Anthropic has committed $150 million over three years to the Genesis Mission, the US federal programme for applying AI to scientific research, announced at a White House science summit in Washington.
In practice the commitment is access: Claude, Claude Code and API credits for what the company calls several hundred research projects, plus training for scientists and closer work with national laboratories on priorities including fusion energy and quantum computing. Genesis Mission spans more than fifteen agencies, of which the announcement names NASA, the National Institutes of Health and the National Science Foundation. Anthropic has worked with the Department of Energy on this since December.
What the announcement does not do is break the figure down. There is no split between cash and product access, no milestones, and no statement of whether work already delivered counts towards the total. That earlier work includes 10,000 free and discounted seats for academic scientists and a science workbench launched this year, so the question is not academic.
One item in the list is worth watching separately: a research preview of what Anthropic calls a Model Hardware Standard, a shared specification for AI agents to operate laboratory instruments safely. That sits alongside the physical-hardware conditions the company added to its usage policy the same week, and it is the more concrete of the two.
Source: Anthropic, 8 October 2026.