Absci

Phase A — Understand the business

Lens 1 · Company Overview

Absci is a clinical-stage, AI-native biopharma that designs antibody therapeutics in silico and validates them in a "lab-in-the-loop" wet lab. The platform is branded Integrated Drug Creation (generative design model "Origin-1" + high-throughput validation). The pitch: take a target from AI design to IND in ~2 years for ~$15M/program, versus an industry average of 4–6 years and >$50M. Founded 2011 in Oregon (McClain's basement), converted to a Delaware C-corp in 2020, IPO'd on Nasdaq July 2021.

Two-engine model:

1. Partner programs — drug-creation agreements where pharma pays research fees + downstream milestones + royalties. This is the only revenue line today, and it is tiny: $2.8M FY2025, $215K Q1 2026. The company has never received a milestone or royalty payment — those are years away "if at all" (its own words).
2. Internally developed programs — the real value story. Absci now self-funds wholly-owned candidates to a value-inflection point, then partners/out-licenses. Lead = ABS-201 (anti-PRLR antibody) for androgenetic alopecia (AGA) and endometriosis.

Customers/partners: disclosed names include Merck, AstraZeneca, Almirall; web adds historical collaborations with Memorial Sloan Kettering, Twist Bioscience, Owkin, Oracle, PrecisionLife. Compute partner: AMD (Jan 2025 strategic collaboration + $20M PIPE). Concentration is extreme: two partners = 100% of partner revenue in Q1 2026, and substantially all of it is foreign. HQ Vancouver WA (78k sq ft); satellite AI lab NYC, innovation center Zug CH, R&D presence Belgrade RS.

Contract structure: upfront + research fees + project-based technical milestones + (downstream) clinical/regulatory/commercial milestones + royalties on net sales. Revenue is lumpy and milestone-timing-driven by design — management warns it "will fluctuate from period to period".

Lens 2 · Supply Chain (→ CDMO / manufacturing, per +clinical)

For a platform biotech the "supply chain" is compute → wet-lab validation → CDMO manufacture → CRO-run trials → (eventual) partner commercialization.

• Upstream compute: Cloud + accelerated compute. AMD Instinct accelerators + ROCm software is now a named, contracted input via the Jan 2025 collaboration — Absci is explicitly optimizing its de novo design models on AMD silicon. (Notable: AMD, not Nvidia, is the compute partner — a differentiator vs. the broader AI-bio field that defaults to CUDA.)
• Lab inputs: "We rely on a limited number of suppliers for laboratory equipment and materials and may not be able to find replacements… on a timely basis" — a disclosed single-/limited-source risk.
• Clinical execution: Absci depends on third-party CROs/CDMOs for preclinical studies, contract manufacturing, and trials — a named dependency in the risk factors ("we rely and expect to continue to rely on third parties to conduct our preclinical studies and clinical trials"). The HEADLINE trial runs at multiple sites in Australia (a common cost/speed choice for early-phase biotech).
• Downstream: For partnered assets, the partner owns the commercial chain — Absci has "no control over the clinical development plan, regulatory strategy or commercialization efforts". That is the structural cost of the partner model: the upside is leveraged but the steering wheel is someone else's.

Chokepoint: the wet-lab validation + CDMO capacity for its own programs. No foundry-style single point of catastrophic failure, but the company is small and externally dependent at every physical step.

Lens 3 · Competitive Advantages (moats)

The bull moat claim is a generative-AI antibody platform that compresses time and cost (2yr/$15M vs 4–6yr/$50M) and can attack "underexplored mechanisms where unmet need is high and competition is low". Concretely:

• Process/IP moat (the real one): a portfolio across proprietary cell lines, protein-expression tech, screening assays, antibody-discovery methods, and generative AI models. Built partly by acqui-hiring Totient (synthetic biology) and Denovium (deep learning).
• Target-selection moat (the strategy bet): rather than race crowded targets, pick underexplored ones (PRLR for hair loss/endometriosis) where being first matters more than being incrementally better.
• Compute-partnership signal: AMD's equity-backed collaboration is a (modest) third-party validation that the platform is worth optimizing silicon for.

