Long-Term Thesis
Long-Term Thesis — Onyx Biotec Limited
1. Long-Term Thesis in One Page
The long-term thesis is that Onyx is not a long-duration compounder — it is a sub-scale Indian sterile-injectables contract manufacturer whose 5-to-10-year case requires four things to land: Unit II's cephalosporin segregation suite rebuilding to FY25-era operating margins (16-17%), Unit I's IPO-funded large-volume parenterals upgrade lifting product mix into the ₹/unit tier above commodity sterile water, working capital releasing from its current 144-day debtor stretch, and management using the next governance window (main-board migration eligibility around FY28) to close the institutional credibility gap. None of those four pre-conditions is observable today; one of them (the OPM rebuild) prints in November 2026. Set against this is the structural ceiling — no US FDA or EU EMA pathway, single-site (Solan) operations, no long-term customer contracts, a customer base where the #1 competitor (Akums) is simultaneously a major loan-licensee, and a peer set in which every other listed Indian sterile CDMO held or expanded margin through FY22-FY26 while Onyx broke the wrong way. The honest 5-to-10-year frame is "tier-2 surviving Indian sterile CDMO compounding modestly off a low base if Unit II fills" — an operating-leverage option whose duration value is contingent on the first margin-recovery print.
The single most important multi-year conclusion. Even the friendliest reading of every long-term driver — Unit II fills, LVP commercialises, customer base diversifies further, working capital normalises — produces an Indian tier-2 sterile CDMO compounding at 12-14% revenue with 12-15% OPM. That is a respectable middle-tier outcome at the company's current sub-₹100 Cr revenue scale; it is not the 25-35% OPM profile of Caplin Point or Gland Pharma without an FDA-grade aseptic line that the ₹58 Cr market cap cannot fund from internal accruals. The 5-to-10-year ceiling is structurally lower than the peer group's, and the floor is lower than book.
2. The 5-to-10-Year Underwriting Map
The map below isolates the durable drivers from the near-term noise. Six things have to be at least partially true through FY2031 for ONYX to be a superior investment; six things would mark the thesis as failing.
The driver that matters most is Unit II utilisation. Every other driver in the table either (a) depends on the cost stack first being absorbed (LVP commercialisation routes through the same WHO-GMP umbrella; customer diversification follows volume; working capital release follows operating leverage; governance upgrade requires sustained profitability for main-board migration), or (b) is structurally absent on a 5-year view (FDA/EMA pathway). If H1FY27 OPM prints above 12%, four of the six drivers become incrementally underwriteable on a 3-5 year view. If it prints below 7%, all six drivers become harder to underwrite because the operating-leverage geometry that the whole long-term case rests on has been falsified by its own equipment.
3. Compounding Path
The compounding question is not "can Onyx grow revenue" — revenue has compounded at ~16% per year over FY22-FY26 — but "can it convert revenue growth into owner cash and durable returns on capital." The historical answer is no: across five years, ₹13.56 Cr of cumulative net profit produced ₹3.90 Cr of cumulative operating cash and -₹32.35 Cr of cumulative free cash flow, with the gap funded entirely by IPO proceeds and bank borrowings.
The two charts side by side tell the durable story: the top line is real and compounding, the operating profit line is increasingly disconnected from it, and the cash line never participated in the headline growth at all. ROCE peaked at 12.85% in FY24 and 12.16% in FY25 — within touching distance of mid-tier CDMO peers — then collapsed below the company's own ~8-9% cost of debt in FY26.
Three things to read from the scenarios. First, even the Base case (an honestly underwritten "Unit II fills, LVP commercialises, governance upgrades on main-board migration") implies roughly ₹55/share over five years against ₹32 today — an 11% IRR before dilution risk and execution friction. Second, the Bull case requires a step-function regulatory upgrade (FDA-pathway optionality) that the company has not articulated and the market cap cannot fund organically. Third, the Bear case ₹19/share is roughly book value haircut for sub-cost-of-capital returns — what the stock is mechanically worth if the next two halves confirm FY26 is the new normal. The distribution is wider than the multiples suggest because at this scale, a single tier-1 customer loss or a Schedule M audit failure can collapse the operating base independently of the margin curve.
