Anthropic, the company behind the widely used Claude AI chatbot, is expanding beyond tasks such as writing emails and coding into healthcare, with a focus on developing new medicines. On October 27, the AI company announced its entry into the sector, acquiring Mithril Medicine, a biotech startup operated by a single person. It also unveiled joint research partnerships with six organisations, including BigHat Biosciences and Bristol Myers Squibb.
The one-person biotech operation
Mithril Medicine is run entirely by its founder, Cody Tranbarger. Using Claude as an AI assistant, Tranbarger managed contracts, logistics and experiments conducted across 11 external laboratories. The startup is working to develop treatments for two rare forms of muscular dystrophy.
With Claude handling much of the work, the company took a potential drug from the initial concept stage to successful testing in mice in less than 18 months. If the early findings prove promising in further research, human clinical trials could begin as early as late 2028.
Reducing months of work to hours
The announcement highlighted the considerable time savings AI can deliver to researchers. Biotechnology company Enveda used Claude to examine a lengthy, 1,061-page manufacturing document. The AI identified all 11 errors previously detected by human reviewers and uncovered seven additional problems that they had overlooked.
The technology reduced verification time by 89 per cent, bringing the task down from more than 25 hours to less than three. In another example, Claude prepared a complicated regulatory document in just four hours, compared with the approximately three months required to complete the same task manually.
Pharmaceutical major Bristol Myers Squibb also used Claude to assess plans for early-stage clinical trials. The company said that determining the likelihood of a medical trial succeeding typically takes several days, whereas Claude produced an initial assessment from a single prompt.
Developing more effective treatments
Claude is also assisting researchers with the design of potential medicines. BigHat Biosciences used Anthropic's AI to improve antibody fragments, producing designs that bound to their targets more than 10 times as effectively while also demonstrating substantial gains in stability.
Divergence Labs used the technology to generate approximately 35,000 antibody designs for food allergy treatments within a month. Meanwhile, Generate Biomedicines employed Claude in cell therapy development, producing proposed gene designs that were 70 per cent shorter than conventional models.
The future of AI in healthcare
To support the growing use of AI in scientific research, Anthropic has introduced Claude Science, a dedicated AI workspace for researchers, along with a Life Sciences Verification Programme.
However, the company stressed that these developments remain at the preclinical stage. The findings are based on laboratory experiments conducted by the participating companies and have not yet established that the treatments are safe or effective in humans.
Despite these limitations, the effort to combine AI with medicine is gathering momentum. On October 20, 2026, a week before Anthropic's announcement, rival OpenAI disclosed its acquisition of Orchestra Bio, signalling that leading technology companies view drug discovery as a major emerging field for artificial intelligence.
