NVIDIA donates $250 million to AI safety research in UK, Imperial amongst benefiters

NVIDIA, leader in chipmaking and the world’s most valuable company, announced 15 October a donation of $250 million in computing power towards research in AI safety. 

The donation is in the form of $50 million of computing power for the next five years, accessed through U.K.-based NVIDIA Cloud Partners ecosystem, as they aim to “support agentic safety and security research across the UK”, according to Nvidia’s press release. As increasingly powerful AI models are developed, NVIDIA commits to build and optimise safeguards: “The U.K. brings world-class scientific rigor to this essential engineering challenge”, says NVIDIA CEO and founder Jensen Huang. 

Benefiters of this donation include major UK AI research initiative British Open-Ended Learning and Discovery Lab (BOLD), founded by Imperial, Oxford, and UCL, the Science of Fundamental AI Research Lab, and Imperial College London. The College has not provided any details of the projects or departments that will benefit from the computing power. 

This initiative comes as the UK’s AI Minister announced a set of measures for AI defence last week at an Imperial event. The UK’s AI Security Institute (AISI) also recently confirmed they will conduct pre-release safety tests of cutting-edge AI models with leading developers such as OpenAI. 

The rise of increasingly independent agentic models, and a need to slow development to prevent loss of human control, has been warned by industry leaders themselves. Anthropic released a report on Monday that describes models could become self-replicating, which could be “challenging to shut down, since there’s no centralised point to shut off.” 

Perhaps make this weekend's movie watch anything but Terminator.