Absci Corporation

Monitor, do not commit. Absci has crossed from preclinical to clinical stage with one internally advancing asset (ABS-201, anti-PRLR) and has handed its former lead asset (ABS-101, anti-TL1A) to the outlicensing desk. The platform story is legitimate but unvalidated in humans. The clinical story is twelve months old. The competitive story is unfavorable on both assets.

• The company dosed its first-ever patient in May 2025 ¹ and has run exactly two clinical trials ². Its entire clinical history fits inside eleven months. ABS-101 entered a class that Merck, Roche, and Teva/Sanofi now defend with Phase 3 programs ³; in November 2025 Absci halted internal development and pivoted to ABS-201 ⁴. - ABS-201 is the only internally advancing asset, and its most direct competitor — Hope Medicine's HMI-115, licensed from Bayer ⁵ — has already completed Phase 2 endometriosis and cleared a US AGA IND ⁶. ABS-201 has dosed three SAD cohorts with emerging safety data only ⁷. - Financial runway reaches 1H 2028 on $144.3M cash + marketables against a $115M FY2025 net loss ⁸⁹; the weighted-average share count grew 24% in 2025 to 136.8M ¹⁰ and Q3 2025 weighted-average already stood at 143.8M ¹¹. Emerging-growth-company status expires 2026-12-31 ¹², adding Sarbanes-Oxley 404(b) audit cost into the next budget cycle. - Insiders bought $903K net on the open market in the trailing twelve months ¹³, with Board director Sir Mene Pangalos ($461K) and retiring Chief Innovation Officer Andreas Busch ($381K) as the two largest buyers ¹⁴. Pangalos himself joined the board in January 2024, one month after the AstraZeneca collaboration announcement (up to $247M in company-disclosed headline value, Tier-3 — upfront and near-term cash not separately disclosed) ¹⁵¹⁶; his accumulation reads as continued personal conviction rather than relationship theater. - Reopen the file on three conditions: (i) ABS-201 HEADLINE interim proof-of-concept in H2 2026 shows target area hair count gains that would credibly outperform the HMI-115 Phase 1b open-label, single-arm, N=12 male AGA result (approximately +14 hairs/cm² mean non-vellus TAHC gain per Hope Medicine's January 2024 press release, uncontrolled). Note: the HMI-115 Phase 2 AGA trial NCT06118866 (N=192, Peking University People's Hospital) completed 28 November 2024 with no published results as of 16 April 2026; there is no peer-reviewed Phase 2 AGA benchmark in the public record, so ABS-201's efficacy floor rests on uncontrolled Phase 1b data only; (ii) ABS-201 generates a US composition-of-matter patent grant around the current WO application, closing the Bayer WO2019011719A1 freedom-to-operate exposure ¹⁷¹⁸; (iii) Absci signs a Large Pharma drug-creation partnership that converts platform activity into disclosed upfront economics, rather than the Tier-3 company-disclosed milestone-weighted headline totals that have characterized prior deals (AstraZeneca up to $247M, Almirall up to $650M — upfront and near-term cash portions remain undisclosed) ¹⁹²⁰.

---

2.1 ABS-101 (anti-TL1A, IBD) — from lead to outlicensing in six months

ABS-101 is a subcutaneous anti-TL1A monoclonal antibody engineered for extended half-life. Absci developed it to compete in inflammatory bowel disease on dosing convenience rather than novelty of target. The Australian subsidiary Absci Pty Ltd sponsored a first-in-human Phase 1 trial at Nucleus Network in Melbourne — registry ACTRN12625000212459 — which dosed its first participant on 12 May 2025 (ANZCTR actual-enrolment date; the 13 May 2025 press-release date reflects the post-dosing announcement) ²¹²². The US IND cleared but was never used ²³; the single Melbourne site and the Australian registry narrowed US-investor visibility on the program.

The trial design was a single-ascending-dose study: four subcutaneous cohorts (150, 300, 600, 1000 mg) plus one intravenous bridge at 300 mg, with eight participants per cohort for a target of 40 ²⁴²⁵. By the last public registry update on 22 July 2025, fifteen of forty target participants had been accrued ²⁶; the November 2025 interim readout therefore rested on fewer than twenty participants.

That readout — released alongside the Q3 2025 results — claimed "extended half-life as compared to first-generation anti-TL1A competitor programs, with no apparent impact of ADA on PK" ²⁷. The comparison elided the second-generation programs: duvakitug (Teva/Sanofi) and afimkibart (Roche, t½ 18–19 days). Absci disclosed no numeric ABS-101 half-life value. Within one week of the interim release, the company announced its decision "not to initiate additional later-stage development trials for ABS-101 internally at this time" and shifted resources to ABS-201 ²⁸. Absci now seeks a partner ²⁹.

This pivot tracks the class. Merck paid $10.8B for Prometheus in April 2023 to secure tulisokibart ³⁰, Roche paid $7.1B upfront plus $150M in milestones for Telavant's afimkibart in October 2023, and Sanofi paid Teva $500M upfront plus up to $1B in milestones for duvakitug. Three Phase 3 programs now run in parallel. Tulisokibart's Phase 2 ARTEMIS-UC posted 26% clinical remission versus 1% placebo at 12 weeks (NEJM 2024). ABS-101's differentiation was half-life; engineering differentiation attracts partners only when the class leader has not yet set a benchmark, and tulisokibart has set one. Absci's capital position — roughly $115M in annual losses against $144.3M in liquidity ³¹³² — could not fund a competing Phase 3.

The outlicensing decision is defensible capital allocation. The open question is whether the extended-half-life story alone, with no Phase 2 data, attracts a licensee willing to pay meaningful upfront economics. The precedent here is unhelpful: BB-TL1A-VIAL-HLE, another half-life-extended TL1A program, terminated its Phase 1 in 2026 when the PK endpoint was unlikely to be met.

2.2 ABS-201 (anti-PRLR, AGA + endometriosis) — the only asset Absci now develops itself

ABS-201 is an AI-designed anti-prolactin receptor (PRLR) monoclonal antibody developed through Absci's Origin-1 platform (prior brand: Integrated Drug Creation). The company first disclosed the target and lead indication at its December 2024 R&D Day ³³; prior quarters had described it only as "an undisclosed dermatology target" ³⁴. First-in-human dosing occurred on 3 December 2025 in the HEADLINE Phase 1/2a trial (NCT07317544), a quadruple-masked, 227-participant study at Sinclair Dermatology and Nucleus Network in Melbourne ³⁵³⁶. The design comprises four to six SAD IV cohorts (150–1800 mg, a twelve-fold dose span) followed by three to four MAD subcutaneous cohorts (300–1200 mg) ³⁷; the MAD portion is "powered to demonstrate human proof-of-concept for the use of ABS-201 to treat androgenetic alopecia by stimulating significant hair regrowth" ³⁸. Secondary endpoints include target area hair count (TAHC), hair width, darkening, and investigator-reported outcomes ³⁹ — standard AGA readouts.

As of 24 March 2026, Absci had completed three of the four-to-six SAD cohorts with favorable emerging safety data ⁴⁰. The company's pipeline page, however, shows ABS-201 AGA as "Phase 1/2a at 85%" ⁴¹; the gap between three dosed SAD cohorts and the 85% progress bar is a presentation choice that overstates asset momentum. The pipeline page makes the same 85% claim for ABS-201 in endometriosis ⁴², for which no registered clinical trial exists and Phase 2 initiation is guided to Q4 2026 ⁴³. The website progress bars and the regulatory footprint do not match.

The preclinical package behind ABS-201 rests on company disclosures rather than peer-reviewed publications. Non-human primate studies demonstrated "extended half life" with greater than 90% subcutaneous bioavailability ⁴⁴ and projected Q8W–Q12W dosing in humans ⁴⁵; Absci also reports a ≥150 mg/mL high-concentration formulation target ⁴⁶. In December 2025 the company claimed "statistically significant superior hair regrowth compared to minoxidil in a preclinical mouse model" ⁴⁷. No effect size was disclosed, no peer-reviewed publication exists, and the telogen-to-anagen-transition mouse AGA model has low translatability to human androgen-driven miniaturization. March 2026 added human ex vivo data showing STAT5 phosphorylation inhibition and anagen prolongation in isolated follicles ⁴⁸; mechanistic coherence with the PRLR literature is genuine (follicle PRLR expression is documented in Int J Mol Sci 2021) ⁴⁹.