Honest moat assessment: thin and unproven. The platform's only objective external validation to date is (a) a handful of small partner deals that have produced $2.8M lifetime revenue and zero downstream payments, and (b) — damningly — its own lead clinical asset ABS-101 missing on half-life (Lens 5). A design platform whose flagship in-house molecule underperforms "next-generation comparators" on a core PK property has not yet demonstrated that the AI confers durable advantage. Bargaining power is weak: Absci needs partners more than partners need Absci (extreme revenue concentration, foreign-partner dependence, downstream-only pricing model that partners may dislike — the 10-Q flags that partners "may prefer… upfront payments rather than downstream revenues"). The moat is a hypothesis that ABS-201 data will either validate or destroy.

Lens 4 · Segments

One reportable segment — the CODM (CEO) manages on a consolidated basis. No product/geographic P&L breakout exists because there is essentially no product revenue. The only meaningful "segmentation" is the expense decomposition the CODM actually reviews (Q1 2026 vs Q1 2025, $000s):

The signal: spend is rotating into the clinic (external dev +60%, drug-creation +52%) while personnel/G&A are held flat — exactly what you'd want from a company concentrating its bets on ABS-201. Geographic revenue mix: "substantially all" partner revenue is foreign.

---

Phase B — Measure performance (pipeline-swapped, per +clinical)

Lens 5 · Pipeline by phase (replaces Earnings Result)

The asset table is the company. The most important fact in this entire dossier sits here.

The print itself (for the record, since the +clinical overlay still tracks the cash burn):

• Q1 2026 revenue $215K vs $1.18M Q1'25 (−82%); missed consensus of ~$1.43M badly. Revenue is noise — milestone timing.
• Net loss $(29.6)M, EPS $(0.19) (vs −$0.21 PY; slightly beat the −$0.1943 EPS est).
• R&D $19.3M (+18% YoY), driven by +$1.6M ABS-201 external dev and +$2.1M platform. SG&A $9.1M (−4%).
• Operating loss $(30.8)M; FY2025 operating loss $(120.3)M, FY2025 net loss $(115.2)M, FY2025 EPS $(0.84).
• Cash + marketable securities $125.7M at 3/31/26 ($8.6M cash + $117.1M U.S. treasuries), down from $144.3M at 12/31/25. Accumulated deficit $(654.4)M.
• Stock reaction to the print: −3.1% after-hours — a non-event, because the only thing that matters is the H2'26 ABS-201 readout, not Q1 revenue.

The read: ABS-101's half-life miss is the quiet bombshell. Eighteen months ago the TL1A asset was a co-lead in a validated, white-hot class (every big player is chasing TL1A — Lens 7). Absci's AI-designed molecule couldn't match the comparators on a basic PK metric, and management shelved it to "partnering" (biotech for "we couldn't justify the spend"). Simultaneously they killed ABS-301/-501. The company has deliberately narrowed itself to a one-asset story: ABS-201 in alopecia. Everything now rides on H2 2026.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts in the research layer (transcripts/ empty), so this is ``. Across 2025→Q1 2026 the narrative arc is a forced pivot dressed as focus:

• Mid-2025: dual-lead messaging (ABS-101 TL1A "potential best-in-class" + ABS-201). Tone confident, "best-in-class," cash "into 2028".
• Q4 2025 / Jan 2026: the half-life disappointment lands; language shifts to "prioritization," "capital-efficient," ABS-101 "seeking a partner".
• Q1 2026 call (2026-05-07): explicitly redirects resources away from ABS-301/ABS-501, leans hard into the "prolactin-based portfolio" (ABS-201 + ABS-202), reaffirms runway into H1 2028 and the H2'26/early-2027 ABS-201 cadence.

Recurring phrases: "underexplored mechanisms," "capital-efficient," "AI-native," "first-in-class/disease-modifying," "$15M per program." Stopped saying: "best-in-class TL1A," dual-lead. The tonal tell is classic single-asset biotech — when a pipeline narrows to one shot, management amplifies "focus" and "discipline." Believe the discipline (the expense data backs it, Lens 4); discount the spin.

