How to Analyse a CDMO / CRAMS Pharma Company

Here is a number that should stop you before you buy any contract-research stock on its order pipeline. When an analyst pressed Syngene on why a strong run of incoming RFPs wasn't showing up in revenue, management explained the gestation plainly: these are "generally for pilot projects… it would not be atypical for a pilot program on conversion to a full program to have an increase of 4x on an FTE" (Syngene Q4 FY25 concall). Read that twice. A win in the funnel can quadruple in value when it converts — or it can sit as a pilot for a year and convert to nothing. The pipeline is real, and it is also a lottery ticket with a long fuse. (Illustration of how to read the model, not a view on the stock.)

A CDMO/CRAMS company — contract development and manufacturing, or contract research and manufacturing services — does not sell a drug. It sells capacity and skill: a slice of a scientist's time, a reactor, a clean room, a fermentation tank, rented to a pharma or biotech client who owns the molecule. That single fact rewires the whole analysis. There is no own-product approval to win, no price erosion on a generic to model, no USFDA warning letter that freezes your revenue — those are the risks you read in a pharma company, and you should read that guide for them. A services company's risk is different and quieter: will the client keep coming back, will the expensive asset stay full, and is the molecule it is paid to make heading for the market or the bin?

This piece is about the handful of things that actually decide a CDMO/CRAMS outcome — the RFP funnel and pipeline mix, the capex-to-asset-turn engine, client concentration, capacity utilisation, gross margin, and the biotech-funding cycle that sits under all of it — and how to value a business the market pays a fat premium for the privilege of owning. A boundary first: judging whether a client's molecule will succeed in trials is drug science most of us, including me, can't do. What you can do is read the asset, the funnel, the customer book and the cash, and test whether management's account survives the down-leg.

What actually drives the economics of a CDMO?

Picture a company that builds a fleet of bespoke kitchens, each tuned to one chef's recipe, and rents them out. It can only charge well if the kitchens stay busy; it spends years and a fortune building a kitchen before the chef shows up; and if its biggest chef walks, an entire wing goes dark. That is a CDMO. The kitchen is a multi-hundred-crore plant or a panel of PhD chemists; the chef is the client who owns the molecule; the rent is a fee per FTE or per kilo.

Three consequences fall out of that, and they govern everything.

Capacity utilisation is the margin. Most of a CDMO's cost is fixed — the plant, the depreciation, the scientists who don't get laid off between projects. So profitability is overwhelmingly a function of how full the asset runs. A new facility commissioned ahead of demand is a guaranteed margin drag until it fills. Syngene said exactly this when its margin slipped: "the majority of the drop… comes from the new facilities that we are going to take on board… as and when these sites get operationalized and its capacity starts getting utilized, you will start seeing the drag come down" (Syngene Q4 FY25 concall).

Revenue is built, then it lags. A CDMO grows by spending capex first — building the kitchen — then waiting twelve to twenty-four months for it to validate, win regulatory clearance and fill. Syngene put the lag at "12, 18, 24 months depending upon the nature of the site" (Syngene Q4 FY26 concall). You are always paying for capacity before it earns. That is why the capex-to-asset-turn relationship, below, is the real business model.

The demand is someone else's R&D budget. A CDMO's order flow ultimately comes from how much money is flowing into drug development worldwide — and a large slice of that, especially in early discovery, is venture-funded biotech. When biotech funding dries up, the funnel thins, regardless of how good the operator is. Syngene named the macro directly: "there was a significant reduction in US funding into biotech that created a sort of sector-wide drawback" (Syngene Q1 FY26 concall).

Hold those three — utilisation, the capex-then-revenue lag, and the funding cycle — and the metrics below stop being a list and become one story.

The metrics that matter — and where they hide

The uncomfortable part for anyone used to a P&L: the numbers that decide a CDMO investment mostly aren't on the income statement. The RFP funnel, the pipeline phase mix, asset turnover, client concentration and utilisation live in the investor deck and the concall. The income statement gives you revenue, EBITDA and gross margin; it cannot tell you whether revenue came from a durable commercial contract or a one-off pilot, and those mean opposite things.

