Roivant Sciences Ltd. (NASDAQ: ROIV) — A Fortress Balance Sheet and a Real Serial Dealmaker, Now Priced for the Next Windfall Before It Lands

Independent equity research — an analyst’s published view. As-of date: 2026-07-04. Fiscal year ends March 31; “FY26” = year ended 2026-03-31.

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows takes no position, carries no price target, and confines itself to embedded expectations and scenarios.

Verdict: HOLD at ~$35 / accumulate-on-weakness toward base-case NAV (~$26–28) / not-a-short. Conviction: medium. Framing: a high-quality capital allocator and fortress balance sheet whose genuinely-de-risked lead asset is real — but whose stock now trades between base- and bull-case NAV, at its own all-time high, with ~63% of the price resting on optionality that will not be adjudicated until 2027.

Roivant is one of the better-run vehicles in biopharma: management has turned pharma cast-offs into ~$10B of realized and contingent value (Telavant→Roche ~$7.1B, Myovant→Sumitomo, Dermavant→Organon, and now a $2.25B Genevant/Moderna patent settlement), bought back ~$1.5B of its own stock at ~$11–13 when the market gave the pipeline away for free, and deployed another ~$700M increasing its Immunovant stake — a cash-backed conviction bet, not a press-release buyback. The lead asset, brepocitinib in dermatomyositis, is the first drug ever to win a placebo-controlled Phase 3 in that orphan disease and carries an FDA Priority Review with a Q3-CY2026 decision. That is not hope; that is a de-risked, near-term product. Underneath it sits a debt-free ~$4.3B net-cash pile stepping to ~$5.06B this month as the $950M Moderna cash lands. This is not a company you short: the balance sheet is a floor, the catalysts are real, and management is aligned on high-hurdle equity that only paid because the stock tripled.

But the tape has already paid for most of the good news. My sum-of-the-parts marks the visible pieces — parent cash, the 56% public Immunovant stake, Datavant/Arbutus, the settlement — at roughly $9B (~$13/share, ~37% of the price). The other ~$16B (~$22/share) is the market’s price for brepocitinib’s multi-indication franchise (not just its de-risked orphan lead) plus a “next-Telavant” premium on windfalls not yet sourced. That residual embeds the Goldman-style >$4–5B brepo peak, not the ~$2.3B dermatomyositis-only number, and it assigns real value to 2027 IMVT-1402 read-outs in Graves’ and myasthenia gravis that have not been run — in a platform that just failed a Phase 3 (batoclimab in thyroid eye disease, April 2026). The momentum is an extreme, low-drawdown, idiosyncratic one-way street (+220% in 12 months, Sharpe 5.2, now at the highs) — which cuts both ways: it is not a crowded systematic-momentum trade that unwinds on a factor rotation, but it is a single name priced for flawless catalyst execution, with a −79% five-year max-drawdown as the reminder of the downside distribution. The single fact that flips me bullish: a positive IMVT-1402 Graves’ or MG registrational read in 2027 (re-rates Immunovant toward ~$13B and validates the whole FcRn leg → bull-case ~$41). The single fact that flips me bearish: a broad JAK boxed-warning label that throttles the DM launch, or any 2027 IMVT miss (back toward ~$16–17). Tag: “A cash fortress at the top of its own rope.”

📈 Stock Price Action — Five-Year Event Map

Factual price history — no recommendation, no price target. Price moves are FACT (AZI adjusted 5-year series); attributed drivers are INTERPRETATION.

Roivant has completed a full biotech round-trip and then some. It traded near its ~$10 SPAC value through 2021, collapsed ~72% to an all-time low of $2.81 (2022-05-12) in the biotech bear market, then compounded roughly 12.5× to an all-time high of $35.39 (2026-06-30). It closed $35.13 (2026-07-02) — essentially at its record (relative-strength “peak” reading −0.7%). The 52-week range is $10.74 → $35.39; the stock is up ~+220% over the trailing twelve months and sits ~0.7% off its high.

Two nuances the rest of this memo relies on. First, the Telavant/Roche windfall (#4) produced a one-day decline, not a pop — the stock had already run ~4× into the news, a textbook priced-in event, and a warning about how much of Roivant’s good news tends to be anticipated. Second, the entire ~$12→$35 leg (#6–8) is catalyst-driven idiosyncratic — anchored on the VALOR read-out, the compounding Immunovant mark, and the Moderna settlement — not a broad biotech-beta move. This stock is repriced by events, not by the tape.

1. Executive Summary

Roivant Sciences is not a drug company in the ordinary sense; it is a develop-and-divest holding company (“holdco”). It in-licenses or acquires early-to-mid-stage assets that big pharma has undervalued, houses each in a nimble single-asset subsidiary it brands a “Vant,” advances it through the clinic, and monetizes the winners either by selling the Vant to a strategic acquirer or by retaining a majority equity stake in one that lists publicly. Revenue from selling medicine is negligible — $8.3M in FY26 against R&D of $681.8M and SG&A of $610.5M — so conventional margin, ROIC, and P/E analysis is meaningless. The company’s economic output is created option value, realized through M&A exits and marked-to-market equity stakes.

The model has, so far, worked unusually well. Three monetizations in roughly three years — Myovant→Sumitomo, Telavant (RVT-3101/anti-TL1A)→Roche for ~$7.1B in October 2023, and Dermavant (VTAMA/tapinarof)→Organon in October 2024 — seeded a debt-free ~$4.3B net-cash balance sheet (stepping to ~$5.06B this month as a $950M Moderna patent-settlement payment lands). Management then bought back ~$1.5B of deeply discounted stock at ~$11–13 and deployed a further ~$700M increasing its Immunovant stake.

The current portfolio is anchored by two clinical franchises. Brepocitinib (Priovant, ~72% owned; oral TYK2/JAK1 inhibitor) delivered the first-ever positive Phase 3 in dermatomyositis (VALOR: week-52 TIS 46.5 vs 31.2 placebo, p=0.0006, all key secondaries positive), holds an FDA Priority Review with a Q3-CY2026 PDUFA, and — if approved — becomes the first targeted oral therapy in that orphan disease, with follow-on Phase 3 read-outs in non-infectious uveitis and cutaneous sarcoidosis. IMVT-1402 (Immunovant, ~56% owned, NASDAQ: IMVT; next-generation anti-FcRn antibody) carries the larger but later opportunity, with registrational data in Graves’ disease and myasthenia gravis due in 2027 — though its predecessor, batoclimab, was fully discontinued in April 2026 after both Phase 3 thyroid-eye-disease trials failed. A third leg, the Genevant (~83%) LNP patent estate, produced a $2.25B settlement with Moderna ($950M fixed by July 8, 2026 + $1.3B contingent) with litigation against Pfizer/BioNTech still live.

At ~$35 (~$25.3B market capitalization) the stock has re-rated +220% in a year to its all-time high. Our sum-of-the-parts marks the visible/liquid assets at ~$9B (~$13/share, ~37% of the price); the residual ~$16B (~$22/share) is the market’s price for brepocitinib’s multi-indication potential plus a premium for future, un-sourced monetizations. That is an optionality-priced holdco, not an asset-value discount — a fortress balance sheet and a genuinely skilled allocator, but with little margin of safety in the marks and negligible weight on the bear case. No recommendation or price target appears below; this summary and the body discuss valuation only as embedded expectations and scenarios.

