy than one-time EPC execution. Management has been quite clear that it wants O&M to move beyond 20% of revenue over time, and it also repeatedly emphasized that international business is better not just for growth, but for working capital, collections, and overall cash flow quality.

There is also a strong quality angle here. Over the last five years, management highlighted EBITDA CAGR of 19% and PAT CAGR of 30%, while return ratios have improved to roughly 19% RoCE and above 15% RoE.

Another subtle value point is that Wabag remains relatively asset-light. It is not pouring capital into factories or commodity manufacturing lines. Growth comes from engineering capability, execution depth, process know-how, and bidding discipline rather than heavy fixed-asset intensity. Management also stressed that most projects are backed by multilaterals, sovereign counterparties, or letters of credit, which lowers the receivables and payment-risk profile versus what investors might normally fear from infra names.

On competition, Wabag argued that in the kinds of projects it is choosing to pursue, namely high-technology desalination, recycle and reuse, industrial water, and complex wastewater treatment, competition is limited and is mostly international rather than domestic, with the main rivals being European players such as Spanish, Israeli, and French companies rather than Chinese firms. Management explicitly said that Chinese companies are generally not the main competition on the technology side. Where they do show up, it is more often as construction partners or subcontractors rather than direct competitors. The broader point management was making is that Wabag is trying to stay away from commoditized, purely low-bid water EPC and instead focus on specialized water projects where technical qualification matters, competition is narrower, and pricing discipline is better.

Growth

The growth story in Wabag is about a better order book. Management has guided for 15–20% medium-term growth, and current execution is already tracking in that range, with 9M FY26 revenue up 18.3% YoY. More importantly, the backlog now stands at ₹16,342 crore, giving the company strong visibility for the next few years. What’s better is the mix: the closing order book is 64% EPC and 36% O&M, versus the current revenue mix where O&M is only 18%. Geographically too, the business is becoming more balanced, with the order book split roughly 52% India and 48% overseas, while international operations already contributed 50% of 9M FY26 revenue.

Management repeatedly emphasized that international projects are better for margins, cash flow, and working capital, while O&M is more recurring and less lumpy than plain EPC. The Middle East in particular is emerging as the next major engine, with projects like Yanbu and Al Haer already under execution, and management seeing sustained opportunities across desalination, wastewater treatment, and reuse in the GCC. Also, given the ongoing conflict, Middle Eastern countries have definitely learned their lesson on heavily depending on a few desalination plants! Africa is another medium-term lever, supported by multilateral and government-backed water projects, while Southeast Asia and Europe offer selective expansion in high-technology water treatment.

There is also a second layer of growth: Wabag is beginning to participate in what management calls Future Energy Solutions, which includes ultra-pure water, effluent treatment, and ZLD systems for sectors like solar manufacturing, semiconductors, green hydrogen, compressed biogas, and data centers. The RenewSys order is the clearest proof point so far, but management also said they are in active discussions with hydrogen developers, semiconductor manufacturers, and data-center players. This means Wabag is also becoming a picks-and-shovels provider to newer industrial themes where water treatment is mission-critical.

Put together: a large backlog, improving mix, rising international exposure, growing O&M share, and a new optionality layer in high-tech and energy-transition water infrastructure.

Valuation

Assumptions:

1. Revenue growth of 15% YoY until FY27

2. Increasing EBITDA Margin due to Future Energy Solutions, increasing international presence, and higher potential contribution of O&M.

3. 3 year median EV/EBITDA for base case

4. Net Debt to stay roughly the same

Risks

1. Execution and working-capital risk - Even though Wabag’s balance sheet is much cleaner now, it is still an EPC-led business executing large, multi-year projects across multiple geographies. That means delays in construction, customer approvals, billing milestones, or collections can still affect quarterly revenue recognition and cash conversion. Not to mention that most of their business is still municipal.

2. Order inflow timing can remain lumpy - The longer-term order pipeline looks healthy, but management was very deliberate in refusing to give short-term order guidance and repeatedly framed the business on a multi-year basis rather than a quarter-to-quarter basis. That usually means investors should be prepared for uneven order inflow or temporary pauses in awards, even if the broader pipeline remains intact. This is especially relevant in domestic municipal projects, where tender timing and project approvals can slip.

3. Some of the newer growth verticals are promising, but still early - Management itself made clear that parts of this opportunity set are still in discussion or early conversion stages. In hydrogen, the key issue is commercial viability; in semiconductors, fabrication activity in India is taking time and much of the near-term activity is lower-water-intensity assembly. So while these areas add upside, they should not yet be treated as fully proven growth engines.

Disclaimer: Not a buy/sell recommendation. Please do your own research. I am not invested right now but may decide to do so in the future.

Before diving in, a brief note. I typically write about markets, commodities, liquidity, and capital flows, but this piece is slightly different. AI has real-world macro consequences, and ignoring it would mean missing a structural shift that increasingly shapes capital flows, policy decisions, and market outcomes.

Section I: US vs China Is an Electricity War

(Ritesh Jain & Luke Gromen)

Electricity as the Real Battleground

The U.S. vs. China contest is often discussed through narratives like Chips, export controls, technology leadership, ideology. That framing misses the underlying variable that determines whether any of those ambitions can scale: electricity.

For decades, the Western growth model has been services-led. Services are high margin but low power intensity. You can run a large services economy without building a massive electricity surplus. Manufacturing is different, and so is AI. Once an economy shifts toward large-scale industrial policy and large-scale compute, the limiting input stops being capital and starts being power.

This is why electricity is treated as the macro theme of the decade. It is the constraint that sits underneath reshoring, defense production, grid-scale data centers, and any attempt to rebuild industrial capacity. If you cannot generate and transmit power cheaply at scale, you cannot win the AI race. You can only talk about winning it.

Why China Might Be Ahead

On this basis, the claim is that China has already built the winning platform. It has secured the inputs needed to deploy both at national scale. Across the measurable drivers of industrial dominance, China screens stronger.

Start with labor and engineering depth: Tesla employs roughly 85,000 people in total. BYD alone is said to employ roughly 125,000 engineers. Whether the exact figures are directionally perfect is less important than what they intend to capture. China is operating at a different scale of engineering labor.

Power economics: The cost of building one gigawatt of nuclear capacity in China is roughly 1/6th of the cost in the United States. In a world where AI and industrial policy are power-hungry, the country that can build generation cheaply and quickly has a durable edge.

China’s only long-duration uncertainty in this framework is not is demographics. Whether there is enough population depth in 50 years to sustain the machine it has built. Everything else, power cost, engineering scale, industrial capacity, is presented as already in place.

The Deflation-to-Inflation Regime Shift via China’s Capacity Cuts

For the past two decades, China’s overcapacity was one of the most powerful deflationary forces in the global system. Excess supply suppressed global prices and allowed the West to consume cheaply while hollowing out domestic production.

That dynamic is now changing.

China is already cutting steel capacity, with further reductions planned. If those cuts materialize meaningfully, global steel prices will rise. Construction costs in the West will increase. Across industries, China is reducing capacity after years of overproduction.

If China’s overcapacity helped anchor global disinflation for twenty years, the withdrawal of that capacity points in the opposite direction. The next decade may not look like the last one. Instead of importing deflation, the world may import higher input costs. That shift links back to electricity. Industrial production cannot simply relocate without cheap and reliable power.

The US Electricity Stress Story as a Race Against Time

The U.S. side of the ledger looks more fragile.

The distribution network is roughly 25 years old. Grid expansion has lagged demand growth, and data centers are already clustering in hotspots like Northern Virginia where utilities are struggling to meet contracted load. In the coming years, capacity may already be insufficient for projected AI demand.

AI is uniquely power intensive. Large training clusters require continuous base-load power. Unlike many legacy industries, they cannot tolerate interruptions. The more compute is centralized, the more acute the pressure on local grids. If AI demand were to scale to current projections, it could consume a very large share of existing U.S. electricity production.

This reframes the AI race as a race against time. Governments have effectively regained control in moments like this, and capital shifts from private optimization to national priority. We’ll likely see large-scale energy projects, nuclear restarts, grid buildouts, and electrification drives. The outcome may be uncertain, but the spending is not.

Investment Frame

The debate over which AI model wins is secondary. Every model requires electricity, every training cluster requires generation and transmission, and every industrial reshoring plan requires power.

The principle articulated here is simple: do not invest in what is obvious on the surface, rather in the derivative effects. Electrification, grid expansion, transmission equipment, and generation capacity represent the second- and third-order plays on the AI and industrialization cycle.

Even if the United States fails to fully close the gap with China, the journey will involve enormous capital deployment. Historically, more money is often made during the build phase than at the final equilibrium.

Section II: China’s AI Model and the Coming Margin Compression

(Dr. Michael Power & The Monetary Matters Network)

China’s huge AI potential:

In the U.S., AI is treated as a product. It is built with massive capital expenditure and expected to generate direct monetization through subscriptions, API usage, and enterprise contracts.

In China, AI is increasingly treated as a utility. The analogy is electricity. No one asks how to monetize electricity itself at high margins, the value lies in what electricity enables, like factories, logistics, payments, e-commerce, industrial automation. Electricity is low margin at the generation level but enormously valuable at the application level. The argument is that Chinese AI is following that model.

Fewer than 5% of Chinese AI models charge users in any meaningful way, and even those are niche applications. For most use cases, AI is effectively free. That immediately alters adoption curves.

The U.S. model is closed and monetizable, whereas the Chinese model is open-weight and ecosystem-driven. The closest analogy is Android versus Apple, or Linux versus proprietary systems. Linux runs every top supercomputer globally. Android dominates global handset market share. Apple captures a disproportionate share of profits but in a relatively small set of countries.

The projection is that AI may bifurcate in a similar way. A high-margin, closed ecosystem anchored in the U.S. and a few allied economies. A broader, open, low-cost ecosystem across much of the rest of the world.

The funding model is also different. Take Alibaba’s Qwen. It is not monetized as a standalone AI product, instead, it is embedded across Taobao, Alipay, logistics, and other internal platforms. Those divisions effectively subsidize the model through internal transfer pricing. AI becomes shared infrastructure rather than a profit center.

How China could take the lead in the AI race

What does the world look like in five years if AI adoption scales under a low-cost, open model rather than a high-margin, closed one?

The first adjustment is to redefine what “world” means. Much of U.S. market commentary equates “world” with U.S.-centric ecosystems. In reality, the global market is far larger and more price-sensitive. Adoption does not need to dominate Silicon Valley to dominate globally.

There are already signs of leakage. Startups outside the U.S. ecosystem are increasingly using Chinese open-weight models because they are free and sufficiently powerful. If the software performs at a competitive level and costs nothing, experimentation accelerates.

To analyze the impact properly, hardware and software must be separated. On hardware, Nvidia’s dominance rests on two pillars: cutting-edge nodes and ecosystem lock-in through CUDA. Both are being challenged.

Inside the U.S., hyperscalers are already reducing dependence on Nvidia by developing alternatives. Amazon is building Trainium. Google continues to develop TPUs. Outside the U.S., diversification away from high-cost GPUs is accelerating, driven by cost pressure and strategic necessity.

At the same time, the traditional engine of chip progress (shrinking node sizes under Moore’s Law) is running into limits of physics, material science, and economics. As transistors approach extreme miniaturization, heat, electron tunneling, defect sensitivity, and cost escalate nonlinearly. The incremental performance gain from 3nm to 2nm does not justify the exponential increase in manufacturing complexity.

