rxrx
Recursion is a $1.85B AI drug discovery platform with $754M cash and runway into early 2028 after a disciplined 2025 reprioritization. But the cleanup was deeper than the market initially absorbed: REC-994 (CCM) was discontinued May 2025 after long-term extension data evaporated the Sept 2024 efficacy signal; REC-2282 (NF2) failed Phase 2 futility; REC-3964 (C.diff) was deprioritized for out-licensing. The wholly-owned clinical pipeline is now FIVE programs (REC-4881 FAP Ph2, plus REC-617, REC-1245, REC-3565, REC-4539 in Phase 1/early-Phase 2) — narrower than the platform multiple implies. REC-4881 is the single remaining clinical validation point in 2026.
We initiate Neutral at 50/100 conviction; fair value approximately $3.25. REC-994's long-term failure is additional evidence — not against — the variant perception that AI discovery has not yet differentiated Phase 2 hit rates from base rates. REC-4881 (43% polyp burden reduction at wk 13) remains the single positive clinical PoC. The 35% YoY opex discipline + runway to early 2028 cap dilution-spiral tail risk; the upside path now requires REC-4881 FDA registration progress + clean Phase 1 reads on REC-1245/REC-617.
report snapshot
Recursion is a $1.85B AI drug discovery platform with $754M cash, runway into early 2028 after the May 2025 reprioritization. The cleanup was deeper than initially absorbed: REC-994 (CCM) discontinued after LTE efficacy failure, REC-2282 (NF2) halted at Phase 2 futility, REC-3964 (C.diff) deprioritized for out-licensing. The surviving wholly-owned clinical pipeline is REC-4881 (FAP Phase 2 — the lone positive PoC) plus four Phase 1 oncology programs. Neutral at $3.25 PT, 50/100 conviction; fair near current.
Recursion is a $1.85B AI drug discovery platform with $754M cash, runway into early 2028 after the May 2025 reprioritization...
$3.51 · $1.85B · as of Apr 26, 2026.
Neutral RXRX at $3.51. 12-month target $3.25; intrinsic $3.25. Conviction 50/100...
Numbers can look similar while narrative labels diverge — focus on which spreadsheet row the market is pricing.
variant perception & thesis
Our thesis on RXRX has shifted with 2024-25 readouts. The market still prices RXRX as a binary-readout AI drug discovery platform with ~$1.1B in pipeline value beyond net cash. Two facts compound to flatten the asymmetry: 2024-25 Phase 2 readouts have largely already happened (REC-994 met safety primary, REC-4881 FAP showed positive 25-week data, REC-2282 NF2 discontinued), and the catalyst path through 2026 is dominated by Phase 1 readouts (REC-1245, REC-617) plus FDA registration engagement on REC-4881.
1. Phase 2 Readout Risk
5/102. Cash Runway & Dilution
4/103. Platform Optionality (BD/Partnerships)
7/104. AI Discovery Differentiation
6/10Read the pillar scores as conviction on each leg of the variant view; low scores are where consensus could be right.
financial analysis
Recursion's financials read like a Phase 2 biotech, not a platform: $74.7M of revenue against $475M of R&D and a $648M operating loss in FY25. There is no merchandise margin, no unit economics, no operating leverage to model. The financial story is mechanical — cash, burn, dilution — and the load-bearing fact is that 2025 share count grew 30% (406M → 528M) to fund the $371.8M of operating cash burn while Exscientia integration stepped FY25 loss above FY24's $479M.
Income Statement Summary (FY2022-FY2025)
read first| FY2025 | FY2024 | FY2023 | FY2022 | |
|---|---|---|---|---|
Revenue (collaboration) | $74.7M | $58.8M | $44.6M | $39.8M |
R&D Expense | $475.3M | $314.4M | $241.2M | $155.7M |
Total Operating Expense | $722.8M | $537.8M | $394.6M | $285.6M |
Operating Loss | $(648.1M) | $(479.0M) | $(350.1M) | $(245.7M) |
Net Loss | $(644.8M) | $(463.7M) | $(328.1M) | $(239.5M) |
Diluted EPS | $(1.44) | $(1.69) | $(1.58) | $(1.36) |
Cash Flow Summary (FY2023-FY2025)
read first| FY2025 | FY2024 | FY2023 | |
|---|---|---|---|
Operating Cash Flow | $(371.8M) | $(359.2M) | $(287.8M) |
Capex | $(6.5M) | $(13.7M) | $(12.0M) |
Free Cash Flow | $(378.3M) | $(372.9M) | $(299.7M) |
Financing (equity issuance, net) | ~$430M | ~$555M (Exscientia stock + ATM) | ~$330M (incl. NVIDIA $50M) |
Capex Intensity | 8.7% of revenue | 23.3% of revenue | 26.9% of revenue |
These numbers ground the thesis in reported economics; the debate is durability and cycle, not obvious accounting gaps.