The RFP funnel and order book

This is the leading indicator: requests-for-proposal received, proposals won, and the dollar value of the contracted pipeline. It matters because revenue is built ahead, so the funnel today is the revenue eighteen months out. Where it hides: the concall, almost never quantified cleanly on the income statement — and you must read it with the gestation caveat. A fat RFP count is encouraging and not bankable: as Syngene's "4x on an FTE" line shows, a pilot can balloon on conversion or stall indefinitely (Syngene Q4 FY25 concall). What "good" looks like is conversion you can see flowing into capitalised sites and commercial revenue, not just rising RFP tallies management quotes to soothe a soft quarter.

Pipeline mix: development vs commercial molecules

A CDMO's book of molecules sits at different stages — early discovery, development, and commercial (the molecule is approved and being made at scale). Commercial molecules are the prize: sticky, high-volume, multi-year. Development-stage work is lumpier and can vanish if the molecule fails a trial. Where it hides: the deck and the call, often as "number of commercial molecules" or segment splits. Laurus Labs frames its CDMO ambition around exactly this shift — CDMO reached ~30% of company revenue (₹2,080 crore in FY26), with management targeting 50% by FY30, and its small-molecule CDMO grew 38% to ₹1,896 crore (Laurus Q4 FY26 concall) — the migration from a commodity ARV base toward contracted CDMO work is the structural story. The same migration toward contract work is visible across India's large integrated pharma — Dr. Reddy's Laboratories, Cipla and Zydus Lifesciences are all building out CDMO and contract-API arms alongside their own-product books. The thing to watch: is the mix tilting toward commercial, validated, repeat volume, or is reported growth leaning on one big development program that could end?

Capex and asset turnover — the actual business model

This is the engine. A CDMO converts capex into revenue at a ratio — asset turnover — and that ratio, against the cost of capital, is whether the model creates value. Where it hides: capex in the cash-flow statement and the deck; asset turnover is usually a management-stated target on the call. Divi's Labs, building out Kakinada, told analysts it expects an asset turnover of "1.5 to 1.6 in the next 4-5 years" on its new capacity (Divi's Q3 FY26 concall) against an FY26 capex revised to about ₹1,900 crore (Divi's Q3 FY26 concall). Laurus is heavier still — guiding "around INR3,000 crores in the next 2 years" of capex (Laurus Q4 FY26 concall). The discipline: a CDMO guiding for growth is really committing to spend capex now and earn the asset-turn later. If the turn comes in below plan, you funded a low-return asset. Track the gap between capex guided and revenue delivered, with a two-year delay built in.

Capacity utilisation

The percentage of the asset actually running — the single biggest swing factor on margin. Where it hides: the deck and the call, often qualitatively ("ramping," "near full"). It is the bridge between capex and profit: Syngene's margin drag was new sites not yet utilised (Syngene Q4 FY25 concall), and the recovery thesis is those sites filling. A CDMO commissioning capacity into a weak funding environment is the dangerous combination — fixed cost arriving with no demand to absorb it.

Gross margin and EBITDA margin

Gross margin reflects the value-add of the work (high-science discovery and complex synthesis earn more than commodity manufacturing); EBITDA margin then absorbs the fixed-cost and utilisation story. Where it hides: EBITDA margin is reported; gross margin and the reasons behind the move live in the call. Laurus printed a 61.4% gross margin and 28.9% EBITDA margin in Q4 FY26 (Laurus Q4 FY26 concall); Neuland ran an EBITDA margin of about 30% in stronger quarters but fell to 18% in Q3 FY26 (Inve data) — that volatility is the lumpiness of a smaller CRAMS book in one chart. Read margin alongside utilisation and mix, never alone.