2.3 ABS-301 (immuno-oncology, undisclosed target)

ABS-301 remains at Lead ID stage, displayed as 25% on the pipeline page ⁵⁰. The target has been undisclosed since its December 2024 unveiling ⁵¹; a Q3 2025 update reported completion of "the first in vivo target validation study" with signaling-pathway-driven anti-tumor response ⁵². December 2024 guidance called for candidate selection in 1H 2025 ⁵³; sixteen months later the asset still shows Lead ID. The timeline has slipped without acknowledgement.

2.4 ABS-501 (anti-HER2)

ABS-501 is an AI-designed anti-HER2 antibody generated through the same platform, displayed as Candidate at 0% ⁵⁴. Absci characterizes it as having "novel epitope interactions, affinity equal or greater than trastuzumab in preclinical, efficacious against trastuzumab-resistant xenograft" ⁵⁵⁵⁶. The international patent application WO2025122885A1 was filed 8 December 2023 ⁵⁷. Trastuzumab is a twenty-five-year-old benchmark; preclinical affinity parity is the table stakes for entry into the anti-HER2 space, not differentiation.

2.5 History — from SoluPro to Origin-1

Absci began in a 400-square-foot basement lab in 2011 ⁵⁸. Founder Sean McClain holds inventorship on the original SoluPro/SoluPure cytoplasmic expression patent (US12157766B2, priority 2012-08-05, granted 2024-12-03) ⁵⁹, and the earliest granted Absci US patent (US9617335B2) dates to 2017 ⁶⁰. The company's original IP base was microbial protein expression. The pivot to generative-AI antibody design came later, accelerated by the June 2021 Totient acquisition (which supplied the early oncology antibody patents US12285484 and US12134655) ⁶¹⁶². The Origin-1 platform brand launched in 2026 ⁶³; its predecessor brand was "Integrated Drug Creation". The older and current platform URLs both return 404s and have been relocated to `/denovo` without redirects ⁶⁴⁶⁵ — a minor hygiene signal on a company now claiming AI leadership.

---

This report scores target validation on a 0–100 scale: TL1A = 85, PRLR = 42.

3.1 TL1A (target validation score: 85)

TL1A (tumor necrosis factor-like ligand 1A) drives T-cell activation and fibroblast-mediated tissue damage in inflammatory bowel disease. Three independent lines of evidence establish the target: (i) GWAS signals in TNFSF15 (the gene encoding TL1A) associated with both ulcerative colitis and Crohn's disease; (ii) elevated TL1A expression in inflamed gut mucosa; (iii) preclinical knockout and antibody-blockade models demonstrating attenuation of colitis. Clinical validation has now replaced preclinical inference. Merck's tulisokibart delivered 26% remission versus 1% placebo at 12 weeks in the Phase 2 ARTEMIS-UC trial (NEJM 2024); Teva/Sanofi's duvakitug posted 48% UC remission at 900 mg versus 20% placebo in RELIEVE UCCD; Roche/Genentech's afimkibart reached 36% remission at Week 56 in TUSCANY-2.

The class's commercial validation matches its biological validation. Merck paid $10.8B for Prometheus in April 2023 to secure tulisokibart ⁶⁶; Roche paid $7.1B upfront for Telavant's afimkibart six months later. Peak-sales consensus for tulisokibart alone sits at $4–5B per year. Against this backdrop, TL1A warrants a Target Validation Score of 85 — the mechanism is biologically established, clinically proven, and commercially priced.

ABS-101 inherits all of this validation. Its question is not whether TL1A is a real target but whether a fourth-place half-life-extended program can earn Phase 2/3 capital from a licensee.

3.2 PRLR (target validation score: 42)

Prolactin receptor (PRLR) signals through JAK2-STAT5 in multiple tissues including hair follicles, breast, ovary, bone, and pituitary. PRLR engagement by prolactin drives the catagen (regression) phase of the hair cycle; PRLR blockade accordingly promotes anagen prolongation in mouse models. Absci's human ex vivo data confirm STAT5 inhibition and anagen prolongation in isolated follicles ⁶⁷; the mechanistic biology is coherent with the published literature ⁶⁸. In endometriosis, PRLR signaling is upregulated locally in ectopic lesions and drives inflammation independently of the hormonal axis targeted by current GnRH-based therapies ⁶⁹.

Clinical validation of PRLR as a non-oncology target rests almost entirely on one competitor program. Hope Medicine's HMI-115 (licensed from Bayer) completed a Phase 2 trial in endometriosis-associated pain, published in Lancet Obstetrics, Gynaecology, and Women's Health in November 2025 (N=108 completed; randomized 1:1:1:1 to 60/120/240 mg HMI-115 vs placebo, SC Q2W × 12 weeks). Only the 240 mg arm reached statistical significance against placebo: dysmenorrhea NRS reduction of 41.57% (240 mg) vs 18.61% (placebo) at Week 13; non-menstrual pelvic pain reduction of approximately 52% (LSM) versus placebo in the final readout. No sex-hormone disruption, no change in BMD, estradiol, LH, FSH, or progesterone ⁷⁰. The earlier 2024 interim release (N=102 of 142 planned) reported 42% / 50% figures that were subsequently refined in the peer-reviewed publication; the shipped numbers in the first cut of this report matched the interim, not the Lancet final, and have been reconciled here. An AGA Phase 1b (N=12 males, open-label, single-arm, uncontrolled, Sinclair Dermatology, Australia) has been described in a January 2024 Hope Medicine press release as delivering approximately +14 hairs/cm² mean non-vellus TAHC gain; this is not a Phase 2 benchmark. The Phase 2 AGA trial NCT06118866 (N=192, single site at Peking University People's Hospital) completed 28 November 2024 with `hasResults: false` and no Lancet, JAAD, JID, or conference-abstract publication as of 16 April 2026. Bayer's own mouse endometriosis data (Otto et al., 2022) demonstrated efficacy equivalent to faslodex and cetrorelix without anti-estrogenic effects.

Two cautions temper this validation. First, the Novartis anti-PRLR LFA102 failed to show efficacy in Phase 2 breast cancer — a different indication, but the only prior anti-PRLR in a large efficacy study. Second, Bayer's own PRLR program BAY 1158061 completed Phase 1 in 2018 with favorable PK and has posted no Phase 2 registry entry in the eight years since. The silence is unusual for a drug with positive preclinical efficacy data. Absci is not entering a fully de-risked class; it is entering a class whose lead clinical data sit in one paper from one sponsor.

PRLR accordingly warrants a Target Validation Score of 42 — the mechanism is biologically supported and one competitor has generated supportive Phase 2 data, but no approved anti-PRLR therapy exists and two prior industry programs have stalled silently.

3.3 Origin-1 platform claims

Absci's platform asserts de novo antibody design against zero-prior epitopes: targets without published protein-protein complex structures or homologs above 60% sequence identity ⁷¹. The bioRxiv preprint (Levine et al., 2026) reports successful developable binders for four of ten targets — COL6A3, AZGP1, CHI3L2, and IL36RA — with cryo-EM atomic accuracy (3.0–3.1 Å resolution, DockQ 0.73–0.83) and an IL36RA binder optimized to 104 nM potency ⁷²⁷³. Absci characterizes this as "the first demonstration of de novo design of full-length, monoclonal antibodies against 'zero-prior' epitopes with atomically accurate complex structures and functional activity" ⁷⁴.

Two constraints bound the platform claim. The 4/10 success rate is a 40% hit rate, meaning six of ten targets failed to produce developable binders ⁷⁵. And all 54 authors on the Origin-1 preprint are Absci employees or contractors; no independent validation has been published. The 2023 flagship de novo antibody preprint (Shanehsazzadeh et al.) remains a preprint after three years, with bioRxiv view count at 1,134 ⁷⁶⁷⁷ — low for a claimed flagship. Absci has never published its core antibody-design work in a peer-reviewed journal ⁷⁸. The platform's most distinctive claim — zero-prior de novo design — has not been tested against a human clinical target. ABS-201 uses PRLR, which has extensive prior structural knowledge; it is not a zero-prior case.

---

4.1 ABS-101 Phase 1 — the half-life comparison that was not made

Absci's November 2025 interim disclosure claimed "extended half-life as compared to first-generation anti-TL1A competitor programs, with no apparent impact of ADA on PK" and a favorable overall safety profile with no serious adverse events ⁷⁹. The claim is verbatim accurate. Three things it does not state:

1. The numeric half-life of ABS-101. No company disclosure provides t½ in days. 2. The comparator. "First-generation anti-TL1A" omits the second-generation programs that currently define the class — duvakitug (SC Q2W) and afimkibart (t½ 18–19 days) [cid:88 verification note]. 3. The sample size. The ANZCTR registry showed 15 of 40 target participants accrued as of 22 July 2025 ⁸⁰; the interim readout published three and a half months later therefore rested on fewer than twenty participants.