valuation
Standard P/E and EV/EBITDA multiples do not apply — RXRX is loss-making across all reported periods (FY25 EPS $(1.44) diluted, operating loss $(648.1M)). The relevant valuation lens is sum-of-the-parts: $743M of cash plus $1.11B of pipeline-and-platform value implied by the current $1.85B market cap. The question is whether the platform deserves $1.11B in EV without a single Phase 2 efficacy hit yet on the board. Our base case answer: roughly fair. Bull case answer: significantly underpriced. Bear case answer: trades to a discount to net cash.
| Assumption | Base Value | Note |
|---|---|---|
WACC | 12-13% | High discount rate reflects clinical and dilution risk; small-cap biotech equity risk premium |
Revenue CAGR (5yr) | ~30% | BD-driven; assumes Sanofi / Roche / Bayer milestone payments materialize on schedule |
Terminal Operating Margin | -50% | Still loss-making; pipeline maturity required before margin inflection |
Capex / Revenue | 2-3% | Asset-light platform model; FY25 capex only $6.5M against $74.7M revenue |
Effective Tax Rate | 0% | Significant NOL carry-forwards from sustained losses; no near-term tax burden |
Phase 2 Probability Weighting | 25-30% per asset | Rare-disease base rate; not credit given to AI platform until proven clinically |
$7.00
REC-994 (CCM) Phase 2 hits primary endpoint cleanly plus one new BD deal with greater than $500M headline value...
$3.25
Mixed Phase 2 readouts...
$1.25
Both REC-994 and REC-4881 miss primary endpoints...
what breaks the thesis
RXRX trades within ±15% of fair value because the asymmetry of binary Phase 2 outcomes is offset by the certainty of continued dilution. Two risks are load-bearing: (1) Phase 2 efficacy miss on REC-994 or REC-4881 removes the platform's first clinical validation; (2) forced raise at depressed prices compounds the cap-structure problem. Sector de-rating, partner loss, burn acceleration, FDA delay are secondary.
This is not generic macro risk language — it is a short list of observable thresholds that would force us to change the view.
Risk Matrix
read first| Risk | Probability | Impact | Mitigant |
|---|---|---|---|
Phase 2 efficacy miss (REC-994 / REC-4881) | 40-50% | -40% to -60% stock | Multiple shots on goal; 4 Phase 2 programs running parallel |
Dilution death spiral (raise at <$2.50) | 30-40% | Permanent cap impairment | BD milestone cash; selective ATM timing post-readout |
AI drug discovery thematic re-rating | 30% | -25% multiple compression | Diversified partner cash flow; not pure-play sentiment trade |
Loss of Tier-1 partner (Sanofi/Roche/Bayer) | 15-20% | -15% to -30% | Long-dated contracts; multi-program portfolios; switching costs |
Cash burn acceleration above $400M/yr | 25% | Pulls forward dilution event | Active opex management; Exscientia integration synergies |
FDA hold / regulatory delay | 10-15% per program | -20% on affected program | Multiple Phase 2 programs across non-overlapping indications |
Watch for drawdowns driven by fundamentals where funds de-risk faster than the business narrative updates.
fundamentals & operations
Recursion is not a unit-economics business. No products, no customers in the conventional sense, no inventory turns. The operating units are pipeline programs and partnership slots. Platform inputs — BioHive2, ~50PB of phenomic imaging data, billions of perturbed-cell measurements — produce outputs measured in BD deal value and clinical milestones, not revenue per FTE. Revenue is 100% collaboration; operating leverage exists only conceptually until a Phase 3 product clears.
Why "Unit Economics" Doesn't Apply Here
FRAMEWORKThe instinct to compute revenue-per-employee or gross margin breaks down here. The right way to read RXRX as an operating system: Inputs: Capital, compute, data, scientific talent, IP licenses. Throughput: Validated targets per year, leads into IND, INDs into Phase 1, Phase 1s into Phase 2...
FY25 Revenue Decomposition ($74.7M)
read first| Revenue Component | Estimated $M | % of Total | Recognition Pattern |
|---|---|---|---|
Sanofi collaboration (BD program) | ~$30-35M | ~45% | Amortized upfronts + milestones |
Roche / Genentech collaboration | ~$15-20M | ~25% | Amortized upfronts + research fees |
Bayer collaboration (fibrotic) | ~$5-8M | ~10% | Research fees |
Other partners + Exscientia legacy | ~$10-15M | ~15-20% | Mixed milestones / fees |
Product Sales | $0 | 0% | No approved products |
Total Revenue FY25 | $74.7M | 100% | All collaboration; lumpy quarterly |
Growth Drivers: Where the Next $50M of Revenue Comes From
DRIVERS1. Sanofi expansion. 2022 deal structured at up to $1.5B across multiple programs...
