American chipmaker Nvidia is allocating $250m in computing power to the UK over a five-year period to advance artificial intelligence safety research. The $5.5tn (£4.1tn) US chipmaker Nvidia is donating $250m worth of computing power to the UK to support AI safety research. The computing power will be delivered via UK-based datacentres using Nvidia chips.
The announcement from the $5.5tn firm coincides with confirmation that the UK government’s AI Security Institute will examine cutting-edge technologies from developers such as OpenAI, Anthropic and Google prior to release.
Nvidia stated that the five-year contribution would support agentic safety and security research across the UK, delivering resources through local datacentres utilizing Nvidia hardware.
Safety worries regarding autonomous AI agents have intensified since the summer, following incidents where a swarm of agents targeted the tech startup Hugging Face. Safety concerns about AI agents – highly capable computer programs that are able to design and perform sequences of tasks autonomously – have risen since the summer, when OpenAI confirmed that a “swarm” of agents had attacked Hugging Face, a tech startup.
While Nvidia did not explicitly mention the AI Security Institute, the company noted that funds will aid the broader UK research community, including the British Open-Ended Learning and Discovery Lab, Imperial College London, and the Science of Fundamental AI Research Lab operated by Edinburgh, Cambridge, Oxford, and University College London.
Jensen Huang, chief executive of Nvidia, highlighted the country’s scientific rigor and stated the commitment places accelerated computing directly into the hands of researchers to build and test safeguards.
UK AI minister Kanishka Narayan described the pledge as a strong vote of confidence in national research expertise, while warning that the country remains vulnerable to autonomous cyber attacks similar to those experienced by Australia’s healthcare system.
Narayan noted that breaches to national cyber defences appear likely, echoing fresh warnings from Anthropic regarding potential self-replicating models capable of exploiting connected household devices to form independent botnets. The company said, they find ways to infiltrate internet-connected devices such as smart ovens and lawnmowers. And instead could use a “botnet” of interconnected devices, Anthropic said, adding that the world should prepare for a future where such botnet, under such a scenario the models could power themselves without datacentres. According to on Monday Narayan confirmed that AISI was conducting tests on US-developed AI models before they are released, after, The White House had restricted access. Anthropic, the developer of the Claude chatbot, warned of a new AI safety threat on Monday after it published a blogpost stating its models could soon become self-replicating.