2. Business Overview

FACT. Roivant describes its own model plainly: “We advance our pipeline by creating nimble subsidiaries or ‘Vants’ to develop and commercialize our medicines and technologies” (10-K FY26, filed 2026-05-20). It was registered in Bermuda in 2014 as Valor Biotechnology Ltd., renamed Roivant Sciences Ltd. later that year, and is headquartered across London and New York. It files as a U.S. domestic registrant (10-K/10-Q; CIK 1635088). Founder Vivek Ramaswamy departed in early 2024 (to politics) and has effectively fully exited the stock; Matthew Gline is CEO, with Mayukh Sukhatme (President & Chief Investment Officer) and Eric Venker (President & COO, and Immunovant’s CEO).

How it makes money (INTERPRETATION). Not by selling drugs — FY26 product/royalty revenue was $8.3M. The three levers are: (i) exits — selling a de-risked Vant to a strategic acquirer at a large step-up over the cost to build it; (ii) stake appreciation — retaining majority ownership of a Vant (most importantly the publicly-listed Immunovant) whose value compounds as the pipeline de-risks; and (iii) non-operating monetizations — chiefly the Genevant LNP patent-litigation estate. Everything else on the income statement is cost: R&D to advance the Vants and a heavy corporate/SG&A layer to run the platform.

The asset map (FACT; ownership per FY26 10-K).

• Priovant — brepocitinib (oral TYK2/JAK1 inhibitor). Roivant ~72% basic / 66% fully-diluted. The near-term value driver; in-licensed from Pfizer.
• Immunovant (NASDAQ: IMVT) — IMVT-1402 (next-generation anti-FcRn monoclonal antibody). Roivant ~56% basic / 52% FD; ~44% public. Because IMVT is listed, a large slice of Roivant’s value is directly observable in IMVT’s market capitalization (~$7.9B). Core IP in-licensed from HanAll (Korea).
• Pulmovant — mosliciguat (inhaled soluble-guanylate-cyclase activator for pulmonary hypertension). ~97%; Phase 2. Licensed from Bayer.
• Genevant (~83%) — lipid-nanoparticle (LNP) delivery platform plus the Arbutus royalty interest and the COVID-vaccine LNP patent-litigation estate.
• Minority/associate holdings, carried at fair value: Datavant (private health-data company; 9% of Class A units, $233.2M mark), Arbutus Biopharma (public; 20% / 38.85M shares, $174.8M mark), plus renamed platform stakes Proxima (formerly VantAI) and PsiThera (formerly Psivant), and residual Dermavant milestone/royalty rights from Organon.

Platforms (FACT/INTERPRETATION). Roivant markets a technology layer atop the deal machine — Roivant Discovery (physics-based computational chemistry) and the Proxima AI protein-design stake — and points to its Datavant health-data heritage. These are real, but they are not yet the source of value: no platform-originated molecule has produced an exit. The two realized big wins (RVT-3101 and tapinarof) were in-licensed assets; the value came from capital allocation and clinical execution, not from proprietary discovery IP.

Verdict. Roivant is a levered, actively-managed portfolio of majority biotech stakes plus a ~$4.3B cash pile — an M&A option-generation engine, not an operating pharmaceutical franchise. “Revenue” is irrelevant; value is entirely forward, binary, and marked-to-market. Judge it as a capital-allocation vehicle, not a drug company.

3. Industry Dynamics

Three “industries” bear on the thesis: the FcRn autoimmune market (Immunovant), the oral TYK2/JAK immuno-dermatology market (Priovant/brepocitinib), and the capital-markets business of biotech incubation itself.

(a) FcRn inhibitors — a fast-growing but rapidly-crowding class (FACT). The reference product is argenx’s VYVGART (efgartigimod), the first approved FcRn blocker and the runaway leader: ~$4.2B in global product sales in FY2025, +90% year-over-year, across ~19,000 patients (argenx FY2025 release, 2026). Two more FcRn drugs are already approved in generalized myasthenia gravis — UCB’s rozanolixizumab (RYSTIGGO, 2023) and J&J’s nipocalimab (IMAAVY, FDA-approved April 2025), the latter positioned as a “pipeline-in-a-pathway” across multiple indications. Third-party sizing puts the class at roughly $2.1–2.4B (2024–25), growing toward ~$3.5–3.9B by the early 2030s on approved indications — an estimate that likely understates the ceiling, because the mechanism (broad IgG lowering) is being pushed into 20+ IgG-mediated autoimmune diseases. INTERPRETATION: the addressable market is large and expanding, but the competitive market is filling fast, with four programs converging on the same anchor indications (gMG, CIDP, Graves’, Sjögren’s, RA). Immunovant’s IMVT-1402 arrives as a fast-follower, and its own competitive edge — deeper, cleaner IgG suppression — must be proven in head-to-head-relevant data, not asserted.

(b) Oral TYK2/JAK immuno-dermatology (FACT). The reference product is BMS’s Sotyktu (deucravacitinib), a highly selective allosteric TYK2 inhibitor that uniquely escaped the JAK-class boxed warning because of its selectivity. This is the critical structural fact for brepocitinib: brepocitinib is a dual TYK2/JAK1 inhibitor — a broader, less-selective agent that will almost certainly carry the JAK-class boxed warning (major adverse cardiac events, thrombosis, malignancy, mortality — the 2021 FDA class action stemming from the ORAL Surveillance study). That is a genuine handicap in broad, competitive indications. But brepocitinib’s lead indication is dermatomyositis, an orphan disease with no approved therapies, managed for decades with steroids and off-label immunosuppressants — a setting where clinicians and regulators tolerate a boxed warning that would be punitive in mild psoriasis, and where being first-to-market confers real standing.

© The biotech-incubation “industry” (INTERPRETATION). As a capital-markets business, drug incubation has almost no barriers to entry — anyone with capital can in-license assets and stand up single-asset SPVs, and the field is populated with variants (BridgeBio, Ligand, PureTech, Centessa, and the historical Roivant-clone cohort). Through Marathon’s capital-cycle lens, the read is instructive: the 2020–21 biotech bubble flooded the sector with capital, and Roivant timed its two big exits (2023–24) into the recovery window — selling assets when strategic acquirers were paying up. That is timing discipline, not a structural barrier. Within the drug markets, the FcRn field is a textbook late-cycle setup — argenx’s spectacular returns have attracted a flood of me-too capital (J&J, UCB, Immunovant, others), which the framework predicts compresses pricing and returns over time. Conversely, dermatomyositis is under-capitalized (no approved drugs, orphan population) — an early-cycle, uncontested niche where brepocitinib is the first mover.

Verdict. Structurally mixed and, on balance, only moderately attractive. The underlying autoimmune drug economics are strong for winners — patent-protected, high-price specialty products — but the FcRn arena is a crowded, capital-attracting late-cycle field where Immunovant is a follower, while brepocitinib’s best market (orphan dermatomyositis) is attractive precisely because it is small and uncontested. The holdco business layer itself has no barrier to entry.

4. Competitive Position

Does a holdco have a moat? We pressure-test each candidate against Greenwald’s taxonomy (supply/cost advantage, demand/captivity, economies-of-scale-plus-captivity, with intangibles as the practical fourth).