Instead of competing at 2nm or 3nm, architectural innovation becomes the lever. China’s approach uses larger nodes, roughly 14–18nm, combined through advanced packaging, stacking, and system-level optimization. Performance is achieved through structure rather than pure transistor density, and these “cognitive towers” create scale through coordination.

Chinese AI vs. US AI:

The capital comparison between U.S. and Chinese AI efforts is often framed in absolute numbers: Billions committed to training runs, massive infrastructure projects, and multi-year data center expansions.

Let’s focus on how much performance was achieved per dollar spent.

Even if one accepts higher estimates of capital deployed by Chinese model developers, the ratio relative to U.S. spending remains small. Yet benchmark performance has been competitive. In some cases, incremental releases have climbed to the top of evaluation tables in mathematics, reasoning, and multilingual fluency. External validation, including peer review and independent testing, has not dismissed these claims outright.

The core innovation attributed to DeepSeek and similar efforts centers on training architecture.

A longstanding challenge in large models is catastrophic forgetfulness. As new data is introduced, previously learned information degrades. The solution traditionally has been brute force retraining with large memory and compute budgets. China’s new architectural approach attempts to stabilize memory so that models can continuously incorporate new data without losing prior capability.

If successful, this does two things simultaneously.

First, it reduces the need for repeated large-scale retraining runs. That lowers compute intensity. Second, it allows models to accumulate knowledge over time without resetting or decaying. That changes the economics of scaling.

The more controversial claim relates to memory efficiency. Under traditional assumptions, a chip must allocate near 100 percent of certain memory and compute resources to achieve a given output. The architectural shift suggests that far lower resource allocation may achieve similar results. Whether the ratio is precisely 100 to 7 is less important than the direction. If training efficiency improves by multiples rather than percentages, the assumed demand trajectory for high-end GPUs must be revisited.

This directly challenges the high-capex model dominating U.S. AI discourse.

U.S. players have largely pursued a centralized, capital-intensive approach with massive clusters, concentrated compute, and enormous energy requirements. The thesis assumes that model quality scales with capital intensity and that pricing power will justify the investment.

The alternative path emphasizes distributed intelligence and architectural optimization. Larger-node chips. Stacking and packaging. Efficient memory management. Lower incremental energy requirements.

The difference becomes clearer when comparing spending ratios. OpenAI’s projected commitments, including participation in large infrastructure initiatives, run into tens of billions, whereas Chinese efforts are described as operating at a fraction of that scale while achieving competitive benchmarks. If the efficiency curve bends, valuation assumptions tied to perpetual capex growth weaken.

The question for investors is whether the capital deployed into centralized AI infrastructure will earn its cost of capital in a world where compute efficiency improves rapidly and open-weight models proliferate.

If AI becomes cheaper to train and cheaper to deploy, margins compress. If margins compress, valuation multiples adjust. That is the structural risk embedded in the current U.S.-centric AI thesis.

The Three Assassins of the US AI system:

1. Physics: At extremely small geometries, electrons do not behave cleanly. Transistors are meant to function as binary switches (On or off, Zero or one). But as nodes approach 3nm and below, electron tunneling becomes more common. Electrons pass through barriers that are supposed to block them, increasing heat density and weakening signal integrity. There is a point at which miniaturization yields diminishing stability rather than meaningful performance gains.

2. Material science: At atomic scales, imperfections matter. Protective layers are only a few atoms thick, and a single defect can degrade yield. As chips become denser, the percentage of fully functional dies declines. Lifespans shorten, and the useful life of a high-end chip may even compress to three to five years before performance degradation or obsolescence sets in.

3. Economics: Even if physics and material challenges can be managed, the cost of doing so rises exponentially. Extreme ultraviolet machines are extraordinarily complex and expensive, and the incremental performance improvement from 3nm to 2nm may not justify the additional capital intensity.

China’s response has been to avoid competing at the bleeding edge and instead reframe the problem. Rather than chasing the smallest transistor, it focuses on architecture. Larger nodes combined through advanced packaging. Chips stacked vertically. Systems optimized for parallel processing. Performance derived from coordination rather than density.

Nvidia itself has begun moving in this direction through more advanced packaging and system-level design, acknowledging implicitly that node shrinkage alone is insufficient. But this shift carries margin implications, as larger-node chips and architectural efficiency typically support lower pricing power than monopoly control over the smallest node.

The future of US AI models:

If hardware margins face pressure, the natural assumption is that value shifts to the model layer.

Google appears structurally advantaged as it combines a competitive model with embedded distribution through Search, Android, and its broader product suite. If AI becomes a utility layer rather than a standalone product, distribution becomes decisive. The company that controls where intelligence is deployed captures value, even if the intelligence itself carries thin margins. Google can embed AI across search, advertising, productivity, and mobile at scale, monetizing indirectly rather than charging explicitly.

Microsoft occupies a similar position through enterprise software and cloud, though it remains more directly exposed to infrastructure economics. Its ability to integrate AI into Office, Azure, and enterprise workflows provides insulation, but not immunity, if model-layer pricing collapses.

OpenAI sits in a more fragile position. Its cost structure is large and highly visible and its business model depends on monetizing AI usage directly or through premium enterprise access. If open-weight alternatives become sufficiently competitive and materially cheaper, pricing power erodes and even if demand grows, margins may compress. High capital intensity combined with uncertain monetization creates structural vulnerability.

Anthropic and other independent model providers face an even narrower path. Without embedded distribution or deep internal cash flows, they are dependent on sustained external funding and strategic partnerships. Over a three-to-five-year horizon, consolidation becomes likely. Infrastructure providers such as Amazon, which has strong cloud capabilities but lacks a dominant proprietary model, may look to integrate vertically through acquisition.

Let’s look across the value chain.

Chip suppliers are profitable. Cloud providers remain profitable through adjacent businesses. Large platforms can absorb AI losses within broader ecosystems, but many end users and startups building on AI could be deeply unprofitable. Capital expenditure across the stack assumes that future demand will justify today’s infrastructure buildout.

Over the next three to five years, the likely outcome is structural compression rather than collapse.

A handful of integrated ecosystem players will remain dominant because they control distribution and can monetize indirectly, with several independent model providers consolidating/disappearing.

Counterarguments

The most obvious counterargument is historical. Open systems have often won scale without eliminating highly profitable closed ecosystems. For example, Linux dominates supercomputers and Android dominates global handset share. Yet Apple remains one of the most profitable companies in the world.

The same pattern could emerge in AI.A closed, high-margin ecosystem may persist in the United States and a handful of aligned economies. In those markets, pricing power, enterprise integration, and regulatory alignment can support durable margins. The U.S. “world” remains large, wealthy, and technologically sophisticated.

But the broader “world world” is different. Emerging markets are more price-sensitive, and growth in users and applications is likely to be faster outside the core Western bloc. If open-weight Chinese models are effectively free and sufficiently capable, adoption curves in these markets may tilt decisively in their favor. Sovereign AI projects in countries such as Indonesia signal that price and flexibility often outweigh brand or origin.

The monetization model also changes under this scenario. AI becomes less like enterprise SaaS and more like language or electricity. No one pays to use English. Value accrues to what is built on top of it. If models commoditize, the economic surplus migrates downstream to applications, logistics, fintech, e-commerce, robotics, and industrial optimization.

Two reinforcing dynamics strengthen this counterpoint.

First, memory and architectural efficiency breakthroughs weaken the case for ever-rising compute intensity. Even U.S. firms can adopt open architectural improvements, but adoption compresses margins across the industry. When training and inference become more efficient, fewer high-end chips are required to achieve similar outputs. That does not eliminate demand, but it caps its growth rate relative to current expectations.

Second, data advantages may fragment. U.S. models increasingly rely on synthetic data as high-quality training data becomes scarcer. Distributed or edge-based systems can generate real-time data locally without central aggregation. If open systems enable continuous learning at the edge, they may accumulate practical advantages in applied use cases without requiring centralized control.

The refined outcome: U.S. leaders may survive and remain globally relevant, but with lower margins than currently implied. Open ecosystems may dominate global adoption, particularly in emerging markets, and AI shifts from product to infrastructure layer.

In that world, the central debate is no longer who builds the smartest model. It is who captures value when the model itself approaches zero price.

I would like to credit most of this work to my teachers Ritesh Jain, Luke Gromen, Jack Farley, and Michael Power. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

(Ritesh Jain, Vijay Vaitheeswaran)

Venezuela has re-entered the global conversation with a familiar narrative attached to it: Drugs, instability, humanitarian concern. None of these explain the timing or the intensity of recent U.S. actions. As Ritesh Jain argues, those explanations are largely for public consumption. The real issue is strategic neglect and its reversal.

Over the past decade, the United States allowed its immediate sphere of influence to thin out. China, Russia, and Iran expanded their presence across Latin America quietly, through financing, military cooperation, and energy partnerships. Venezuela became the most visible symbol of that drift. The current response is less about punishment and more about reassertion. This is Monroe Doctrine logic updated for a multipolar world. The “backyard” is not negotiable.

From this perspective, Venezuela was the easiest place to act. It is close, diplomatically isolated, economically fragile, and symbolically powerful. Re-engaging there allows the U.S. to push rival powers out of its immediate neighborhood, demonstrate decisiveness at relatively low cost, and signal that the era of passivity is over. It is a message directed as much at Beijing and Moscow as it is at Caracas.

Oil sits at the center of this strategy, but not in the way it is often framed. The objective is not simply to add barrels to global supply or lower gasoline prices, but to control energy flows in a way that constrains rivals. China dominates critical materials. It does not dominate global oil. Venezuela, along with Iran and Russia, sits on energy resources that matter strategically even if they are not economically attractive at the margin. By squeezing influence in Venezuela, the U.S. indirectly pressures China’s access to energy rather than confronting it head-on.

Why the Strategy Works Politically but Fails Economically

Venezuelan oil is not light, cheap, or easy. It is heavy, carbon-intensive crude that requires specialized extraction and refining. Restoring production from roughly 1 million barrels per day back to the 3-4 million barrels Venezuela once produced would require well over $100bn of sustained investment. More importantly, it would require confidence in property rights and political stability that Venezuela has repeatedly destroyed.

History matters here. The expropriation waves under Chávez and Maduro wiped out decades of trust. Major oil companies exited after bitter legal battles, and only one remains meaningfully exposed. Chevron’s position is unique and should not be generalized, it operates under special dispensations granted by both Democratic and Republican administrations, and today a material share of its production comes from Venezuela. That exception underscores the rule, and no other major is likely to commit fresh capital at scale.

Even if capital were available, economics remain unfavorable. Venezuelan projects require sustained oil prices closer to $80 per barrel over long horizons to justify investment. Current market discussions around $60 oil make that hurdle implausible. For an industry now focused on capital discipline and shareholder returns, Venezuela is uninvestable.

There is, however, a narrow pocket where the strategy does make sense. U.S. Gulf Coast refineries were historically optimized for heavy Venezuelan crude. Before expropriation, that relationship was deep. Any renewed flow would benefit those refineries disproportionately, creating a short-term windfall for specific operators.

Greenland: Strategic Asset or Strategic Error

(Peter Zeinhan, Ritesh Jain)

Greenland has two very different strategic arguments. One views it as an unnecessary provocation that destroys alliances for little gain, the other sees it as an under-appreciated asset.

