According to reports, during an interview in London, NVIDIA CEO Jensen Huang stated that under hypothetical scenarios, NVIDIA might cease supplying chips to clients developing unsafe AI software, but would first communicate with the clients before taking action. He also emphasized that NVIDIA provides general-purpose technology and thus has limited ability to judge how clients utilize the technology.

NVIDIA further committed to providing computing resources worth a total of $250 million over the next five years to support AI safety research in the UK, with annual value amounting to approximately $50 million. The focus areas include agent safety testing, supporting initiatives such as the BOLD project led by the University of Oxford, as well as research efforts at Imperial College London and University College London. Huang stressed the need to accelerate the integration of safety constraints and monitoring technologies into AI systems, promoting the synchronized advancement of AI capabilities and security.