(i) Capital-allocation and deal-making skill (INTERPRETATION). This is Roivant’s actual differentiator and the source of every dollar of realized value. Telavant→Roche (~$7.1B), Myovant→Sumitomo, and Dermavant→Organon are strong evidence that management can identify de-risked mid-stage assets others undervalue, run efficient trials, and sell into strategic demand. But it is not a Greenwald barrier to entry. Deal-making is an execution skill; it is person-dependent (Ramaswamy is gone; the current core is Gline/Sukhatme/Venker); it is not protected by patents, scale, or customer captivity; and its repeatability is an open question — a handful of exits is a small sample, and survivorship bias is severe, because the many quiet or wound-down Vants are not in the highlight reel. A skilled allocator can outperform without possessing a moat, and the market re-underwrites that skill every cycle.

(ii) Asset-level intangible moats — real, but binary and patent-life-limited.

• Brepocitinib in dermatomyositis is the strongest single moat in the portfolio. FACT: VALOR hit hard — week-52 mean TIS 46.5 vs 31.2 placebo (p=0.0006, n=241), superior on all key secondaries, published in the New England Journal of Medicine — and the FDA granted Priority Review with a Q3-CY2026 PDUFA. If approved, brepocitinib is potentially the first-ever approved therapy in DM: a genuine first-mover intangible moat (orphan exclusivity, patents, standard-of-care reference status). The offset: the dual-JAK boxed warning that will likely cap its expansion into larger, more competitive indications (uveitis, sarcoidosis) where cleaner mechanisms compete.
• IMVT-1402 vs argenx (FACT). IMVT-1402 is a fast-follower into a field where argenx already has a ~$4.2B, first-mover, 19,000-patient franchise and two other approved competitors. Immunovant’s pitch — best-in-class deep IgG lowering with a clean lipid/albumin profile — is credible but unproven at registrational scale, and its own predecessor batoclimab raised LDL, lowered albumin, and failed Phase 3 in thyroid eye disease (both trials, April 2026). Absent a differentiated new indication — Graves’ hyperthyroidism, where no FcRn is yet approved and batoclimab’s Phase 2 showed durable, disease-modifying responses — IMVT-1402 is a second/third entrant fighting for share of an incumbent’s market: a demand-captivity disadvantage, not a moat. Its edge, if any, is being first in net-new indications, not in displacing VYVGART.
• HanAll-licensed FcRn IP is a contractual position (up to $420M in milestones plus mid-single-to-mid-teens royalties owed to HanAll), not proprietary intangible ownership.
• Genevant LNP patents are the standout hard-IP asset. FACT: Genevant (~83% Roivant) and Arbutus reached a $2.25B global settlement with Moderna (March 2026): $950M by July 8, 2026 plus $1.3B contingent on a favorable appellate ruling; the Pfizer/BioNTech case continues after a favorable September-2025 claim-construction ruling. This is a genuine patent-based intangible producing near-certain cash — but it is a one-off legal/royalty windfall, not a recurring franchise.

(iii) Scale / cost / network effects at the holdco level: none. Roivant has no manufacturing scale, no at-scale commercial infrastructure (it divests before scale), no customer captivity, and no network effects. The computational platforms could someday create a supply-side edge, but no evidence yet ties them to a financial outcome that would deteriorate without them — by our own test, not yet a moat.

Verdict. No durable enterprise moat. Roivant is a levered bet on (a) management’s non-durable, person-dependent deal-making skill; (b) a few binary, patent-life-limited asset moats — brepocitinib’s genuine first-mover position in orphan dermatomyositis being the best, IMVT-1402 being a fast-follower into a crowded FcRn field; and © the one hard, cash-generating Genevant LNP-litigation IP. It is a portfolio of options, not a franchise: real value and real skill, but not a business whose returns are protected by a structural barrier to entry.

5. Growth History and Forward Opportunities

Framing (INTERPRETATION). There is no revenue-growth story to analyze — Roivant is an option-value engine, not an operating company. The “growth” that matters is (a) the historical cadence of monetizing Vants at large step-ups, and (b) a dense 2026–2027 catalyst stack that converts balance-sheet cash into approved-asset value.

Historical value-creation cadence (FACT). The model has printed large liquidity events in ~three years: Myovant→Sumitomo (2023); Telavant (RVT-3101/anti-TL1A)→Roche (~$7.1B upfront, Oct-2023) — an asset in-licensed from Pfizer in December 2022 for a modest upfront plus a Pfizer equity stake, sold ~10 months later for roughly ~$5.1B to Roivant; and Dermavant (VTAMA)→Organon (Oct-2024, up to ~$1.2B: $175M upfront + milestones + royalties). The “growth algorithm” is therefore: in-license cheaply → de-risk → sell or partner at a step-up → recycle the cash. It has worked — but each data point is discrete and deal-dependent, not a compounding revenue base.

Forward catalyst calendar — the real “growth” (FACT unless noted):

Adjacencies (INTERPRETATION). Two genuine “pipeline-in-a-product” optionalities: brepocitinib multi-indication expansion — management calls it “an embarrassment of riches,” noting it “has worked almost everywhere it’s been tested” (positive Phase 2 lichen planopilaris, positive cutaneous sarcoidosis Phase 2), with DM/uveitis/sarcoidosis only the opening set; and IMVT-1402’s broad FcRn TAM across Graves’ (management cites ~330,000 prevalent refractory U.S. patients), MG, CIDP, difficult-to-treat RA, Sjögren’s, and CLE. Mosliciguat (a 38% PVR reduction shown in PAH) and the Genevant royalty/settlement stream are smaller, discrete adders.

Quality of growth (INTERPRETATION). High-optionality but low-visibility and binary. Every value node is a catalyst — a PDUFA, a topline, an appellate ruling — not a durable, recurring franchise, and it is funded by the balance sheet, not by operations (~$750M/yr consolidated burn against ~$4.3B cash). Positives: a real, approvable lead asset (VALOR hit all ten ranked endpoints), a de-risked FcRn platform with human validation, and a fortress balance sheet that removes financing risk through the catalysts. Negatives: DM is an orphan market (~38,500 U.S. prevalent; consensus peak-sales estimates range from ~$350M to “mid-to-high billions”); the 2027 IMVT reads are unhedged binaries; and the serial-exit engine has no guarantee of repeating.

Verdict. High-optionality but low-certainty growth. This is not compounding; it is a laddered sequence of binary bets financed by a war chest. The lead node (brepo/DM) is now near-certain to reach market, which raises the floor materially versus a year ago — but the bulk of the market-implied value still rests on unproven 2027 outcomes and on management’s ability to monetize them.

6. Financial Quality

Reframe (INTERPRETATION). Roivant has effectively no operating business to analyze on a margin basis. FY26 revenue of $8.3M against $681.8M R&D + $610.5M SG&A produced a −$1,285.3M operating loss; the ROIC-computed operating margin of −15,561% underscores how meaningless conventional ratios are here. Three things matter: how much net cash sits on the balance sheet, how fast it burns, and the value of the stakes.

(a) Balance sheet — best-in-class for a clinical-stage name (FACT). At 3/31/26: cash + short-term investments $4,291.8M ($1,419.2M cash + $2,872.6M marketable securities), plus $408.0M of long-term investments at fair value (Datavant $233.2M + Arbutus $174.8M). Total debt is only $107.4M of capital leases — there is no funded debt. Net cash ≈ $4.29B. Current ratio 18.4×; tangible common equity ~93% of assets; equity before minority interest $4,527.4M, with minority interest of $765.0M (up from $499.6M, reflecting the 44%-public Immunovant and the Genevant minority). This is one of the strongest balance sheets in clinical-stage biopharma, and it is about to get stronger: the $950M Moderna fixed payment lands on/before July 8, 2026, taking pro-forma cash to ~$5.06B.

