Until now AI for Science had one face: a researcher with a high-impact project and free Claude credits. Now the list has organizations too.
Fact
The facts: on 28 October 2026, Anthropic announced an expansion of AI for Science. In the company's words the program started in May 2025 and gives “free Claude credits to researchers working on high-impact scientific projects”, and it is now “highlighting 20 scientific foundations, institutions, and research groups”. They include the funders Alfred P. Sloan Foundation, Simons Foundation, Schmidt Sciences, the Open Source for Science Fund of Renaissance Philanthropy, the US National Science Foundation and CAS; the institutes Space Telescope Science Institute, Wellcome Sanger Institute, Francis Crick Institute, EMBL, Gladstone Institutes and Imperial College London; and the projects FlyWire, LEAP, AstroPath, gnomAD, Tau Ceti and Centre for Population Genomics. Separately there is the Model Hardware Standard, developed with HHMI Janelia: a specification that “lets AI agents safely operate lab instruments”. It is in research preview. The page does not state the value of the credits. Source: Anthropic, 28 October 2026.
No money is counted on the page: no sum for the credits, no terms per partner. So the list shows only a direction: besides individual scientists, Anthropic now names the ones who pay for science and the ones who do it every day.
The second part is smaller in size and heavier in meaning. With the Model Hardware Standard the agent leaves the screen. Articles and databases you can read over and over, but an instrument in a lab is something else.
On a screen, a mistake gets erased. On a lab bench, not always.
Watch the preview, not the list. If the standard leaves it and a lab with real instruments says it uses it, then it is news. Until then it is a specification.
Official primary sources
Anthropic - AI for Science expands to foundations and research institutions, 28.10.2026The visual is generated code-art, no third-party images.