The first is that the Greenland focus makes little sense. Denmark is arguably one of the most supportive allies the United States has ever had. Despite its small size, it has consistently shown up for U.S. military operations, from the Gulf War to Afghanistan, fully aware that its own geography makes external alliances essential. Greenland itself has never been an obstacle. The United States already has full access to the island for any military or strategic activity it wants to conduct, and Denmark helps pay for that presence. From a narrow cost benefit perspective, this is an extraordinarily favorable arrangement.

Economically, the case is even weaker. Roughly 80% of Greenland is covered by ice, and even aggressive global warming scenarios do not change that reality in any meaningful timeframe. Greenland lacks natural ports, sits in one of the stormiest maritime environments in the world, and would require vast infrastructure investment just to make extraction feasible. Ports, roads, power, housing, and processing facilities would need to be built almost from scratch. The cost would likely run into the trillions. Rare Earths are typically byproducts of other mining processes, not standalone opportunities, and transporting raw material elsewhere would be prohibitively expensive, forcing local smelting and further raising costs.

The second argument starts from a different premise. The Arctic is changing. Ice is melting, opening new shipping routes and increasing the strategic relevance of the region. Russia has been investing heavily in Arctic militarization, building bases and infrastructure along its northern frontier. From this angle, Greenland is the first early warning point for any missile launched from Russia toward North America, making it critical for missile defense and surveillance in a world of renewed great power competition.

There is also a psychological dimension. Europe has consistently underinvested in defense despite repeated warnings. Public pressure and polite diplomacy have failed to change that behavior. The Greenland rhetoric can be read as an attempt to jolt Europe out of complacency by making the costs of dependency more explicit.

Why Percent Returns Are Becoming Meaningless

(Luke Gromen)

For roughly 40 years, the global system rewarded financial engineering over physical production. Control of money printing, capital markets, and dollar plumbing was treated as the primary source of power. Factories were offshored, grids were neglected, refining capacity was allowed to concentrate elsewhere, and balance sheets grew faster than productive capacity. That trade worked as long as physical abundance could be assumed. It no longer can.

The shift arguably began with the 2008 Global Financial Crisis. The response (zero rates, QE, and repeated liquidity backstops) made clear that the supply of “safe” dollar claims, and their long-run real return, are ultimately policy variables. For reserve managers who had spent prior decades treating gold as a legacy holding, the crisis helped flip the equation: official-sector selling dried up and, by 2010, central banks became consistent net buyers. In that sense, gold’s re-monetisation started as a hedge against the gradual dilution of nominal claims rather than a bet on any single scenario.

The current moment can be described as a hard bifurcation between the paper world and the physical world. On one side sit financial claims: Bonds, equities, swap lines, reserves, and percentage returns. On the other side sit factories, power generation, grids, mines, refineries, and labor. For decades these two worlds moved broadly in sync, but today they are splitting apart.

The limitation of paper claims becomes obvious when stress rises. You cannot eat percentage returns. You cannot build a factory with a favorable internal rate of return if you lack steel, copper, power, or skilled labor. You cannot refine rare earths without refining capacity, regardless of how many dollars you can print. At some point, financial abundance runs into physical scarcity, and the latter always wins.

This is why refining capacity has emerged as one of the most critical bottlenecks. Raw materials are only valuable if they can be processed. Control of mines without control of refining is incomplete power. Control of financial assets without control of production is leverage without output.

AI, war, and energy have all exposed this split simultaneously. AI highlights how dependent advanced technology is on power, cooling, water, and specialized hardware. Military conflict exposes the fragility of supply chains and the limits of just-in-time production. Energy transition shows us how long it takes to build physical infrastructure relative to how quickly capital can be allocated on a screen. In each case, the paper world moves faster than the physical world, until it cannot.

This bifurcation also reveals itself through inversion. Luke Gromen often asks investors to look for the dogs that did not bark. If financial power were decisive, why do nations with the deepest capital markets struggle to build grids, refineries, or shipyards on reasonable timelines? If monetary dominance were sufficient, why do sanctions fail to produce desired outcomes when physical supply remains intact? What is absent from the data is often more revealing than what is present.

Debt-heavy systems are fragile because debt presumes stable cash flows generated by physical output. When productivity gains, automation, or geopolitical shocks disrupt that output, the claims stacked on top become unstable. Physical scarcity does not need to be extreme to cause problems. It only needs to bind at the margin.

We are moving to a system where financial claims and physical reality are no longer aligned. As that gap widens, capital will be forced to reprice what actually matters.

Gold as the Response

(Luke Gromen)

Gold is increasingly functioning as a neutral settlement asset in a system where every other major asset represents a claim on someone else’s balance sheet. A Treasury bond is a claim on future fiscal capacity. A bank deposit is a claim on a leveraged intermediary. Equities are claims on earnings that depend on stable demand, policy continuity, and functioning credit markets. Gold is none of these. It has no counterparty, no maturity, and no dependency on policy credibility.

Central banks have recognized this faster than private investors. Over the past several years, reserve managers have been steadily recycling surpluses out of dollar assets and into gold as a response to misalignment. Reserves built for a rules-based, low-volatility world are poorly suited to an environment defined by sanctions, fragmentation, and political risk, and gold sits outside those frictions. On a broad definition, gold already rivals U.S. Treasuries as a reserve asset.

Gold is a 0% yielding bond of infinite duration with finite issuance and no counterparty risk. Let’s consider the system’s extremes: In an inflationary outcome, gold protects purchasing power as nominal claims are diluted. In a deflationary or default-driven outcome, it protects against credit risk as revenues fall faster than obligations. In both situations, gold survives without needing policy support.

The China angle fits naturally into this framework, without requiring a full geopolitical overlay. China runs persistent trade surpluses, which gives it optionality that deficit nations lack. Those surpluses must be recycled somewhere. Belt and Road investments have faced rising scrutiny and constraints, particularly in the Western Hemisphere. Recycling into domestic projects can only go so far before diminishing returns set in. Gold provides a politically neutral outlet that strengthens sovereign, banking, and household balance sheets simultaneously.

Here is Luke’s model for valuing gold: Looking at the market value of America’s official gold position as a percentage of the foreign held treasuries outstanding, it historically has never been below 20% (in 1989). As of recent, it is 14%, implying gold would have to rise another 50% just to reach the previous low. Luke claims the long term average is between 40-60%, suggesting a fair value of roughly $15,000.

Again, these are super long term views. For gold to reach anywhere above $10,000, we would need to see major moves towards a dollar crisis in my opinion.

Silver & Copper

(Ritesh Jain)

Silver and copper sit further down the risk spectrum than gold, but they express the same underlying pressures with greater volatility. They respond not just to monetary stress, but to the physical constraints that monetary systems are increasingly running into.

Silver carries monetary characteristics similar to gold, but with a much smaller market and far higher volatility. Historically, when gold becomes expensive or inaccessible, investors rotate into silver. It is the poor cousin effect. Capital moves down the quality ladder in search of exposure to the same forces at a lower nominal price.

Over long periods, silver has traded at roughly 2% of gold’s value, with brief episodes of excess treated as outliers rather than norms. What is new is the industrial overlay.

Silver is becoming more relevant in energy storage and electrification than is widely appreciated. Battery technology, particularly in next-generation applications for grid storage and electric vehicles, appears to be far more silver-intensive than prior designs. Samsung’s decision to take over a silver mine in Mexico is a signal, companies do not vertically integrate into mining unless they see structural supply risk.

This creates a second demand channel layered on top of the traditional investment cycle. When gold rises, silver attracts speculative and monetary demand. When energy storage and electrification accelerate, silver attracts industrial demand.

Copper is different, it is the physical backbone of the energy transition.

Electrification is happening: Electric vehicles, charging infrastructure, renewable generation, grid expansion, and industrial electrification all require copper in large quantities. There is no substitute at scale, and every version of the transition runs through the same metal.

The issue sits on the supply side. A new copper mine takes roughly 17 years from discovery to production under optimistic assumptions. Some industry estimates stretch that timeline closer to 20+ years once permitting, financing, and community opposition are factored in. Either way, meaningful new supply is not arriving this decade. Existing mines are aging, grades are declining, and capital discipline remains tight.

Copper lagged gold and silver initially because it is tied more directly to real activity than to monetary hedging. That lag is now closing. Ritesh Jain’s view is that this dynamic becomes more visible in 2026, when the gap between energy ambition and material reality can no longer be ignored. Copper does not need speculative excess to rise. It only needs demand to persist and supply to remain constrained.

What This Means for Markets

(Ritesh Jain, Luke Gromen)

The most important market risk in the current environment is not direction, but sequencing. Many of the forces discussed in this memo point in the same long-term direction, but the path to get there is unlikely to be smooth. Order of operations matters.

A short-term deflationary impulse is plausible. AI is beginning to compress costs and displace labor faster than most forecasts anticipated. Early signs are visible in labor-market data, particularly among recent graduates and white-collar roles. Productivity gains arrive immediately, while income adjustment lags.

Debt systems cannot tolerate sustained deflation. Deflation raises the real burden of fixed liabilities while shrinking the cash flows used to service them. As revenues fall and defaults rise, stress migrates from households to corporates, then to banks and sovereign balance sheets. What begins as efficiency-driven disinflation can quickly turn into a credit problem.

The historical pattern is consistent: Deflation leads to defaults —> defaults trigger panic —> panic forces intervention —> liquidity is injected at scale to prevent systemic collapse. The result is often a sharp reversal from deflationary pressure to inflationary or even hyperinflationary outcomes.

This is where inversion becomes essential, it is a practical discipline in environments where linear thinking fails. Most analysis looks at only one side of the ledger: what policymakers intend, what markets are pricing, what narratives emphasize. Far fewer ask how the other side can respond, or whether it already has.

Luke Gromen frames inversion through concrete questions. If actions have been taken, did the other side react? And if so, where is that reaction visible. In recent conflicts, it has been common to conclude that nothing happened because markets did not move and headlines did not escalate. That conclusion is often wrong.

Russia did respond. Coordinated strikes hit Ukrainian energy infrastructure. Power generation and transmission facilities were damaged. Large parts of Kyiv lost electricity and water, heating systems failed during winter conditions, and public transport was disrupted. The city’s mayor warned residents to prepare for prolonged outages and urged vulnerable populations to temporarily leave because restoration timelines were uncertain. These actions were targeted at the physical systems that sustain daily life.

The next inversion follows directly. If prevailing narratives emphasize overwhelming radar coverage and missile-defense superiority, why did those strikes land at all? Why were energy nodes disabled? Why did essential services go offline?

This discipline will matter more, not less, going forward. In 2026, many of the most consequential moves are likely to occur off the main stage, before they are widely acknowledged. The dogs that did not bark often matter more than the ones that did.

Bitcoin illustrates this risk clearly. It may eventually behave like a reserve asset, but today it still trades as a high-beta technology proxy. In a deflationary shock, it is vulnerable to sharp drawdowns, even if long-term liquidity trends remain supportive.

I would like to credit most of this work to my teachers Ritesh Jain, Luke Gromen, Peter Zeihan, and The Kobeissi Letter. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

Business Segments

Neogen today is best understood as a specialty chemistry platform with three profit pools at different stages of maturity:

1. A large legacy specialty chemicals engine (organic + inorganic) that pays the bills today,

2. A newer organolithium / custom synthesis layer that improves mix and stickiness, and

3. A high-capex battery chemicals option (Neogen Ionics) that is strategically valuable but still in the “approval + ramp” phase.