(b) Runway (FACT + INTERPRETATION). FY26 cash from operations was −$750.3M; capex only −$8.2M; FCF −$758.6M. Naïve runway = $4.29B ÷ ~$0.75B ≈ ~5.7 years before any new exit or milestone. Two adjustments extend this materially. First, Immunovant self-funds: IMVT held ~$902M of its own cash and finances its own anti-FcRn program via its own balance sheet and equity raises, so parent/ex-Immunovant burn is meaningfully lower than the ~$750M consolidated figure (we estimate parent burn in the ~$400–550M range — the 10-K does not cleanly segment it, an open question — implying parent-only runway well beyond a decade). Second, the ~$5.06B pro-forma cash comfortably funds every catalyst above. Management’s claim that current cash “will get our current pipeline to profitability” is, on these numbers, credible.

© Quality of earnings — two enormous one-time gains distort the record (FACT). GAAP net income is uninformative here without normalization:

• FY24 net income of +$4,349M was the Telavant→Roche gain. Underlying FY24 was a pre-revenue R&D burn (operating loss −$824.9M). Anyone anchoring on FY24 “profitability” is anchoring on an asset sale.
• FY26 net loss to common of −$299.8M is flattered by a +$770.2M Genevant/Moderna litigation gain (recognized as income with an offsetting settlement receivable at 3/31/26 — non-cash in FY26, cash in July 2026). Strip it out and normalized FY26 net loss ≈ −$1,070M — the truer read of the underlying burn. Reported −$300M looks far better than operating reality.
• Between the bookends: FY23 net loss −$1,009M; FY25 net loss −$172M (with $373M of discontinued-ops income from Dermavant). There is no run-rate earnings number here — only a run-rate burn.

(d) Interest income — a real, high-quality, but shrinking offset (FACT). Net interest income was +$178.1M in FY26 (essentially all income; no meaningful interest expense), down from a $258.4M peak in FY25. On ~$4.3B of cash this offsets roughly a quarter of gross opex — a legitimate, durable cushion so long as the cash is not fully deployed, but a declining one as rates and the cash balance fall.

(e) Tax (FACT/OPEN QUESTION). FY26 recorded income-tax expense of $133.3M despite a pretax loss, driven by taxable subsidiary-level income (the Genevant GmbH settlement is taxable). Roivant pays real cash tax on its litigation monetizations — a drag on the net proceeds from the Moderna cash.

(f) SBC — the main value leak (FACT). Stock-based compensation was $346.2M in FY26 (up from $289.0M and $199.6M in the prior two years), equal to ~46% of the cash burn. Combined with option exercises, SBC drove a rise in the share count despite the buyback (below). This is the standard biotech-platform tax on shareholders and materially blunts buyback accretion.

Verdict. High-quality balance sheet; low-quality (and largely irrelevant) income statement. The correct frame is a de-risked, essentially debt-free ~$4.3B net-cash pile (pro-forma ~$5.06B) against a consolidated ~$750M burn that overstates parent cash need because Immunovant self-funds. “Do economics improve with scale?” does not apply — this is a capital-allocation vehicle whose value is deal IRR plus stake appreciation, both strongly positive to date. Normalize out the FY24 Telavant and FY26 Genevant gains before any valuation, and treat the ~$346M annual SBC as the principal leakage.

7. Capital Allocation

This is the most important section for this name — capital allocation is arguably the entire thesis.

(a) The flywheel (INTERPRETATION). Source assets cheaply (often from big pharma’s discard pile) → build a Vant → sell or partner at a large markup → recycle the cash into (i) buybacks when the stock trades below the value of cash + pipeline, and (ii) doubling down on the best remaining Vant. FY24–FY26 was a textbook execution: bank the Telavant/Dermavant proceeds, then simultaneously buy back deeply discounted ROIV stock and accumulate more Immunovant.

(b) Buybacks — genuinely elite, executed into maximum pessimism (FACT). Two authorizations: $1.5B (April 2, 2024, fully exhausted by ~6/30/25) and $1.0B (June 25, 2025). FY25 repurchases totaled $1.18B when ROIV traded ~$10–12; FY26 repurchases were ~24.0M shares for ~$314.9M (blended ~$13.1), leaving $890.3M remaining at 3/31/26. Cumulatively Roivant retired on the order of ~130M+ shares for ~$1.5B at a blended ~$11–13, when the market ascribed almost no value to the pipeline behind the cash. At $35 those repurchased shares are worth ~$4.5B+, a mark-to-market gain on the order of ~$3B (~$4 per current share of accretion). This is one of the more accretive biotech buyback programs on record — buying a dollar of net cash for ~50–60 cents.

• Caveat (FACT): the buyback did not shrink the share count in FY26 — shares outstanding rose from 695.9M to 720.4M because SBC and option exercises (+$193.6M of option proceeds) more than offset the $314.9M repurchased (net repurchase only ~$126M). In FY25 the buyback overwhelmed dilution (806.7M→695.9M). So the accretive-buyback story is real historically but is now running roughly in place against dilution at higher prices.

© Immunovant accumulation — a cash-backed conviction bet, easily misread (FACT). The Form 4 corpus shows “Roivant Sciences Ltd.” as a code-P open-market purchaser of large IMVT blocks: 16,845,010 shares @ $20.00 (Jan-13-2025, ~$337M) and 16,666,666 @ $21.00 (Dec-12-2025, ~$350M), plus earlier buys. These are not ROIV buybacks — they are Roivant buying Immunovant shares (filed because Roivant is an IMVT insider), deploying ~$700M+ to increase its economic interest in its highest-value Vant. A direct, cash-backed conviction signal in the FcRn franchise — materially more bullish than a generic buyback, and a common misread.

(d) M&A / divestiture track record — elite, with a luck component (FACT/INTERPRETATION). The wins are extraordinary: Telavant (in-licensed from Pfizer Dec-2022 for a modest upfront + Pfizer equity, sold to Roche Oct-2023 for ~$7.1B upfront — roughly ~$5.1B to Roivant in ~10 months, one of the great trades in biotech history); Myovant to Sumitomo (~$1.7B, 2023); Dermavant/VTAMA to Organon (up to $1.2B, Oct-2024). Four large monetizations make the pattern more skill than luck — but the Telavant multiple rode a right-place-right-time TL1A frenzy that will not recur at that magnitude, and the model depends on a continuous supply of cheap assets and receptive M&A windows, both cyclical.

(e) R&D / SG&A intensity (FACT). R&D rose from $550.4M to $681.8M (+24%) as Priovant and Immunovant advanced; SG&A from $591.4M to $610.5M. For a holdco with $8M of revenue, SG&A of $610M is heavy — the cost of running a multi-Vant platform plus stock-comp-laden corporate overhead, and the line most exposed to the criticism that the structure is expensive.

