Coefficient Bio
Last reviewed
Jun 7, 2026
Sources
16 citations
Review status
Source-backed
Revision
v1 · 1,697 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 7, 2026
Sources
16 citations
Review status
Source-backed
Revision
v1 · 1,697 words
Add missing citations, update stale details, or suggest a clearer explanation.
Coefficient Bio was a stealth-stage biotechnology and artificial-intelligence startup that, according to multiple press reports, was acquired by Anthropic in an all-stock deal reported at just over $400 million and disclosed in early April 2026. The company, founded only about eight months earlier by two former computational-drug-discovery researchers from Genentech's Prescient Design unit, was building AI models for biological research and drug discovery. The acquisition was widely read as Anthropic's most direct move yet into AI for the life sciences, extending a strategy anchored by the company's Claude for Life Sciences product. Anthropic did not publicly confirm the purchase price, and the $400 million figure rests on reporting by The Information and the journalist Eric Newcomer rather than on any official Anthropic disclosure.
Coefficient Bio operated in stealth, with no publicly released product, disclosed revenue, or conventional traction metrics at the time the deal became public. Press accounts describe a small, research-heavy team of fewer than ten people, most of them computational biologists, working on AI systems intended to make drug discovery and broader biological research more efficient. According to TechCrunch and FierceBiotech, the platform the team was developing aimed to let AI draft drug research-and-development plans, manage clinical and regulatory strategy, and help identify new drug candidates. In framing obtained by Newcomer from a letter to investors, the company described its ambition in expansive terms, positioning its work toward "artificial superintelligence for science" and the goal of bringing biopharma "into the Intelligence Age." Those phrases reflect the company's own marketing language and should be read as aspiration rather than as a description of shipped capability.
The acquisition was reported on April 3, 2026, first by The Information, with TechCrunch, FierceBiotech, BioSpace, and Newcomer following. Sources close to the deal told TechCrunch that the transaction had closed but declined to comment on the amount, so the dollar figure remains attributed rather than confirmed. The incoming team was reported to be joining Anthropic's healthcare and life-sciences organization, which is led by Eric Kauderer-Abrams.
Coefficient Bio was co-founded by Samuel Stanton and Nathan C. Frey, both of whom previously worked in computational drug discovery at Prescient Design, the AI-focused computational-biology unit inside Genentech (part of Roche). Reports place the company's formal founding roughly eight months before the April 2026 announcement, which points to a launch around the second half of 2025.
Frey's public research profile is substantial and independently verifiable: he holds a PhD in materials science and engineering from the University of Pennsylvania, has published extensively in venues such as Nature Machine Intelligence and Science Advances, and his work on machine learning for biomolecule design and biological foundation models received an Outstanding Paper Award at the ICLR conference in 2024. Stanton's background is likewise in machine learning, experimental design, and drug-discovery research, with work spanning Bayesian optimization and generative methods for biological sequence design carried out during his time at Genentech. The pairing of two researchers from the same elite industrial computational-biology group is part of what made the deal notable: the value was understood by observers to lie primarily in the team and its expertise rather than in any revenue-generating product, a structure often described as an acqui-hire.
Reporting by Newcomer identified the venture firm Dimension as the company's lead backer, holding roughly a 50 percent stake. In a letter to its limited partners disclosing the outcome, Dimension reported an internal rate of return of 38,513 percent on the investment, a figure that reflects both the all-stock nature of the deal and the extremely short interval between founding and acquisition. That eye-catching IRR became one of the most-shared details of the story, though it derives from a single investor letter and should be treated as a reported figure.
The deal was structured as an all-stock acquisition. Newcomer's account, based on the Dimension letter, put the value at "just over $400 million in stock," and several outlets noted that a deal of that size represented only a small fraction of Anthropic's valuation. At the time, Anthropic had recently been associated with a post-money valuation in the range of roughly $350 billion to $380 billion, so a $400 million stock transaction implied dilution on the order of about a tenth of one percent. Because Anthropic declined to comment publicly on the price, and because the figure traces to The Information's reporting and the investor letter rather than to a company filing or press release, this article treats the $400 million number as reported and unconfirmed.