1) Organic Chemicals (~88% of Q2FY26 revenue)

This is Neogen’s largest segment and the one that still drives the majority of revenues today. “Organic chemicals” here means carbon-based specialty molecules, typically made through complex reactions and sold as intermediates (inputs) to other industries.

What they sell

• Bromine-based specialty compounds (bromination chemistry)

• Advanced intermediates for pharma, agrochemicals, flavours & fragrances, and electronics

• Custom Synthesis / Contract Manufacturing (CSM/CDMO-like) products where molecules are built to customer specs

• Organolithium reagents (post BuLi acquisition), such as n-Butyl Lithium, used heavily in pharma/agro synthesis

How this business works (simple)
Neogen is not selling a “finished product” to consumers. It’s selling a critical ingredient that sits inside someone else’s supply chain. The customer (say a pharma or agro player) qualifies Neogen’s molecule on parameters like purity, consistency, and documentation. Once approved, switching suppliers becomes annoying and risky so relationships tend to be sticky.

Why this segment matters

• It is the cash engine today (volume + customer relationships + diversified end markets)

• It benefits from process know-how and approvals, which reduce easy competition

• It has a meaningful mix-up lever via CSM (higher value, more embedded with customers)

2) Inorganic Chemicals (~12% of Q2FY26 revenue)

This segment is largely lithium-based inorganic specialty products used across:

• pharmaceuticals and intermediates,

• specialty polymers and construction chemicals,

• and selectively, battery-adjacent applications (outside the Ionics vertical).

This is where Neogen historically built its India lithium ecosystem, including its long-standing position as a large importer/handler of lithium carbonate/hydroxide and related derivatives.

How this business behaves

• More exposed to global lithium derivative pricing cycles

• More vulnerable to China-led price pressure in certain products

• Typically lower “story value” than battery chemicals, but still important for:

  • customer relationships,

  • know-how in lithium handling/purification,

  • and being a feeder capability for higher value chemistries.

Think of it as the lithium foundation layer that sits underneath the more exciting battery narrative.

3) Battery Chemicals (Neogen Ionics)

This is the segment most investors focus on because it can change Neogen’s scale and relevance. Ionics is not yet a mature earnings contributor, it is currently in the “qualification + commissioning + market timing” phase.

Ionics primarily targets two product buckets:

A) Electrolyte (liquid formulation)

Electrolyte is essentially a salt dissolved in solvents, designed to enable ion movement in a lithium-ion cell. It is safety-critical and very sensitive to contamination and consistency.

• Neogen has highlighted achieving a major Indian customer milestone (PPAP/plant approval) for electrolyte supply.

• However, demand ramp is linked to cell plant commissioning, which has shifted timelines by quarters.

Electrolyte has a practical localization advantage because it is sensitive and time-bound in handling (short shelf life), which supports local sourcing once customers are operational.

B) Electrolyte salt (LiPF6)

LiPF6 is the key salt used in most conventional Li-ion electrolytes. It is also the product where approvals matter most as customers want audits, quality systems, and stable long-term assurance. Neogen is still working through final approvals and upgrades for global customer requirements.

How Ionics should be understood

• It is capex-heavy with high operating leverage

• The “success condition” is:

  1. customer qualification,

  2. customer commissioning,

  3. sustainable price umbrella (global),

  4. utilization ramp.

If those align, Ionics can become a meaningful revenue and margin driver. If they don’t, it remains a fixed-cost drag.

What Went Wrong: The LiPF6 Bet

Neogen’s recent underperformance is best explained by one central issue: its aggressive LiPF6 battery-chemicals investment arrived too early, both for India and for global markets.

LiPF6 is a high fixed-cost, safety-critical product. Neogen invested roughly ₹400–500 crore assuming two things would hold: that Indian cell manufacturing would ramp on schedule, and that export markets could absorb volumes in the interim. Neither materialized.

Domestically, India simply did not have operating gigafactories. Without cells being produced, there was no real demand for electrolyte salts. LiPF6 cannot be meaningfully stockpiled, and customers do not commit volumes until their own plants are live.

At the same time, the global market turned hostile. From late 2023 through most of 2024 and into early 2025, China-led overcapacity crushed LiPF6 prices. Chinese suppliers sold at levels that were uneconomical for new, non-Chinese entrants. As a result, export customers either refused to sample Neogen’s material or benchmarked pricing so low that sales were impossible. Capacity stayed idle while depreciation, manpower, utilities, and compliance costs continued to hit the P&L.

This had knock-on effects beyond Ionics. Fixed-cost absorption worsened at the consolidated level, compressing EBITDA despite the base business remaining largely intact. Management attention shifted heavily toward the battery platform at a time when the legacy bromine and intermediate businesses were already facing pressure from softer demand, bromine price volatility, and Chinese competition. Instead of improving mix, the transition temporarily weakened it.

Why LiPF6 Can Still Work

The most important shift has occurred on the supply side in China. The severity of the LiPF6 price collapse was such that it eliminated marginal capacity. Prolonged periods of uneconomical pricing forced many mid-sized and smaller Chinese producers to shut down permanently. As a result, a once fragmented market has consolidated sharply, with roughly three large players now controlling around 65% of China’s LiPF6 capacity. This matters because extreme price undercutting becomes harder in a concentrated market, and pricing tends to gravitate back toward sustainable levels once loss-making capacity exits.

This structural reset is already visible in pricing behavior. After bottoming out around $6.5–8 per kg through much of 2024, LiPF6 prices rebounded meaningfully in late 2025, reaching the $15–17 per kg range at recent peaks. While volatility will remain, this is a fundamentally different pricing environment from the one Neogen faced when it initially attempted to commercialize capacity.

In this new environment, export opportunities are likely to emerge before domestic demand meaningfully ramps. Global battery manufacturers, particularly those supplying the US and Europe, are increasingly prioritizing supply-chain diversification to reduce concentration risk. With regulatory and policy frameworks pushing toward non-FEOC sourcing by 2027, reliability, compliance, and provenance are becoming important. Neogen’s investments over the past two years in audit readiness, systems, and process upgrades position it more favorably under these criteria than during the earlier downcycle.

Domestic demand, meanwhile, has been delayed rather than destroyed. Importantly, LiPF6 demand is highly non-linear. It remains close to zero until cell plants stabilize, after which volumes ramp quickly and become sticky due to the difficulty of requalification and the critical nature of electrolyte salts. Neogen’s disadvantage earlier was building capacity ahead of this curve but its advantage now is that the asset already exists when the curve finally steepens.

Operating leverage, which punished the company when capacity sat idle, also works in reverse. LiPF6 economics are defined by high fixed costs and relatively modest variable costs. Once utilization crosses a threshold, incremental volumes can significantly improve margins. Even partial utilization can meaningfully change consolidated profitability, because the most painful phase is already behind the company.

Other Growth Drivers

1. Electrolytes - Electrolyte and electrolyte salts are not the same business, either operationally or commercially. Based on management disclosures, Neogen’s electrolyte platform has already crossed a meaningful hurdle that LiPF6 has not: plant and process approval with a leading Indian gigascale customer.

   This matters because electrolyte qualification is operationally demanding. Clearing PPAP-level approvals signals that Neogen’s manufacturing systems meet customer standards. In contrast to LiPF6, electrolyte demand also has a natural localization advantage. Shelf life is limited, handling risks are high, and customers prefer short supply chains once cell plants are operational. As a result, even modest domestic cell production can translate into recurring electrolyte demand far sooner than it does for salts.

2. Base Business Stabilizing Under Pressure - Another important takeaway from recent quarters is that Neogen’s legacy chemistry platform has not structurally deteriorated, despite the drag from Ionics and the Dahej incident. Organic revenues grew 12% YoY even through outages and job-work inefficiencies. Management commentary also pointed to demand-led volume growth across the base portfolio, including organolithium products.

Is Neogen a FY27 Story?

Management’s messaging across the concall makes one thing clear: FY26 is not the year they want to be judged on. Profitability is distorted by incident-related costs, delayed ramps, and fixed-cost absorption. Instead, every major milestone (commercial production timelines, insurance recoveries, and meaningful battery revenues) has been explicitly pushed into FY27.

FY27 is positioned as the first year where multiple moving parts converge: electrolyte production scaling in H1, electrolyte salt commercialization following in H2, normalization of the Dahej cost base, and loss-of-profit insurance recoveries beginning to flow. By that point, the base business is expected to be running near capacity, while Ionics either starts contributing meaningfully or clearly fails to do so.

Risks

1. LiPF6 Approval Slippage - The most obvious risk is also the most consequential: LiPF6 approvals do not translate into meaningful volumes, even after audits and provisional clearances. Electrolyte salts are deeply embedded in customer qualification systems, and even small gaps in instrumentation, software, or documentation can push final approvals by quarters. If customers continue to delay commitments or require incremental upgrades beyond current expectations, Neogen could find itself in the same position as the previous years.

2. Execution Discipline - The recent shift toward more conditional, modular expansion is encouraging but it needs to persist. Any temptation to accelerate capacity additions before demand visibility returns would recreate the original mistake. The turnaround thesis implicitly assumes capital discipline has structurally improved. If that assumption proves false, incremental capex could dilute returns even if demand eventually arrives.

3. Base Business Strength - Bromine price volatility, intermittent Chinese dumping in intermediates, or softer pharma/agro demand could re-emerge. If the base business weakens at the same time Ionics fails to scale, Neogen loses the time buffer that currently allows it to wait for the battery thesis to play out.

Disclaimer: Not a buy/sell recommendation. Please do your own research. I am not invested right now but may decide to do so in the future.

Wishing everyone a happy new year and continued compounding of knowledge in 2026!

Today, headline indicators still look reassuring. However, there are many cracks in global economies.

In this month’s memo, I cover electricity and grid capacity in the U.S., issues with AI-driven productivity, gold, and India’s position in this environment.

Electricity is the Bottleneck

(Peter Zeihan)

The United States is running into a hard physical limit: electricity. Even without AI or a data center buildout, simply doubling the industrial base would require roughly 50% more power generation. Data centers sit on top of that requirement, adding another 25-50% depending on adoption paths.

The weakness is not generation, rather it is the grid itself. The US power system is fragmented, state bound, and poorly interconnected. Long range, high voltage transmission barely exists outside a small Midwestern corridor built decades ago to move coal generated electricity more efficiently than rail could move coal itself. Without national scale transmission, electricity must be generated close to where it is consumed. That forces overbuilding of local capacity, lower utilization rates, and higher costs.

This matters because data centers do not tolerate intermittency, they run 24/7. Power sources that cannot deliver continuously are not viable for base load demand. Solar fails at night, wind fails where currents are unreliable, natural gas works best for surge capacity, not constant draw. What remains are nuclear and coal. That conclusion is uncomfortable but unavoidable. A digital future built on data centers requires base load power.

The investment implications are enormous. Expanding the grid within state boundaries would require roughly a trillion dollars in new generation and another half trillion in distribution. Building a national transmission backbone would cost more upfront, but allow fewer plants to run more efficiently. The problem is time: The US has not meaningfully expanded grid capacity in over twenty years. Skilled labor is aging out and timelines are stretching into decades. Even under ideal conditions, meaningful progress would take 5-10 years.