Platform Infrastructure: Real Assets, Unproven Returns
PLATFORMThe platform investments are tangible and substantial. Whether they translate into a clinical-readout edge is the open question. BioHive2: NVIDIA-partnered top-50 supercomputer for AI workloads...
competitive position
RXRX competes across public AI-native peers (SDGR, BAI, TEM), private specialists (Insitro, Atomwise) and Big Pharma in-house AI groups. The Nov 2024 Exscientia merger consolidated two of the most-cited names but did not answer the core question: does the platform change P2/P3 outcomes, or just compress P1 entry?
AI Drug Discovery Landscape
read first| Company | Approach | Cap Stack | Pipeline Stage | Differentiator |
|---|---|---|---|---|
Recursion (RXRX) | Phenomics + generative chem | ~$1.85B cap; $743M cash; ~no debt | 5 P2 wholly-owned + Exscientia P1 | 50PB+ phenomics, BioHive2, Tier-1 BD book |
Schrodinger (SDGR) | Physics-based + ML | ~$2-3B cap; software cash offsets burn | Multiple P1/2 + licensing book | Standalone software funds discovery |
Insitro (private) | ML + functional genomics + iPSC | Private; >$600M raised | Pre-clinical to early clinical | Deep wet-lab/ML; BMS, Gilead deals |
BenevolentAI (BAI/AMS) | Knowledge-graph target ID | Sub-$200M cap; restructured 23-24 | Pipeline thinned post-restructure | Cautionary AI-narrative template |
Atomwise (private) | CNN small-molecule docking | Private; partnership-funded | Pre-clinical / partner-program | AIMS platform; broad pharma deals |
Exscientia (in RXRX) | Centaur generative chem | Merged into RXRX Nov 2024 | GTAEXS617, MALT1, A2A, others | Folded in; expands clinical bench |
Moat: What RXRX Owns vs Yet to Prove
MOAT ASSESSMENTAdvantages: 50PB+ phenomics dataset - hardest asset to replicate. Recursion OS - phenomics, target ID, lead opt, translational biology in one stack. BioHive2 with NVIDIA ($50M strategic equity, Sept 2023)...
Where RXRX Sits vs Peers
POSITIONINGvs SDGR: Schrodinger's software funds discovery; RXRX is pure-bet biotech with a more concentrated P2 portfolio. vs Insitro/Atomwise: Private peers iterate without quarterly dilution accountability; RXRX pays the public-cap-stack tax in down markets. vs BenevolentAI: BAI's 2023-24 restructuring is the cautionary template; RXRX has more cash and pipeline but the same risk vector if P2 disappoints...
market size & tam
Post May-2025 reprioritization, RXRX's wholly-owned clinical TAM concentrates in REC-4881 (FAP — rare disease, ~50K US prevalence, no approved therapies, $1-2B peak rNPV potential at high PoS) plus oncology Phase 1 optionality. The original 'many shots on goal' framing has narrowed: CCM (~360K US) and NF2 (~33K) are no longer addressable through wholly-owned programs.
TAM by Indication (Wholly-Owned P2)
read first| Indication | Patient Pop (US) | Pricing Frame | Peak Sales Est. | Probability Lens |
|---|---|---|---|---|
FAP (REC-4881) | ~50K | $200K-500K/yr; peak $500M-1.5B | 25-40% (Ph2 positive surrogate) | ACTIVE — H1 2026 FDA Type B |
Solid tumors / lymphoma (REC-1245 RBM39) | Multiple biomarker-enriched indications | Variable; oncology peaks $300M-2B per indication | 10-15% (Ph1) | ACTIVE — Ph1 readout H1 2026 |
Solid tumor combinations (REC-617 CDK7) | Multiple | Combination context; peak depends on label | 10-15% (Ph1) | ACTIVE — combo data H1 2026 |
B-cell malignancies (REC-3565) | Multiple lymphoma/leukemia subsets | $500M-2B per indication | 10% (Ph1) | ACTIVE — early Ph1 |
Solid + heme onc (REC-4539) | Multiple | Variable | 10% (Ph1) | ACTIVE — early Ph1 |
CCM (REC-994) | — | — | — | DISCONTINUED May 2025 |
TAM Reframed: Many Shots, No Megadeal
ADDRESSABLE FRAMENo wholly-owned asset is a $5B+ blockbuster. TAM framing must be portfolio-level: Aggregate wholly-owned peak: Bull-stack risk-unadjusted ~$1.5-2.5B across REC-994/2282/4881/3964 plus early-stage. ~10-30% approval probability shrinks risk-adjusted materially...
