How to Analyse a Pharma Stock in India

The first thing I got wrong about pharma was trusting the margin. Two companies can post the same operating margin in the same quarter and not be the same business at all. In the quarter ended December 2025, Aurobindo Pharma ran a 21% operating margin and Ipca Laboratories ran 22% (Inve data, Q3 FY26). Near-identical at the headline. But 43.2% of Aurobindo's revenue came from the United States that quarter (Aurobindo Q3 FY26 investor PPT), against a US contribution at Ipca of roughly one-sixth of sales — about ₹395 crore of ₹2,392 crore (Inve data, Q3 FY26). One of those margins is hostage to a regulator in Maryland; the other mostly is not. The headline can't tell them apart. (Illustration of how to read the mix — not a view on either stock.)

An Indian pharma company is rarely one business. It is several stapled together — a high-margin, high-risk US generics operation, a steady domestic-branded franchise, an API or contract-manufacturing arm — each with its own economics and its own way of going wrong. Read the consolidated number as a single story and you will misprice the risk in exactly the way the December margins above invite you to.

There's a second habit worth knowing before you start. Across more than 15,700 management commitments tracked on Inve, barely half are delivered as stated (Inve data, as of 2026-06-12), and pharma carries a special version of this: management routinely guides on USFDA resolution timelines — and these are among the most frequently slipped commitments in the sector, because the regulator, not management, controls the clock. Natco's Kothur plant is a useful case precisely because it did eventually resolve — the warning letter was lifted, with the USFDA close-out letter dated 25 September 2025 (Inve Promise Tracker; FDA close-out record) — but only after several quarters of remediation that management could guide but not guarantee. A timeline guidance here is a forecast about a third party's decision, and it should be weighted as one.

This piece is about the handful of things that actually drive a pharma company's value — the revenue mix, the regulatory file, the R&D engine, and price erosion — plus the concall questions that test management's candour. A boundary first: properly judging a drug pipeline or the science behind it takes domain expertise most of us don't have, and I don't have it for the molecules below. What you can do is read the business mix, the regulatory track record, and the direction of the key numbers, and test whether management's account survives the Q&A. That is the circle this piece stays inside.

Why must you split the business before you read it?

The single biggest mistake in pharma analysis is treating consolidated revenue as one stream. The sub-businesses inside a typical Indian pharma firm have wildly different margins, risks, and durability:

• US generics is the high-stakes engine: large market, attractive margins on a freshly approved drug, but brutal price erosion over time and total dependence on USFDA approval and plant compliance. One bad inspection can freeze a facility's US revenue.
• Domestic (India) branded formulations is the steady, defensive base: prescription-driven, sticky brands, pricing power within regulatory limits, far lower tail-risk. This is the part that holds up when the US engine stalls. Ipca's domestic business grew 12% in Q3 FY26 (Inve data, Q3 FY26) — unspectacular, volume-led, and almost entirely indifferent to what the FDA does.
• APIs and contract manufacturing (CDMO/CRAMS) are lower-margin, capital-intensive, tied to raw-material costs (often Chinese supply) — but can be a sticky relationship business. Aurobindo's management flagged "pricing pressure in the APIs as well" the same year its formulations held up (Aurobindo Q1 FY26 concall) — a reminder that the arms move independently.

The relationship that matters: the US mix gives a pharma company its upside and almost all its tail-risk; the domestic mix lets it survive a US setback. Picture a household where one earner is a salaried teacher and the other works on commission that swings with one big client's mood. Same total income some months. Completely different fragility. A firm with 43% of revenue in US generics is structurally more exposed than one a sixth dependent — even when the December margins match. Before reading anything else, find the split and ask: how much of this depends on a regulator in Maryland staying happy?

How do you read a USFDA inspection outcome?

For any pharma company with US exposure, the regulatory file is the risk profile, and there is a clear escalation ladder to read literally.

• EIR (Establishment Inspection Report) with "VAI" or "NAI" — the plant was inspected and the outcome was benign (Voluntary Action Indicated, or No Action Indicated). The clean outcome, often under-reported because it's expected.
• Form 483 — the inspector left a list of observations. A 483 is not a verdict; it's a list of concerns the company gets to respond to. When Natco's Kothur plant drew a Form 483 with eight observations from the October 2023 USFDA inspection (FDA inspection record), the natural first question on the analyst call was the right one: by when can these observations be cleared? A handful of minor, procedural observations is routine. The questions: how many, how serious (procedural vs data-integrity), repeat offender?
• Warning Letter — the FDA was not satisfied with the company's response to the 483. Serious: it can stall new approvals from that facility and signals systemic issues; resolution typically takes several quarters to years, on the FDA's schedule. Kothur escalated to exactly this.
• Import Alert — the most severe: products from that plant can be detained at the US border, directly cutting that facility's revenue.