(f) Insider behavior — steady selling, near-zero conviction buying (FACT). Parsing the full 408-filing Form 4 corpus (2021-08 to 2026-06):

• Open-market insider purchases (code P): essentially none. The only management open-market buy in five years is CEO Matt Gline — 3,315 shares @ $15.07 (Sep-2025, ~$50K), a token. Every other “P” is either Roivant buying Immunovant (above) or a 10%-holder accumulating. No executive or director has bought ROIV with their own cash near $35.
• Insider/holder sales (code S): ~175.6M shares for ~$2.21B (2021–2026), scaling up with the price (avg realized $4.75 in 2022 → ~$30 in 2026 YTD). The largest sellers are early financial backers exiting — the QVT complex ~$850M+, founder Ramaswamy ~$225M (fully exited), SoftBank Vision Fund ~$243M, Viking Global ~$257M, Sumitomo ~$130M — a natural post-deSPAC VC/founder unwind, an overhang rather than a fundamentals red flag. Management sales are programmatic 10b5-1 (Venker in a metronomic 200,000-share monthly clip; President Frank Torti 3.0M shares/~$82M in Feb-2026; Sukhatme ~1.34M; Gline ~290K).
• Read (INTERPRETATION): the selling is a natural VC/founder unwind into strength, not a pipeline warning — but the complete absence of open-market management buying at $35 is a genuine, if mild, cautionary tell. Insiders are monetizing, not accumulating, at today’s price.

(g) Compensation — genuinely shareholder-aligned, high-hurdle equity (FACT). The FY2024 Senior Executive Program for Gline, Sukhatme, Venker (and a matching Torti grant) is built on PSUs earned only on rigorous 30-day-VWAP share-price hurdles over a 5-year performance period with a 2-year post-vesting hold, granted “all-in” with the explicit expectation of no further equity for five years:

At 3/31/25 (stock $10.09) none had vested — the hurdles represented +49% to +197% moves. At $35 today, all six hurdles (including the $30 top tranche) are cleared and the awards are effectively fully earned. Venker also holds a contingent value award that settles into Immunovant stock above a $14.46 hurdle — creative cross-Vant alignment. Management got paid only because the stock tripled off the buyback-era lows — a well-designed program, though now fully in the money (removing some forward incentive tension).

Verdict. Capital allocation is genuinely elite — arguably the whole thesis. Management turned pharma cast-offs into ~$10B of realized/contingent value, bought back ~$1.5B of deeply discounted stock at ~$11–13, deployed ~$700M+ increasing its Immunovant stake, and pays itself on 5-year, high-hurdle, long-hold PSUs. The two blemishes: ~$346M/yr SBC that now offsets the buyback, and heavy insider/VC selling with zero open-market management buying at $35. On balance, capital has been allocated intelligently and with unusual discipline; the open question is whether the source-cheap/sell-high engine keeps running now that the balance sheet is fully valued.

8. Changes and Headwinds — Last Two Years

Portfolio monetization (FACT, thesis-strengthening). Telavant→Roche (~$7.1B, Oct-2023) and Dermavant→Organon (Oct-2024) converted two Vants to cash and milestones, seeding the ~$4.3B balance sheet and the buyback program, and validating the model.

Brepocitinib VALOR — the single biggest positive change (FACT). The September-2025 topline was the first positive Phase 3 in dermatomyositis, ever, hitting the primary (week-52 mean TIS) and all key secondaries at 30 mg, with a clean dose response and a meaningful steroid-taper benefit. It moved brepocitinib from “pipeline” to “product”: NDA accepted with Priority Review, PDUFA Q3-CY2026, launch guided to ~end-September 2026.

Moderna LNP settlement — major, recent, positive (FACT). On March 3, 2026, Genevant and Arbutus reached a $2.25B global settlement with Moderna — $950M by July 8, 2026 (Genevant’s allocated share ≈ $770.2M, the balance-sheet receivable) plus $1.3B contingent on a favorable appeal — described as the largest disclosed pharma patent settlement. It drove Q4-FY26 income from continuing operations of +$355.7M and narrowed the FY26 GAAP loss to ~$299.8M. The Pfizer/BioNTech LNP case remains live (favorable September-2025 claim-construction ruling), plus OUS actions with 2026 hearings — residual optionality on top of the $1.3B contingent tail.

Immunovant pivot — batoclimab→IMVT-1402 (FACT, net-negative on batoclimab, strategy intact). The pivot was validated the hard way: on April 2, 2026, both Phase 3 batoclimab thyroid-eye-disease trials FAILED the primary proptosis endpoint, and Roivant/Immunovant discontinued batoclimab across all indications. This is a real negative — a Phase 3 miss on a lead antibody and the retirement of the near-term batoclimab-launch option some had modeled — that pushes the entire FcRn value onto the 2027 IMVT-1402 read-outs. Management reframes it as de-risking (the high-dose data still supported “deeper-is-better,” now carried by next-generation IMVT-1402), but the skeptic should note it followed pre-emptive hedging on the call a quarter earlier.

Leadership (FACT). Founder Vivek Ramaswamy departed (early 2024); Matt Gline is CEO. Continuity of the deal-making core is intact, but the highest-profile promoter is gone.

Capital return (FACT). The $1.5B repurchase program (April 2024) was exhausted and a $1.0B program authorized (June 2025), cumulatively ~$1.5B+ repurchased at ~$11–13.

JAK-class label overhang — the key structural headwind (FACT/INTERPRETATION). Because brepocitinib inhibits JAK1 (unlike pure-TYK2 deucravacitinib, which escaped the box), it is likely to carry the JAK-class boxed warning (MACE, VTE, malignancy, mortality — the 2021 FDA class action). Mitigant: brepo’s own psoriatic-arthritis data showed no MACE/VTE at 52 weeks, and DM is a severe orphan disease where clinicians tolerate more risk. Open question: the exact label language at the Q3-2026 approval — a broad box vs. narrow language materially affects uptake.

Verdict. On net, these changes STRENGTHEN the thesis. The two largest — VALOR/Priority Review and the $2.25B Moderna settlement — are unambiguous positives that de-risk both the lead asset and the balance sheet. The batoclimab TED failure and the JAK-label overhang are real offsets, but they degrade option branches rather than the core thesis. The floor is higher than two years ago; the ceiling now depends on the 2027 IMVT-1402 binaries.

9. Risk Analysis

Catastrophic-loss / total-loss assessment (INTERPRETATION). A total loss is highly improbable: ~$9B of hard/visible value (cash + liquid public stakes + settlement receivable) sets a floor at roughly $12–13/share even if every private-pipeline option expired worthless. The realistic downside is a de-rating to that hard-value floor (~$16–17 including a modest option value), i.e., roughly −50% from $35 — not a wipeout. The asymmetry is set by the binary catalysts, not by solvency.

10. Valuation Discussion (Embedded Expectations)

Roivant is a holdco, not an operating company, so EV/EBITDA, P/E, and P/S are structurally meaningless (ROIC shows FY26 GAAP P/E ~−40×, P/S ~2,300×, EV/EBITDA ~−9× — all noise). The only defensible frame is a sum-of-the-parts NAV: mark the visible/liquid pieces, then solve for the residual the market pays for the wholly-owned private pipeline.

Hard / visible value (base case):

Solve for the implied brepocitinib value. Market cap $25.3B − ~$9.1B ≈ ~$16.2B (~$22/share) is what the tape pays for the wholly-owned private pipeline dominated by brepocitinib (Priovant, 72%) plus serial-value-creation optionality. Grossing the residual up for the 72% Priovant stake implies a ~$22B gross enterprise value on brepocitinib alone. Is that defensible? Only under a multi-indication bull. Sell-side peak-sales estimates span an enormous range: ~$2.3B by 2032 (Evaluate, DM-centric) versus >$10B unadjusted (Goldman) across DM + uveitis + sarcoidosis + others. A ~$2.3B DM-only peak supports maybe $6–9B of gross value (3–4× peak, risk-adjusted for post-VALOR ~80%+ DM approval odds); a ~$22B implied gross value therefore requires the multi-indication $8–10B franchise — i.e., the market is already underwriting close to the Goldman bull.

