NVIDIA Pledges $250M for UK Agentic Safety and Security Research
A five-year commitment valued at $250 million will support studies of AI agent safeguards, model security and robotics safety.
- By Chelsey Arries
- Oct 15, 2026
NVIDIA plans to provide $250 million in computing access over five years to support AI researchers in the United Kingdom studying agentic AI safety and security. The initiative will provide $50 million in access annually.
The company said most of the resources will serve the broader research community. A portion will support the University of Oxford-led British Open-Ended Learning and Discovery (BOLD) Lab, the University College London-led Science of Fundamental AI Research (SOFAIR) Lab and Imperial College London.
BOLD’s work spans AI systems that adapt alongside people; SOFAIR’s efforts include open-source architectures; and Imperial College London’s research focuses on safety engineering and robotics, according to NVIDIA.
Other research will examine agent security, security for open models and threat mitigation, NVIDIA said. The research will support secure testing environments, reproducible research and benchmarks for assessing AI safety, with some research using NVIDIA’s OpenShell software and the Sentry reference design.
NVIDIA will pay for access through NVIDIA Cloud Partners using infrastructure in the U.K. Providers, system designs and contract terms remain undecided. The company said findings are intended to help improve protections beyond its own technology.
About the Author
Chelsey Arries is assistant editor of Security Today, Campus Security Today, Occupational Health & Safety and Environmental Protection Online.