Signs of strain are already visible. Utilities are failing to deliver on contracted power. Data center projects are being delayed into the next decade because grid hookups are unavailable, electricity prices have risen sharply over the last fifteen years, and the lowest income households now spend close to a third of take home pay on power.

AI Meets a Debt-Based Economy

(Ruchir Sharma, Luke Gromen)

AI is widely framed as the next great productivity leap, a force that will lower costs, raise output, and expand economic potential. That framing is not wrong, but it is incomplete. Productivity gains do not exist in a vacuum. They operate within a monetary system built on wages, debt, and consumption, and it is within that system that the tension emerges.

Why Productivity Can Be Contractionary in a Debt System

Modern economies are financed through debt that assumes stable employment, growing wages, and predictable cash flows. Mortgages, auto loans, credit cards, and student debt are all serviced from labor income. AI, by design, raises output while reducing the need for labor across wide swathes of the economy. That trade-off works cleanly in theory, but in practice it compresses the very income streams that keep a leveraged system functioning.

The first stress point is consumption. As labor income weakens or becomes less secure, households pull back spending well before they miss payments. Consumption slows, corporate revenues soften, and margins come under pressure. Only later do credit losses appear. In that sense, productivity shocks destabilize debt systems not because they destroy output, but because they erode income velocity.

This is the second-derivative problem often highlighted by Luke Gromen. Efficiency improves, but the cash flow required to service existing liabilities does not scale down in parallel. Debt is fixed in nominal terms, while wages are not. When efficiency gains arrive faster than balance sheets can adjust, stress accumulates quietly.

The AI Capex Surge and the Four O’s

Ruchir Sharma’s framework helps place the current AI cycle in context. Over-investment is visible in the sheer scale of capital expenditure, now approaching levels historically associated with bubbles as a share of GDP. Overvaluation is evident in AI-linked equities trading at prices that assume long-dated, high-margin outcomes. Over-ownership shows up in crowding, with a narrow set of names absorbing an outsized share of capital flows. The final leg, Over-leverage, is the most recent development.

Earlier in the cycle, AI expansion was largely funded through retained earnings and equity. That has changed. The arms race is increasingly debt-funded, with major platforms issuing bonds and relying on vendor financing to sustain spending momentum. This matters because the underlying assets have short economic lives. Chips turn over every few years, and data center hardware depreciates quickly. Yet the liabilities financing them are long-dated.

This mismatch echoes earlier cycles, notably telecom fiber in the late 1990s and US shale in the 2010s. Both combined massive capex, borrowed money, and optimistic demand assumptions. Both produced real technological advances, and both destroyed capital when cash flows failed to match the investment pace. AI shares elements of those cycles, but with an added constraint: its true bottlenecks are not chips alone.

Power, water, grid access, and skilled labor are emerging as binding limits. Data centers require continuous electricity and large volumes of water for cooling. Grid connections are delayed for years in several regions. Utilities cannot meet contracted demand. These constraints slow monetization even as spending continues, extending the period during which capital is consumed without commensurate returns.

This is where productivity’s dark side becomes clear. AI may raise output potential, but in a debt-based economy it can weaken the income foundation before it strengthens growth. If spending slows or capex falters, the same AI investment that supported GDP becomes a source of downside risk. The system becomes more efficient, but also more fragile.

The paradox is that AI can be both indispensable and destabilizing. If it succeeds, it pressures wages and tax bases. If it slows, it exposes the debt used to build it.

China, Reshoring, and Triffin’s Dilemma

(Luke Gromen, Ritesh Jain)

For years, the debate around U.S.–China competition has focused on chips, software, and capital markets. That framing misses the core advantage: China’s edge is not compute, it is electricity and grid scale.

China crossed U.S. electricity generation levels around 2008–09. Since then, it has added grid capacity roughly equal to the entire existing U.S. grid. Over the same period, U.S. grid capacity has barely grown. Financialization, offshoring, and regulatory inertia left the physical backbone of the economy stagnant. The result is that the bottleneck for the next industrial and digital wave is no longer capital or technology, but power.

This matters directly for reshoring. Manufacturing is energy-intensive and data centers and defense supply chains are energy-intensive. You cannot rebuild an industrial base without first rebuilding the grid that feeds it.

Cost structures make this brutally clear. The cost of a gigawatt of nuclear capacity in China is roughly one-sixth of the U.S. cost. No tariff regime can bridge that gap, a 50% or even 100% tariff does not offset a sixfold cost disadvantage embedded in electricity, labor coordination, and construction timelines. To equalize competitiveness purely through tariffs would require levels closer to 300–500% according to Luke Gromen, which would be instantly inflationary and politically unsustainable.

This is where the reshoring narrative collides with macro reality. To rebuild the grid and industrial base at scale, the United States must accept three conditions simultaneously:

1. Inflation tolerance - Grid expansion, nuclear buildouts, skilled labor shortages, and domestic manufacturing all raise input costs. There is no version of rapid reshoring that is disinflationary.

2. Yield suppression - Financing trillions of dollars of long-dated infrastructure requires keeping nominal rates below what a free market would otherwise demand. Left unchecked, higher yields would choke off the very investment reshoring depends on.

3. Dollar weakness - An overvalued reserve currency makes domestic production structurally uncompetitive so a weaker dollar is a prerequisite.

These three conditions are incompatible with preserving U.S. Treasuries as the dominant global reserve asset.

This is Triffin’s dilemma. A country that issues the world’s reserve currency must run external deficits to supply safe assets to the global system. A country that wants to reshore manufacturing must reduce deficits, weaken its currency, and redirect capital toward domestic production. You cannot do both.

Policy over the last decade has tried to postpone this choice by supporting asset prices while talking about industrial revival, and it worked when capital was the constraint but won’t when electricity is.

As Luke Gromen has argued, every path forward imposes stress somewhere else. Funding grid expansion through monetary accommodation weakens the dollar and debases bonds. Preserving bond credibility constrains investment and forces industrial leakage abroad. Trying to do both results in half-measures that solve neither problem.

Ritesh Jain frames the implication cleanly: in a world where real production is energy-bound and capital is abundant, the reserve asset must rise against finite output rather than promise fixed nominal returns. That is why gold increasingly sits at the center of this adjustment as a balance-sheet release valve.

Why Gold Is Becoming the Neutral Release Valve

(Ritesh Jain)

Gold today is more about neutrality. A Treasury bond is a claim on future fiscal capacity, a bank deposit is a claim on a leveraged balance sheet, equities are claims on earnings that depend on stable demand, credit availability, and political continuity. Gold is different, it is a monetary asset with no counterparty, no maturity, and no dependency on policy credibility. In an environment where every major adjustment path creates stress somewhere else, that neutrality matters.

Roughly 60–70% of above-ground gold now sits outside the G7, concentrated in emerging markets and non-aligned countries. Central banks are not interested in short-term spikes but they are buying it because their existing reserves are structurally misaligned with the world they are operating in.

On a broad definition, gold already rivals, and in some measures exceeds, U.S. Treasuries as a reserve asset.

Gold’s appeal is that it performs across regimes. In inflation, it preserves purchasing power as fiat claims are diluted. In deflation, it protects against sovereign credit risk when revenues fall faster than obligations.

Silver operates as a higher beta expression of the same forces. It carries monetary characteristics similar to gold, but with an added layer of industrial demand tied to electrification, electronics, and energy infrastructure.

Where Stress Is Showing Up in Real Data

Here’s a compilation of posts from The Kobeissi Letter showing where stress is already visible:

• Labor market cracks are widening - U.S. employers announced more than 1.1 million layoffs this year, the first time that threshold has been crossed since the pandemic period. Heavy truck sales, a reliable forward indicator for employment, have fallen to the lowest level outside COVID in eight years, historically consistent with a meaningful rise in unemployment.

• Corporate stress is rising quietly - Large bankruptcies in the U.S. are running at the highest pace in roughly 15 years. This reflects weaker demand, tighter financing conditions, and margin compression across multiple sectors.

• Consumers are pulling back - Surveyed holiday spending expectations fell by more in a single month than during the 2008 financial crisis. The decline spans income brackets, suggesting broad-based pressure rather than isolated strain at the margins.

• Physical bottlenecks are intensifying - Memory-chip inventories have collapsed, with DRAM supply down significantly. The shortage now extends from advanced AI-related components to consumer electronics, reinforcing the point that infrastructure and supply chains are struggling to keep pace with capital deployment.

• Data quality itself is deteriorating - A record share of U.S. CPI components is now estimated rather than directly observed, particularly in shelter and services. This reduces confidence in precision at a time when policy decisions are increasingly sensitive to small changes.

Section on Indian Markets

The Market Everyone Gave Up On (Ridham Desai)

India’s equity market has been the weakest-performing large market globally over the past year. Among the top twenty markets by size, it ranked last, while peers delivered gains ranging from 20% to 70%.

The slowdown was driven by a cluster of idiosyncratic factors. Election-related pauses in government spending slowed activity at the margin, excessive rainfall disrupted parts of the rural economy, monetary policy remained tight for longer than expected. Together, these forces weighed on GDP growth and corporate earnings momentum, and that softness fed directly into share prices. That phase is now turning. The RBI has pivoted decisively, cutting the cash reserve ratio and interest rates in April in a move that is rare outside periods of acute stress. This shift has been reinforced by fiscal and regulatory actions from the government, creating an unusually coordinated policy impulse aimed at restoring growth. The breadth and timing of that response suggest the cyclical downswing is ending.

Valuations have also done a large part of the adjustment. India’s relative multiples had reached extremes that made foreign capital increasingly cautious, but over the last year, that premium has compressed sharply. On few measures, India now trades near lows relative to global peers. As valuation pressure eases, the marginal incentive for foreign selling diminishes.

A separate headwind came from the global obsession with AI. Capital crowded aggressively into markets and companies offering a direct AI narrative, leaving India structurally underrepresented in that trade.

What makes this cycle different from earlier slowdowns is what has changed beneath the surface since 2007 and 2013. India’s historical vulnerability lay in its external balance sheet. Oil shocks translated directly into balance-of-payments stress, currency weakness, and policy tightening. That channel has been structurally altered. Oil intensity has fallen sharply through a combination of logistics efficiency, highway expansion, GST-led removal of border delays, near-complete railway electrification, rural electrification, and ethanol blending. Even when oil prices spiked in 2022, India avoided the crises that defined earlier cycles.

Capital flows have also become more stable. Foreign direct investment has risen as a share of inflows, reducing dependence on volatile portfolio capital. In prior downturns, FPI outflows amplified domestic slowdowns and forced abrupt adjustment, and that amplification mechanism is weaker today. Capital committed to long-term capacity, services, and manufacturing is less sensitive to short-term sentiment and currency moves.

The income structure of the economy has transformed as well, with extreme poverty being largely receded. India now has a meaningful cohort of households with global-level purchasing power, layered above tens of millions of consumers whose spending responds quickly to incremental income gains and price changes. Small shifts in policy, taxation, or financing conditions translate into large changes in demand. As a result, India already contributes close to a fifth of global growth, and that share continues to rise. For many multinational companies, India now accounts for a disproportionate share of incremental revenue growth.

