Nvidia Weighs Chip Cuts as It Backs UK AI Safety Research

Nvidia CEO Jensen Huang said the company could stop supplying chips to customers developing unsafe AI, but would speak with them before taking action. He added that Nvidia has limited ability to judge how customers use its general-purpose technology. Nvidia controls about 90% of the AI training chip market, giving its supply decisions significant industry influence, though Huang described any cutoff as conditional and case-specific. Nvidia also pledged $250 million in computing resources over five years for AI agent safety and cybersecurity research in the UK—equivalent to $50 million a year. Most resources will support the wider UK research community, with additional capacity for projects led by Oxford University, University College London and Imperial College London. Providers and commercial terms are not yet finalised. Research will focus on safer AI agents, open-source model security and real-world threat prevention. Huang called safety concerns growing pains and supported using AI to monitor other AI systems. The announcement broadens Nvidia’s role in AI safety, but does not directly change chip supply or its near-term business outlook. It has no clear direct effect on cryptocurrency prices.