Bear / Base / Bull SOTP:

The current $35 sits between base and bull. Notably, the base case (~$19–20B) is almost exactly the FYE-3/31/26 market cap of $19.22B — i.e., the ~$6B of market value added since March is the market moving from “base” toward “bull” on the back of the rising IMVT mark ($5.06B→$7.91B) and continued brepocitinib de-risking.

Embedded-expectations narrative — what must be true at $35. The market is underwriting, in order of load-bearing weight: (1) brepocitinib approval (de-risked; correctly priced); (2) brepocitinib becoming a multi-indication franchise (the aggressively-priced piece — the residual implies the Goldman peak, not the Evaluate DM number); (3) the FcRn class working and IMVT holding its ~$7.9B mark against an entrenched argenx (a single-name mark that can gap either way; the 2027 Graves’/MG reads are the swing factor); and (4) “next-Telavant” optionality priced partly as free. Key sensitivity: the IMVT stake alone is ~17% of market cap — every ±20% move in IMVT is ~±$1.2/share, before any brepo or deal news.

AZI own-history percentiles (context, not verdict). P/B 6.3× sits at the 83rd percentile of the stock’s own range and the composite valuation at the 91.5th — “priced at the richest end of its history.” But book value is the wrong yardstick for a stake-holding biopharma (consolidated IMVT, Level-3 Datavant, no product revenue), so the percentile is corroborating color, not proof. NAV is the right frame.

Verdict. Richly valued on a sum-of-the-parts basis. The market pays roughly base-case NAV plus a large call-option premium on brepocitinib’s multi-indication potential and management’s serial-monetization record. Hard/visible assets (~$9.1B, ~$12.6/share) are only ~37% of the price; the other ~63% is a levered bet on pipeline execution and continued IMVT re-rating. This is an optionality-priced holdco, not an asset-value discount — little margin of safety in the marks, and negligible weight on the bear case (~$16–17).

11. Variant Perception

Consensus. The market has re-rated ROIV as a de-risked, catalyst-rich, cash-fortress biopharma with an approvable lead asset and a validated serial-monetizer management team — hence +220% in a year to a ~$25.3B cap. Sell-side is broadly constructive (Leerink, JPMorgan, Citi, TD Cowen, Goldman, Jefferies, Piper, Wells all engaged on the last call), treating the balance sheet as a floor and the pipeline as free-ish optionality.

Strongest bull case. A serial value-creation machine holding its best-ever hand: (1) brepo/DM launching Q3-26 as the first targeted therapy in DM with category-defining data, plus a pipeline-in-a-product (uveitis, sarcoidosis, LPP); (2) best/first-in-class FcRn optionality in IMVT-1402 across Graves’, MG, CIDP, Sjögren’s, RA; (3) a ~$4.3B net-cash fortress (~$5.06B pro-forma) that funds everything to profitability with no financing risk; (4) aligned, buyback-happy management (~$1.5B repurchased, ~$700M more into Immunovant); (5) residual litigation upside (Pfizer case, $1.3B contingent). Bull framing: you’re paying ~$21B EV for brepo + a de-risked FcRn platform + Datavant + a settlement receivable — cheap for the option set.

Strongest bear case. Strip the visible assets and the market capitalizes ~$16B for a still-largely-unproven private pipeline — pinned on brepo peak sales in an orphan market (consensus as low as ~$350M) and on 2027 IMVT-1402 binaries that have not been run, in a platform that just failed a Phase 3 (batoclimab/TED, April 2026). It is a levered bet on binary readouts plus a deal-making streak that may not repeat (no recurring cash engine; the “next Telavant” is a hope, not a plan), with a JAK boxed-warning overhang capping brepo and a credible fast-follower (argenx) entering Graves’/MG. Bear framing: great balance sheet, priced as if the options are already in the money.

The 3–5 assumptions that matter most: (1) brepo is approved on time with a commercially workable label, and DM peak sales justify a multi-billion asset; (2) IMVT-1402 hits in Graves’ and/or MG in 2027 — the single largest swing factor; (3) management monetizes at least one more Vant at a Telavant-scale step-up; (4) the $1.3B contingent Moderna payment and/or the Pfizer case resolve favorably; (5) argenx/competition does not commoditize the FcRn Graves’ opportunity before Roivant reaches market.

What would falsify each side. Falsifies the bull: an IMVT-1402 Graves’/MG miss in 2027; a broad brepo box + slow DM launch; failure to sign another value-creating deal within 12–24 months; adverse /Pfizer rulings. Falsifies the bear: brepo launches cleanly with a narrow label and strong early scripts; a positive 2027 IMVT read; another Telavant-scale exit; full $2.25B Moderna cash realized.

Momentum / factor read (feeds the framing). An extreme, low-drawdown, idiosyncratic one-way street: +220% over 12 months, Sharpe 5.23, Sortino 12.7, max drawdown only −12.8% over the run, now at the all-time high. Zoom out and the cyclicality reasserts: −36.5% max-DD over three years, −79.2% over five years (the 2021–22 crash) — a stock that both compounds in a straight line and halves. Critically, despite the +220% run the model reads market beta ~0.86–0.95, a large biotech-industry loading (~1.32), R² only ~37% (so ~63% of variance is idiosyncratic), and a slightly negative style-momentum loading — i.e., this is single-name catalyst risk, not a crowded systematic-momentum trade. That is a double-edged read: it is less exposed to a factor-rotation unwind, but more exposed to a single disappointing binary, with little market-beta cushion. Factor-similar peers (XBI, LABU, DNLI, IMTX, IDYA) confirm the model treats ROIV as a leveraged clinical-stage biotech, not a defensive holdco. Consensus is priced near the top of its own range on the good scenario continuing — the asymmetry from here is set by binary events, not by trend.

12. Fact vs. Interpretation

13. Open Questions

1. Parent-only cash burn. The 10-K does not cleanly segment parent/ex-Immunovant burn from the ~$750M consolidated figure. Our ~$400–550M estimate materially changes the runway story; the December Investor Day slides and 10-K segment note are the places to pin it.
2. Exact brepo label at Q3-2026 approval. Broad JAK box vs. narrow language is the single biggest determinant of DM (and follow-on) uptake.
3. Datavant realizable value. A Level-3 mark ($233.2M at 9%); Datavant’s ~$1.43B revenue and private-market comps could imply a higher or lower realizable value on any liquidity event.
4. Is the exit engine repeatable? Two great exits are a small sample; the “next Telavant” is priced partly as free optionality. What is the current BD pipeline, and at what cost of capital?
5. IMVT-1402 differentiation vs. argenx. Does “deeper IgG lowering” translate into a registrational-scale efficacy or durability edge in Graves’/MG, or is it a marketing distinction?
6. Fully-diluted share count. Options/RSUs (and now-earned PSUs) push above 720.35M, modestly lowering per-share NAV; confirm the FD count and the PSU share issuance.
7. ** appeal and Pfizer case timelines** — the magnitude and probability of the $1.3B contingent payment and the separate Pfizer/BioNTech recovery.

14. What Must Be True

Bull case — what must be true, and its falsification test.

• Thesis: brepocitinib becomes a multi-indication franchise (DM approved and launched cleanly, then uveitis/sarcoidosis), and IMVT-1402 delivers at least one positive registrational read in Graves’ or MG in 2027 (re-rating Immunovant toward ~$13B), and management sources another value-creating deal — collectively pulling NAV toward the ~$41 bull case.
• Falsification test: a broad JAK boxed warning that throttles the DM launch (first-year scripts well below the trajectory implied by a $4–5B peak), or a negative IMVT-1402 Graves’/MG read in 2027. Either single event breaks the bull case and pulls the stock toward base or below.