Manufacturing, long discussed but rarely delivered, is also becoming viable in a sustained way. The constraints that once held it back (complex taxation, weak infrastructure, rigid labor laws, and high logistics costs) have been systematically addressed. The combination of physical infrastructure buildout, tax reform, digital public goods, and regulatory simplification has lowered the threshold for scale. This does not produce instant results, but definitely changes the trajectory.

Is 10% Nominal Growth Enough? (Sahil Kapoor)

India’s nominal GDP growth has slowed to roughly 9–10%. On the surface, this appears reasonable due to uneven global growth. The problem lies in the asymmetry. Inflation has declined from around 7% to closer to 5%, a compression of roughly 200 basis points. Nominal growth, however, has fallen by more, closer to 260–300 basis points.

At this stage of India’s development, 10% nominal growth is not sufficient. Sustained over the next 15–20 years, it would fail to fully monetize the demographic dividend, complicate the transition to middle-income status, and leave the economy vulnerable to stalling just as the demographic window begins to close. For a country still converging toward higher income levels, nominal growth needs to be meaningfully higher to absorb labor, build capital stock, and compound incomes fast enough.

The key point is that 10% nominal is an outcome of deeper balance-sheet dynamics.

The first constraint sits with households. India’s growth model relies on a simple structure, and households are the only net savers in the economy. Corporates and the government are net borrowers, making household balance sheets the foundation of sustainable growth.

During the prior strong cycle from FY01 to FY13, consumption growth was driven by rising incomes rather than leverage. Wage growth was robust, household savings were healthy, and consumption loans were falling as a share of spending. This created the most durable form of demand expansion: income-led consumption that reinforced savings rather than eroding them.

In the current cycle, the composition has shifted. Income growth has slowed, while household debt accretion has risen. Consumption has weakened because incremental spending is increasingly debt-funded rather than income-funded. This dynamic caps how fast demand can grow. Even moderate leverage growth, when combined with slower income expansion, places a ceiling on nominal GDP growth. The result is an economy that grows, but not fast enough for its stage of development.

The second constraint is investment. To test whether capex could offset softer consumption, a broad-based tracker covering all major sources of investment was constructed by DSP. The conclusion is unambiguous: Central government capex is the only component outperforming the previous cycle. Every other driver is growing more slowly, often below nominal GDP.

The most striking datapoint comes from listed corporates. Capex by BSE 500 companies compounded at roughly 26% during FY01–FY13. In the current cycle, that figure has fallen to around 9%. Even 9% growth is not weak in isolation, but it is far below what India historically delivered when it was successfully accelerating up the income curve. Most other capex indicators now sit in the mid–single digits, compared with double-digit or 20% plus growth previously. Government spending is filling part of the gap, but it cannot substitute for broad-based private investment.

The third pillar, exports, has also underperformed. Global demand constraints and shifting trade dynamics have limited export growth, preventing it from compensating for weaker household demand or subdued private capex. Exports are contributing, but they are not acting as a swing factor.

Taken together, India’s nominal growth rate reflects a three-way slowdown. Household income growth has softened, pulling down consumption momentum. Private capex has reset sharply lower relative to the prior cycle. Exports have remained modest. The arithmetic of these three engines leads naturally to nominal growth settling around 10%.

For India, that pace is not enough. Raising nominal growth meaningfully requires repairing household income dynamics and reigniting private investment, not merely sustaining government spending. Until those engines regain traction, nominal GDP growth is likely to remain capped near current levels, leaving a significant portion of the demographic opportunity underutilized.

Behavioral Insight: Buying at the “Worst Time” (Indian Markets)

(Ishmohit Arora & Siddhant Bhandari)

Timing worries most investors. The fear of being wrong at precisely the wrong moment. My teachers Ishmohit Arora and Siddhant Bhandari conducted a useful thought experiment that reframes this concern and grounds the broader macro discussion in actual investor behavior.

The exercise starts in January 2018, the exact peak of a prior Indian market cycle. Assume an investor who had missed the rally leading up to that point. Valuations looked stretched, markets were rising daily, and hesitation felt prudent. Eventually, frustration overtook caution and capital was deployed at the worst possible time. Importantly, this investor did not buy the index blindly. They did what most real investors do when entering late: they gravitated toward already “discovered” quality companies with strong narratives, visible earnings, and perceived durability.

The resulting basket included businesses such as Info Edge, Astral, Berger Paints, Motilal Oswal, DLF, Prestige, and similar names that were widely owned, widely discussed, and widely considered expensive at the time. From a psychological standpoint, this was the least comfortable entry point imaginable.

Yet the outcome challenges the conventional lesson drawn from market peaks. Despite buying at the top, the median return from this diversified basket of quality small and mid-cap companies compounded to roughly 3-4x capital over the next 6 years.

This matters because it reframes risk. The dominant fear during periods of uncertainty is that valuation errors permanently impair capital. What this experiment shows is that high-quality companies with strong competitive positions, clean balance sheets, and reinvestment runways tend to compound through cycles, even when purchased during moments of maximum discomfort.

In the current environment, noise is abundant. This translates into waiting for clarity that rarely arrives in real time. The lesson from my teachers’ work is not to ignore risk, but to redirect focus. Long-term compounding still accrues to businesses that execute well, reinvest intelligently, and survive difficult periods.

Way Forward: Position, Don’t Predict

The common thread running through this memo is constraint. Electricity, balance sheets, labor, and fiscal capacity are all becoming binding at the same time. When systems tighten around constraints, outcomes become harder to forecast and easier to misread. In that environment, precision is less valuable than alignment. The task is not to predict turning points, but to position around the forces that are already visible.

The first anchor is collateral over credit. Assets that sit at the top of the capital structure, or that do not rely on someone else’s promise to pay, become more valuable as leverage accumulates and policy choices narrow. Hard collateral and businesses that self-finance their growth offer resilience in regimes where financial claims expand faster than the income supporting them.

The second anchor is power and energy infrastructure. Electrification has become the limiting input to growth across industry, technology, and defense. Countries and companies with reliable generation, surplus capacity, and efficient transmission will attract disproportionate capital over the next decade.

A related theme sits within artificial intelligence, but with a specific lens. Power availability, cooling, water, grid access, and specialized engineering are where friction is accumulating. Businesses that ease these constraints are likely to see steadier demand.

India warrants selective attention rather than broad enthusiasm. The policy impulse has turned supportive, valuations are reseting, and domestic demand is stabilizing from a position of structural strength. The opportunity lies in consumption and financial channels that benefit directly from incremental household cash flow. Selectivity matters, but the direction of travel is improving.

I would like to credit most of this work to my teachers Ritesh Jain, Luke Gromen, Ishmohit Arora, Siddhant Bhandari, Peter Zeihan, Ruchir Sharma, Sahil Kapoor, and The Kobeissi Letter. I wrote this piece simply as a means of consolidating all their knowledge to paint a digestible picture for you all.

Business Segments

1. Xanthine Derivates & Allied Products (50% of Q2FY26 revenue)

This is Aarti Pharmalabs’ largest and most stable business.

At its core, the company manufactures xanthine derivatives, the most familiar of which is caffeine. These molecules are not end products but functional ingredients that go into beverages, medicines, and nutrition products.

If you’ve ever had a cola or an energy drink, the stimulant effect comes from caffeine (1-3% of total ingredient cost) molecules that were likely produced by a handful of global manufacturers. Aarti Pharmalabs is one of them.

Think of caffeine manufacturing as a scale-and-reliability business. Customers don’t want novelty, they want consistent quality, uninterrupted supply, and competitive pricing. Once approved, they rarely switch suppliers.

APL manufactures:

• Caffeine (India’s largest capacity)

• Theophylline and related xanthine compounds used in respiratory drugs and formulations

The company operates two dedicated plants at Tarapur, Maharashtra, with a combined capacity of over 5,000 MTPA, currently under expansion to 9,000+ MTPA in phased commissioning during H2 FY26.

What makes this business strong is not just volume, but positioning:

• 15–20% global market share in xanthines

• Fully backward integrated, giving control over costs and quality

• Minimal dependence on China, benefiting from the global “China+1” sourcing shift

On the demand side, sales are diversified:

• ~65–70% goes into beverages (colas, energy drinks)

• The rest into pharma and nutraceuticals

Geographically, over half the revenue comes from exports, reducing reliance on any single market.

In simple terms, this segment is the cash engine of the company: high utilization, predictable demand, and steady returns that fund growth elsewhere.

2. APIs & Intermediates (40% of Q2FY26 revenue)

The second segment moves up the value chain into pharmaceutical manufacturing.

Here, Aarti Pharmalabs produces APIs (Active Pharmaceutical Ingredients) and key intermediates used in finished medicines. APIs are the chemicals that actually make a drug work; tablets and syrups are just delivery formats around them.

APL specializes in complex and regulated molecules, particularly:

• Oncology and anti-cancer drugs

• Corticosteroids

• Cytotoxic and high-potency APIs (HPAPIs)

These are difficult to manufacture safely and consistently, which naturally limits competition.

Manufacturing is spread across Vapi, Tarapur, and Dombivli, with facilities approved by USFDA and other global regulators. The company has built serious regulatory depth, with 50+ US DMFs and 35+ CEP approvals, enabling it to sell directly into the US, Europe, and Japan.

From a scale perspective:

• 1,100+ kiloliters of multipurpose reactor capacity

• 14 finished API product lines

• 60 APIs already commercialized

• 11 new APIs under development

• 140+ intermediates offered

APIs are sticky businesses. Once a pharma company files a drug with a specific API supplier, switching becomes expensive and time-consuming due to regulatory re-approvals. This creates long product lifecycles and repeat business. However, API business is lumpy.

3. CDMO/CMO Services (10% of Q2FY26 revenue)

Under its CDMO/CMO business, the company does not just sell molecules, it sells chemistry capability. It partners with global pharma and biotech companies to design synthetic routes, develop processes for new drug molecules, manufacture intermediates and APIs from clinical trials all the way to commercial launch.

Think of this as APL becoming the outsourced manufacturing arm for drug innovators.

The company has built expertise in: HPAPIs, Cryogenic reactions, Complex chemistries such as cyanation, and Corticosteroid development

As of Q2 FY26, APL works with 21 active customers across ~60 projects, of which ~33 are already commercial, and the rest are in various stages of development at the customer’s end

Why this business matters is the payoff structure. Early-stage projects contribute modest revenues, but if a drug succeeds and scales commercially, volumes can rise sharply, turning a small project into a multi-year, high-margin revenue stream.

This explains the rapid growth trajectory of the segment, with CDMO revenues scaling multiple times over the last few years.

In the long run, management views CDMO as a strategic growth engine, while xanthines and APIs provide the stability to fund it.

Industry

Who buys? Across all three segments, buying decisions follow a similar logic: qualification first, continuity second, price third.

In xanthines, buyers are beverage companies, ingredient distributors, and nutraceutical formulators. Although caffeine looks like a commodity, in practice it is not traded spot. Once a supplier is approved, switching is avoided unless pricing becomes meaningfully uncompetitive or supply reliability falters. This is why global market share in xanthines is concentrated among a small set of producers despite low technical complexity.

In APIs, the buying process is even more rigid. When a pharma company files a drug with regulators in the US, Europe, or Japan, it references a specific API manufacturer in its regulatory dossier. That link creates de facto lock-in. Changing suppliers later requires re-validation, new filings, and regulatory approval, which can take years and introduce risk to the finished drug’s supply. As a result, API relationships are slow to form but also slow to break. Reliability, documentation quality, and regulatory track record matter far more than headline pricing.