It is worth being precise about how this acquisition fits Anthropic's corporate history. Some coverage described it as Anthropic's first major acquisition, but reporting by Newcomer noted that Anthropic had previously made acquisition moves, including the team behind the Bun JavaScript runtime and the startup Vercept. The cleaner characterization, and the one supported across sources, is that Coefficient Bio was Anthropic's most significant move into the life sciences specifically, and one of its largest talent-focused deals to date, rather than its first acquisition outright.
The purchase sits within a deliberate, roughly half-year expansion of Anthropic's presence in healthcare and the life sciences. In October 2025 the company launched Claude for Life Sciences, a configuration of its Claude models aimed at supporting researchers across the discovery-to-commercialization pipeline. That product added connectors to scientific platforms including Benchling, PubMed, BioRender, Synapse.org, and 10x Genomics, alongside Agent Skills and a life-sciences prompt library. Anthropic reported that its Claude Sonnet 4.5 model scored 0.83 on the Protocol QA benchmark, above a human baseline of 0.79, with comparable gains on the BixBench bioinformatics evaluation. In January 2026 the company followed with a broader healthcare push, sometimes branded as Claude for Healthcare, extending capabilities from preclinical research through clinical operations and regulatory affairs.
Against that backdrop, acquiring a team of computational biologists who had built tooling for protein modeling, biomolecular representation, and R&D planning is a natural complement. Where Claude for Life Sciences applies a general-purpose model to scientific workflows, Coefficient Bio's researchers bring domain depth in building biology-specific AI models, the kind of capability associated with the emerging class of biological foundation models and "virtual cell" research. Several commentators framed the deal as part of a wider race among frontier-AI labs to secure proprietary biological data and specialized scientific talent, on the view that progress in AI for drug discovery depends as much on domain expertise and data as on raw model scale. Anthropic itself has long argued, including in writing by its leadership, that accelerating biology and health is among the most valuable potential applications of advanced AI, which makes a life-sciences talent acquisition consistent with its stated public-benefit mission.
For Anthropic, Coefficient Bio signaled that the company intends to compete not only on general-purpose generative AI but also in specialized scientific domains where rivals such as Google DeepMind, through protein-structure work like AlphaFold, and OpenAI have invested. Bringing in researchers from Genentech's Prescient Design gave Anthropic credibility and capability in computational biology that a horizontal model provider would otherwise lack.
For the broader market, the deal became a talking point well beyond its size. The combination of a roughly eight-month-old, sub-ten-person startup, a reported $400 million all-stock valuation, and a five-figure-percentage investor IRR crystallized a pattern that observers had been describing throughout 2025 and into 2026: frontier labs paying large sums to absorb tiny, elite research teams before those teams ever ship a product. Whether Coefficient Bio's specific technology proves durable inside Anthropic remains to be seen, since much of what is publicly known rests on a single round of reporting and an investor letter rather than on detailed technical disclosure. What is clear is that the acquisition marked a concrete, well-covered step in the convergence of frontier AI and biotechnology.
| Item | Detail |
|---|---|
| Company | Coefficient Bio |
| Type | Stealth-stage biotech / AI startup |
| Acquirer | Anthropic |
| Reported deal value | Just over $400 million, all stock (reported by The Information and Eric Newcomer; not officially confirmed by Anthropic) |
| Deal disclosed | April 3, 2026 |
| Co-founders | Samuel Stanton and Nathan C. Frey |
| Prior employer | Genentech, Prescient Design (computational drug discovery) |
| Founded | Approximately mid-2025 (about eight months before the deal) |
| Team size | Fewer than 10 employees |
| Focus | AI models for biological research: drug R&D planning, clinical/regulatory strategy, identifying drug candidates |
| Lead investor | Dimension (reported ~50% stake; reported 38,513% IRR) |
| Joins | Anthropic's healthcare and life-sciences group (led by Eric Kauderer-Abrams) |
| Strategic context | Extends Claude for Life Sciences (Oct 2025) and Claude for Healthcare (Jan 2026) |