Computer chip maker Nvidia’s $250 million donation to UK AI security research will be partly distributed to Oxford AI labs over five years.
The donation by the world’s first $5 trillion company is a compute donation, meaning no money will be transferred. Instead, UK laboratories, universities, and government projects will be able to use Nvidia’s computing power up to that value.
Part of the donation is expected to be allocated to the British Open-Ended Learning and Discovery Lab (BOLD), a laboratory hosted and led by the University of Oxford in partnership with University College London (UCL) and Imperial College London.
BOLD’s founding director, Jakob Foerster, is an associate Professor in Engineering Science and a Fellow of St Anne’s College. He has emphasised that the UK cannot compete in AI by outspending big tech firms and urged focus on “new paradigms for learning, rather than only scaling existing methods.”
Nvidia has not confirmed which labs will receive various proportions of the donation.
According to a University announcement, BOLD plans to recruit 12 fellows and 18 PhD students across Oxford, UCL, and Imperial in its first phase. Nvidia’s donation will give students involved access to Nvidia’s compute for parts of their research.
The announcement comes as concerns over the security of AI systems grow. In July, OpenAI announced that an AI model had gone beyond the isolated test environment and broken into the systems of another major AI startup. That same month, Anthropic also announced that its AI agents had broken into the systems of outside organisations during safety testing.
It remains unclear whether Nvidia will control the direction of AI safety research or whether control over research will be fully devolved to labs and the government.
The donation will also go to the Science of Fundamental AI Research Lab, led by UCL in collaboration with Cambridge, Oxford, and Edinburgh. While Nvidia did not discuss the role the UK AI Security Institute (AISI) will play with respect to the donation in their statement, the £4.1tn firm pledged more broadly to “support the wider UK AI safety and security research community”. Nvidia CEO Jensen Huang said: “The UK brings world-class scientific rigour to this essential engineering challenge.”
The UK AI Security Institute were approached for comment.