The non-obvious skill is reading severity and repeat-offence pattern, not the headline. Note how Natco's management de-risked the Kothur warning letter in their own words: "all our top five, six revenue items have an approval from our Vizag side in addition to the Hyderabad Kothur side... our impact will be minimal" (Natco Pharma Q3 FY25 concall). That is the right thing to check — not whether a 483 exists, but whether the affected plant makes the products that matter. A first-ever 483 at a minor plant is noise; a warning letter at the only facility cleared to make your top sellers is a thesis-changer.

The fullest history of a plant's compliance record lives in the regulatory disclosures of the annual report, which is why reading the annual report like an analyst pays off here. The thing to track is the gap between the stated resolution timeline and what actually happens. In Q1 FY26 Natco's management guided a "positive resolution within 90 days" of its July response on Kothur — and that one landed (Inve Promise Tracker, Q1 FY26). The next quarter, on a separate Chennai plant, they guided FDA "classification in the next 30, 40 days" (Natco Pharma Q3 FY26 concall). Same management, same regulator, two clocks. Inve's Promise Tracker is built for this: pinning each "we'll resolve it in two quarters" to the quarter it was made, then attaching the verdict as later calls reveal whether it held, slipped, or quietly went silent.

What does R&D spend actually buy — and how do you judge it?

R&D in pharma is not a cost to minimise; it is the seed corn of the future pipeline, and the question is never "how much?" but "what kind, and is it converting?" Read R&D as a percentage of sales against the company's strategy. Aurobindo spent ₹409 crore on R&D in Q3 FY26 — 4.7% of sales (Aurobindo Q3 FY26 investor PPT). Ipca guides its own R&D to about 4% of turnover, with management saying "this 4% may go to around 4.5% or 4.75% in next financial year" (Ipca Q2 FY26 concall). Two firms, near-identical intensity. The intensity tells you almost nothing on its own.

What it buys is the question. The relationship to watch is R&D intensity versus the quality of what it produces: a thinning, commoditising pipeline of plain oral solids is value destruction dressed as investment; complex, limited-competition approvals are the engine of durable margin. Aurobindo's spend is partly funding a biosimilars push — denosumab and an omalizumab candidate with FDA/EMA submissions planned during CY2026 (omalizumab guided to Q2/Q3 CY2026) (Aurobindo Q3 FY26 investor PPT) — which is a different animal from R&D that merely refills a generic oral-solids hopper. Read the output, not the line item.

Why is US price erosion the headwind that never stops?

Here is the structural fact that humbles every US-generics growth story: prices only go one way over a drug's life — down. Once a generic loses exclusivity, competitors pile in and the price erodes. The whole business is a treadmill: keep launching new approvals fast enough to outrun erosion on the existing portfolio.

You can watch the treadmill run in Aurobindo's own commentary. In Q1 FY26 management described it almost mechanically: gRevlimid was rolling off, but "overall sales were partially offset by steady demand in our overall base business and a series of new product launches during the quarter... We launched 15 new products in the U.S. this quarter, filed 4 ANDAs, received 14 approvals" (Aurobindo Q1 FY26 concall). Base erodes; new launches plug the hole. By Q2 FY26 base-portfolio price erosion was running, in their words, "closer to 1... very low single digit" (Inve data, Q2 FY26) — benign for now, which is precisely when the launch cadence matters most, because the day erosion accelerates, the launches have to already be there.

This reframes US revenue growth. Flat US revenue is not flat — it is new launches exactly offsetting erosion on the base; growing means launches are outrunning erosion; falling means the launch engine has stalled while erosion grinds on. So the real question behind any US number is: what's the pace of new approvals versus the rate of base erosion? And the mix matters as much as the count: a limited-competition or complex product stays profitable for years, while a plain generic facing fifteen competitors is a race to the bottom. Because reported US revenue can be flattered by one-off opportunities or channel stocking, it's worth pairing this with a quality-of-earnings check on profit versus cash flow.

The pieces, side by side

Two real Indian pharma companies, same quarter, near-identical operating margin — and opposite risk profiles. The point is not which is "better"; it is that the headline margin hides the thing that actually matters. (Illustration of how to read the mix, not a view on either stock.)

The lesson isn't that one company is safe and the other dangerous. It's that a 43%-US business with a plant under re-inspection carries a fundamentally different shape of risk than a domestic-anchored one — even when this quarter's margin says they're twins. Reading pharma well is learning to find the first column less reassuring than its margin suggests.

