Summary
NVIDIA has announced a five-year compute donation valued at $250 million for AI safety and security research in the UK. The commitment provides for $50 million in compute access annually, with most intended for the wider research community.
NVIDIA announced on October 13, 2026, a five-year compute donation valued at $250 million for AI safety and security research in the United Kingdom. The company says the commitment will provide $50 million in compute access annually. Most of that access is intended for the wider UK research community, with a proportion supporting the British Open-Ended Learning and Discovery Lab (BOLD), the Science of Fundamental AI Research Lab (SOFAIR) and Imperial College London.
The contribution is compute access rather than a stated cash grant. NVIDIA says it will fund access through UK-based NVIDIA Cloud Partners, with the infrastructure hosted in the UK. The initiative is intended to support research into safeguards for AI systems that can act, adapt or interact with people and the physical world.
Research spans AI agents and robotics
The programme brings together several areas of research. BOLD, led by the University of Oxford, is developing approaches for AI systems to learn and adapt while operating alongside people. SOFAIR, led by UCL, is exploring open-source AI architectures and ways to make AI development and deployment more accessible. Imperial College London’s work will include safety engineering and robotics.
The announcement also identifies research priorities including secure agents, security for open models and ways to mitigate real-world threats. NVIDIA says the work will support secure sandboxes, reproducible research and safety-assessment benchmarks. Some research is expected to use NVIDIA OpenShell, described as open, secure runtime software, and the NVIDIA Sentry reference design. The company says findings are intended to inform safeguards across the wider computing ecosystem.
These efforts address different parts of the safety problem: how AI systems are built, how they are tested in controlled settings, and how their risks are assessed when they are used. In robotics and other physical AI, safety engineering also has to account for systems that perceive and act in the real world, not only produce digital outputs.
Delivery arrangements are still being set
The compute will be delivered through UK-based cloud partners, but NVIDIA says provider selection, technical architecture and commercial terms are still to be agreed. The announcement therefore sets out the programme’s intended scale and research focus, while leaving those implementation arrangements for further agreement.
NVIDIA describes the commitment as building on its work with the UK research community, including the University of Bristol’s Isambard-AI supercomputer, powered by NVIDIA Grace Hopper Superchips. The new initiative is specifically dedicated to agentic AI safety and security research, and its stated aim is to expand researchers’ access to the computing resources needed for that work.