Bear case — what must be true, and its falsification test.

• Thesis: the ~$16B residual for the private pipeline is unwarranted — brepo is a DM-orphan drug worth ~$2–2.5B peak (not a multi-indication blockbuster), the 2027 IMVT binaries disappoint or the FcRn field commoditizes, and no new Telavant materializes — so the stock de-rates toward hard-value NAV (~$16–17).
• Falsification test: brepo launches with a workable label and strong early uptake, and a 2027 IMVT-1402 read hits — validating the multi-indication + FcRn legs and confirming the market’s forward pricing. Either event breaks the bear case.

The pivot both cases share: the 2027 IMVT-1402 Graves’/MG read-outs are the fulcrum. Until then, brepo’s launch trajectory and label are the near-term swing, and the ~$9B hard-value floor caps the downside at roughly −50%.

This article takes no investment position and sets no price target; the sole opinion and valuation zone appear, by design, in the labeled “Claude’s Take” block at the top. Management commentary throughout is treated as hypothesis and validated against filings, financials, and external evidence.

APPENDIX A — Standard Diligence Questionnaire

Roivant Sciences Ltd. (NASDAQ: ROIV) — as-of 2026-07-04

Supplemental to the memo. Fact / Interpretation / Assumption labeled where it matters. Where a question does not map to a holdco/clinical-stage biopharma business model, the correct sector analog is given.

General

What thoughtful questions have other investors asked? The recurring buy-side questions on the Q2-FY26 call (Leerink, JPMorgan, Citi, TD Cowen, Goldman, Jefferies, Piper, Wells) were: (1) brepocitinib DM launch cadence and analogs (management deliberately sandbagged — “no great analogs… guide cautiously”); (2) whether argenx entering Graves’ commoditizes IMVT-1402 (Gline: “imitation is the finest form of flattery… a rising tide lifts all boats”); (3) the Genevant/Moderna exposure (“less than half of less than half of less than half”); (4) whether the Investor Day would unveil new strategy/data; (5) BD/“next Telavant” plans (“we’ve got some exciting ideas,” unspecified). Interpretation: the sophisticated debate is not whether brepo works (VALOR settled that) but how big it becomes and whether the FcRn and deal-making legs justify the residual.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Not applicable in the usual sense — there are no operating earnings; FY26 was a −$1.29B operating loss. GAAP net income is dominated by one-time monetizations (FY24 +$4.35B Telavant; FY26 +$770M Genevant). The correct analog: the value-creation cycle is arguably near a high — the stock is at an all-time high, the IMVT mark near a peak, and the balance sheet fully valued after two big exits. Driven by external environment or internal actions? Internal (clinical read-outs, deal execution) far more than macro; the factor model shows ~63% idiosyncratic variance and a low market beta (~0.86). How stable are revenues? Product revenue is negligible and irrelevant (~$8M). “Revenue” for this business is episodic deal proceeds and litigation settlements — inherently lumpy and unpredictable. Outlook for products/services / how big is the market? Brepocitinib TAM: DM orphan (~38,500 U.S. prevalent) plus uveitis, sarcoidosis, LPP — a multi-indication franchise if the pipeline-in-a-product thesis holds. IMVT-1402 TAM: the broad FcRn autoimmune market (Graves’ ~330,000 refractory U.S. patients, MG, CIDP, Sjögren’s, RA) — large and growing (~$2.1–2.4B class today → ~$3.5–3.9B early-2030s on approved indications, with meaningful upside from new indications).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? The FcRn field is getting more competitive (argenx, UCB, J&J all in-market; argenx entering Graves’). Orphan dermatomyositis is essentially uncontested (no approved therapies). How profitable is the business (ROIC/ROE)? Not meaningful — pre-revenue; ROIC/ROE are negative artifacts. The correct analog is deal IRR (Telavant ~$5.1B return on a modest upfront in ~10 months) plus stake appreciation (IMVT mark $5.06B→$7.91B) net of ~$750M/yr burn. How profitable is the industry / barriers to entry? Winning autoimmune drugs are highly profitable (patent-protected specialty biologics). The holdco layer has no barriers to entry — incubation is a crowded, easily-replicated model. Can the business be easily understood? No — the holdco/minority-interest structure (consolidated Immunovant, 72/97/83% private Vants, Level-3 Datavant) makes look-through economics genuinely hard. This complexity is itself a risk (and, for the patient, a source of mispricing). Undermined by foreign low-cost labor? Do brands matter? No/labor-irrelevant. “Brands” matter only as clinical reputation and KOL standing (first-mover in DM confers this). Nature of competition / switching costs? In autoimmune drugs, competition is on efficacy/safety/route (oral vs infused) and payer access; physician/patient switching costs are moderate once a therapy is established (first-mover advantage is real — cf. VYVGART in MG).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? Yes — the Genevant/Moderna $1.3B contingent payment and the Pfizer/BioNTech litigation are off-balance-sheet optionality; the Datavant stake is a Level-3 mark ($233.2M) that may understate realizable value; brepocitinib’s approval-stage value is not on the balance sheet at all. Off-balance-sheet liabilities? In-license royalty/milestone obligations: Priovant→Pfizer (sub-teens royalty + milestone on brepo), Immunovant→HanAll (up to $420M + mid-single-to-mid-teens royalty), Pulmovant→Bayer (up to $280M + high-single royalty). These burden net economics on any commercial success. How conservative is the accounting? Mixed. The fair-value option on Datavant/Arbutus introduces mark-to-market earnings volatility (unrealized gains flow through net loss). The Genevant gain is booked as income ahead of cash collection. Interpretation: aggressive in timing recognition, but transparent about it. How CapEx-hungry? Not at all — capex was $8.2M in FY26. This is an IP/clinical business; “capex” is R&D ($681.8M).

Capital Allocation & Management

How much FCF, and how is it used? FCF is negative (−$758.6M FY26) — the business consumes cash to fund pipeline. The sources of cash are exits/settlements; uses are R&D, SG&A, buybacks, and Immunovant accumulation. Significant acquisitions recently? Roivant deployed ~$700M+ buying Immunovant shares (Jan-2025 ~$337M @ $20; Dec-2025 ~$350M @ $21) — accumulating its own listed Vant. No large third-party M&A recently; management signals continued BD appetite. Buying back shares? Yes — ~$1.5B repurchased at ~$11–13 (highly accretive; ~$3B mark-to-market gain), though net share count is now rising as ~$346M SBC + option exercises outrun the buyback. Issuing large amounts of new shares to insiders? SBC is heavy ($346M/yr, ~46% of burn) — the principal value leak. But grants are structured as high-hurdle PSUs. Compensation policy? Genuinely aligned: 5-year, 30-day-VWAP price-hurdle PSUs ($15–$30 tranches, all now cleared at $35), 2-year post-vest holds, “all-in” awards. Directors paid modestly, mostly equity. Venker holds a cross-Vant CVAR settling into IMVT stock. Motivations of management? Aligned to share price via PSUs — but note zero open-market management buying at $35 and steady 10b5-1 selling (Venker 200k/mo; Torti 3M/$82M Feb-2026; Sukhatme; Gline). A mild cautionary tell.

