NVIDIA pledges $250M compute for UK AI safety

What happened: NVIDIA committed a five-year compute donation valued at $250 million, or roughly $50 million a year, to advance UK research on agentic AI safety and security. The majority goes to the wider UK research community, with a portion supporting the BOLD lab at Oxford, SOFAIR at UCL and Imperial College London. The compute is delivered through UK-based NVIDIA Cloud Partners on infrastructure hosted in the country, and the work covers secure agents, open-model security, real-world threat mitigations, secure sandboxes, reproducible research and safety benchmarks, drawing on NVIDIA’s OpenShell and its Sentry reference design.

Context: The announcement carries quotes from NVIDIA chief executive Jensen Huang, UK AI minister Kanishka Narayan and Business Secretary Jonathan Reynolds, and builds on the UK’s national Isambard-AI system at Bristol. It follows NVIDIA’s roughly $1 billion commitment to US science earlier in October, extending the same donate-compute playbook to Britain.

Why it matters: Serious safety research increasingly needs frontier-scale compute, and donating it lets NVIDIA shape which labs and methods get studied. It also frames the UK as a hub for agentic safety work at a moment when governments are deciding how to test and secure autonomous systems.

(Original source: NVIDIA)