BD Market - Other Revenue Stream
PARTNERSHIPSSanofi: Up to $1.5B across oncology/immunology (2022; expanded with Exscientia). Roche/Genentech: Up to ~$300M+/program (neuro, onc). Bayer: ~$50M up-front + milestones (fibrotic)...
product & technology
RXRX's clinical pipeline post-May 2025 reprioritization is materially narrower than the platform-multiple thesis implies. Five wholly-owned clinical programs remain (REC-4881 Ph2 FAP, REC-617 Ph1/2 CDK7, REC-1245 Ph1 RBM39, REC-3565 Ph1 B-cell, REC-4539 Ph1 oncology) plus REC-7735 and REC-102 in preclinical. Three Phase 2 programs (REC-994, REC-2282, REC-3964) were halted or deprioritized. Partner programs (Sanofi 15 programs, Roche maps) remain in discovery / preclinical.
1 placeholder-heavy block(s) remained in the source pane; inspect against the original json before publishing.
The Recursion OS Stack
PLATFORMRecursion OS: 50PB+ in-house phenomics dataset (high-content cellular imaging at scale), ML-driven target identification + lead optimization. BioHive2 supercomputer: NVIDIA H100 cluster (deployed 2024) — one of the largest dedicated ML training clusters in pharma. NVIDIA contributed $50M equity (Sept 2023) plus collaboration on biology-focused foundation models...
Does the Platform Have Standalone Value?
PLATFORM ECONOMICS1. Pipeline-only: Value = risk-adjusted NPV of molecules; platform is a cost center. 2...
supply chain
Pre-commercial: no real commercial supply chain; clinical supply runs through standard CDMOs and CROs. Real dependencies are compute (NVIDIA GPU access, BioHive2 capacity), phenomics wet-lab consumables, and CRO/site capacity for the P2 portfolio. Compute and CRO capacity are the two to watch.
Key Operating Dependencies
DEPENDENCIES1. NVIDIA GPU compute: BioHive2 on NVIDIA H100-class hardware co-developed with the $50M Sept 2023 strategic equity. Priority access in tight cycles is a real dependency...
Supply / Dependency Risks
RISKCompute supply risk: NVIDIA hyperscaler priority in a tight GPU cycle slips BioHive2 and training timelines. Public-cloud GPU at this scale is not a short-notice substitute. CRO concentration risk: Rare-disease trials cluster among a small pool of qualified sites; Tier-1 CRO disruption can push a P2 readout - functionally a near-term miss for the share price...
catalyst map
RXRX's 2026 catalyst path crystallized after the May 2025 reprioritization removed three Phase 2 programs (REC-994 CCM long-term failure, REC-2282 NF2 futility, REC-3964 C.diff strategic deprioritization). Remaining wholly-owned 2026 catalysts: REC-4881 FDA registration engagement; Phase 1 dose-escalation readouts on REC-1245 (RBM39) and REC-617 (CDK7); plus partner milestones (Sanofi/Roche).
| Date | Catalyst | Impact | Probability | Price Sensitivity |
|---|---|---|---|---|
Sept 2024 (already) | REC-994 Phase 2 SYCAMORE topline — met safety primary, modest 12-month efficacy trend | Realized — initially positive | 100% | Initially +; reversed by LTE |
May 2025 (already) | REC-994 DISCONTINUED — long-term extension showed efficacy loss | Realized — negative | 100% | Negative — clinical failure |
May 2025 (already) (2) | REC-2282 (NF2) Phase 2 halted on futility | Realized — negative | 100% | Negative — Phase 2 fail |
May 2025 (already) (3) | REC-3964 (C.diff) deprioritized — out-license path | Realized — strategic | 100% | Neutral / mildly negative |
Dec 2025 (already) | REC-4881 Phase 1b/2 TUPELO 25-wk data — 43% polyp reduction | Realized — positive | 100% | Positive (small-cohort surrogate) |
H1 2026 | REC-4881 FDA Type B meeting on FAP registration pathway | Very High | 85% | +/- 10-20% |
Catalyst Color & Path Dependency
PATHWhat May 2025 showed: Three Phase 2 programs cleared from the wholly-owned pipeline in a single reprioritization round — REC-994 (LTE efficacy failure), REC-2282 (Phase 2 futility), REC-3964 (strategic deprioritization). REC-994 is the cleanest data point yet against the AI-platform-changes-Phase-2-hit-rates thesis. What Dec 2025 showed: REC-4881 TUPELO Phase 1b/2 25-week data — 43% median polyp burden reduction at wk 13, 53% at wk 25 (12 weeks off therapy)...
street expectations
Sell-side coverage is mixed. Bulls anchor on 2026 P2 catalysts, the Sanofi/Roche/Bayer BD book and Exscientia synergies; bears anchor on cash burn, 30% 2025 dilution and zero AI-discovered P3 success. Per public commentary, consensus 12-month PT sits in the $5-8 zone vs the $3.51 print - dispersion reflects binary readouts, not conviction.
