What happened
Anthropic is inviting members of the life sciences community to apply as guest investigators. It wants proposals from principal investigators and their research groups at academic institutions, nonprofits and government labs.
The topics are broad: causal human biology, structural and computational molecular biology, medicinal chemistry, drug discovery and clinical science. The common thread is using Claude to chase real research questions.
Projects run 8–12 weeks. Guest investigators work directly with Anthropic researchers on a problem they care about and come out with the analyses, artifacts and tools that help their own work.
Anthropic says it will run its models and harnesses at no cost to the guests. The compute bill, in other words, is not your problem.
Why it matters
Plenty of labs have a question that a frontier model might help with and no budget to find out. This removes the cost barrier and adds hands-on help from the people who build the thing.
The terms are also notably researcher-friendly. Anthropic says guest investigators keep the rights to their data at every step. They also hold the rights to any independent follow-on research that grows out of the collaboration.
The output is meant to travel. Anthropic says it will work with the guests to publish results and open-source the accompanying data and code.
Now the caveat. This is a selective program, not a product launch. Anthropic is looking for applicants who can explain clearly why direct collaboration with its researchers would help them, so a vague "AI for biology" pitch will not cut it.
It is also a smart way to see where Claude actually earns its keep in the lab, with outside experts doing the stress testing. The machines get real-world biology homework. The humans get free compute. Fair trade, for now.
What to do with it
- Pitch a specific problem.Anthropic wants a clear case for why working directly with its researchers helps you, so name the bottleneck.
- Plan for 8–12 weeks.Scope a question that can yield analyses, artifacts or tools inside that window, not a five-year grant in disguise.
- Know your rights.You keep your data and any independent follow-on research, so say what you would open-source and what you would not.
- Check eligibility first.The call targets academic institutions, nonprofits and government labs, so confirm your group fits before drafting anything.
Ciao, sapiens. Apply early; the pipettes are not going to hold themselves.