4. Durability and Moat Tests
A durable thesis survives stress tests that look across the cycle. Five tests below — three competitive, two financial — are the ones a 5-to-10-year investor should care about, with the validation and refutation signals separated.
The cleanest five-year test is Financial test #5: returns on capital clearing the cost of debt for three consecutive years. If Onyx earns a sustained 12% ROCE through FY27-FY29, every other thesis driver mechanically becomes plausible (Unit II is absorbed, LVP is contributing, working capital is releasing, the IPO economics are working). If it cannot earn 8% ROCE across three years, the franchise is value-destructive at the equity level regardless of how the revenue line behaves — and at that point book value is a generous floor rather than a structural one.
5. Management and Capital Allocation Over a Cycle
Management credibility over a 5-to-10-year cycle is the most subjective input to the thesis and the one where the public record is thinnest. Founders Sanjay Jain (MD) and Naresh Kumar (WTD) have run Onyx for 20 years and built two operational units, one WHO-GMP-certified cephalosporin segregation suite, and a 100+ customer roster from a single Solan site — that is real operating heritage. Promoter holding has held at 65.10% since the November 2024 IPO through a 47% drawdown from the listing price, which is meaningful capital at risk. But the structure around the founders is the issue: the board has used every SEBI SME exemption available to it, the audit committee seats the CFO who is the audited executive, three independent directors were installed four months before the IPO and have served less than two years, ₹1.27 Cr of unsecured promoter loans remain on the books after the company used IPO proceeds in part to repay its own promoters, and ₹19.57 lakh annually flows to a promoter-controlled vendor (Imperial India) for "machinery repair."
Capital allocation through one IPO cycle is the longest data set available. The November 2024 IPO raised ₹25.38 Cr net; through 31 March 2026, ₹23.25 Cr (~92%) has been deployed against the prospectus objects. Three deployments tell the story: (1) ₹12 Cr to prepay debt — kept; (2) ₹5.99 Cr to general corporate purposes — kept; (3) ₹4.34 Cr (of ₹6.08 Cr earmarked) toward the Unit I LVP upgrade and ₹0.92 Cr (of ₹1.24 Cr) toward Unit II's cartooning line — partially deployed, machinery not commissioned 18 months in. The growth-capex tranches that the IPO equity story depended on are the parts that have slipped. No dividend has been declared, no buyback has been authorised at sub-book prices, and promoters have not added to their stake during the drawdown — three quiet capital-allocation signals that taken together suggest management is in survival mode on the existing platform rather than opportunistically deploying public-market capital.
On the 5-to-10-year question, the management ledger reads "competent operators, governance-lite shareholders." The franchise can be built — they have built it once already — but the structure does not have an external check on management decisions, the disclosure regime obscures the kind of mid-cycle communication a long-duration investor depends on, and the one explicit forward statement they have made on record (the May 2025 "better performance in FY26" guidance) was broken without explanation. A long-duration investor needs more credibility evidence than 18 months of public-company life has provided.
6. Failure Modes
Failure modes for a 5-to-10-year position in ONYX are concrete and observable. Six below are the ones that actually break the thesis, not generic execution risk.
The four high-severity failure modes share a common signature: they are all reversible-revenue or franchise-integrity events with no contractual or structural cushion. A long-duration position in this name is therefore not a "buy and forget" — it is "buy and read every half-yearly disclosure cover-to-cover," which is itself an argument against owning it for a multi-year hold without dedicated coverage capacity.
7. What To Watch Over Years, Not Just Quarters
Five observable milestones below would update the long-term thesis in either direction. Each is multi-year in horizon — not a near-term catalyst — and tied to a specific disclosure or external signal.
The long-term thesis changes most if Onyx files for NSE Main Board migration in or before FY29 — that single act would force the governance upgrade (audit committee restructure, quarterly Ind-AS results, Reg.17-18 compliance), prove the company has reached the profitability and scale floor required for it, and reset the institutional ownership profile that has shrunk from 8.7% to 1.1% in 18 months. Main-board migration is the closest thing to a step-function re-rating event the long-term thesis has, because it is observable, binary, and converts most of the governance discount into a measurable multiple expansion overnight.