Client concentration

How much revenue rides on the top one, five, or ten clients. It matters because the flip side of a sticky, multi-year contract is dependence: lose the anchor and a wing goes dark. Where it hides: the annual report's segment and customer notes, and the call when an analyst pushes. Syngene's long relationship with Bristol-Myers Squibb is "already a substantial business in scale," and management cautioned growth "will not be linear… it will grow around U.S. inflation at some level" (Syngene Q4 FY26 concall) — telling you both that the anchor is large and that you shouldn't model it compounding fast. Concentration isn't automatically bad; an undisclosed, deepening dependence on one client whose program could end is.

Biotech-funding cyclicality

Not a company metric — a macro one that sets the weather for the whole sector. Early-discovery demand tracks venture funding into biotech; when that contracts, the discovery funnel thins for everyone. Where it hides: management commentary and external funding data. Syngene named the "significant reduction in US funding into biotech" as a sector-wide drag (Syngene Q1 FY26 concall). The lesson: a soft year at a well-run CDMO may be the cycle, not the company — and a great year may be the cycle flattering it. Judge the operator against the funding backdrop it faced.

How do you value a CDMO?

The market pays a premium for CDMOs, and the premium has a logic: a long-dated, contracted revenue book looks like visibility, and visibility deserves a higher multiple than a price-taking generic. The danger is paying for visibility you don't actually have.

Look at what the market charges. Divi's Labs trades at a P/E of about 67.9x with an ROCE of 22% (Screener.in); Syngene at about 47.0x with an ROCE of 10.1% (Screener.in). Both are rich multiples by any absolute standard — but they are not the same bet. Divi's earns a high return on the capital it deploys; Syngene, mid-build with new sites not yet full, earns a single-digit ROCE while paying for tomorrow's revenue today. The same "CDMO premium" is being applied to a business compounding capital well and a business in the trough of its capex cycle.

So the right lens is not the headline P/E. It is capex-to-asset-turn against the cost of capital, read through the cycle. Ask: what asset turnover does this capex realistically earn (management's stated target, discounted for slippage), what ROCE does that imply once the asset fills, and am I paying a multiple that assumes the filled state already exists? A CDMO mid-build will always look expensive on trailing earnings because the E is depressed by under-utilised new assets — that is the cyclical version of the trap, and it cuts both ways: the premium can be deserved (the asset will fill and the turn will come) or a mirage (the funnel was a pilot that never converted).

The owner's frame: don't ask "is this cheap on this year's earnings?" Ask "what does a rupee of this company's capex earn once the plant is full, how confident am I that it fills, and what am I paying today for an asset-turn that arrives in two years?"

A worked case: Syngene, set in one quarter and cut three later

The cleanest way to feel the CDMO model is to watch guidance set ahead of the cycle and then cut as the cycle moved under it. In Q4 FY25, Syngene set its FY26 guidance trio; three quarters later, in Q3 FY26, all three came down together. (Illustration, not a view on the stock; figures as reported.)

The FY26 guidance trio was set in one window and cut in another. Set in Q4 FY25: revenue growth guided to mid-single-digits, EBITDA margin to the mid-20s, capex of $55 million (Inve data, set Q4 FY25) — and that $55 million was itself a trim from an earlier $85 million. Then in Q3 FY26 all three were revised down: revenue to -3% to -5%, EBITDA margin to 22-23%, capex to $45 million (Inve data, cut Q3 FY26). Three forecasts, three cuts — and they rhyme, because they share one cause. The biotech-funding drawback thinned the discovery funnel ("significant reduction in US funding into biotech," Syngene Q1 FY26 concall), revenue softened, and management trimmed the spend that builds tomorrow's capacity. FY26 closed with revenue up just 3% and reported PAT down about 36% to ₹317 crore (Inve data, FY26) — management framed the fall as nearer 20% on a before-exceptionals basis (Syngene Q4 FY26 concall).

Now read what held. The new-facility drag management warned about — "as and when these sites get operationalized… you will start seeing the drag come down" (Syngene Q4 FY25 concall) — played out: Q4 FY26 operating EBITDA margin recovered to 29% (Inve data, Q4 FY26), with management attributing the swing partly to "product mix" (Syngene Q4 FY26 concall). And the Bayview biologics site guided for H2 FY26 operationalisation stayed on track (Inve data). The capex-then-revenue lag was honoured in both directions: the spend that depressed margins is the same spend that, filled, lifted them.

