Overview
Coefficient Bio was a New York biotechnology AI startup founded in September 2025 and acquired by Anthropic in April 2026. Co-founder Samuel Stanton confirms both dates in his biography. The company worked on AI for biological research before its team moved into a larger AI laboratory.[1]
Its short history suggests that scientific operating knowledge can attract a buyer before a standalone business becomes publicly measurable. Journalist Eric Newcomer reported a purchase worth just over $400 million in stock, citing an investor letter. The consideration remains a reported figure; the underlying agreement is not public.[2]
Image 1 / 1
Founding Story
Stanton worked at Genentech's Prescient Design from 2022 to 2025, studying machine learning methods and antibody design. His biography records a doctorate from NYU's Center for Data Science and an operations research master's degree from Cornell.[1] Co-founder Nathan Frey describes leading a Genentech group that worked on antibody design and models for drug discovery. Their shared professional background connected model research with experimental biology.[3]
Fierce Biotech identified Aris Theologis as CEO and co-founder, Nathan Frey as CTO and co-founder, and Joyce Hong as another co-founder. Theologis previously worked at Evozyne and Paragon Biosciences; Hong came from Roivant Sciences.[4] The shorter founder lists in other accounts are incomplete; no public legal roster establishes completeness.
Newcomer reproduced Stanton's January recruiting claim: “It will change everything about how the industry learns and makes decisions.”[2] This is a statement of ambition.
Timeline
- September 2025: Stanton says he co-founded Coefficient Bio.[1]
- December 2025: Frey publicly links biotech's operating problems to the need for clearer goals and faster feedback.[5]
- April 2026: Anthropic acquires Coefficient Bio, according to Stanton's subsequent biography. Newcomer reports the transaction on April 3.[1][2]
- October 2026: Frey's website describes his current Anthropic role leading drug discovery and development research.[3]
What They Built
The company profile described its purpose as “Building for the future” and linked to coefficientbio.com.[6] Newcomer's reporting described AI models for biological research.[2] No verified product interface or pricing establishes how customers used it.
Frey's current description offers a clearer view of the continuing research direction. His Anthropic team aims to let one skilled biotech operator use multiple Claude instances to create and operate drug programs.[3] The description establishes today's research direction. A shipped Coefficient product remains unverified.
Market Position
Target Customers
The disclosed focus places Coefficient Bio in biological research and biotech operations. Named customers and purchasing criteria remain unknown.
Market Size
No defensible company-specific market estimate is public. An acquisition price cannot establish customer demand.
Competition
Anthropic already had a route into this field. It launched Claude for Life Sciences in October 2025, connecting its assistant to scientific tools and content.[7] In June 2026, it introduced Claude Science, a workbench connecting research tools and computing resources.[8] This creates a plausible fit between specialist scientific knowledge and a buyer with an existing product platform. Neither announcement assigns a specific feature to Coefficient Bio.
Business Model
Newcomer reported that Dimension backed the startup and owned roughly half of it. The reported acquisition consideration was stock.[2] Operating revenue, burn, retention, and employees' proceeds remain unknown. Undisclosed revenue does not establish zero revenue. Stock consideration does not establish an immediate cash payout.
Post-Mortem
In a December 11, 2025 essay, Nathan Frey wrote that “we need faster feedback loops.”[5]
The plausible mechanism is the transfer of scientific operating knowledge before a public commercial launch. The founders' experience connects model research, experimental design, and drug-program operations. Those capabilities could be valuable inside Anthropic's existing scientific products even without a mature independent sales operation.[1][3]
This interpretation remains untested. The public record cannot separate the value Anthropic assigned to personnel, software, research assets, or future work. A pure team-hire label would exceed the disclosed terms.
Proprietary technology could instead explain the price. No disclosed benchmark, asset inventory, or contract lets readers test that explanation. The acquisition's eventual financial return remains open.
Key Lessons
- Coefficient Bio combined research and biotech operating backgrounds. Its story suggests that scientific AI businesses can draw value from how experts organize work, alongside model development.
- The acquisition arrived before public commercial evidence could establish repeatable demand. Buyers and founders should evaluate strategic fit separately from proof of an independent business.
- Frey's current research description provides a useful sequel. Future evidence should be judged through observed scientific results and product use, without attributing every Anthropic life-sciences launch to the acquired team.
Sources
- Samuel Stanton's biography.
- Newcomer's acquisition report, April 3, 2026. The public article describes an investor letter; the letter requires a subscription.
- Nathan Frey's personal website, checked October 23, 2026.
- Fierce Biotech's acquisition coverage.
- Frey on biotech's operating problems, December 11, 2025.
- Coefficient Bio's company profile.
- Claude for Life Sciences, October 20, 2025.
- Claude Science, June 30, 2026.