Anthropic Hands Claude Credits to Sloan, Simons and NSF for Science

Anthropic's AI for Science program has a new way of handing out credits. Until now, researchers applied for free Claude API credits themselves. Now Anthropic is giving credits to science funders, which will distribute them to the researchers they already support. The expansion names 20 partners at once.

The most significant group is the funders and societies: the Alfred P. Sloan Foundation, founded in 1934; the Simons Foundation, known for mathematics and basic science and home to the Flatiron Institute; Schmidt Sciences, founded in 2024 by Eric and Wendy Schmidt; the Open Source for Science Fund at Renaissance Philanthropy; CAS, a division of the American Chemical Society; and the US National Science Foundation, a roughly $9 billion agency that accounts for about a quarter of federal support for basic research at US colleges and universities.

From who applies to who chooses

It looks like a channel change, but it's a smart one. An AI company has neither the expertise nor the scale to judge which biology or astronomy project deserves support. Finding and vetting high-impact research is these foundations' core job. Letting them distribute credits borrows an existing review system, so Claude reaches people doing frontier work faster and more accurately.

The list also includes institutions that will use credits directly: the Space Telescope Science Institute, which runs science operations for Hubble, Webb and Roman, plus the Wellcome Sanger Institute, the Francis Crick Institute, Imperial College London, EMBL-EBI and the Gladstone Institutes. Some projects got direct grants, such as Princeton's FlyWire fruit-fly connectome and the Broad Institute's gnomAD human variation database. HHMI's Janelia Research Campus co-developed a hardware standard for AI agents to operate lab instruments, and the Allen Institute is building teams of AI agents to analyze open brain, cell and immune datasets.

Science is a new front for AI companies

Since launching AI for Science in May 2025, Anthropic has supported hundreds of projects across biology, medicine, physics, chemistry, climate science and mathematics, and added a discounted Team plan for scientists. The business logic is clear. Research is one of the few high-value settings that truly needs long-chain reasoning, long context and tool use, and it's the best stage for the larger claim that AI can speed up discovery.

OpenAI has a dedicated science team, and Google carries DeepMind's AlphaFold halo. Anthropic's angle is more infrastructural: rather than chase a Nobel-grade discovery itself, it wants thousands of researchers using Claude in daily work. Once a generation of scientists gets used to reading papers, writing analysis code and running data pipelines with Claude, that habit will outlast any benchmark lead.

Working with a federal agency like the NSF is also a form of endorsement. For an AI company increasingly focused on government and the public sector, the return on these credits may go well beyond the API bill.