Claude's AI for Science program adds 20 research partners
Anthropic just expanded AI for Science beyond individual researchers to 20 foundations, institutions, and research projects, from the Sloan Foundation to FlyWire at Princeton. Learn how working scientists are actually prompting Claude today: to generate hypotheses, design experiments, analyze data, and write up findings for non-expert readers.
Anthropic's AI for Science program just got a lot bigger. Started in May 2025 to give free Claude credits to researchers, it's now expanding to 20 scientific foundations, institutions, and research projects worldwide, announced October 28, 2026.
The partners fall into three groups. Funders who will redistribute Claude credits to researchers they fund: the Alfred P. Sloan Foundation, the Simons Foundation, Schmidt Sciences, the Open Source for Science Fund (Renaissance Philanthropy), the US National Science Foundation, and CAS (a division of the American Chemical Society).
Institutions using Claude credits internally: the Space Telescope Science Institute (which runs science operations for Hubble, James Webb, and Roman), the Wellcome Sanger Institute, the Francis Crick Institute, Imperial College London's Department of Life Sciences, EMBL, and the Gladstone Institutes.
Research projects getting direct credits: FlyWire (Princeton, neuroscience), LEAP (Columbia, Earth system science), AstroPath (Johns Hopkins, cancer imaging), gnomAD (Broad Institute, genomics), Tau Ceti (a mathematics project formalizing proofs in Lean), and the Centre for Population Genomics (Garvan Institute and Murdoch Children's Research Institute).
Anthropic also highlighted two deeper collaborations: with HHMI Janelia Research Campus, it co-developed the Model Hardware Standard (MHS), a specification that lets AI agents safely operate lab instruments (currently in research preview). With the Allen Institute, Claude-based agent teams now analyze brain, cell, and immune-system datasets, cutting work that took months down to hours.
According to Anthropic, researchers across these partnerships are using Claude for four things: generating hypotheses, designing experiments, analyzing complex data quickly, and communicating findings. That's the actual skill worth practicing, whether or not you're one of these 20 partners.
Notice the order in that list: hypotheses, then experiment design, then analysis, then communication. Skipping straight to "just tell me the answer" throws away the part of the conversation where Claude can surface alternatives you hadn't considered, or flag a control you forgot. Treat Claude as a collaborator you walk through your reasoning step by step, not an oracle you query once.
The last mile of a research project is explaining it to someone who isn't in your field: a grant officer, a journal editor, a lab director reviewing budgets. Many of these projects (gnomAD, FlyWire, Tau Ceti) are explicitly building open datasets, open-source tools, and open standards meant for exactly that kind of outside audience. The prompt you write for that handoff matters as much as the analysis itself.
AI for Science now spans 20 funders, institutions, and research projects giving researchers free Claude credits. The underlying workflow Anthropic describes, hypotheses, experiment design, data analysis, and communicating results, is one you can apply to any research or analysis task today: build up context with Claude step by step instead of asking for a single final answer.