Consensus Frame (Indicative)
read first| Metric | FY26E | FY27E | FY28E | |
|---|---|---|---|---|
Revenue (consensus) | ~$85M | $120-160M | $180-250M | Consensus tracking ~$85M FY26 (per Insider Monkey 2026 priorities update) |
Net Loss | $(550)-(650)M | $(450)-(550)M | $(400)-(500)M | Trend: declining as opex discipline holds |
EPS Diluted | $(1.10)-(1.30) | $(0.85)-(1.05) | $(0.70)-(0.85) | Improving on op leverage + buyback-free dilution |
Cash Year-End | ~$420M | ~$200M | <$100M (raise required) | Implies further ATM raise H2 2027 or deal milestone |
Analyst PT (LSEG) | $3-$11 range | Avg ~$7 | — | 8 analysts; HOLD consensus rating |
Bull / Bear Sell-Side View
STREET VIEWPer LSEG / public commentary (April 2026): 8 analysts cover RXRX with a HOLD consensus rating. PT range $3-$11 with average ~$7. Our Neutral $3.25 PT sits below the consensus average, reflecting our skepticism that earlier-stage 2026 catalysts (Phase 1 dose-escalation, FDA registration discussion) will drive a step-function re-rate to the consensus number...
earnings scorecard
No GAAP earnings story — loss-making by design while building platform and pipeline. The quarterly print is read for three things: revenue lumpiness from milestone-driven BD contracts, R&D trajectory, and the rate at which cash and share count diverge. FY2025 is the first post-Exscientia year, which is why expense and dilution both step-change versus FY2024.
Quarterly Print — FY2025 (and FY-end anchors)
read first| Quarter | Revenue ($M) | R&D ($M) | Net Loss ($M) | Cash ($M) | Shares Out (M) |
|---|---|---|---|---|---|
Q1 2025 | $14.7 | $129.6 | ($202.5) | n/d | n/d |
Q2 2025 | $19.2 | $128.6 | ($171.9) | n/d | n/d |
Q3 2025 | $5.2 | $121.1 | ($162.3) | n/d | n/d |
Q4 2025 (implied) | $35.5 | $95.9 | ($108.1) | $743.3 | 528.2 |
FY2025 total | $74.7 | $475.3 | ($644.8) | $743.3 | 528.2 |
FY2024 total (memo) | $58.8 | $314.4 | ($463.7) | n/d | ~406 |
What the FY2025 Tape Actually Says
PATTERNRevenue is milestone-driven. Q3 printed $5.2M, Q4 implied $35M — timing of partner milestones (Sanofi, Roche, Bayer), not demand. FY revenue +27% to $74.7M, but a smooth quarterly curve does not exist...
Earnings Quality — What Counts as a 'Beat'
QUALITYWithout earnings, the scoring rubric is four lines: Burn vs guide: the headline. Inside guide is a beat regardless of reported loss; a miss reprices dilution risk. Pipeline progression: enrollment milestones, IND filings, Phase 2 timing carry more multiple impact than any P&L line...
alternative data
Textbook small-cap biotech signal mix: rich Form 4 cadence, structurally large strategic shareholder base (Roivant via Exscientia, NVIDIA stake), meaningful retail float, and elevated short interest that makes the name a two-way squeeze candidate around binary catalysts. None of these signals is dispositive alone; together they describe a stock that punishes static positioning and rewards explicit views on the next readout window.
Insider Activity — Read the Form, Not the Headline
INSIDERSForm 4 filings have run regularly through 2025–2026, most recent landing 2026-04-24. Cadence reflects standard biotech executive comp, not a directional shift: 10b5-1 plan sales are scheduled and disclosed in advance; reading them as bearish is almost always wrong because they're decoupled from current information. RSU/PSU vesting drives mechanical sell-to-cover that shows up as Form 4 sales but isn't discretionary...
Institutional Holding Map
OWNERSHIPInstitutional ownership is meaningful but concentrated in strategic, not financial, hands: Roivant Sciences rolled in via the Nov 2024 Exscientia merger (largest Exscientia holder pre-deal) — strategic, sticky, creates a float floor. NVIDIA retains the Sept 2023 $50M strategic stake — small economic, anchors the AI-compute narrative. Biotech specialists and long-only funds hold conventional float...