The concall questions that actually matter

The financials give the what; the call tests whether management's account of the regulatory file and pipeline is candid. The questions worth listening for cluster tightly:

• "What's the current status and expected resolution timeline on the [facility] observations?" Then note, next quarter, whether it held — the best recurring credibility test in the sector. Natco's analysts asked it about Kothur quarter after quarter, and got a moving answer each time.
• "What's the price erosion on the US base portfolio, and the launch cadence offsetting it?" A management comfortable with its US business quantifies both — as Aurobindo did, putting erosion at "closer to 1" and naming the launch count. One that talks only about "a healthy pipeline" is steering you off the treadmill.
• "How much of this quarter's US growth is one-off limited-competition opportunities versus the durable base?" A number flattered by a short-lived exclusivity is not a run-rate.
• "What's the complexity mix of your filings and approvals?" This separates R&D buying a future from R&D maintaining a commoditising present.
• "What's the domestic growth ex-acquisitions and ex-price increases?" Organic, volume-led growth is the durable kind.

Read these across quarters, not once: one call tells you about a quarter, a sequence tells you about management.

The one pattern that should worry you

If you watch one thing, watch this: the resolution timeline that keeps moving, paired with the metric that stops being mentioned. A clean way to see the FDA-clock problem is to listen to the same management on the same plant across calls — tracking how the remediation language evolves, and whether the verdict ever actually arrives. Kothur is the constructive version of this pattern: Natco guided resolution across several quarters, the timeline genuinely moved as it waited on the agency, and the warning letter was ultimately lifted with the USFDA close-out letter dated 25 September 2025 (Inve Promise Tracker; FDA close-out record). The cautionary version is the same shape without the ending: when remediation language stretches across quarters and the facility's revenue disclosure quietly thins and no close-out ever lands, you may be watching a problem managed in words rather than fixed in fact. The discipline is the same in both cases — pin each timeline guidance to the quarter it was made and wait for the agency's verdict, rather than taking the language as the outcome.

Stack the forces and you see why pharma blindsides people. Take a single US-heavy name: incentive-caused optimism (management is paid to sound confident about a plant they can't control), anchoring (the market fixes on a benign current erosion number like "closer to 1" and extrapolates it), and the launch-deferral trap ("our key launch will drive H2," pushed out for several halves running). Any one is survivable. Converging on one facility in one quarter, they produce the 15% single-session drop on a one-line 483 filing — the market repricing the next two years of that plant's US revenue, not the quarter just reported.

Where this read can be wrong. The honest steelman cuts against everything above. A warning letter at a US-heavy plant can be the buy signal, not the warning, precisely because the market over-punishes the headline: if the affected facility doesn't make the top products — exactly Natco's "all our top five, six revenue items have an approval from our Vizag side" point — the revenue hit is small and the de-rating is the opportunity. Disaggregation cuts both ways; sometimes the consolidated panic is the mispricing. And reading management well does not let you forecast a regulator's next inspection: a clean file and candid commentary do not immunise a company against a surprise failure at a previously clean plant, a patent-litigation loss, a pricing-policy shock, or a China raw-material disruption. Many of these tail risks never touch the balance sheet — outstanding lawsuits, patent litigation, and USFDA-related provisions live in the notes under contingent liabilities, so locate them there and size them before judging the risk. Reading management lowers your odds of being blindsided by avoidable deterioration; it does not abolish the unavoidable kind.

A repeatable workflow

The sequence: split the business and quantify US dependence; read each plant on the EIR → 483 → warning letter → import alert ladder, and check whether the affected plant makes the products that matter; judge R&D by the complexity mix of filings, not the amount; read US revenue as launches minus erosion; and audit the commentary against what actually happened.

Inve's KPI Screener lines up pharma operational metrics — US and domestic growth, margin, R&D intensity where disclosed — across companies with YoY/QoQ trends and a data-confidence flag, so the peer comparison is quick. Listed names worth running through this lens — spanning US-heavy generics, domestic-anchored franchises, and API/CDMO models — include Sun Pharma, Dr. Reddy's, Cipla, Divi's Laboratories and Mankind Pharma. The concall summaries pull every forward commitment into one guidance table per quarter, with speaker and exact quote, so tracking resolution timelines is a read rather than a re-listen.

Frequently asked questions

Reading a pharma company is the discipline of disaggregation. The numbers tell you the state of the business; the concall, read across years, tells you whether to trust the people describing it. So here is the owner's question to leave with: if the US engine were frozen tomorrow by a single inspection, what would you still own — and is that enough?

Related reading

• Capacity Utilisation: Pre-Capex Signal for Stocks
• How to Analyse a Cement Company (India)
• How to Analyse a Real Estate Developer Stock