In CDMO, the process is relationship-led rather than procurement-led. Projects are awarded based on chemistry capability, problem-solving ability, and trust built during early development stages. Most projects begin with small volumes and modest revenues. Only a minority reach commercial scale, but those that do can turn into long-duration, high-margin relationships. This makes CDMO a portfolio business rather than a linear growth business.

Unit Economics

Xanthines operate on a classic manufacturing logic: utilization drives profitability. Fixed costs are meaningful, and margins expand sharply once plants run close to capacity. Backward integration into raw materials, control over utilities, and yield optimization are the main differentiators. Demand is steady, but pricing fluctuates with capacity additions and input costs, making margins cyclical even when volumes are not.

APIs sit one level higher. Here, margins are driven by molecule complexity, process efficiency, regulatory compliance costs, and plant utilization. High-potency APIs add additional safety and containment costs, raising entry barriers. Once approved and scaled, APIs generate stable cash flows, but the path to approval is long and capital-intensive. Importantly, APIs are not immune to competition: genericization leads to gradual price erosion over time, making portfolio renewal essential.

CDMO economics are fundamentally different. Early-stage development work is low-volume and service-heavy, with limited profitability. The real value emerges if and when a molecule succeeds commercially. At that point, existing assets are leveraged over much higher volumes, leading to strong operating leverage. The key variables are conversion rates from development to commercial, customer concentration, and utilization of specialized infrastructure.

Industry Tailwinds

1. Global supply-chain de-risking - For decades, China was the default supplier for bulk chemicals, intermediates, and APIs. That model is being rethought. Regulatory scrutiny, geopolitical risk, environmental enforcement, and pandemic-era disruptions have pushed global customers to diversify sourcing. This is not a short-term trend. For regulated industries, dual-sourcing has become a board-level mandate rather than a cost optimization exercise. India, with its chemistry talent, regulatory familiarity, and cost base, is the primary beneficiary.

2. Structural growth of chronic and oncology therapies - Unlike acute treatments, chronic and oncology drugs are consumed over long periods and tend to see stable or rising demand as populations age and diagnosis improves. This supports sustained API volumes even when individual molecules face price pressure.

3. Outsourcing bias of global pharma - Large pharma companies increasingly prefer to outsource manufacturing rather than invest capital into facilities that may be underutilized as pipelines evolve. Biotech companies, which drive much of drug innovation today, often lack manufacturing capabilities altogether. This structurally expands the addressable market for CDMO players, especially those that can handle complex chemistries and scale reliably.

4. Regulatory tightening - As compliance standards rise, smaller or poorly capitalized manufacturers drop out. This gradually concentrates market share among companies with proven systems, approvals, and financial staying power.

Competitive landscape and positioning

Globally, China remains dominant in many bulk chemicals and intermediates due to scale and cost. Europe and the US retain high-quality capacity but at significantly higher costs. The result is a bifurcated market: China competes on price and scale, the West on quality and proximity, and India increasingly occupies the middle ground, offering a combination of cost efficiency and regulatory credibility. A notable global xanthine competitor is BASF.

Within India, the competitive set includes large, highly specialized players such as Divi’s Laboratories in custom synthesis, Laurus Labs and Neuland in APIs and CDMO, and integrated pharma companies like Dr Reddy’s and Aurobindo that are closer to formulations. Aarti Pharmalabs’ differentiation lies in its combination of xanthine leadership, regulated API depth, and a growing CDMO platform. Most peers excel in one or two of these areas, but not all three.

Cyclicality

These industries are not uniformly cyclical, but they are not immune to cycles either. Xanthines experience pricing cycles driven by capacity additions and input costs. APIs are subject to molecule-level cycles, regulatory events, and competitive entry. CDMO is influenced by drug development cycles and biotech funding conditions.

APL’s Growth Drivers

1. Xanthine Capacity Expansion

The xanthine (caffeine) business is already mature in scale. Management has repeatedly said that Aarti represents 20-25% of the global market, which fundamentally changes how growth should be interpreted. At this level of penetration, growth is not about aggressive share grabs but selective expansion.

The ongoing expansion from ~5,000 TPA to 9,000 TPA by end-FY26 is therefore not a bet on immediate volume absorption. Management has been careful to frame this as installed and available capacity, not capacity that will automatically run at peak utilization.

Large global customers increasingly require dual-site qualification and supply redundancy, especially for critical ingredients like caffeine. The additional capacity allows Aarti to qualify with customers who demand multi-site sourcing and to deepen wallet share with existing accounts without straining reliability. Parallel to this, finish-line debottlenecking has improved throughput and product mix, lifting realizations rather than just volumes.

Pricing in xanthines has largely bottomed and stabilized, removing downside risk but also limiting near-term upside. However, if mean reversion were to take place, we could see significant inflation in xanthine prices to further improve growth.

2. API Normalization

The API and intermediates business has been the primary drag on near-term profitability, but the drivers of weakness are cyclical and structural to the industry, not indicative of erosion in Aarti’s competitive position.

Two forces dominated H1 FY26. First, second-year genericization pressure. After a successful launch year, additional players typically enter, prices soften, and margins compress. This is a known and recurring pattern in generic APIs. Second, customer inventory behavior distorted order flows. Management disclosed instances where customers stocked up to 1–1.5 years of inventory early in the launch cycle, leading to delayed replenishment and intensified price competition once inventories normalized.

Management expects H2 normalization, supported by both demand recovery and internal capacity rebalancing. As CDMO projects move out of intermediate-heavy phases, fungible assets free up, improving API utilization. In parallel, Aarti plans anti-cancer block debottlenecking in late FY26 / early FY27, explicitly timed for oncology launches in FY27–FY28. Management described these as “very good products,” signaling confidence in their durability and scale potential.

Rather than forcing all intermediates through internal assets, Aarti is actively qualifying external intermediate suppliers where it makes economic sense. This allows the company to allocate internal capacity to the highest value uses, reduces internal congestion during peak CDMO runs, and improves return on capital.

3. CDMO Scale-up

CDMO is the most important medium-term growth driver, and also the most easily misunderstood if viewed through a quarterly lens.

As of Q2 FY26, Aarti works with 21 customers across 59 active projects, of which 39 are already commercial. In recent quarters, 7–8 projects transitioned to commercial, while only two dropped, reflecting normal attrition. This shift materially improves the risk profile of the CDMO book. Once projects clear FDA or EDQM approvals, binary regulatory risk falls away, and demand visibility improves, even if early commercial volumes remain variable.

Management has been confident on this front, explicitly stating that they expect to achieve and exceed the previously guided 30–40% YoY CDMO growth, and emphasizing that this confidence is based on orders already in hand, not aspirational pipeline commentary.

Short-term margins have been pressured because CDMO consumes capacity before revenue is recognized. Plants remain occupied for months producing intermediates and batches, but invoicing only occurs once goods are shipped. Management has been clear that these costs cannot be capitalized as WIP profit. The result is higher utilization and costs today and revenue booked later. This explains the rise in inventory and the temporary margin compression.

Structurally, however, management reaffirmed that CDMO carries higher margins than APIs once commercialized. Growth will be further supported by a deliberate business development push, with a Europe representative office is already operational, a dedicated North America CDMO BD hire is planned, and the company is targeting Phase-2 to Phase-3 entry points, where success probabilities are higher.

4. Atali Facility

The Atali facility is best understood as a capacity enabler, not an immediate earnings driver. Inaugurated in September, it adds ~440 KL of capacity, with ~340 KL already commercialized. Currently, the plant is running trial batches, customer audits, and qualification programs. Management expects it to be fully operational within 2-3 quarters, with meaningful revenue contribution beginning in FY27.

In FY26, Atali will act as a P&L drag. Management has guided to incremental quarterly costs of ~INR 4–5 crore as power charges, minimum charges, and depreciation begin to flow through, with depreciation stepping up further in Q4 as capitalization progresses. This impact is explicitly embedded in the revised 8–12% FY26 EBITDA growth guidance.

Strategically, Atali expands fungible capacity for CDMO and intermediates, reduces bottlenecks during overlapping demand cycles, and enables participation in larger, later-stage CDMO programs that require assured scale and flexibility. Without Atali, the next phase of CDMO growth would be constrained.

Valuation Napkin Math

I have used a SOTP-style approach to project revenue for FY28:

1. Xanthine: ~1000 cr right now —> 2000 cr. This is assuming full utilization of the new capacity by FY28, plus a little price hike in xanthine prices

2. API: ~800 cr right now —> 1000 cr. I am being conservative here and assuming that there will be nominal growth in this segment.

3. CDMO: ~200 cr right now —> 400 cr. This is in line with management’s 30-40% growth guidelines.

So, roughly, we arrive at ~3500 cr revenue for FY28.

3500 cr revenue * 23% EBITDA * 15 EV/EBITDA = 12000 cr EV - 700 cr debt = 11300 cr MC. With the current MC of 6600 cr, this brings the IRR to 24%.

Risks

1. CDMO Execution - If customer launches slip, regulatory approvals take longer, or early commercial uptake is slower than expected, fixed costs hit the P&L without offsetting revenue. Even late-stage projects can underperform due to factors outside Aarti’s control, such as competitive therapies or pricing pressure at the customer level. With Atali adding further fixed capacity, any delay in CDMO commercialization risks prolonged operating leverage deferral rather than outright failure, showing up as weaker margins and returns over multiple years.

2. API Segment Erosion - While management has framed recent API pressure as typical post-launch genericization, the risk is that price erosion proves deeper or more persistent than assumed. As additional suppliers enter and customers consolidate purchasing power, certain molecules can see a structural reset at lower profitability levels rather than a temporary dip. This risk is amplified in oncology and HPAPIs, where high margins depend on smooth scale-up, continuous debottlenecking, and timely customer launches. If expected volume ramps slip or competition intensifies faster than anticipated, normalization may fall short, reducing the payoff from recent and planned capacity investments.

3. Xanthine Under-utilization - self explanatory.

Disclaimer: Not a buy/sell recommendation. Please do your own research. I am not invested right now but may decide to do so in the future.

I didn’t include valuations for this one, just trying to teach a unique business!

Business Segments

BlackBuck operates through six tightly connected verticals that together form a digital operating system for India’s trucking ecosystem. At the center are tolling, payments, and vehicle tracking: high-frequency, high-stickiness products that truckers use every single day. Around this core, the company has layered lending, hardware, and a freight marketplace to create a full-stack platform for small fleet owners.

Tolling Services

This is BlackBuck’s foundation. Every time a truck crosses a toll plaza, the payment is processed through FASTag, and BlackBuck acts as the digital backbone for that transaction.

For a trucker, tolls are unavoidable daily expenses. For BlackBuck, that means recurring revenue and high-frequency engagement.

Think of FASTag like your Netflix subscription, but instead of monthly payments, it runs on thousands of small, automatic debits every day across India’s highways. Once a fleet signs up, switching is rare because the entire rhythm of their operations depends on seamless toll payments.

Vehicle Tracking (Telematics)

BlackBuck sells ICAT-certified GPS devices and pairs them with a tracking software layer that helps operators monitor trucks in real time.

Every device shows live truck location, speed and route deviations, stoppage patterns, and trip history

For fleet owners, this is like getting a “live dashboard” of their entire business. For BlackBuck, it’s a high-margin, subscription-style service.

