Jensen Huang tells London audience the US and China must eventually align on “potentially dangerous” AI

NVIDIA CEO Jensen Huang said the United States “absolutely must” work with China on advanced artificial intelligence that he described as potentially dangerous, warning that competing national interests between the two countries will have to be reconciled, according to a report published by Yahoo Finance and syndicated via Stocktwits on October 15.

The remarks were attributed to Huang speaking at an event in London. The report characterizes his message as less about near-term agreement and more about the inevitability of cooperation, framed around the need to manage the risks of highly capable AI systems rather than focusing solely on technology competition.

Huang’s comments highlighted a central tension for the AI industry: vendors and regulators increasingly confront the same question across borders, namely how to prevent misuse and unsafe deployment when powerful models can be built, sold, and scaled globally. In the framing cited by the report, the US and China will not only need to compete, but will also need a path to coordination because the consequences of uncontrolled deployment do not stop at national lines.

While the report indicates Huang used the phrase “potentially dangerous AI,” it does not spell out specific policy proposals, technical standards, or enforcement mechanisms in the information available here. It also does not identify whether he was referring to particular AI capabilities, specific application categories, or concrete governance frameworks. No details were provided on what forms of collaboration he believes would be workable.

For NVIDIA, the comments land in the context of its role as a supplier of accelerated computing, including GPUs and related software stacks that are widely used to build and train AI models. That market position can make the company a focal point in geopolitical debates, because AI progress depends on the availability of compute, networking, and developer tooling.

Sectorwide, the risk-and-governance conversation is also intensifying as AI systems become more capable. Even when governments take different approaches, leaders and executives often return to the same practical theme: if two major AI powers do not coordinate on safeguards, the global system can still be exposed to unsafe practices that originate in either jurisdiction.

The report does not disclose how Huang’s remarks relate to NVIDIA’s current compliance posture, customer mix, or any specific China-related product constraints. It also does not indicate whether he was speaking directly about export controls, licensing, or internal safety programs. As presented in the accessible information here, the statement is primarily described as a strategic and diplomatic call for cooperation, not a company policy update.