A December 2024 R&D Day disclosure reported "reduced internalization of TL1A complexes in in vitro THP-1 immunogenicity tests compared to a competitor molecule with a high clinical anti-drug antibody (ADA) rate" ⁸¹. The unnamed comparator is likely PF-06480605 (tulisokibart precursor), which generated an 82% ADA rate in its Phase 1 trial. That context validates the engineering choice but does not translate to a head-to-head ADA comparison in humans. The 13-week GLP toxicology reported "no treatment-related adverse findings during in-life phase and necropsy were observed; histopathology is pending" ⁸²; the "histopathology is pending" qualification has not been updated since the December 2024 disclosure.

The Phase 1 outlicensing decision followed the interim readout by six days. If the data had cleanly beaten the second-generation competitors on head-to-head-comparable PK metrics, the narrative and the outlicensing timing would likely have differed.

4.2 ABS-201 HEADLINE — three SAD cohorts and a proof-of-concept that is still to come

HEADLINE's three completed SAD IV cohorts have reported favorable emerging safety data and no serious adverse events ⁸³. The proof-of-concept readout requires the MAD SC portion ⁸⁴, which has not yet initiated. Company guidance calls for preliminary safety/tolerability/PK data in 1H 2026, interim POC in H2 2026, and full POC in early 2027 ⁸⁵. The guidance cadence is reasonable for a Phase 1/2a that began dosing on 3 December 2025.

Two design features warrant note. The SAD IV dose range of 150–1800 mg spans twelve-fold ⁸⁶; for a molecule destined for subcutaneous administration, the 1800 mg top IV dose reads as a supratherapeutic-exposure safety probe more than a clinical development step. And the primary endpoint is safety alone ⁸⁷; efficacy endpoints are secondary — standard for Phase 1/2a but relevant when the MAD arm is described as "powered for human proof-of-concept" ⁸⁸. Efficacy will emerge from a secondary endpoint analysis.

4.3 Anti-anchoring on preclinical claims

Several Absci disclosures require calibration:

• "80 million Americans with AGA" ⁸⁹⁹⁰. The figure includes all degrees of hair loss; the addressable moderate-to-severe subset is materially smaller. Absci uses "80 million" in its TAM framing without segmentation. - "Approximately 9 million women in the U.S." with endometriosis and "prevalent in up to 10% of women worldwide" ⁹¹⁹². The 5–10% range cited in Absci's own December 2025 disclosure ⁹³ was truncated to the upper bound in subsequent messaging. Flagged as embellished. - "No FDA-approved disease-modifying therapy" for endometriosis ⁹⁴. Orilissa (elagolix, FDA-approved 2018) and Myfembree (relugolix combination, 2022) are FDA-approved for endometriosis-associated pain. Absci's framing is defensible only on a narrow "non-hormonal disease-modifying" reading; the blanket claim contradicts the FDA label record. Flagged CONTRADICTED. - "Statistically significant superior hair regrowth compared to minoxidil in a preclinical mouse model" ⁹⁵. No effect size, no peer-reviewed publication, and a model with low human translatability. Flagged embellished. - "2–3x extended half-life as compared to antibodies in clinical development" (August 2024 ABS-101 NHP data) ⁹⁶. Specific comparator antibodies not named. No peer-reviewed corroboration. - Pipeline-page progress bars. ABS-101 "Phase 1 at 85%" despite the November 2025 internal halt ⁹⁷; ABS-201 AGA "Phase 1/2a at 85%" with three SAD cohorts of four-to-six dosed ⁹⁸; ABS-201 endometriosis "Phase 1 at 85%" with no clinical trial registered anywhere ⁹⁹. The website UI consistently overstates asset momentum.

4.4 The peer-review gap

Absci has published zero peer-reviewed journal articles on its flagship de novo antibody design work ¹⁰⁰. The 2023 Shanehsazzadeh et al. preprint and the 2026 Origin-1 preprint both remain unreviewed ¹⁰¹¹⁰². The publications page lists ten titles ¹⁰³; none is confirmed as peer-reviewed in the cross-checked inventory. For an AI-platform company raising capital and signing partnerships on the basis of platform superiority, the continued reliance on bioRxiv-only evidence is a material trust constraint on the science.

---

This report scores safety signals on a 0–100 scale where higher = higher concern: TL1A class = 22, PRLR class = 58.

5.1 TL1A class safety signal score: 22 (LOW)

The TL1A class has been de-risked clinically. Tulisokibart's Phase 2 ARTEMIS-UC trial reported a clean safety profile comparable to placebo. Duvakitug and afimkibart Phase 2 data showed no class-defining safety signals. FAERS pharmacovigilance reporting on the class to date has not generated disproportionality signals against standard IBD-biologic reference drugs. ABS-101's Phase 1 interim report — "no serious adverse events" across approximately fifteen participants ¹⁰⁴ — is consistent with the class. The safety signal score of 22 reflects a biological mechanism well-understood in inflammation, three Phase 2/3 programs with clean readouts, and no regulatory holds on any TL1A program.

5.2 PRLR class safety signal score: 58 (MODERATE-HIGH)

The PRLR safety picture is less mature. The knockout phenotype in mice is documented: female homozygous mice are completely infertile and lack mammary development, hemizygous females cannot lactate after their first pregnancy, pituitary prolactin levels rise 30–100-fold with pituitary hyperplasia in some cases, and bone formation is reduced in both sexes at all ages ¹⁰⁵. Each phenotype predicts a potential on-target liability for chronic anti-PRLR administration in humans: fertility impairment, lactation failure, hyperprolactinemia, and bone-density reduction.

Absci's Phase 1/2a protocol (NCT07317544) monitors inclusion criteria including BMI and healthy-subject status ¹⁰⁶, but the registered protocol does not specify DEXA bone-density measurement or bone-turnover biomarker monitoring. In a chronic-use asset aimed at androgenetic alopecia — a non-life-threatening indication with strong competition from existing SOC (minoxidil, finasteride) ¹⁰⁷ — the absence of bone-turnover monitoring in the Phase 1/2a readout is a protocol gap that Phase 2 design will have to address.

Clinical precedent tempers the theoretical risk. HMI-115's Phase 2 endometriosis trial reported no sex-hormone disruption and no fertility signal ¹⁰⁸; Bayer's Phase 1 BAY 1158061 posted favorable PK without disclosed safety concerns. But neither program has generated long-term (>12 months) human safety data. The safety signal score of 58 reflects on-target phenotype predictions that have not been ruled out in humans, plus the absence of bone-turnover monitoring in the current protocol, tempered by near-term HMI-115 Phase 2 evidence of tolerability.

5.3 Implications for development

For ABS-101, the low class safety signal supports the outlicensing thesis: a partner would not face unusual regulatory headwinds. For ABS-201, the moderate-high class safety signal means the company cannot rely on safety alone to create differentiation. It must generate efficacy differentiation in HEADLINE — relative to HMI-115 and to minoxidil/finasteride SOC — to justify chronic use in a cosmetic-adjacent indication. Endometriosis changes the calculus: the indication is severe, the unmet need is genuine ¹⁰⁹, and the safety bar is higher in theory but lower in practice given the pain severity.

---

Absci's patent estate is three things in one: a legacy microbial-expression platform that McClain founded ¹¹⁰, an acquired oncology-antibody stack from the June 2021 Totient merger ¹¹¹¹¹², and a pending generative-AI method portfolio that has not yet generated granted US patents ¹¹³. The company's own 10-K claims "79 issued or granted patents and 159 pending patent applications worldwide, which includes ten issued U.S. patents and 24 pending regular U.S. patent applications" ¹¹⁴.

6.1 Nine granted US patents verified (one short of the 10-K)

Independent Google Patents and USPTO verification confirms nine granted US patents assigned to Absci Corp or Absci LLC as of 16 April 2026:

The 10-K's "ten issued U.S. patents" exceeds the nine verifiable via public registries by one. The gap could reflect a recently-granted tenth patent indexed after this DD pull, or a 10-K counting inconsistency; it warrants a direct company inquiry and is flagged for reopening.

6.2 ABS-201 has no granted US composition-of-matter patent

The most material IP gap sits on the only internally-advancing asset. ABS-201 has no publicly discoverable granted US composition-of-matter patent as of 16 April 2026 ¹²⁴, despite the HEADLINE Phase 1/2a trial having begun dosing on 3 December 2025. The international application is expected pending, but the public record does not yet confirm a US grant. An AI-designed antibody that enters clinical development without a granted US CoM patent in its destination indications carries meaningful freedom-to-operate risk — a risk that sharpens in view of the Bayer prior art described below.

