Nvidia, the US semiconductor giant valued at $5.5tn (£4.1tn), is donating $250m worth of computing power to the UK to support artificial intelligence safety research.
The contribution from the world’s most valuable company comes as the UK Government confirmed that its AI Security Institute (AISI), which evaluates advanced AI models, will have access to cutting-edge systems developed by companies including OpenAI, Anthropic and Google before their public release.
Nvidia said the five-year commitment would “support agentic safety and security research across the UK”.
Concerns about AI agents, sophisticated computer programmes capable of independently planning and carrying out sequences of tasks, have increased since the summer. OpenAI confirmed that a “swarm” of AI agents had attacked technology startup Hugging Face, raising further questions about the security risks associated with increasingly autonomous systems.
Although Nvidia did not specifically mention the AISI in its announcement, it said most of the funding would “support the wider UK AI safety and security research community”. Some of the computing resources will go to the British Open-Ended Learning and Discovery Lab, the Science of Fundamental AI Research Lab, operated by the universities of Edinburgh, Cambridge, Oxford and University College London, and Imperial College London.
The computing resources will be supplied through UK-based data centres equipped with Nvidia chips, giving researchers access to the processing capacity needed to evaluate advanced AI systems and develop safety measures.
“The UK brings world-class scientific rigour to this essential engineering challenge,” said Jensen Huang, Nvidia’s chief executive. “With this $250m commitment, Nvidia will put accelerated computing in the hands of UK researchers to develop, test and strengthen safeguards for increasingly powerful AI.”
Kanishka Narayan, the UK’s AI minister, described the announcement as a “vote of confidence in our world-leading strength and expertise in AI research”.
On Monday, Narayan confirmed that the AISI was testing AI models developed in the US before their public release, following reports that the White House had restricted the institute’s access to certain systems.
The minister also warned that the UK could face AI-driven cyberattacks similar to the incident involving OpenAI agents and Australia’s healthcare system earlier this year.
“It seems likely, if not inevitable, that our own cyber defences will be breached,” said Narayan.
Anthropic, the developer of the Claude chatbot, raised further concerns about AI safety on Monday after publishing a blog post warning that its models could eventually become self-replicating. The company said such systems might find ways to infiltrate internet-connected devices, including smart ovens and lawnmowers.
Under this scenario, AI models could potentially operate without relying on conventional data centres, instead using a network of interconnected devices known as a “botnet”, Anthropic said. The company warned that preparations would be needed for a future in which such networks could persist.
“A replicator like this would be challenging to shut down since there’s no centralised point to shut off,” Anthropic’s researchers wrote.