Valuation & Market Data

ADR/MLP/K-1? No — Roivant is a Bermuda-incorporated company filing as a U.S. domestic registrant; common shares, NASDAQ: ROIV. No K-1. Dividend policy? None — no dividend; capital returned via buybacks. How profitable / net income vs. cash from operations? Not profitable; net income is dominated by one-time gains and mark-to-market swings, and diverges wildly from operating cash flow. Use the ~$750M cash burn (adjusted for the non-cash Genevant gain) as the operating truth, not GAAP net income.

Risks & Downside

What would cause the stock to decline? A broad JAK box or slow DM launch; a 2027 IMVT-1402 Graves’/MG miss; failure to sign another deal; adverse /Pfizer rulings; a decline in the Immunovant share price (~17% of ROIV’s cap); a broad biotech drawdown (biotech-industry beta 1.32). Risk of catastrophic loss / total loss? Catastrophic drawdown to the ~$9B hard-value floor (~−50%, to ~$16–17) is the realistic bad case. Total loss is highly improbable — ~$9B of cash + liquid public stakes + the settlement receivable underpin the equity even if every private option expired worthless.

Recent News & Events

Has the business environment changed recently? Yes, materially and mostly positively: VALOR positive → brepo Priority Review (Q3-26 PDUFA); $2.25B Moderna settlement (Mar-2026, $950M this month); offset by batoclimab TED Phase 3 failure and discontinuation (Apr-2026) and continued argenx encroachment on Graves’. Significant acquisitions / accounting changes / new markets? ~$700M Immunovant accumulation; fair-value-option treatment of Datavant/Arbutus; brepocitinib moving from clinical to commercial (first launch ~end-Sep-2026); the December-2025 Investor Day reframed the story from serial-dealmaker to multi-asset commercial-stage company.

APPENDIX B — Source Appendix

Roivant Sciences Ltd. (NASDAQ: ROIV) — as-of 2026-07-04

Primary sources prioritized. Facts reconcile to filings; third-party aggregators (ROIC.ai, AZI, FactorsToday) used for cross-checks and computed ratios, reconciled to filings. All accessed 2026-07-04 unless noted.

Primary — SEC filings (EDGAR, CIK 0001635088)

• Form 10-K, FY ended 2026-03-31 (filed 2026-05-20; roiv-20260331.htm) — business/Vant model, ownership table (Priovant 72%/66%FD, Immunovant 56%/52%FD, Pulmovant 97%, Genevant 83%), pipeline & PDUFA table, income statement, balance sheet, cash flow, equity-method investments (Datavant $233.2M/9%, Arbutus $174.8M/20%), Moderna settlement terms ($950M by 2026-07-08 + $1.3B contingent), buyback disclosure ($890.3M remaining; Q4 repurchases incl. 3,535,907 sh @ $28.50 in March-2026), in-license royalty/milestone obligations (Pfizer/HanAll/Bayer), CNPV/RDEP discussion.
• Prior Form 10-Ks FY2022–FY2025 — multi-year burn, Telavant/Dermavant/Myovant history.
• Form 10-Q for quarter ended 2025-09-30 and prior — interim balance-sheet and stake detail.
• Form 4 corpus (408 filings, 2021-08 → 2026-06) — insider transactions: Roivant’s code-P Immunovant purchases (16,845,010 @ $20.00 Jan-13-2025; 16,666,666 @ $21.00 Dec-12-2025); insider/VC sales (~175.6M sh/~$2.21B; QVT, Ramaswamy, SoftBank, Viking, Sumitomo); management 10b5-1 sales (Venker 200k/mo; Torti 3.0M/$82M Feb-2026; Sukhatme; Gline); Gline’s sole open-market buy (3,315 @ $15.07 Sep-2025).
• DEF 14A (latest, filed 2025-07-29) — Senior Executive PSU program (30-day-VWAP hurdles $15/$17.50/$20/$22.50/$25/$30; 5-yr performance + 2-yr hold), director comp, Venker Immunovant CVAR ($14.46 hurdle).
• 8-K material events — Telavant/Roche closing (Oct-2023), Dermavant/Organon (Oct-2024), buyback authorizations (Apr-2-2024 $1.5B; Jun-25-2025 $1.0B), Moderna settlement (Mar-2026), VALOR and batoclimab TED data.

Primary — Company disclosures / press releases (investor.roivant.com, immunovant.com, arbutusbio.com)

• Roivant/Priovant VALOR Phase 3 positive topline (2025-09-17) — DM week-52 TIS 46.5 vs 31.2 placebo, p=0.0006, n=241; all key secondaries positive.
• NEJM publication of VALOR results (2025).
• Priovant FDA acceptance & Priority Review for brepocitinib DM NDA; PDUFA Q3-CY2026.
• Roivant Q4/FY26 results & business update (2026-05-20) — cash ~$4.3B, Q4 continuing-ops income +$355.7M, FY26 net loss ~$299.8M.
• Genevant/Arbutus–Moderna $2.25B settlement announcement (2026-03-03).
• Immunovant batoclimab Phase 3 TED results / discontinuation (2026-04-02).
• Roivant Q2-FY26 earnings call transcript (held 2025-11-10, quarter ended 2025-09-30) — capital-allocation, brepo launch prep, IMVT-1402, Genevant litigation, “next Telavant”/BD commentary.

Industry & competitive

• argenx FY2025 results (argenx.com, 2026) — VYVGART/efgartigimod ~$4.2B sales, +90% YoY, ~19,000 patients (FcRn benchmark).
• FcRn class overview & nipocalimab (drugdiscoverytrends.com; grandviewresearch) — J&J IMAAVY (Apr-2025), UCB RYSTIGGO (2023).
• BMS Sotyktu (deucravacitinib) clean-label TYK2 selectivity (biospace.com; bms.com) — the JAK-boxed-warning contrast for brepocitinib.
• FDA JAK-inhibitor class boxed-warning Drug Safety Communication (fda.gov, 2021) — MACE/VTE/malignancy/mortality class action.
• Brepocitinib peak-sales estimates — Evaluate (~$2.3B by 2032, DM-centric) and Goldman (>$10B unadjusted multi-indication), via ainvest.com (Sep-2025); DelveInsight DM market insights.
• Dermatomyositis prevalence & unmet need — DelveInsight; management commentary (Q2-FY26 call).

Quantitative cross-checks (public market data; reconciled to filings)

• Aggregated financial data — income statement, balance sheet, cash flow, enterprise value, valuation multiples (ROIV and IMVT), earnings-call transcripts. IMVT market cap $7.91B (2026-07-02); ROIV FY26 figures.
• Adjusted price history — 5-year daily series (2020-12-08 → 2026-07-02); valuation_index own-history percentiles (P/B 6.3× = 83rd pct; composite 91.5th). News feed thin (4 articles).
• Factor/risk model — leaderboard (y1 +220%, Sharpe 5.23, max-DD −12.8% / 5y −79.2%), stock-loadings (biotech-SPDR beta 1.32, market ~0.86–0.95, R² ~37%, negative momentum loading), stock-info (rs_12m, rs_peak −0.7%), related-stocks (XBI, LABU, DNLI, IMTX, IDYA).

Note on authority: for U.S.-filer facts, SEC filings (10-K/10-Q/8-K/DEF 14A/Form 4 on EDGAR) are primary; third-party market-data providers were used only to cross-check computed ratios and prices and do not replace the filing. Where a data provider and a filing disagree on a material number, the filing governs. No analyst target is used as a price target anywhere in this article.