6.3 The ABS-101 freedom-to-operate blocker (Merck/Prometheus)

US11999789B2 (Cedars-Sinai, Prometheus/Merck) — "Humanized antibodies to TL1A" — granted 4 June 2024, expires 23 October 2040 ¹²⁵. This is the primary freedom-to-operate concern for ABS-101. Absci's own CoM patent US12473370B2 granted November 2025 ¹²⁶ covers its specific ABS-101 sequences, but overlap analysis with the Merck/Prometheus claims has not been independently published. For an outlicensing candidate, FTO becomes the licensee's diligence problem; for a company developing the asset itself, it would be a material program risk. The outlicensing pivot reduces Absci's direct FTO exposure on ABS-101 but does not eliminate it as a constraint on the deal value the company can extract.

6.4 The ABS-201 freedom-to-operate blocker (Bayer / Hope Medicine)

WO2019011719A1 — "Prolactin receptor antibody for male and female pattern hair loss," Bayer IP GmbH, priority 10 July 2017 — is the primary FTO concern for ABS-201 ¹²⁷. Bayer has licensed the asset to Hope Medicine, which developed HMI-115 and completed a Phase 2 AGA program in August 2024 (N=192 males) ¹²⁸. Bayer's WO2011069799A1 earlier patent on PRLR-neutralizing antibodies expands the prior-art surface. Absci's ABS-201 may clear around these claims on specific sequence and epitope grounds, but a competitor with licensed Bayer IP is already running Phase 2 trials in ABS-201's two target indications. The competitive landscape and the FTO landscape are the same landscape.

6.5 Pending method claims — the core AI platform IP is not yet granted

US20230268026A1, "Designing biomolecule sequence variants with pre-specified attributes," filed 7 January 2022, represents the core generative-AI method claim and remains in pending status ¹²⁹¹³⁰. Absci holds no granted US method patent for its de novo antibody design algorithm. The international application WO2025122885A1 (ABS-501 anti-HER2) filed 8 December 2023 ¹³¹ and five other PCT applications ¹³² represent the global filing strategy, but PCT applications do not become enforceable patents until national-phase entry and grant.

6.6 Portfolio expiration window

The 10-K reports an expected expiration range of August 2033 through December 2044 ¹³³. The earliest-expiring patents — the 2012-priority SoluPro family — protect platform manufacturing technology more than antibody composition. The composition-of-matter patents protecting specific assets (US12473370 for ABS-101, pending for ABS-201) extend to 2044. The gap between platform-IP and asset-IP expiration narrows the company's long-horizon defensibility against biosimilar entry on individual products.

---

7.1 Executive team

Sean McClain — Founder and CEO has run Absci since inception in August 2011 ¹³⁴. He holds a BS in Molecular and Cellular Biology from the University of Arizona. No prior exits or CEO history outside Absci. His beneficial ownership as of 31 March 2025 was 10,760,728 shares (8.30%) ¹³⁵. Across fourteen years he has taken one company from basement lab through IPO into clinical stage. The pattern of long tenure without visible governance conflicts is favorable; the absence of prior exits leaves his operating-CEO track record unbenchmarked outside Absci.

Zach Jonasson, PhD — CFO and Chief Business Officer assumed the role on 31 August 2023 ¹³⁶. Before taking the CFO/CBO seat he served on the Absci board as an independent director from October 2020 to December 2023, and on the Absci LLC predecessor board from April 2016 to October 2020 ¹³⁷. He co-founded Phoenix Venture Partners and Convergent Ventures. His Harvard PhD is in behavioral neuroscience. The governance fact that requires disclosure: Phoenix Venture Partners II LP — Jonasson's prior firm — is a Schedule 13G 10%+ beneficial owner of ABSI (9,354,695 shares filed 19 February 2025) ¹³⁸. Jonasson transitioned from independent board chair into an executive officer role at a company where his former fund is a 10% holder. The double-trigger change-in-control package (12 months base + 1.0× target bonus + 12 months COBRA) ¹³⁹ is standard. Late Section 16(a) filing disclosed in the 2025 proxy (one Form 4 filed late for a 5,194-share RSU settlement) ¹⁴⁰ is a controls irregularity, modest in scope.

Ransi Somaratne, MD, FACC, MBA — Chief Medical Officer started 3 March 2026 ¹⁴¹¹⁴². He joined from Vertex Pharmaceuticals, where he was SVP Clinical Development; earlier roles at BioMarin (Roctavian, hemophilia A gene therapy) and Amgen (Repatha PCSK9) and on Journavx (NaV1.8). Thirty-plus peer-reviewed publications including NEJM and JAMA ¹⁴³; MD from Albany Medical College, MBA from UNC Chapel Hill. Somaratne was Absci's first full-time CMO, hired ten months after ABS-101 entered Phase 1 ¹⁴⁴. His inducement grant — 650,000 options at $2.57 (the 3 March 2026 close), 4-year vest ¹⁴⁵ — sits 14% below the July 2025 offering price of $3.00, tying compensation to a depressed strike.

Andreas Busch, PhD — Chief Innovation Officer retires effective 31 March 2026 and transitions to Scientific Advisory Board co-chair ¹⁴⁶. The 8-K discloses "Dr. Busch's resignation was not the result of any disagreement with the operations, policies or practices of the Company" ¹⁴⁷. Busch is credited by the company as "a driving force in building Absci's AI-designed therapeutics pipeline, most notably in the discovery and development of ABS-201" ¹⁴⁸. His $381K open-market purchase over the trailing twelve months ¹⁴⁹ — on the way out the operational door — is a meaningful positive signal on personal conviction. Compensation was paid via Swiss subsidiary Absci GmbH in CHF ¹⁵⁰.

Amir Shanehsazzadeh — Chief AI Officer ascended to the CAIO title in April 2025, replacing Amaro Taylor-Weiner (who had held the title at the December 2024 R&D Day) ¹⁵¹. Harvard AB Mathematics and SM Statistics. Shanehsazzadeh is an inventor on the ABS-101 CoM patent US12473370B2 ¹⁵² and first author of the 2023 flagship de novo antibody bioRxiv preprint. His tenure began as an AI Scientist in June 2022 — a fast internal promotion track for a company where the AI platform is the central thesis.

Shelby Walker, JD — Chief Legal Officer joined June 2024 from Korro Bio GC/Corporate Secretary (May 2023 – June 2024); prior SVP Intellectual Property at CRISPR Therapeutics (March 2018 – April 2023) and GC at Ginkgo Bioworks (May 2016 – March 2018) ¹⁵³. Deep IP track record for a company whose valuation depends on IP and platform defensibility.

Todd Bedrick, CPA — Chief Accounting Officer has served since June 2022 ¹⁵⁴. He acquired 124,200 shares and disposed of 36,173 net over the trailing twelve months ¹⁵⁵ — small purchases but directionally consistent with insider bullishness.

7.2 Board of Directors (seven members)

The board was fixed at seven seats on 7 July 2025 upon Mary Szela's appointment ¹⁵⁶. Current composition: Sean McClain (CEO), Sir Mene Pangalos (Class II, since Jan 2024), Daniel Rabinovitsj (Class II, since Nov 2022), Karen McGinnis (Class I, audit chair, since Aug 2020), Frans van Houten (Class I, since June 2023), Joseph Sirosh (Class III, since Jan 2022), Mary Szela (Class III, since July 2025) ¹⁵⁷¹⁵⁸. The 8-K class split is 2/2/3, not the 3/3/1 the ledger source originally recorded — the DEF14A and 8-K reconcile on Class I: McGinnis/van Houten; Class II: Pangalos/Rabinovitsj; Class III: McClain/Sirosh/Szela.

Sir Mene Pangalos, PhD served as EVP BioPharmaceuticals R&D at AstraZeneca from January 2019 to March 2024 ¹⁵⁹. He joined the Absci board one month after the December 2023 AstraZeneca $247M collaboration ¹⁶⁰¹⁶¹. His trailing twelve-month open-market purchases of $461K ¹⁶² are the largest among insiders. Biogen board service runs in parallel. The AstraZeneca connection is visible; the personal capital commitment is material.

Mary Szela joined 7 July 2025 ¹⁶³. Current CEO of TriSalus Life Sciences since January 2018; previously CEO of Novelion Therapeutics and Melinta Therapeutics; earlier ran the US pharma business at Abbott Laboratories ($8B) ¹⁶⁴. Also on the boards of Kura Oncology, Prometheus Biosciences (the company Merck acquired for $10.8B), Omega Therapeutics, and Senda Biosciences. Commercial-stage CEO experience at three biotechs adds a lane not otherwise represented on the board.

