Asked in London by Bloomberg's Emily Chang if he'd halt shipments to a customer developing dangerous AI, Nvidia CEO Jensen Huang called it a hypothetical and answered yes. He added that he would start with a conversation, saying Nvidia does not manage other firms, applies only so much judgment over how its technology is used, and that its products are general purpose.
While on a trip through Europe in the run-up to a Berlin company gathering, Huang characterized much of the worry about AI as the natural aches of a rapidly scaling field. "Many of the young companies are going through growing pains, and they're growing incredibly fast," he said, adding that some of the concern is "self-imposed" and that the industry is doing a "horrible job" explaining AI's benefits. He said he believes the software can be kept in check and rolled out safely.
As part of its safety commitments, Nvidia will provide $250 million over five years to support UK work on AI safeguards. Rather than cash grants, the backing will come as access to cutting edge computing time. Universities in the UK, including the University of Oxford and University College London, will use it to study controls on AI agents and related guardrails.
Huang also argued that AI can help police AI. "We should accelerate the integration of this containment technology into AI systems," he said. "We should accelerate the development of monitoring technologies, and AI is that monitor. So we should accelerate AI technology."
When technologies and regulations evolve quickly, staying informed matters, so sign up for the free Market Briefs to catch up daily
Nvidia is the leading seller of AI accelerators, the processors that power the training and operation of AI models. That central role puts it in the crosshairs of the debate over whether AI could become too powerful and slip out of human control. Some leaders, including Anthropic PBC CEO Dario Amodei, have floated a coordinated slowdown across the sector to tackle those risks.
On a trading day when Nvidia and other chip names were dragging the US market lower, Huang reiterated that talk of an AI bubble is overdone. At a London event, he said, "Eventually supply and demand will harmonize," adding, "but not today."
The takeaway for regular investors: Nvidia is signaling it will engage with customers on safety and, if necessary, walk away, while pouring compute time into academic research that could make AI safer to deploy. It is also telling you that the appetite for AI gear still exceeds what can be built, which often supports pricing power. If you follow AI's ripple effects on everything from cloud spending to university research, this is one more sign the buildout is still early.
With complex developments across tech and finance, get clear context each morning by subscribing to our free Market Briefs newsletter