Lens 7 · Catalyst calendar + mechanism comps (replaces Comps)

Catalyst calendar — the only thing that sets the stock:

Mechanism comps — not P/E, by class:

• PRLR for hair loss: essentially empty competitively — this is the bull's whole point. Antibody literature names PRLR, IL-6R, CXCL12, DKK1 as AGA antibody targets, but "there are only few therapeutic antibodies for AGA" and no approved antibody. Standard of care (minoxidil/finasteride) is decades old, modestly effective, compliance-limited. If ABS-201 works, it is plausibly first-in-class in an 80M-patient US market. Guggenheim/analyst peak-sales estimate ~$2.2B for ABS-201.
• TL1A (ABS-101's dead class): brutally crowded and validated — Merck MK-7240 (acquired Prometheus for ~$10.8B), Roche/Roivant RVT-3101, Sanofi/Teva TEV-48574, Spyre SPY002, AbbVie ABBV-701, Xencor XmAb942. Absci entered a proven class and lost on the molecule — the single most important comp datapoint for judging the platform.
• AI-antibody platform peers (for cap/cash context, multiples n/a — pre-revenue):

Provenance discipline: traditional multiples (EV/Sales, P/E) are n/a and would be meaningless for a $215K-revenue pre-revenue name. The honest valuation frame is rNPV (Lens 11), not comps. Note Absci trades at a richer cap than AbCellera on less cash — the market is paying up specifically for the ABS-201 alopecia optionality.

Lens 8 · Stock-Price Catalysts (5-yr pattern)

``, with the 10-Q/8-K context:

• IPO July 2021 at a near-$2B valuation → multi-year grind down to a $2.24 52-wk low as the platform-only story failed to produce revenue.
• July 2025: −18.8% on a $50M public offering at $3.00 (14.5% discount, ~28.5% dilution). Dilution is the recurring down-catalyst.
• Oct 2025 → Dec 2025: sharp re-rate up as ABS-201 HEADLINE dosing began and the alopecia narrative took hold.
• Jan 2026: ABS-101 half-life miss — notably did not break the stock, because the market had already moved its attention to ABS-201.
• 2026 YTD: ~3x off the low to $7.39 (2026-06-20), 52-wk range $2.24–$8.17, market cap ~$1.15B.

Pattern read: this stock trades on (1) ABS-201 alopecia news flow (up) and (2) equity raises (down). It is not an earnings stock — revenue prints are ignored. The market has decided ABS-201 is the whole company, and has paid up ~3x in anticipation of the H2'26 readout. That cuts both ways: a lot of good news is already in the price (Lens 12/13).

---

Phase C — Judge people & books

Lens 9 · Management

• Sean McClain — Founder & CEO since 2011. Started Absci at 22, straight out of U. Arizona (molecular & cellular biology), in his parents' basement. Archetype: founder-operator, "more businessman than scientist" (his own framing), Angela-Duckworth-"Grit" devotee. Built the company to 170–200+ employees and raised >$525–530M across its life. Capital-allocation track record includes two tuck-in acqui-hires — Totient (synthetic biology) and Denovium (deep learning) — that built the platform.
• Zachariah ("Zach") Jonasson — CFO (also a director); McClain credits him as a mentor and he is tied to Phoenix Venture Partners.
• Skin in the game: executive officers, directors, and 5% holders beneficially own >41% of the stock as of 2026-04-30. That is high insider/concentrated ownership — aligns McClain with shareholders but also means a small group controls outcomes.
• Comp / incentive structure: the 2024 founder market-award (1.5M PSUs to McClain) vests only on share-price thresholds; as of 3/31/26 none of the thresholds had been met (zero vested), and unvested tranches expire March 2027. Read that carefully — the CEO's own price-based equity is underwater and about to expire worthless. It's a genuine alignment signal (he's only paid if the stock works) and a tell that the stock has badly lagged the bar set in 2024.
• Capital-allocation judgment: The decisive 2026 act — killing ABS-101 on its half-life miss and de-prioritizing ABS-301/-501 to concentrate on ABS-201 — is, on balance, good discipline (don't throw money after a molecule that lost on PK in a crowded class). But it is also a confession that the platform's earlier picks didn't pan out, and it leaves the company with no diversification.
• Red flags: none forensic (clean Item 3, no enforcement — Lens 10). The soft flags are (a) a decade as CEO with zero approved products and zero milestone/royalty revenue, (b) a founder who self-describes as a businessman running a deep-science company, and (c) heavy, repeated dilution as the financing model.