Frans van Houten, MSc — former CEO of Royal Philips (March 2011 – October 2022) ¹⁶⁵. Novartis AG independent director since February 2017. The audit committee financial expert designation is shared with Karen McGinnis ¹⁶⁶.

Karen McGinnis, CPA — ex-VP and Chief Accounting Officer at Illumina (November 2017 – April 2021) ¹⁶⁷. Audit committee chair.

Daniel Rabinovitsj — VP at Meta since August 2018 ¹⁶⁸; prior COO Ruckus Wireless, SVP Qualcomm Atheros. Tech-sector experience; no biotech-specific credentials.

Joseph Sirosh, PhD — CEO of CreatorsAGI Inc; former VP Amazon Alexa Shopping (October 2022 – November 2023); prior CTO Compass, CVP and CTO of AI at Microsoft (July 2013 – December 2018) ¹⁶⁹. AI-sector board bridge.

Amrit Nagpal (Managing Director, Redmile Group LLC) resigned from the board on 25 April 2025 without disclosed disagreement ¹⁷⁰. The original ledger misstated the first name; the 8-K confirms Amrit Nagpal. Redmile is a healthcare-focused hedge fund; the departure of a Redmile MD may reflect reduced dedicated-biotech-fund engagement.

7.3 Insider open-market activity — net +$903K over twelve months

The trailing twelve months show net insider open-market purchases of approximately $903K across four buyers ¹⁷¹: Pangalos $461K, Busch $381K, van Houten $149K, Bedrick $27K. Two modest sellers: McClain $80K and Jonasson $52K, same-day 2 February 2026 — a pattern consistent with pre-scheduled tax-related sales rather than directional signaling. The gross open-market purchase total of $1.02M across 37 Form 4 filings in nine and a half months is a net-bullish insider signal, led by the two directors (Pangalos, van Houten) and the retiring CIO (Busch), none of whom needed to commit personal capital and all of whom did.

7.4 Scientific Advisory Board and endometriosis advisory board

The SAB comprises Andreas Busch (co-chair post-retirement), Mene Pangalos, Iain McInnes (Vice Principal, University of Glasgow), Hubert Truebel (CMO AiCuris), and Victor Greiff (Associate Professor, University of Oslo) ¹⁷². The March 2026 endometriosis advisory board added Hugh Taylor (Yale), Linda Giudice (UCSF), Gaurang Daftary (Inception; Yale/Mayo), Steven Young (Duke), Zaraq Khan (Mayo), and Joan-Carles Arce (Repronovo) ¹⁷³ — credible academic coverage for an indication Absci has not yet registered a trial in.

7.5 Team summary

The executive bench has the credentials it needs: a commercially-seasoned CMO (Somaratne), a veteran finance/BD operator in the CFO role (Jonasson, with a governance flag on his former-firm relationship), a heavyweight IP generalist as CLO (Walker), and an in-house AI lead (Shanehsazzadeh) whose promotion trajectory matches the company's central thesis. The board has pharma commercial experience (Szela), Big Pharma R&D experience (Pangalos), deep tech/AI experience (Sirosh, Rabinovitsj), and two audit-committee financial experts (McGinnis, van Houten). Two simultaneous C-suite transitions — a new CMO and the CIO's retirement — during a clinical-ramp period ¹⁷⁴¹⁷⁵ invite execution-risk scrutiny. The net-bullish insider purchase pattern, led by the directors and the outgoing CIO, signals personal conviction that outlasts the re-org.

---

Absci ended FY2025 with $144.3M in cash, cash equivalents, and marketable securities against a $115.2M net loss ¹⁷⁶¹⁷⁷. Management tells investors the balance funds operations "into the first half of 2028" ¹⁷⁸. The math holds only if the company spends faster than it earns and then finds new capital before the wall — a pattern the last three years document clearly.

Revenue tells the first-order story. Partner-program revenue fell to $2.8M in FY2025 from $4.5M in FY2024, a 38% decline ¹⁷⁹. Nine-month 2025 revenue fell 44% YoY to $2.15M ¹⁸⁰. Q3 2025 revenue collapsed 78% YoY to $0.378M ¹⁸¹. The four-year arc — $5.7M (2022) ¹⁸², $5.7M (2023), $4.5M (2024), $2.8M (2025) — shows two consecutive years of contraction while headline partnership announcements accumulated.

R&D spend moved the opposite direction. FY2025 R&D reached $81.4M, up 27% from $63.9M, driven by ABS-101 Phase 1 costs and ABS-201 IND-enabling work ¹⁸³. SG&A held flat at $35.1M vs. $36.2M ¹⁸⁴. Total operating expenses hit $123.1M, producing a $120.3M operating loss and $115.2M net loss ^185186187. The accumulated deficit grew to $624.8M at year-end ¹⁸⁸.

Q4 2025 cash burn appears deceptively modest at $8.2M because Absci recognized a $5.1M gain on settlement of Totient contingent consideration ¹⁸⁹. The Totient escrow resolution returned $8.7M unrestricted cash to Absci and $7.6M to former Totient sellers from the original June 2021 merger ¹⁹⁰. Strip the one-time gain and the true operating burn runs closer to $13.3M for the quarter — roughly $53M annualized, or consistent with the ~$115M full-year cash-loss pattern. The Totient earn-out failed to trigger its full payout, indicating the post-acquisition milestones were not hit.

Share count tells the dilution story. Weighted-average diluted shares grew 24% YoY, from 110.2M (FY2024) to 136.8M (FY2025) ¹⁹¹. Q3 2025 weighted-average already stood at 143.8M, and year-end 2025 shares approached 150M ¹⁹²¹⁹³. Three capital raises inside 18 months drove the move: $86.4M follow-on closed March 1, 2024 (days after the AstraZeneca $247M deal announcement) ¹⁹⁴; $20M AMD PIPE at $3.50/share in January 2025 ¹⁹⁵; and a $64M July 2025 raise — $50M underwritten offering of 16.67M shares at $3.00 plus $14M through the ATM ¹⁹⁶¹⁹⁷.

The July 2025 raise extended runway guidance from 1H2027 to 1H2028 ¹⁹⁸¹⁹⁹. The shelf pattern is clear: Absci raises opportunistically on news-driven equity windows rather than on predictable milestones. The August 2025 S-3 base shelf paired with a $100M ATM facility managed by TD Cowen remains available — a standing option to issue whenever market conditions allow ²⁰⁰²⁰¹.

Emerging-growth-company exemptions expire December 31, 2026 ²⁰²²⁰³. FY2027 will require full SOX 404(b) auditor attestation on internal controls, a cost and disclosure step that hits at the same time ABS-201 interim data is expected to drive capital decisions. The EGC revenue threshold is $1.235B ²⁰⁴ — not a binding constraint given $2.8M actual revenue, but the calendar trigger fires regardless.

---

Absci runs both clinical programs outside the US IND framework. The ABS-101 Phase 1 first-in-human trial is registered at ACTRN12625000212459 on the Australian New Zealand Clinical Trials Registry, sponsored by Absci Pty Ltd, sited at Nucleus Network in Melbourne ²⁰⁵²⁰⁶. The ABS-201 HEADLINE Phase 1/2a trial NCT07317544 shows `isFdaRegulatedDrug: false` in the ClinicalTrials.gov oversight module; the sponsor is Absci Pty Ltd., both sites sit in Melbourne, and the study runs under an Australian CTN ²⁰⁷²⁰⁸.

The 10-K acknowledges the consequence directly: the FDA "will generally not approve the application on the basis of foreign data alone." An ABS-101 IND cleared in the US but the Phase 1 ran only in Australia ²⁰⁹. The company captured Australia's cash-refundable R&D tax incentive and faster start-up, but a BLA strategy will require bridging Phase 1 or repeating key studies on US sites and US patients. Timelines for any US pivotal program compound accordingly.

IBD endpoint framework. Tulisokibart's ARTEMIS-UC Phase 2 used the standard FDA ulcerative-colitis primary: clinical remission at Week 12, composite of Mayo subscores (stool frequency 0 or 1, no rectal bleeding, endoscopic subscore 0 or 1). Tulisokibart achieved 26% vs 1% placebo (p<0.001). Duvakitug RELIEVE reported 48% UC remission at 900mg vs 20% placebo at Week 14, and 48% Crohn's endoscopic response vs 13% placebo. Roche's afimkibart TUSCANY-2 reported 36% remission and 50% endoscopic improvement at Week 56. Any ABS-101 pivotal will need to match or beat this bar.