The point isn't that management misled anyone — the guidance was cut transparently and the asset story delivered. The point is the texture: in a CDMO, the funding cycle, the capex lag and the guidance reset are one mechanism, not three, and you only see them as one by tracking each commitment against the quarter it was made. That is the job of Promise Tracker — and not the kind of thing anyone reconstructs by re-reading four transcripts by hand.

Red flags specific to a CDMO / CRAMS company

• Revenue growth carried by one development program. A big un-named molecule driving the quarter can be a development project that ends when the trial does. Ask whether growth is commercial, validated, repeat volume.
• Capex rising while asset turnover guidance slips. Spending more to earn a lower turn is value destruction in slow motion. Track guided turn vs delivered, with a two-year lag.
• Capacity commissioned into a weak funding cycle. A new plant is fixed cost. If it ramps as biotech funding contracts, the utilisation drag has nothing to absorb it.
• Undisclosed deepening client concentration. Sticky contracts are good; a quietly growing dependence on one anchor whose program could end is a single point of failure.
• "Strong RFP pipeline" used to paper over soft revenue. RFPs are leads, not orders, with a long conversion fuse. A funnel quoted to explain away a weak quarter, year after year, is a tell.
• Premium multiple justified by trough earnings. A mid-build CDMO looks "cheap" on forward and "expensive" on trailing earnings precisely because new assets aren't full. Don't pay for the filled state before it exists.

Frequently asked questions

A repeatable workflow

1. Read the funnel with its fuse. RFPs and order book from the concall — strong leads, long gestation; watch conversion into capitalised sites, not the count.
2. Check the pipeline mix. Commercial vs development molecules; is growth durable repeat volume or one program that could end?
3. Model the engine: capex-to-asset-turn. Guided turn discounted for slippage, against cost of capital, with a two-year lag from spend to revenue.
4. Watch utilisation and margin together. New sites drag until full; the recovery is them filling, not a one-off.
5. Map client concentration. From the annual report's customer notes — sticky is good, undisclosed dependence is a single point of failure.
6. Frame it against the funding cycle. A soft year may be biotech funding, not the operator. Audit guidance against what actually happened next.

Inve's KPI Screener lines up capex, EBITDA margin, asset metrics and segment mix across CDMO and pharma-services names — value, trend and a data-confidence flag per number — so the deck-mining takes minutes, not an afternoon. For a services business in a completely different shape, see how to analyse a SaaS company, where the recurring book is software, not a reactor; and for the own-product side, how to analyse a pharma company.

Where this lens can be wrong. The strongest case against everything above is that the metric this guide leans on — the capex-to-asset-turn, read through the cycle — depends on a funnel no analyst can verify and a funding cycle no one can forecast. You can read utilisation, asset turnover and client concentration perfectly and still be wrong because a marquee client's molecule failed a Phase III you had no way to assess, or because biotech funding stayed frozen two years longer than the consensus expected, leaving good capacity idle. Reading the operating numbers tells you whether a CDMO is built to fill its assets and compound through a cycle — disciplined capex, a tilting-to-commercial mix, a diversified book. It does not tell you when the funding cycle turns, and a well-run CDMO mid-build will earn a poor return for as long as the funnel stays thin. The honest claim is narrower than it looks: this analysis lowers your odds of paying a visibility premium for capacity that never fills, and raises your odds of owning a durable operator into the recovery. It cannot time the recovery.

The owner's question to sit with before buying any CDMO: across a full funding cycle — not this quarter's RFP pipeline — what does a rupee of this company's capex earn once the asset is full, how confident am I that it fills with repeat commercial work, and is the customer book diversified enough that no single molecule's failure takes a wing dark? If the answer leans on a strong funnel converting on schedule and one anchor client compounding forever, you have read the pitch, not the business.

Related reading

• How to Analyse a Bank Stock (NIM, CASA, GNPA, ROA)
• How to Analyse a Building Products Company (Pipes, Tiles)
• How to Analyse a Cables and Wires Stock