Lens 10 · Forensic Red Flags

Forensic pass on a pre-revenue biotech — the accounting is simple; the risk is going-concern and dilution, not aggressive revenue recognition.

• Revenue recognition: Trivial in size ($215K Q1'26). Standard ASC 606 over-time recognition on drug-creation agreements; contract assets $0.1M, deferred revenue $1.0M. No channel-stuffing surface — there's almost no revenue to manipulate. Clean.
• Cash vs. earnings: Operating cash burn $(26.3)M Q1'26 vs $(29.6)M net loss — the gap is non-cash SBC ($4.4M) + D&A ($2.7M). No suspicious divergence; cash burn tracks the P&L honestly.
• SBC flattering non-GAAP: SBC $4.4M/qtr (~14% of opex), with $27.0M unrecognized option comp + $8.9M unrecognized RSU comp still to flow. Material but disclosed and ordinary for the sector. Absci does not lean on a flattering non-GAAP "adjusted" metric to obscure losses — it reports the GAAP loss straight.
• Intangibles/goodwill: Intangibles $40.8M (from the Totient/Denovium-era acquisitions), amortizing normally; a $5.1M gain on settlement of contingent consideration in FY2025 reduced opex — worth noting as a one-time below-the-line-ish item, but disclosed and immaterial to the thesis.
• Balance sheet flags: No debt of consequence ($0.4M long-term debt). The asset base is ~$118M of U.S. treasuries — pristine. The only balance-sheet risk is the clock: $125.7M cash, ~$26M/qtr burn → runway into H1 2028 on the company's own guidance. No going-concern qualification in the 10-K or 10-Q.
• Dilution (the real "red flag"): APIC grew from $688.7M (12/31/24) to $826.0M (3/31/26); shares outstanding 115.4M → 155.9M in ~15 months — a ~35% share-count increase. Funding sources: July 2025 public offering (16.67M sh @ $3.00, $46.7M net), AMD PIPE ($20M), and an open $100M ATM with TD Securities. The ATM is a permanent, at-market dilution spigot.

Regulatory findings (required sub-section).

• SEC Litigation Releases / AAERs: None. EDGAR EFTS (LR + AAER) returned 0 findings for "Absci" over 2021-06-21 → 2026-06-21.
• 10-K Item 3 (Legal Proceedings): "We are not currently a party to any material litigation or other legal proceedings.".
• Non-SEC enforcement (FTC/DOJ/FDA/etc.): Web search surfaced no material enforcement actions or settlements against Absci. The only FDA touch-point is forward-looking regulatory risk (the agency's draft guidance on AI in drug-development decision-making, and EU AI Act provisions effective Aug 2026) — disclosed as compliance risk, not an action against the company.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-21. The forensic risk here is dilution and runway, not fraud or accounting.

---

Phase D — Project & stress-test

Lens 11 · rNPV + runway-to-catalyst (replaces Forward Projection)

EPS projection is meaningless for a pre-revenue single-asset biotech — the right frame is rNPV of ABS-201 and whether cash reaches the value-inflection readout.

Runway-to-catalyst (the question that actually matters):

• Cash + securities $125.7M (3/31/26); operating burn ~$26.3M/qtr.
• Naïve runway ≈ $125.7M ÷ $26.3M ≈ 4.8 quarters → ~mid/late 2027. Management guides into H1 2028, implying they assume some combination of (a) slowing burn after ABS-101's wind-down, (b) modest ATM top-ups, and (c) potential ABS-101 partnering cash.
• Verdict on the question: yes — cash clears the gating catalyst. The H2 2026 interim PoC and even the early-2027 full PoC both sit inside the runway. Absci does not need to raise before the readout that defines the company. That is the single most important risk-management fact: they can let ABS-201 data print before they're forced back to the market. (A good readout then lets them raise cheap; a bad one means a distressed raise or worse — but the binary resolves first.)

rNPV of ABS-201 (illustrative, every input ``):