Endometriosis endpoint framework. Pain scores (NRS for dysmenorrhea, non-menstrual pelvic pain), symptom diary completeness, and laparoscopic confirmation at entry define the class. HMI-115's Lancet OGWH 2025 paper (N=108 completed; 240 mg arm) reported 41.57% dysmenorrhea NRS reduction vs 18.61% placebo, and approximately 52% non-menstrual pelvic pain reduction (LSM) vs placebo, without sex-hormone disruption ²¹⁰. ABS-201's endometriosis Phase 2 has not started; the company plans initiation Q4 2026.

AGA endpoint framework. Target Area Hair Count (TAHC), Target Area Hair Width (TAHW), and Global Photographic Assessment (GPA) anchor the FDA AGA package for finasteride and minoxidil. NCT07317544 lists TAHC, TAHW, and darkening/pigmentation as secondary efficacy endpoints ²¹¹. The HEADLINE MAD portion is "powered to demonstrate human proof-of-concept" in AGA, according to the company [cid:95, flagged EMBELLISHED pending disclosed statistical plan]. HMI-115 Phase 1b in male AGA reported +14 hairs/cm² mean non-vellus TAHC increase ²¹² — modest by the standard of approved comparators.

PRLR on-target safety. PRLR-knockout mice are infertile (homozygous females), fail to lactate after first pregnancy (hemizygotes), develop pituitary hyperplasia with 30- to 100-fold prolactin elevation, and show reduced bone formation in both sexes ²¹³. The NCT07317544 protocol does not monitor bone turnover or schedule DEXA scans. HMI-115 in its endometriosis Phase 2 reported no sex-hormone disruption, but the published endpoint set did not include bone-mineral-density measurements across treatment duration. ABS-201 inherits this liability and will need a safety package that addresses it before any chronic-dosing indication proceeds.

---

IBD is the largest market Absci addresses. The global IBD therapeutic market reached $25.5B in 2023, projected to $47.3B by 2034 at roughly 6% CAGR ²¹⁴. The US market reached $16.6B in 2024 per GMInsights' August 2025 edition (an earlier draft of this report cited $13.12B from a prior edition; the figure has been reconciled to the current publication). Biologics account for $22.4B of the 2024 global total. AbbVie's Skyrizi + Rinvoq combined reached $25.8B in 2025 sales, growing 14% YoY; J&J's Tremfya hit $5.2B, up 40%. Tulisokibart analyst consensus projects $4–5B peak sales.

ABS-101's role in this market is now an outlicensing asset, not a revenue-bearing product. The November 2025 decision to partner rather than self-develop shifts the valuation question from "what will Absci earn on IBD?" to "what will a partner pay for Phase 1 TL1A with extended half-life?" ²¹⁵. The deal-value bracket is bounded above by Roche's $7.1B upfront for Telavant (Phase 2b/Phase 3-ready) and below by AbbVie's $150M upfront plus $1.56B milestones for FutureGen's preclinical FG-M701 (~$1.7B total). Realistic ABS-101 partnering range: $100–300M upfront, $800M–1.7B total, if 2026–2027 outlicensing completes with supportive Phase 1b data. The FTO wall is Cedars-Sinai/Prometheus/Merck's US11999789B2, granted June 2024, expiring 2040-10-23 — the primary freedom-to-operate risk for any ABS-101 licensee ²¹⁶.

Endometriosis. The global market was $1.97B in 2025, projected to $4.68B by 2033 at 11.28% CAGR. GlobalData's 7MM forecast is narrower: $938M (2024) to $2.5B (2034), with the US 68–88% of the total. Approved hormonal options — elagolix (Orilissa, AbbVie, FDA-approved), relugolix combination (Myfembree/Ryeqo, Pfizer/Myovant, $10–15K/year) — suppress estrogen with bone-loss and fertility liabilities. HMI-115 has already demonstrated the non-hormonal mechanism can work in humans ²¹⁷. ABS-201 peak sales range, unrisked: $300–900M in endometriosis, assuming 15–35% class share against HMI-115 and legacy hormonal options.

Androgenetic alopecia. The market reached $2.7B in 2023, projected to $4.76B by 2030. Biologic pricing faces a hard cash-pay ceiling. Generic finasteride telehealth costs $10–26/month; oral minoxidil 1.25mg runs $28.60/3-month supply ²¹⁸. Ritlecitinib's $49,000/year list price covers alopecia areata, an autoimmune indication with insurance coverage — it does not transfer to AGA. A systemic biologic for pattern hair loss will face prior-auth rejection and must sell cash-pay at a fraction of that price. ABS-201 AGA unrisked peak: $50–200M.

Outlicensing comparable band. AbbVie/FutureGen (TL1A, preclinical, June 2024) set the floor at $1.7B total. Sanofi/Teva duvakitug (Phase 2b, October 2023): $500M upfront + $1B milestones. Merck/Prometheus tulisokibart (Phase 2 UC/CD, April 2023): $10.8B all-cash ²¹⁹. Roche/Telavant afimkibart (Phase 2b UC, October 2023): $7.1B upfront + $150M milestone. The ceiling is not available for Phase 1 TL1A with a single extended-half-life differentiation story.

---

TL1A class scorecard. Three Phase 3 programs now define the IBD TL1A field: Merck tulisokibart (ATLAS-UC NCT06052059 and ARES-CD NCT06430801) ²²⁰; Teva/Sanofi duvakitug (RELIEVE UCCD, Phase 3 initiating H2 2025); Roche/Genentech afimkibart (AMETRINE-1/2, initiated September 2024). One Phase 1 terminated for PK endpoint failure: BB-TL1A-VIAL-HLE (NCT07029971). One adjacent PD-1 depleter (AnaptysBio rosnilimab) failed UC Phase 2. ABS-101 enters as the fifth or sixth TL1A program, Phase 1, with a half-life differentiation narrative that remains untested in Phase 2. The class is validated by $18B+ in M&A and validated against placebo at 26–48% remission rates.

PRLR class — ABS-201 is second-in-class, not first. HMI-115 (Hope Medicine, originally BAY 1158061 from Bayer, licensed 2019) completed Phase 2 in endometriosis with Lancet OGWH publication November 2025 (N=108; 240 mg arm delivered 41.57% dysmenorrhea NRS reduction vs 18.61% placebo, only the 240 mg dose statistically significant). The Phase 2 AGA trial NCT06118866 (N=192, single-site at Peking University People's Hospital) completed 28 November 2024 but has not been published, with `hasResults: false` on the ClinicalTrials.gov registry — HMI-115's AGA Phase 2 benchmark therefore rests on the Hope Medicine Phase 1b open-label N=12 readout, not a controlled Phase 2. HMI-115 received Breakthrough Therapy Designation from China's NMPA for endometriosis. The 2018 Bayer Phase 1 of BAY 1158061 in breast cancer completed (PMID 29806538) with favorable PK; the eight-year gap before the molecule returned in women's health remains unexplained in the public record. A third PRLR program — Novartis LFA102 — failed breast cancer Phase 2 on efficacy. Absci is second-in-class; the company's public language calling ABS-201 "category-defining" should be read against this fact [cid:205, flagged EMBELLISHED].

AI-platform peers (public comp table).

Absci trades at one-third to one-half of AbCellera's market cap despite a later clinical stage and lower revenue. The discount reflects trajectory (ABS-101 outlicensed rather than self-developed, ABS-201 generating no revenue), $115M annual net loss, and the absence of any Phase 2 human efficacy data. EV/cash runway math: RXRX 1.3x, SDGR 1.3x, RLAY 2.7x (zovegalisib Phase 3 BTD premium), ABSI ~1.9x. Absci trades in line with platform peers on EV/cash but below clinical-catalyst peers.

The three biggest competitive threats: HMI-115 establishes PRLR proof-of-concept before ABS-201 reaches Phase 2; TL1A class crowding renders ABS-101 unlicensable at favorable economics; AI antibody platform commoditization compresses partnership economics (Xaira's $1B war chest exceeds Absci's entire market cap).

---

Four CONTRADICTED claims in the ledger.

SEC inventory referenced "Eric Furfine, PhD as CMO (joining April 7, 2026, ex-Apellis/Regeneron)." No 8-K confirms any Furfine appointment; the single CMO appointment on record is Ransi Somaratne, MD FACC MBA, effective March 3, 2026 ²²³. The Furfine entry is a hallucination that entered the initial inventory pass.

An earlier 8-K digest referenced "Ruchir Nagpal resignation." The actual name per 8-K 2025-04-28 is Amrit Nagpal, Managing Director at Redmile Group LLC ²²⁴. A Redmile MD departure still signals reduced healthcare-hedge-fund engagement; the transcription error does not change the signal.