Short Interest — Squeeze Risk Cuts Both Ways
SHORT INTERESTShort interest typically 10–12% of float — high enough to matter, not a focused activist campaign. Mix of fundamental shorts (cash burn / dilution) and event-driven shorts pricing a Phase 2 miss. Clean hit: high SI + heavy retail + small effective float creates squeeze conditions...
Average daily volume runs 10–15M shares (~$35–$50M notional at $3.51) — liquid enough for a 0.5–1% sleeve to enter/exit cleanly even with short-side flow. Block trades above $5M notional should be worked, not swept; binary readout days see ADV multiples of 5–10x and far wider intraday spreads.
historical analogies
The AI-drug-discovery cohort that listed in 2020–2022 (RXRX, SDGR, BAI, ABCL, EXAI pre-merger, smaller names) now has five years of post-IPO history. The pattern is clearer: most have evolved into Schrödinger-like grinders — platform-plus-one-or-two-programs mid-caps that compound modestly without delivering blockbuster Phase 3 outcomes — rather than the clinical compounders the early narrative implied.
AI Drug Discovery Cohort — Comparable Trajectories
read first| Comparable | Outcome | Lesson |
|---|---|---|
Schrödinger (SDGR) | Platform-plus-program grinder; valuation de-rated from peak | AI-drug-discovery names settle into platform-software multiples once clinical hits are slow |
BenevolentAI (BAI) | Severe de-rate post Phase 2 misses; restructuring cycle | Cautionary AI tale — narrative without clinical proof collapses fast |
AbCellera (ABCL) | Cash-rich but multiple compressed once antibody-engine TAM debate set in | Even with strong tech and cash, the platform premium fades absent franchise drug |
Moderna pre-COVID | Platform debate; transformed only by a single mega-catalyst | Platform value can suddenly re-rate, but only when the technology delivers a franchise asset |
Genmab | Founder-led biotech compounder — 25-year compounding via partnered franchises | Patient model: compound BD economics rather than chase platform stories |
Alnylam | RNAi platform compounder; took 15 years from IPO to franchise approval | Platform technology timelines are decadal, not quinquennial |
The Cohort Pattern — Where RXRX Most Likely Lands
PATTERNBase-rate: none of the 2020–2022 AI-drug-discovery cohort has produced an end-to-end Phase 3 success. Market has re-rated downward — most names trade closer to platform-software multiples (3–7x sales) than biotech-platform multiples (10–20x peak); gap to non-AI clinical-stage biotechs has compressed. Implication: the brief's variant perception that AI drug discovery has not yet changed Phase 2-to-approval base rates is cohort-supported...
Reading the Cohort De-Rate — Three Mechanics
MARKET STRUCTUREThree reinforcing mechanisms drive the cohort de-rate, all applying to RXRX: Generalist exit: 2020–2023 long money that bid the cohort on platform-narrative has rotated out. Marginal buyer is biotech specialists underwriting specific assets, not platform stories. Comparable repricing: peers reporting mixed/negative Phase 2 mark down implied PoS for the cohort...
management & leadership
Recursion's leadership transitioned in 2025: co-founder Chris Gibson stepped back from CEO; Najat Khan (ex-Johnson & Johnson Chief Data Officer; joined Recursion Aug 2024 as Chief R&D Officer) is now CEO and President. CFO Ben Taylor (ex-Exscientia CFO; 15 years Goldman/Barclays banking) joined via the Nov 2024 merger. The combination is biotech-operator + finance discipline rather than the founder-narrative pattern that defined RXRX through 2024.
Senior Leadership & Board
read first| Name | Role | Background | Tenure | Notes |
|---|---|---|---|---|
Najat Khan, PhD | CEO & President | Ex-J&J Chief Data Officer; biotech / data science | Joined Aug 2024; CEO 2025 | Operator-led era; led 35% opex cut; clinical/data lens |
Chris Gibson, PhD | Co-founder, Director | Stanford bioengineering PhD; founded RXRX 2013 | 13+ years | Stepped back from CEO 2025; remains on board / co-founder voice |
Ben Taylor | Chief Financial Officer | Ex-Exscientia CFO (4+ yr); 15 yr Goldman / Barclays IB | Via merger Nov 2024 | Drove 35% opex cut; led 2026 capital strategy |
David Mauro, MD, PhD | Chief Medical Officer (likely) | Clinical development background | Tenure varies; check public filings | Oversees Phase 1/2 program execution |
Vivek Ramaswamy | Director (Roivant founder) | Roivant Sciences founder; biotech investor | Multi-year board | Strategic / capital-markets perspective |
Independent directors | Various | Lux Capital / Bayer / NVIDIA observer rights (per public filings) | Various | Confirm individual names via DEF 14A |
CEO Assessment — Najat Khan
Operator-ledNajat Khan (PhD chemistry; ex-J&J Chief Data Officer + ex-McKinsey + ex-PhD biology) joined RXRX as Chief R&D Officer in Aug 2024 and was elevated to CEO + President during 2025. She is the first non-founder CEO in Recursion's history. Strengths: Brought J&J operating discipline; executed the May 2025 reprioritization (halting REC-994/2282 and deprioritizing REC-3964) without halting load-bearing programs; led the 35% YoY cash-opex cut; pivoted strategy to oncology + rare disease focus...