Once the hardware is installed, the customer is locked in for years. Data from these devices also feeds back into other parts of BlackBuck’s ecosystem, such as credit underwriting, fuel analytics, and load matching.

Payments Ecosystem

BlackBuck processes digital payments for:

• fuel purchases

• service station spends

• maintenance and repairs

• tolls, challans, and documentation

• driver settlements

In FY25, this ecosystem processed 55 crore+ transactions with a payment GTV of ₹23,493 crore, numbers that place BlackBuck in the league of India’s largest transaction platforms.

In simple terms: this is UPI for truckers, built specifically for the rhythms and needs of road transportation. BlackBuck earns through merchant commissions, transaction fees, and float benefits.

Vehicle Finance (Used Commercial Vehicle Loans)

India’s trucking sector is dominated by small owners who struggle to get loans from banks due to weak documentation. But BlackBuck already knows a truck’s entire behavior (toll usage, fuel spend, kilometers run) because it processes these daily payments.

This allows it to underwrite risk in a way traditional lenders simply cannot.

The company offers loans for used commercial vehicles through partnerships with NBFCs and a network of 100+ hubs.

The economics are simple: data reduces risk —> risk reduction expands credit access —> more credit increases platform usage

Fuel Management Hardware

Fuel is a truck operator’s largest expense—and also the easiest place for leakages and pilferage.

To solve this, BlackBuck sells proprietary fuel sensors that track fuel levels, consumption rates, refills, and leakages/unauthorized siphoning.

A single device can save thousands per month in losses. It turns fuel from a black box into a transparent, measurable cost. This segment combines hardware sales with recurring software fees.

Loads Marketplace

BlackBuck’s newest vertical connects truckers with shippers looking to move goods. It’s a digital marketplace where loads are listed, matched, and fulfilled. It’s like an Uber for trucks.

As the platform grows, the marketplace can become a high-margin business built on commissions, visibility tools, and data-driven matching.

Industry

India’s trucking sector is one of the largest, most critical, and most troubled parts of the economy. It carries 65–70% of all goods moved in the country, touches every sector from steel to FMCG, and sits on top of a market worth $150–170 billion today, expected to grow close to 8–9% annually this decade. Yet the industry still runs in a way that feels stuck in the early 2000s: phone calls instead of platforms, cash instead of systems, and manual routing instead of software.

The Backbone of Indian Logistics

India has 4–6 million trucks on the road, depending on the source, and nearly 3.5 million small fleet owners. Most own fewer than five trucks, operate from small transport yards, and run their business on experience rather than structured processes. Freight contracts are negotiated over phone calls, ledgers are maintained in notebooks, and cash moves through multiple hands before it reaches the driver.

This fragmentation has consequences. A single shipment often passes through 2–3 layers of middlemen, with 8–12% shaved off as commissions, and another 10–15% lost through intermediaries on the transport side. Drivers wait for cash. Owners struggle with working capital. Loads are matched through personal networks, not algorithms. And empty return trips remain one of the biggest hidden inefficiencies in Indian logistics.

Despite its scale, the industry is profoundly unorganized. That’s the opportunity.

The Driver Crisis

India’s trucking industry is facing a structural driver shortage. The country has 6 million trucks but only about 3.6 million drivers. Even transporters who own 40 trucks report that 2–4 are always parked because there is nobody to drive them. This mismatch has flipped the power dynamic: drivers today can walk away from a job instantly and find another within hours.

But the deeper issue is economic. A driver’s real salary often remains around ₹20,000–25,000 per month, barely enough to support a family. Transporters want to pay more, but they simply cannot because their own margins have collapsed.

A new truck can cost ₹28 lakh with EMIs of ₹30,000+ per month. After paying for fuel (the single biggest cost), tolls, maintenance, and the driver, a small fleet owner may earn ₹10,000–20,000 per truck in a good month. One tyre burst, breakdown, fine or accident wipes out the entire income.

No wonder many drivers quit. No wonder the few that remain sometimes resort to fuel theft, carrying illegal passengers, or overloading to make ends meet. And when they’re caught, the fines go to the transporter, not the driver, pushing small businesses deeper into the red.

Truck sales are rising each year, but the supply of drivers is not. As the gap widens, transporters hire untrained drivers out of desperation, leading to more accidents and more losses. In Delhi alone, the first two months of 2025 saw 122 major accidents involving heavy vehicles.

Industry Tailwinds

1. Regulatory Push

   1. FASTag - Now mandatory on national highways, with 97–98% penetration. This forces every truck to adopt a digital identity.

   2. AIS-140 GPS mandate - All commercial vehicles must be fitted with a certified GPS device and a panic button by October 2025, pushing telematics from “optional” to “required by law.”

   3. e-way bills and GST - These make documentation trackable, standardized, and digital-first.

2. Smartphone + UPI Adoption - India now has nearly 1 billion Internet users, with smartphone penetration sweeping across Tier-2, 3, and 4 towns. UPI has made digital payments normal even at dhabas and transport yards. For the first time, fleet owners have the comfort to recharge FASTags, pay for fuel, track expenses, and settle accounts directly through their phone.

3. Cost Pressure and Efficiency Needs - Fuel + tolls = 80% of a truck operator’s daily spend. Diesel prices keep rising. Toll costs have doubled in many corridors over the past five years. Shippers want visibility, and fleet owners want to reduce leakage, fuel theft, and empty runs. The only scalable way to improve utilization and reduce waste is data, and data comes from digital tools.

Competitive Landscape

India’s trucking industry has multiple players, but most operate in silos. There’s companies for Tolling & fuel payments like banks, but BlackBuck is the dominant truck-focused platform, handling 32.9% of commercial vehicle toll GTV. For Vehicle financing, we have firms like Shriram Finance, Cholamandalam, HDFC Bank, and BlackBuck has entered this place as a loan originator, using its toll, fuel, and telematics data to underwrite credit better than traditional lenders.

Globally, BlackBuck is closest to:

• Manbang (China) – the world’s largest digital truck platform

• Uber Freight (US/EU) – digital brokerage

• Sennder, Project44 – freight visibility and matching platforms
  But BlackBuck has something unique: a distribution network across 80%+ of India’s districts, a necessity in a market where trust still comes from a handshake as much as from an app.

Cyclicality

Yes, trucking is cyclical. When GDP slows or industrial output falls, freight volumes soften. Truck OEMs are highly cyclical businesses.

But the digital share of trucking is a structural, long-term story. FASTag doesn’t go away in a recession. AIS-140 GPS doesn’t get rolled back. Driver shortages don’t disappear. Fuel costs don’t decline structurally. Shippers don’t stop demanding visibility.

This is why companies like BlackBuck can grow even in a soft freight cycle.

Key Growth Drivers for Blackbuck

Payments and Telematics Becoming the Default Infrastructure

The first and most important driver is the deepening of BlackBuck’s core stack: toll payments, fuel payments, GPS tracking, and fuel sensors.

Every truck in India wakes up and does two things: pays tolls and buys diesel. BlackBuck sits right in the middle of this spend and processes 32.9% of all commercial vehicle toll GTV in the country. That dominance matters because tolling and fuel are daily, unavoidable, high-frequency transactions. Once a truck operator uses BlackBuck for these, leaving the platform becomes nearly impossible as every truck movement, payment rhythm, and route pattern is tied into the app.

Telematics is the second half of this core. The AIS-140 mandate (GPS + panic button) coming into force by October 2025 means millions of trucks are being pushed into real-time visibility whether they want to or not. BlackBuck’s physical network (10,000+ touchpoints) gives it the ability to install and service devices at a scale that pure software competitors can’t. The company’s proprietary fuel-sensor product, which grew 55% QoQ recently, sits beautifully on top of this push as when fuel is your biggest cost, knowing where it goes becomes non-negotiable.

This is what makes the core so powerful. Every incremental truck onboarded uses payments. Every truck using payments generates data. Every truck generating data becomes a candidate for telematics. Every truck with telematics creates more reasons to stay.

SuperLoads

The second driver is SuperLoads, which goes after trucking’s most expensive inefficiency: empty return trips. India doesn’t have a shortage of trucks, it has a shortage of trucks running with loads. Every idle trip burns diesel, slows EMI recovery, lowers driver earnings, and destroys transporter margins. Freight rates have been stuck at ₹1–2/kg for decades, but input costs have risen sharply. Improving utilization is arguably the single biggest financial lever in this entire industry.

SuperLoads is scaling into this gap. What makes it compelling is that BlackBuck has better load-matching data than any broker or classifieds-style platform ever had. It knows where a truck went, how frequently it travels, what tolls it paid, how fast it moves, when it stops, and how much diesel it consumes. This enables a fundamentally different kind of matching.

The model is now moving from pilots to rollout. BlackBuck is expanding from 4 hubs to 14–15 hubs within months, targeting the busiest freight corridors first (Gujarat, NCR, Maharashtra, West Bengal). The early numbers already show traction. The growth businesses (SuperLoads + finance) grew 226% YoY, with SuperLoads absorbing more of the existing user base every month. Once a transporter starts filling his return trips through SuperLoads, the platform stops being a utility and becomes an earnings engine. Higher utilisation → more fuel spend → more tolls → more payments → more data → more creditworthiness.

Vehicle Financing

India’s small fleet owners run one of the hardest businesses in the country with almost no access to formal loans. Traditional lenders still assess a borrower through paperwork: bank statements, balance sheets, or broker references. None of these actually reflect how a truck actually performs.

BlackBuck uses what matters. It knows whether a truck runs 18 days or 28 days a month, whether it consistently pays tolls, whether its owner tops up fuel regularly, whether it travels predictable routes, whether its GPS device shows disciplined driving, and whether loads are carried with reliability. This creates a better underwriting model.

That’s why BlackBuck’s vehicle finance business is scaling so rapidly. The company is not trying to be an NBFC. Instead it acts as a loan originator, taking no balance-sheet risk but earning fees from lending partners. For lenders, it’s a dream dataset. For truckers, it’s access that didn’t exist before. For BlackBuck, it’s the stickiest revenue layer possible, because a financed truck is locked into BlackBuck’s tolling, fuel, telematics, and load ecosystem for years.

Every financed customer becomes more active. More activity means more payments. More payments mean better data. Better data means more accurate underwriting. And more accurate underwriting means more customers.

Risks

1. Habits don’t change overnight - Indian trucking is deeply informal, driven by habits built over decades: cash transactions, phone-based load booking, handwritten ledgers, and daily negotiations with brokers. Digital adoption is rising because of FASTag, AIS-140 mandates, and UPI, but a large mass of fleet owners still operate on low trust and low-tech routines. Even today, many transporters don’t track fuel accurately, drivers hide route deviations, and owners hesitate to hand over financial visibility to an app. This behavioural lag can slow BlackBuck’s penetration curve because every new user still needs hand-holding through a physical touchpoint. The company’s phygital network solves this, but it’s still a friction that pure software models never face.

2. Freight Cyclicality + Driver Shortages can disrupt growth - The digital share of trucking is a long-term story. But the underlying freight market remains cyclical. When industrial production slows or infrastructure spending softens, loads drop and kilometers shrink. This affects toll flows, fuel consumption, and transaction volume. BlackBuck cannot escape a sharp freight downturn or prolonged driver supply shock.

Disclaimer: Not a buy/sell recommendation. Please do your own research. I am not invested right now but may decide to do so in the future.