Anthropic has published early results from using Claude in real drug development programmes, reporting progress across cancer, food allergy, manufacturing and clinical planning work with six partner organisations, and announcing that it has begun drug discovery efforts of its own. The work remains mostly preclinical, Anthropic stresses, meaning the potential treatments are generally not yet being tested in people, but the company says it plans to scale the programme up substantially.
Six Partners, Six Different Bottlenecks
Anthropic’s argument is that drug development is slow not because of one hard problem, but because bottlenecks are spread across the whole process, from finding a target and designing a molecule to manufacturing it, drafting regulatory filings and running trials. Its life sciences team has therefore worked inside real programmes with BigHat Biosciences, Bristol Myers Squibb, Divergence Labs, Enveda, Generate Biomedicines and The Column Group, looking for where Claude helps most and where it still needs to improve.
At Generate Biomedicines, the focus is CAR T-cell therapy, a cancer treatment for blood cancers which Generate wants to reengineer for solid tumours. Six Generate scientists used Claude Code to propose new designs for one part of the therapy’s genetic instructions. One design was about 70 percent shorter than the standard version, and in laboratory tests it allowed immune cells to make roughly ten times as many receptors, killed tumour cells more efficiently and kept the immune cells active for longer.
Divergence Labs is working on a food allergy treatment that would detach IgE, the molecule behind allergic reactions, from immune cells entirely, rather than only neutralising the IgE circulating in blood, as standard medications do. Claude generated roughly 35,000 antibody designs in about a month. Early tests found three designs that could bind IgE and 20 that could dislodge it, with three appearing to dislodge IgE more than the standard medication. The next step is to combine the two arms into a single drug.
From Lab Bench To Paperwork
BigHat Biosciences pairs AI with an automated lab to improve drug leads, which must bind their target tightly, stay stable through manufacturing and storage, and avoid clumping. Connected to BigHat’s lab data system, Claude proposed changes to two antibody fragments, and the lab made and tested about 1,010 variants in about ten weeks. The best versions bound more than ten times more tightly than the originals, the cancer-targeting lead stayed in solution at a temperature 18 degrees Celsius higher, and the lead designed for longer-lasting drugs clumped more than 100 times more slowly.
Enveda’s case shows a quieter use of AI: paperwork. Manufacturing chemists there spent up to 75 percent of their time on documentation, so Enveda connected all of its manufacturing documents to Claude and let it flag possible issues. Over tests on 1,061 pages, Claude caught all 11 issues human reviewers had spotted and flagged seven they had missed, cutting human review time by 89 percent, from over 25 hours to less than three. Separately, Claude wrote first drafts of two regulatory documents, one in four hours compared with about three months by hand.
Bristol Myers Squibb is the most recent partner, using Claude at the clinical trial stage to analyse proprietary data, create a clinical study plan and weigh a study’s chance of success. Anthropic and Bristol Myers Squibb plan to expand that work in 2027.
Anthropic’s Own Drug Programmes
Beyond its partnerships, Anthropic has acquired Mithril Medicine, a one-person company developing treatments for two rare forms of muscular dystrophy. Its founder, Cody Tranbarger, used Claude to organise contracts, logistics and experiments with 11 external labs. Mithril took a possible treatment from an idea to promising results in mice in less than 18 months, and Tranbarger is continuing the research at Anthropic. If those candidates hold up, Anthropic expects a drug to be ready to start clinical trials in humans as soon as late 2028, a proposed timeline, not a clinical approval. The Column Group, meanwhile, is establishing Pleiades, a new company that will use Claude to run many more drug programmes than is usual for a small team.
To widen access, Anthropic has rolled out Claude Science, an AI workbench for researchers, and is providing Claude credits for life scientists. It has also created a Life Sciences Verification Program for trusted organisations, and opened a call for guest investigators.
Why It Matters
This matters because it tests AI in medicine where it counts: not in a benchmark, but inside programmes where a wrong molecule, a missed document error or a bad trial plan has real costs. The early results are encouraging, but they are vendor-reported, preclinical results that need independent replication before they become medicines, and Anthropic says so itself. If Claude can shorten design cycles, catch manufacturing errors and draft regulatory documents, it could compress a famously slow industry, much as analysts following AI agents most of the time shows how quickly professionals defer to AI output, which is why verified identity and accountability for AI agents matter in high-stakes work, and why AI systems that prove code correct raise the question of whether verification keeps pace with generation when the stakes are human health.
Background
Anthropic is the AI company behind Claude, and its life sciences team works on accelerating drug discovery. Developing a new drug typically takes years, with most candidates failing before they reach patients.
Source: Anthropic research post
0 Comments