Nvidia CEO says chip supply could be halted for unsafe AI developers
Nvidia CEO Jensen Huang said the company would consider halting chip supplies to customers building unsafe AI, but only after speaking with them directly. He made the comments in a Bloomberg interview on October 14, 2026, and did not identify any specific customer. Nvidia also committed $250 million over five years to AI safety research at UK institutions including Oxford and University College London.
The comments emerged during an Oct. 14, 2026, Bloomberg interview in London with Emily Chang. Asked how Nvidia might react if its hardware were used for risky AI, he said a supply halt was conceivable but would follow direct talks with the buyer. He named no company.
Nvidia separately promised $250 million across five years for AI safety research at British universities, including Oxford and University College London. The funding targets research into keeping powerful systems under control and tracking their behavior. Nvidia supplies about nine in ten AI training chips, and its past buyer checks have turned away more than half of would-be customers in some areas.
Nvidia's stance could affect AI developers, cloud providers, and researchers who depend on its chips, since supply decisions may shape which groups can train advanced models. The UK safety funding may expand monitoring and containment research, potentially informing public safeguards. Yet because any cutoff would be discretionary and discussed privately, oversight may remain uneven. Regulators, customers, and the public could face uncertainty about how consistently safety concerns influence access to critical computing hardware.