An earlier board-composition digest reported a "3/3/1 class split." The actual split per 8-K 2025-07-07 is 2/2/3 — Class I McGinnis and van Houten; Class II Pangalos and Rabinovitsj; Class III McClain, Sirosh, and Szela ²²⁵. Total of seven board seats is correct.

A March 2026 press release stated "No FDA-approved" treatment for endometriosis ²²⁶. This claim omits elagolix (Orilissa, AbbVie, FDA-approved 2018), relugolix combination (Myfembree/Ryeqo, Pfizer/Myovant, FDA-approved 2022 for endometriosis pain), and GnRH agonists. Hormonal therapies are approved; no non-hormonal, disease-modifying therapy exists — those are different statements. The company's framing conflates the two.

Seventeen EMBELLISHED claims. The pipeline page shows "Phase 1 at 85%" for ABS-101, "Phase 1/2a at 85%" and "Phase 1 at 85%" for the two ABS-201 indications ^227228229. Progress bars are not trial milestones; the framing treats marketing artwork as clinical status. The ABS-101 Q3 2025 interim-data language — "extended half-life vs first-generation TL1A, no SAEs" — inflates marketing-friendly properties without publishing the PK curve that would let a reader verify the comparison ²³⁰. The "80 million Americans" AGA TAM ²³¹²³² includes all degrees of hair loss; the moderate-to-severe treatable subset is roughly one-third that figure in published dermatology registries. The "10% of women worldwide" endometriosis prevalence is the upper bound of a 5–10% literature range the same company acknowledged in its December 2025 release ²³³²³⁴. ABS-201's AGA MAD portion "powered to demonstrate human proof-of-concept" ²³⁵ reads as a primary-endpoint commitment without a disclosed statistical plan; the hair-count measurement is secondary per the ClinicalTrials.gov record ²³⁶. Rodney Sinclair's testimonial calling ABS-201 "exciting" ²³⁷ is a paid-advisor endorsement, not independent validation.

One hundred and five UNSUPPORTED claims. The company quotes that appear across its own press releases (partnership economic framings, CEO statements about "24 months concept-to-clinic," IgDesign1 "first in vitro validated inverse folding model," "6-week design-build-test cycle") lack independent external corroboration ^{238239240241242}. They may be true; without external benchmarks they remain company assertions. The $247M AstraZeneca and $650M Almirall deal-value headlines are milestone-weighted; neither upfront nor near-term cash is separately disclosed ²⁴³²⁴⁴. The $1B partnerships framing from December 2023 press ²⁴⁵ is theoretical max milestone aggregate; realized revenue remains <$3M/yr.

Probability-weighted existential risks.

• TL1A outlicensing failure (probability 25–35%): If no partner commits by H1 2027, ABS-101 becomes a stranded asset. Effect: compresses ABS-201 to sole-asset status; reduces platform partnership narrative; forces a capital raise at lower share price. Impact: $150–250M market-cap range. - ABS-201 HEADLINE interim signal weakness (probability 40–50%, given HMI-115 Phase 1b TAHC delta of 14 hairs/cm² as the floor): Interim data in H2 2026 that fails to demonstrate clear TAHC advantage versus HMI-115 or minoxidil-like comparators triggers equity-raise pressure before endometriosis Phase 2 initiation. Effect: forced financing at 2025 trough prices. Impact: dilutive raise at $1.50–2.50/share range. - PRLR on-target bone/fertility signal (probability 15–25% over chronic dosing): Bone turnover markers or pituitary signals in MAD cohort or subsequent chronic-dosing studies force development hold or label restrictions ²⁴⁶. Effect: endometriosis pathway (chronic dosing) compromised; AGA positioning shifts to intermittent. - Running out of cash before ABS-201 Phase 2 POC (probability 20–30%): 1H2028 runway guidance assumes no setbacks; the $100M ATM capacity ²⁴⁷ exists precisely because management anticipates further issuance. Effect: heavy dilution from current ~150M shares ²⁴⁸ toward 200M+. - EGC/SOX 404(b) compliance costs and disclosure surprises (probability 30–40%, low individual impact): December 31, 2026 deadline hits during ABS-201 readout season ²⁴⁹. Low-to-moderate dollar impact, but material controls exceptions would compound investor perception risk. - CMO retention risk (probability 20–30% over 24 months): Ransi Somaratne's pre-Vertex CMO history shows three consecutive sub-two-year tenures — Epirium Bio (20 months), Spring Discovery (12 months), Endpoint Health (7 months). He joins Absci six weeks before this report's date, ten months after ABS-101 entered Phase 1 ²⁵⁰. A CMO departure during the ABS-201 interim-readout window or the first US IND transition would disrupt clinical-development execution and signal instability to partners weighing the ABS-101 outlicense. The Vertex SVP credential and the 4-year vesting on 650K options at $2.57 ²⁵¹ argue for retention; the prior-tenure pattern argues against.

---

Platform beyond ABS-101/ABS-201. Origin-1, the January 2026-rebranded platform (formerly "Integrated Drug Creation"), reported 4/10 zero-prior-epitope success on targets COL6A3, AZGP1, CHI3L2, and IL36RA, with cryo-EM structural confirmation at 3.0–3.1 Å resolution ^252253254. IL36RA optimization reached 104 nM potency after AI-guided affinity maturation ²⁵⁵²⁵⁶. The platform's scientific core is a bioRxiv preprint whose 54 authors are all Absci employees or contractors — internal validation only, with no peer review completed as of retrieval. The 2023 flagship anti-HER2 de novo paper on bioRxiv (DOI 10.1101/2023.01.08.523187) remains unreviewed three years after posting, with 1,134 cumulative views ²⁵⁷²⁵⁸.

Partnership deal flow and cumulative disclosed value.

The $247M + $650M + $20M headline sum understates the reality and overstates it at once. Understates, because multi-year royalties on successful products could exceed the upfront-plus-milestone cap. Overstates, because milestone achievement is low-probability and cash realization thin — partner-program revenue remains $2.8M annually against a $115M net loss. The narrative "four new drug creation partners in 2024" ²⁷¹²⁷² delivered on 2024 guidance. The 2025 guidance — "one or more partnerships, including with a Large Pharma company, in 2025" ²⁷³ — appears to have slipped; the March 2026 Q4 FY2025 release guides the same target into 2026 ²⁷⁴. Partnering momentum measured in cash rather than announcements has slowed.

Origin-1 licensing optionality. If ABS-201 Phase 1/2a interim data supports the half-life and potency claims, Origin-1 itself becomes a licensable asset — not just a source of partnered programs. AstraZeneca's December 2023 deal and Almirall's 2025 expansion both validate the platform at the level of "can design a functional antibody to a specified target." A pharma-scale Origin-1 license (not a program-specific collaboration) has not yet been signed. The scenario's value is highest if ABS-101 licenses successfully in 2026–2027 and ABS-201 AGA interim in H2 2026 delivers a clean signal; lowest if both assets stall.

---

14.1 rNPV and scenario table (ABS-101 + ABS-201 combined, unrisked peak-sales)

Important caveat. The probability and IRR digits in this table are this report's scenario-analysis construction, not primary-source facts. They are grounded in published industry base rates (noted below) plus qualitative adjustments described in text; each number is a judgment call, not a cited data point. The reader should treat the rNPV as a framework for decision, not as a verifiable valuation.

Evidence anchors for PoS inputs:

¹ Bull 0.25 ≈ IBD-weighted antibody Ph1-to-approval upper quintile. Source: BIO/BioMedTracker/Informa Pharma Intelligence, Clinical Development Success Rates and Contributing Factors 2011–2020 (published 2021). The industry-average antibody Ph1→approval rate is approximately 14% across all indications; top-quartile for well-validated targets with Phase 2 precedent runs 20–30%. The Bull anchor assumes ABS-201 endometriosis inherits HMI-115's Phase 2 validation and ABS-101 completes an early-partnering event.

² Base 0.14 = industry-average antibody Ph1→approval rate (Informa 2011–2020). No class-specific premium or discount applied. Source: BIO Industry Analysis, Clinical Development Success Rates 2011–2020, Table 1 (all-indications antibody rate).

³ Bear 0.06 ≈ class-adjusted floor reflecting (a) late-entry penalty for ABS-101 (fifth+ TL1A entrant in a class defended by three Phase 3 programs) and (b) second-in-class commercial-share penalty for ABS-201 PRLR behind HMI-115. The penalty math halves the industry base rate and applies a commercial-share discount; the result is a judgment estimate, not a cited data point.

