Anthropic buys Mithril, a one-person biotech run on Claude

Anthropic disclosed on October 27 that in July it bought Mithril Medicine, a Cambridge, Massachusetts, biotech with one full-time employee. That employee, founder Cody Tranbarger, used Claude to handle contracts, logistics and experiments across 11 outside labs. If you’re reading the deal as proof that Claude can run a drug program, the operations are the part with a record behind them. The mouse results behind Anthropic’s idea-to-results-in-18-months claim rest on only seven treated mice and four controls, which weren’t randomized or tested at the same time, and the report has not yet been peer reviewed.

Mithril’s work on two rare forms of muscular dystrophy became Anthropic’s first internal drug-development programs, the company said in its Anthropic update on drug development. The update also covers partnerships with six drug developers and says all of the work is still mostly preclinical. Tranbarger is continuing the research at Anthropic and plans to start several new programs.

One employee, 11 labs

Eight paid consultants and advisers worked behind the single employee, R&D World reports, covering biology, antibody engineering and research planning. Mithril’s team designed and oversaw the experiments. Contract labs made the antibodies, measured their binding, ran cell-based assays and did the mouse studies. Tranbarger came from the finance and operations side of biotech. He finished Harvard’s joint MS/MBA biotechnology program in 2025, worked at the gene-editing company Prime Medicine and the investment firm Casdin Capital, and then became an entrepreneur-in-residence at Atlas Venture.

According to R&D World’s account of the technical report, the team picked one week as its best recent example. In that week Claude summarized a 341-page production report covering 113 antibody constructs in under eight minutes. Claude recommended measuring how quickly all 113 bound to their targets and let go. A team member overruled that, and Claude rewrote the plan and the drafts that went with it. Claude also made mistakes. Of six errors the team examined, at least three reached outside labs in emails that people had edited and sent.

What the mice show

Mithril’s first results are for LAMA2-related congenital muscular dystrophy. In this disease, mutations in the gene for the laminin-α2 protein weaken the links that anchor muscle fibers to the tissue around them. Mithril built on earlier research into repairing those links and developed bispecific antibodies, which bind two targets at once to form a molecular bridge, R&D World writes. The report describes them as early research molecules.

At about six weeks of age, seven mice given one antibody format had roughly twice the grip strength and peak triceps force of four mice given a control solution. But the mice were assigned to groups by litter without randomization. The antibody group was also tested 126 to 161 days after the control group, and the colony was relocated in between. R&D World notes that these differences make it harder to credit the gains to the treatment, and the report itself calls for a concurrent, blinded replication.

The second small biotech this year

In April, Anthropic bought Coefficient Bio, a company about eight months old with fewer than 10 people and no publicly launched product, in an all-stock deal valued at approximately $400 million. That team joined the life sciences group to build agents for drug discovery, clinical trial planning and regulatory work. In September came a Bay Area wet lab where Claude directs the robotic units that run experiments. Seen next to those moves, Mithril looks like part of an effort to build drug programs Anthropic runs itself, alongside its partner work. In its own programs, Anthropic says, Claude handles much of the scientific and operational work.

The Column Group, a biotech investor and one of the six partners, is testing the same small-team model. It is setting up Pleiades, a new company meant to use Claude to run more drug programs, and more complex ones, than a small team usually manages. Its first Claude skill, which assesses whether drug targets are worth pursuing, has been evaluated on 45 target-disease pairs. Mithril has already taken that setup as far as mouse data.

If Tranbarger’s drug candidates keep holding up, Anthropic expects a drug to be ready to start human clinical trials as soon as late 2028.