Anthropic has announced it will run its own drug discovery programmes to treat diseases that traditional pharmaceutical companies avoid because they are not profitable.
The new initiative focuses on early, preclinical-stage work for neglected conditions. Anthropic states the move supports its nonprofit mission while allowing it to build better AI models through direct industry experience. The news came during a presentation for its new science tool, Claude Science.
During the event, a researcher at the University of California, San Francisco used Claude Science to identify viral contamination in minutes. That same error had gone unnoticed by the team for a year. Anthropic also noted the tool analysed 100 rare genetic diseases in under an hour and flagged 32 candidates for computational screening.
Small gains, massive impact
Novartis CEO Vas Narasimhan explained that moving a finished drug candidate from development to approval currently takes about twelve years. He divided these delays into three categories: information latency, operational latency, and biological latency.
New tools could sharply cut the first two categories, which account for roughly 40 percent of total development time. Biological latency, the time needed for animal testing, cell models, and human clinical trials, will not shrink much. That could bring development timelines down to seven or eight years.
Narasimhan sees room to double success rates from 8 to 16 percent. Better safety predictions and optimized molecular properties could help, though the effect of improved patient selection remains unclear. The biggest challenge is still figuring out whether a drug target is biologically the right one for a given disease.
Even these seemingly modest gains would be huge when scaled across major pharma. Together, the big companies spend $150 to $200 billion a year on R&D and have produced only 800 to 1,000 drugs in 120 years. More diseases could be treated, and drug targets that were previously considered unreachable could become viable.
AI across healthcare
Other AI companies are also pushing into medicine. Deepmind CEO Demis Hassabis co-founded Isomorphic Labs with Alphabet to apply AI directly to drug discovery. Google Deepmind’s protein structure prediction tool AlphaFold remains one of the most prominent examples of AI in biology, and its co-developer John Jumper recently left for Anthropic.
On the clinical side, Google DeepMind introduced an AI Co-Clinician in 2026 built around triadic care. AI agents support patients throughout treatment while the physician retains clinical authority.
OpenAI has also been moving into healthcare over the past few years. In early 2026, it launched ChatGPT Health, a dedicated health section within ChatGPT that lets users connect medical records, Apple Health data, and wellness apps.
Independent experts still urge caution, especially when AI is used in clinical settings for diagnoses, treatment plans, and direct patient care. Catherine Pope of the University of Oxford called the results so far “a piece removed from the messy, complex, human world of everyday healthcare.”
What it means
For researchers and patients, this shift means more focus on diseases that have been ignored by the market. It also suggests a future where development cycles are shorter and more candidates reach the lab, even if the biological hurdles remain difficult.