Peak-sales ranges follow the market-stream build: ABS-101 $500M–$2.0B unrisked; ABS-201 endometriosis $300–900M; ABS-201 AGA $50–200M — each assumes mid-cycle class penetration against named comparators (tulisokibart, HMI-115, elagolix, finasteride). Terminal multiples anchor to precedent M&A: Roche/Telavant $7.1B / ~$2B projected peak = 3.5x; Merck/Prometheus $10.8B / ~$4.5B projected peak = 2.4x ²⁷⁵. Implied EV sensitivity runs ±25% from recent share-price volatility. The scenario probabilities (0.20 / 0.55 / 0.25) are analyst judgment; no historical base-rate source applies.

Current market cap ~$400–450M (CompaniesMarketCap, BusinessQuant, Macroaxis live snapshots 2026-04-10/16) against implied base-case EV $350–500M: market is pricing Absci close to base case. Bull case ($1.0–1.3B EV) implies 2.3–3.0x upside to current cap; Bear case ($90–150M EV) implies 65–80% downside.

14.2 Cap table and dilution path

Reconstructed from SEC Form D history and post-IPO filings.

Cumulative capital raised since founding: ~$216M pre-IPO + ~$225M IPO + $170M post-IPO = ~$611M. Accumulated deficit $624.8M ²⁸⁴ roughly equals total capital raised. Implied: every dollar of capital has been spent, with no retained earnings to cushion the next cycle.

Implied pre-money today: ~$400–450M market cap minus ~$144M cash equals enterprise value of ~$256–306M, consistent with platform-peer EV-to-cash ratios. Cumulative dilution from IPO float (~90M shares) to current (~150M shares ²⁸⁵) exceeds 60%. Burn rate of $115M/year against $144M cash matches 15–18 month hard runway if no further financing occurs; the 1H2028 guidance ²⁸⁶ assumes the $100M ATM is partially drawn and operating efficiencies materialize.

14.3 Comparable transactions (last 24 months, same mechanism class or adjacent)

TL1A deal cascade establishes a $1.5–10.8B band depending on stage. ABS-101 at Phase 1 with one published half-life differentiation falls closest to the FutureGen floor. The Recursion/Exscientia combination signals platform-plus-platform consolidation pressure — relevant pricing anchor for Origin-1 platform deals.

14.4 Forty-eight hour action plan

Decision now: MONITOR. The base-case EV matches current market cap; the Bull case requires two sequential wins (ABS-101 outlicense closes 2026 and ABS-201 AGA interim reads positive H2 2026); the Bear case requires only one miss. Risk-adjusted return at current price is 12–17% IRR over three years — below biotech venture-equivalent hurdle rates and below public-biotech-index benchmarks for the same period. Wait for inflection data.

Triggers to reopen:

1. ABS-201 HEADLINE interim hair-growth readout, H2 2026 (TAHC delta vs baseline, with or without placebo separation). 2. HMI-115 Phase 3 initiation or major pharma licensing deal in endometriosis or AGA (removes "first-mover in class" optionality from ABS-201). 3. ABS-101 outlicensing announcement with disclosed upfront terms (confirms floor of base case). 4. Large-pharma platform deal for Origin-1 beyond target-specific collaboration. 5. Capital raise priced below $2.00/share (signals distress or capitulation).

Technical DD gaps to close before position:

1. ABS-101 Phase 1 interim PK data in full (not press-release framing) — verify "extended half-life" with numbers. 2. ABS-201 NHP PK data behind the 90% subcutaneous bioavailability and Q8–Q12 week dosing claims ^288289290[cid:104, flagged EMBELLISHED]. 3. Bone turnover monitoring plan for chronic-dosing indications — PRLR-KO bone phenotype ²⁹¹ must be addressed before endometriosis Phase 2. 4. Internal ABS-201 vs HMI-115 head-to-head preclinical comparison (if any exists). 5. Peer-review status on the Origin-1 bioRxiv preprint and the 2023 flagship HER2 preprint ²⁹²²⁹³.

Reference calls (3–5):

• Clinical dermatologist running comparator AGA trials — question: is Absci's NCT07317544 MAD protocol adequately powered for TAHC to differentiate from finasteride? - IBD gastroenterologist with TL1A investigator experience — question: how do pharma licensors weigh Phase 1 half-life differentiation against three Phase 3 competitors? - Ex-Hope Medicine executive or ex-Bayer BAY 1158061 program — question: what stopped the Bayer program between 2018 Phase 1 and 2019 HMI-115 licensing? Is there a silent liability? - Public biotech analyst covering AbCellera / Recursion — question: what multiples apply to platform-plus-clinical-asset biotechs with declining partner revenue? - Endometriosis KOL (Hugh Taylor at Yale or Linda Giudice at UCSF, both on Absci's endometriosis advisory board ²⁹⁴) — question: what level of TAHC or pain-score separation would make ABS-201 a credible second-in-class?

Commitment window: 120–180 days. The ABS-201 HEADLINE interim readout in H2 2026 is the dominant catalyst; no action before then unless triggers 2 (HMI-115 Phase 3) or 5 (capital raise below $2.00) fire first.

14.5 What we could not verify (Tier-3 claims whose resolution would flip thesis)

• Claim: ABS-101 "extended half-life vs first-generation TL1A, no SAEs" interim Phase 1 data ²⁹⁵. Why not verified: Company press-release framing only; no PK curves, no comparator half-life definition, no reviewed data. ANZCTR accrual 15 of 40 at 2025-07-22 ²⁹⁶ indicates the interim read likely used <20 participants. What resolution would mean: A clean half-life >30 days supports the outlicensing case toward the $1.7B total deal-value floor. A half-life <20 days with equivocal safety collapses the differentiation story; ABS-101 becomes a stranded asset.

• Claim: "$247M AstraZeneca collaboration" and "$650M Almirall collaboration" headline totals ²⁹⁷²⁹⁸. Why not verified: Upfront and near-term cash portions not separately disclosed. Realized partner revenue remains <$3M/year. What resolution would mean: Upfront-plus-milestone-in-first-three-years >$50M per deal supports the Bull platform-value case. Pure-milestone terms at current realized pace collapse the platform revenue thesis.

• Claim: ABS-201 "Q8–Q12 week dosing intervals in humans" extrapolated from NHP PK ²⁹⁹. Why not verified: NHP-to-human PK extrapolation without published clinical data. HEADLINE SAD dose range 150–1800 mg IV, MAD SC 300–1200 mg ³⁰⁰. What resolution would mean: Human PK supporting Q8W+ dosing materially differentiates from HMI-115 (subcutaneous q2w per Lancet protocol); establishes ABS-201's developability claim. Human PK requiring Q2–Q4W dosing eliminates differentiation from HMI-115.

• Claim: Origin-1 "first demonstration of de novo design of full-length mAbs against zero-prior epitopes with atomically accurate complex structures" ³⁰¹. Why not verified: bioRxiv preprint with 54 authors, all internal; not peer-reviewed. Competing platforms (Chroma, RFdiffusion, ESM3) have published similar de novo claims in related contexts; independent head-to-head benchmarks are absent. What resolution would mean: Peer-review acceptance at a top-tier journal (Nature, Science, Cell, Nat Biotech) converts the platform claim from marketing to scientific consensus and supports Origin-1 out-licensing value. Continued preprint-only status after three years (per the 2023 HER2 preprint pattern ³⁰²) leaves the platform claim contestable.

• Claim: "At least one or more partnerships, including with a Large Pharma company, in 2026" ³⁰³. Why not verified: Guidance language; not a signed term sheet. Carries forward from 2025 guidance ³⁰⁴ that slipped. What resolution would mean: A large-pharma platform (not program-specific) deal in 2026 validates Origin-1 as a licensable asset independent of ABS-101/ABS-201. Absence of such a deal by YE2026 confirms platform revenue plateau.

• Claim: CEO "AI design to clinic in approximately two years at a fraction of industry cost" ³⁰⁵. Why not verified: Speed and cost claims lack quantitative comparator benchmark; no published per-program cost disclosure. What resolution would mean: If a partner discloses per-program economics supporting the claim, Origin-1 commands premium platform-licensing value. If cost-per-candidate matches industry norms at disclosure, Absci's differentiation rests on output quality rather than economics.

---

Report generated 2026-04-16. Distribution: BigBio AI Research — Public. Primary sources cached under `sources/`; complete claims ledger at `company_claims_ledger.yaml`; verification ledger at `canonical.yaml`; numeric audit at `numeric_audit.md`; forensic notes at `forensic_notes.md`; provenance trail at `provenance_log.yaml`.