• Peak unrisked sales: analyst proxy ~$2.2B (AGA + endometriosis combined). Treat as the bull anchor, not gospel.
• Probability of success from Phase 1/2a in a novel mechanism: historically ~10–15% to approval for Ph1 assets; arguably a touch higher if the H2'26 interim shows clean efficacy, lower given novel-target risk. Use ~12%.
• Illustrative rNPV ≈ peak $2.2B × ~3–4x peak-sales-to-NPV-ish multiple is the wrong shortcut; instead: $2.2B peak × 12% PoS × ~0.25 (NPV/peak heuristic for a single-asset, royalty-light, self-commercialized profile, discounted at ~12%) ≈ $0.6–0.7B** risk-adjusted. Against a **$1.15B market cap, that says the market is pricing ABS-201 well above a 12% PoS — i.e., it's already crediting the readout with materially better-than-base odds (or crediting the platform/ABS-202/endo optionality on top). n/a — a precise rNPV is not sourced; this is a labeled order-of-magnitude estimate to frame the binary, not a target.
• Brier forecast to log (do NOT create in --watchlist): the scoreable question is binary — "ABS-201 HEADLINE interim PoC (H2 2026) shows statistically/clinically meaningful exploratory efficacy on hair-growth endpoints, p≈0.45." (Coin-flip-ish: novel mechanism, but encouraging ex-vivo + preclinical-vs-minoxidil data and clean safety so far.) Per watchlist rules, this is noted, not written via forecast.ts.

Lens 12 · Bull vs Bear

Bull case. Absci is a cheap call option on a genuinely novel, potentially first-in-class antibody (ABS-201) in a massive, underserved, decades-stale market (80M US AGA patients; endometriosis on top), where the competitive field of antibodies is nearly empty. The platform de-risks the next assets if this one hits (ABS-202 already follows the PRLR thread). Cash reaches the defining readout without a forced raise. Insiders own >41% and the CEO's equity only pays if the stock works. AMD's equity-backed compute deal is third-party validation. If H2'26 interim efficacy is clean, this re-rates violently toward the $15 Guggenheim case and becomes a partnering/M&A magnet.

Bear case (permanent-impairment risks).

1. Single-asset binary. After ABS-101's half-life failure and the ABS-301/-501 de-prioritization, there is no plan B inside the runway. A failed ABS-201 H2'26 interim doesn't dent the thesis — it is the thesis, and the stock likely halves or worse.
2. The platform's only clinical test so far, failed. ABS-101 lost on PK in a validated class against well-resourced competitors. That is direct evidence the AI edge may not translate to differentiated molecules. Bulls treat the platform as proven; the tape of in-house assets says "unproven, 0-for-1."
3. Dilution is structural. ~35% more shares in 15 months, a live $100M ATM, and a financing model that depends on equity markets staying open to a pre-revenue name. 25.9% of float is short.
4. Expectations are baked in. Up ~3x off the low to a ~$1.15B cap before the readout. The rNPV math (Lens 11) implies the market already credits ABS-201 with better-than-base PoS — so a merely-okay readout could disappoint a stock priced for a clean win.

Pre-mortem (18 months out, thesis broke): It's late 2027. ABS-201's interim efficacy was ambiguous (a placebo-adjusted hair-count delta that didn't separate convincingly), the full PoC confirmed mediocrity, and the novel PRLR mechanism turned out to drive less regrowth in humans than the ex-vivo model promised. With no other clinical asset and the share price down 60–70%, Absci did a distressed raise into the ATM, and the platform reverted to a services-and-partnering shell. The "AI antibody" premium evaporated.

Multiple too high? On any fundamental basis (zero revenue, $115M annual loss), yes — but that's the wrong lens. The question is whether ~$1.15B fairly prices a ~12% PoS, $2.2B-peak single asset, and the honest answer is it's pricing closer to 25–40% PoS, i.e., richly. You are buying after a 3x, before the binary.

Contrarian view (what the market refuses to see): The bull crowd is treating "AI-designed antibody platform" as the durable asset and ABS-201 as proof. The contrarian read is the inverse — the platform is unproven (0-for-1 in the clinic), and ABS-201 is a target-selection bet (pick an empty mechanism) more than an AI-design bet. If ABS-201 works, it may validate Absci's strategy (go where competition is thin) far more than its technology (AI makes better molecules). That distinction matters enormously for whether the next programs are worth anything.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case — and 26% of the float already agrees with me.

