Anthropic Enters Drug Discovery With Claude Science
Anthropic unveiled Claude Science, a workspace for researchers, and launched its own drug discovery program. A researcher grant call runs until July 15.

- Anthropic unveiled Claude Science, a workspace for researchers with more than 60 tools and databases for genomics, proteomics and drug discovery.
- The company entered its own drug discovery program, first aimed at neglected diseases that large pharmaceutical firms tend to avoid.
- It opened a grant call: up to $30,000 (about 27,600 euros) in credits for up to 50 projects, with applications closing on July 15.
- Novo Nordisk and the Allen Institute are listed among the first users.
What happens when you hand scientists the same kind of tool that writes code for software engineers? Anthropic says you get Claude Science, a new workspace that assists research instead of software. It arrived alongside a surprise: the company is stepping into drug discovery itself.
This is not an announcement that can sit on the shelf. The researcher grant call has a firm deadline, applications close on July 15, so it is worth understanding exactly what is on the table before that date passes.
Claude Science is designed as the counterpart to the Claude Code environment, only for science rather than programming. It ships with more than 60 tools and databases covering genomics, proteomics and drug discovery. The system writes its own code, runs it on powerful computing clusters, and keeps results reproducible, which in science is half the job. The beta is available for now to Pro, Max, Team and Enterprise subscribers, on macOS and Linux.
Why is Anthropic getting into drugs when that is not its business?
This is the most interesting part of the story. Anthropic is not just selling a tool, it is also running its own drug discovery program. The first focus is neglected diseases, the ones large pharmaceutical companies usually skip because they do not pay off well enough. That is a concrete choice, not empty talk about good intentions.
At the launch, according to what the company shared, developers showed the system proposing new candidates for a treatment for phenylketonuria, a rare genetic disorder. One Harvard physicist rated Anthropic’s model as roughly at the level of a second-year graduate student when it comes to carrying out scientific projects. That is not a scientist who runs a lab alone, but it is a pair of hands that can do a lot of technical work.
Why now? Anthropic is heading toward a public listing and is looking for revenue that does not depend only on selling chat tokens. Life science is a logical field: large, expensive and slow, exactly where automating part of the work could pay off.
What the grant call means and who the first users are
The part with a deadline is the researcher grant call. Anthropic is offering up to $30,000, or about 27,600 euros, in credits for up to 50 research projects. Applications close on July 15, so if anyone at a university or institute is thinking about testing the tool on their own problem, that window is closing fast.
The list of first users shows this is more than startup enthusiasm. It includes Novo Nordisk, the Danish pharmaceutical company known for diabetes and obesity drugs, and the Allen Institute, an American research institute for brain and cell science. These are serious names, not demo accounts.
Anthropic is not alone in this race. Google and OpenAI are also building their own drug discovery tools, so a separate front is slowly forming among AI companies, this time on the lab floor rather than in writing assistants.
Our advice is to view these announcements soberly. A tool that speeds up part of the work is not the same as a medicine on the pharmacy shelf, and the path from candidate to approved therapy is measured in years and clinical trials. But the direction is clear: artificial intelligence is moving from the office into the laboratory.
Would you trust a drug whose first candidate was proposed by a computer, if it still went through the same clinical checks as before?