Compensation & Dilution Linkage
ALIGNMENT / DILUTIONTypical clinical-stage biotech: modest cash base, dominant stock-based comp tied to platform, BD and clinical milestones. Cash-light P&L preserves runway against $(372M) FY25 OCF on $743M cash. Structural share-count growth: stock comp compounds on top of raise dilution...
macro sensitivity
RXRX is duration-heavy small-cap biotech: a portfolio of long-dated, binary cash flows discounted to today. That structure makes the equity disproportionately sensitive to the discount rate, small-cap biotech flows (XBI), and the capital-market liquidity that prices the next dilutive raise. Macro is not a backdrop here — it is a direct input into the option value of every pipeline asset.
Macro Factor Sensitivity Map
read first| Factor | Sensitivity | Direction |
|---|---|---|
Real rates / 10Y yield | High | Cuts +ve, hikes -ve (DCF duration) |
Small-cap biotech flows (XBI) | High | Tracks sector beta; XBI inflows lift the cohort |
FDA policy & approval cadence | Medium-High | Pro-innovation stance +ve; restrictive review -ve |
Biotech M&A premium environment | Medium | Active large-cap BD lifts platform terminal value |
Risk-on / risk-off sentiment | High | Risk-off compresses unprofitable small-cap multiples |
Equity issuance window (ATM viability) | High | Open window protects runway; closed window forces dilution at lows |
Why RXRX Is a Pure Duration Trade
DURATIONCash flow distance: primary value sits in milestone and royalty streams arriving 2028–2035 plus undefined-date platform terminal value. No near-term cash flows anchor valuation, so a 100bp rate move shifts PV far more than for a typical operating company. XBI as funding spigot: small-cap biotech is gated by ETF flows...
quantitative profile
RXRX screens as low-quality, high-vol, value-trap-flagged biotech — the profile systematic factor funds are programmed to underweight or short. The quant lens cannot represent the variant perception (binary readout optionality + platform terminal value), which is why the name trades where it does. Understanding why the screens fail is more useful than reading the scores.
Factor Score Card
read first| Factor | Score | Comment |
|---|---|---|
Quality | Low | Negative ROIC, deeply unprofitable, no FCF — fails every quality screen |
Value | Mixed | EV/cash discount partial; no P/E (loss); P/S 25x — looks rich on flow, cheap on cash |
Momentum | Poor | Multi-year downtrend from 2023 highs; below most trend-following filters |
Size | Small-Mid | $1.85B mkt cap — above microcap exclusions, below large-cap factor coverage |
Volatility | High | Realised vol elevated; readout-driven spikes amplify risk-parity drag |
Liquidity | Good | 10–15M ADV — sufficient for institutional sleeve sizing |
Why the Screen Says 'Avoid' — and Why That Isn't the End
QUANTSystematic verdict: a multi-factor screen on quality/value/momentum/profitability either excludes or shorts RXRX. Negative ROIC, multi-year downtrend, no earnings or FCF anchor — the name fails every gate systematic capital uses. RXRX is structurally absent from factor-tilt and risk-parity books...
Biotech-Specific Factors Not in the Standard Screen
BIOTECH FACTORSStandard factor frameworks miss the dimensions biotech actually trades on: Pipeline depth: 5 wholly-owned Phase 2 programs plus Exscientia legacy — above median for sub-$2B biotechs, rare-disease-concentrated (raising individual PoS, limiting TAM). BD optionality: Sanofi $1.5B + Roche ~$300M+ + Bayer $50M deals provide platform-validating run-rate; each milestone is a non-dilutive funding event factor models cannot represent. Event density: 2026 contains two binary readouts (REC-994 CCM, REC-4881 FAP) each large enough to re-rate the stock...
options & derivatives
RXRX options trade with a pre-Phase 2 biotech profile: IV typically 60–90 with event-vol bumps into the 100s as readouts approach, and a directional skew that tilts toward calls during BD speculation and toward puts as runway tightens. The regime makes outright long calls expensive into catalysts but makes structured trades (vertical spreads, calendars, put-writes at conviction strikes) materially more attractive than they'd be on a lower-vol single-stock.
