By Amit Chowdhry
Publication Date: 2026-10-14 16:31:00
NVIDIA has announced a five-year compute commitment valued at $250 million to strengthen agentic AI safety and security research across the United Kingdom.
The initiative will provide $50 million in compute access annually to support researchers working on AI safety, security, agentic systems, and robotics. The majority of the commitment will support the broader U.K. AI safety and security research community, with a portion directed toward the British Open-Ended Learning and Discovery Lab, the Science of Fundamental AI Research Lab, and Imperial College London.
The program is intended to expand access to accelerated computing resources for researchers developing, testing, and evaluating safeguards for increasingly capable AI systems. Compute access will be provided through U.K.-based NVIDIA Cloud Partners, with infrastructure hosted domestically to support both research capacity and the broader U.K. compute ecosystem.
BOLD, led by the University of Oxford, is researching approaches that enable AI systems to learn, adapt, and operate alongside people. SOFAIR, led by UCL, is developing open-source AI architectures and methods intended to make AI development and deployment more accessible. Imperial College London will focus in part on safety engineering and robotics as physical AI systems become more capable.
Research supported through the initiative is expected to include secure AI agents, open-model security, and protections against real-world threats. Programs…