• Where revenue is concentrated: two foreign partners = 100% of a $215K revenue line. There is no business here yet — only a science project funded by selling stock.
• The moat is a slogan. "$15M, 2 years vs $50M, 6 years" is a marketing claim; the one molecule Absci took deep enough to test against real competitors (ABS-101) lost on half-life. A platform that can't beat comparators on PK in a validated class has not earned an AI premium.
• Most dangerous competitor bulls underestimate: for ABS-201 specifically, the risk isn't a rival antibody — it's that PRLR blockade simply doesn't produce clinically meaningful hair regrowth in humans, leaving the field to incumbents (minoxidil/finasteride/transplants) and oral JAK/next-gen topicals. The "empty competitive field" cuts both ways: maybe it's empty because the biology is hard.
• Worst capital-allocation reality: a decade, >$525M raised, $654M accumulated deficit, zero approved products, zero milestone revenue, and the financing engine is a perpetual ATM. Every quarter of delay is dilution.
• Assumptions that must hold for $7.39: (1) ABS-201 H2'26 interim shows clean efficacy; (2) the mechanism translates to the full PoC; (3) equity markets stay open for the inevitable raises; (4) the platform produces a second viable asset. Break any one and the stock re-rates down hard.
• If growth/data disappoints 20–30%: there is no "20–30%" — it's binary. A soft interim isn't a haircut to estimates; it's a −50%+ gap because the company is the asset.
• Single scenario that permanently impairs: ABS-201 fails its PoC. Plausibility: moderate-to-high for a novel-mechanism Ph1/2a (base rate ~85–90% of such assets never reach approval). That's the trade.

Lens 14 · Management Questions (ordered by information value)

1. ABS-101 missed on half-life versus next-gen TL1A comparators — what specifically did the AI platform get wrong, and what did you change in the design process for ABS-201 so the same PK shortfall doesn't recur?
2. What exact efficacy endpoint, effect size, and placebo separation in the H2 2026 HEADLINE interim would you consider a clear success versus an ambiguous result?
3. Given the runway into H1 2028 and a ~$26M/qtr burn, what is your raise plan if the H2'26 interim is merely "okay" rather than clean — and at what point do you tap the ATM regardless of data?
4. If ABS-201 works, do you partner/out-license it (your stated model) or build commercial yourself — and what does that decision do to the rNPV you'd capture?
5. The platform is 0-for-1 in differentiated clinical molecules. What is the objective, external evidence — beyond ABS-201 — that the AI confers a durable edge rather than just good target selection?
6. ABS-301 and ABS-501 were de-prioritized. Were those scientific failures, capital-triage decisions, or both — and does that imply the discovery engine's hit rate is lower than advertised?
7. Two foreign partners are 100% of revenue. What is the realistic timeline and probability of a new, sizable partner deal with upfront cash, given partners reportedly prefer upfront over your downstream-royalty model?
8. PRLR for AGA has an "empty" competitive field. What is your conviction that the field is empty because of opportunity, not because the biology has failed others?
9. The CEO market-award PSUs expire worthless in March 2027 with no thresholds met. How is management re-incentivized for the 2026–2028 value-inflection window?
10. What is the IP position on the AI models and the PRLR antibody composition specifically — how defensible is ABS-201 if a larger player wants to fast-follow once data prints?
11. How dependent is the design platform on the AMD compute relationship, and what happens to your roadmap if that collaboration lapses?
12. Endometriosis Ph2 starts Q4 2026 — are you funding that yourselves, and does it compete for cash with ABS-201's full PoC?
13. What is the manufacturing/CDMO path and cost for ABS-201 at commercial scale, and is any of that locked in?
14. What gross-margin / royalty economics do your existing Merck/AstraZeneca/Almirall agreements actually carry if a downstream milestone hits — quantify the option value.
15. What single piece of news in the next 12 months would most change your own view of whether Absci is a platform company or a single-asset biotech?

---