Implied Vol Regime
VOLATILITYBaseline IV 60–90 , standing band for sub-$5B biotechs with a Phase 2 inside 12 months. RXRX prints mid-band — high end during timeline updates or BD chatter, low end during quiet integration periods. Event vol overlay: single-name vol expands 30–100 vol points in the 4–6 weeks before a binary readout...
Strategies That Match the Thesis
STRATEGIESFour strategies map cleanest to the Neutral PT and binary-catalyst variant perception: Long calls into catalyst: direct Bull $7.00 expression. Expensive — 3-month $5 strike runs 15–20% of underlying at peak event vol. Use only with explicit conviction on date and outcome; IV decay on slippage is severe...
Event-Driven Structured Plays
EVENT-DRIVENCalendar spreads through the readout: sell front at full event vol, buy back-month pricing event vol plus post-readout regime. Long-vega term-structure trade, not direction. Profit if front collapses post-event more than back — typical after binary biotech catalysts...
governance & accounting
Governance is clean by the structural metrics that usually flag problems — no restatements, no material weaknesses, no related-party concerns, no dual-class voting. The two signals that matter here are not accounting-fraud red flags but revealed-preference signals about capital structure: heavy stock-based comp as cash-light pay, and aggressive ATM activation into price weakness. Both are ordinary in clinical-stage biotech — but encode a willingness to dilute that the equity holder must price.
Board & Executive Leadership
read first| Name | Role | Notes |
|---|---|---|
Chris Gibson, Ph.D. | CEO, Co-Founder, Director | Co-founded 2013; PhD in Bioengineering. Continues to lead post-Exscientia merger. |
Vivek Ramaswamy | Director | Roivant founder; biotech capital allocation perspective on board. |
NVIDIA representative | Board observer (per public disclosure) | Tied to Sept 2023 $50M strategic investment + BioHive collaboration. |
Other independent directors | Independent majority | Mix of pharma, AI, capital markets backgrounds (per proxy disclosure). |
Auditor | Publicly disclosed (Big-4 firm) | Clean audit opinions; no material weakness or restatement. |
Accounting Quality: Clean GAAP, Heavy Stock-Comp
ACCOUNTINGGAAP discipline: Clean GAAP, no restatements. No aggressive R&D capitalization or unusual revenue recognition. Collaboration revenue follows ASC 606 with straightforward milestone-based recognition...
value framework
RXRX cannot be valued on P/E (no earnings) or P/S (revenue is milestone-driven). The right framework is SOTP: net cash + risk-adjusted pipeline NPV + platform optionality. It is what the institutional buyside actually uses on AI-drug-discovery names, and it makes the Neutral $3.25 PT defensible — SOTP center-of-mass lands within ~10% of the $3.51 print, offering neither margin of safety nor clear overvaluation.
Why P/E and P/S Both Fail Here
FRAMEWORKP/E undefined. R&D ($475M FY25) is 6x revenue ($75M FY25) — funding pipeline maturation aggressively. Forcing a P/E metric here is noise...
Sum-of-the-Parts Build
read first| Component | Value ($M) | Methodology |
|---|---|---|
Net cash | $743 | Cash & equivalents 12/31/25; ~$0 LT debt; lease-only obligations |
Pipeline rNPV — wholly-owned | $300–$700 | 5 Phase 2 assets at 10–25% PoS × $200–$800M peak NPV each, risk-adjusted to today |
Platform optionality | $200–$400 | Run-rate BD deal flow (~$75M/yr recognised, multi-billion contracted) × platform multiple |
Less: 12-mo dilution overhang | ($50–$100) | Expected ATM tranches at depressed prices over the next 12 months |
SOTP total range | $1,200–$1,800 | Sum of components, net of dilution overhang |
Implied price (528.2M shares) | $2.27–$3.41 | SOTP total / current shares outstanding |
How the SOTP Anchors the Neutral $3.25 PT
VALUATIONSOTP range $1.2B–$1.8B straddles the $1.85B market cap thinly on the high side. Per-share $2.27–$3.41 vs $3.51 print and $3.25 PT. Three reads: Market pays a small premium to SOTP center (~10–20% above midpoint) — platform multiple awarded on faith, not proof...
Cross-Check — Peer Multiples and Implied Probabilities
VALIDATIONPeer EV/cash. Sub-$5B biotechs with Phase 2 pipelines and active BD trade 1.0–2.0x net cash. RXRX prints ~2.5x ($1.85B / $743M) — high end, consistent with the platform premium, inconsistent with a cash-defensive margin of safety...
appendix & sources
How we source the tape, verify levels, and align this report with XVARY deep-dive standards.
Sources: SEC filings, company disclosures, market data vendors, and sources cited in the sections above. For investment presentation use only.
standards and pipeline: xvary.com/methodology/