Nvidia said on 13 October that it will donate compute valued at $250 million over five years, about $50 million a year, to British research on the safety and security of agentic AI. Most of it goes to the wider UK research community, with a share earmarked for three groups: BOLD, an open-ended learning lab led by Oxford; SOFAIR, a fundamental AI research lab led by UCL; and Imperial College London, which is working on safety engineering and robotics.
The stated agenda is secure agents, open-model security and real-world threat mitigation, including secure sandboxes, reproducible research and safety assessment benchmarks. Some of the work is to use Nvidia’s own OpenShell and Sentry designs.
Two caveats sit inside the announcement itself. The headline figure is a value placed on compute access rather than cash, and Nvidia says the choice of UK cloud partner, the technical architecture and the commercial terms are all still to be agreed, so what the pledge actually buys is not yet fixed.
The UK AI minister, Kanishka Narayan, called it a vote of confidence in British frontier safety research. It is also a dominant hardware vendor funding, and supplying, the labs that will study the safety of systems built on its chips. The press release is on Nvidia’s newsroom.