Nvidia Chief Executive Jensen Huang has called on the United States, China and other major technology powers to deepen cooperation on artificial intelligence safety, noting that governments and researchers must work together to monitor the risks posed by increasingly powerful AI systems.
Huang said Nvidia “absolutely must” work with Chinese AI companies to help ensure the technology develops safely, making the case for international dialogue at a time when concerns about AI-related harm are intensifying, and relations between Washington and Beijing remain complicated by technological competition.
“We absolutely should work with each other so that we advance the technology safely,” Huang told CNBC during a media briefing in London on Wednesday evening.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
This statement puts him at odds with some prominent AI industry leaders who have advocated stronger regulation and greater caution over the pace of development. While OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have called for increased regulatory oversight, Huang has emerged as a prominent opponent of broad restrictions on AI development.
Rather than treating the technology primarily as a regulatory problem, Huang emphasized dialogue between companies and countries developing advanced AI systems. He said he already maintains conversations with researchers in China and argued that these exchanges should be expanded.
“We ought to enhance the dialogue,” he said. “The United States should do it… I advocated that the U.K. do the same thing.”
While advanced AI development is increasingly shaped by competition among countries and technology companies, the potential risks do not necessarily stop at national borders. Huang’s position is that communication between rival AI ecosystems is necessary to identify and monitor those risks, even as governments compete for technological leadership.
AI safety has gained prominence in discussions between the United States and China as both countries consider how to address the risks associated with rapidly advancing systems.
The debate extends beyond the commercial race to develop more capable models. Researchers have raised concerns about the possibility of AI systems being misused, producing harmful outputs or developing capabilities that existing safeguards may not adequately control. Some researchers at AI laboratories, including OpenAI and Anthropic, have called for slowing development and warned of risks they believe could threaten humanity.
These concerns have intensified pressure on technology companies and governments to establish clearer safety practices. Yet the debate over how to respond remains divided, particularly over whether stronger regulation or continued development accompanied by targeted safeguards offers the most effective approach.
Huang’s emphasis on cooperation reflects the difficulty of managing AI risks through unilateral action alone. If companies in different countries develop more capable systems under different rules and safety practices, governments may struggle to establish common approaches to monitoring and responding to harmful outcomes.
The issue has already entered diplomatic discussions between Washington and Beijing. Before US President Donald Trump’s meeting with Chinese President Xi Jinping, US Treasury Secretary Scott Bessent discussed the possibility of a “U.S.-China AI dialogue” with Chinese Vice Premier He Lifeng in September.
According to the account of those discussions, the proposed channel could cover incidents reaching the level of national security. That scope suggests the dialogue could extend beyond technical questions to potential emergencies involving AI systems and their implications for national security.
A Chinese readout of the meeting said Xi believed the two countries could compete while also expanding AI cooperation.
“The two sides can continue their dialogue on AI, exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI,” Xi said, according to the readout.
The statement signals Beijing’s publicly expressed interest in maintaining communication on AI risks despite the broader rivalry between the two countries. However, the existence of diplomatic discussions does not establish that Washington and Beijing have agreed on specific safety standards, enforcement mechanisms or procedures for responding to incidents.
The practical challenge will be converting calls for dialogue into arrangements that produce useful information-sharing and coordinated responses without requiring either side to abandon its technological or national security priorities.
Huang Sees The UK As a Potential Convener
Huang also argued that the United Kingdom could play a useful role in bringing together governments and researchers to discuss AI safety.
He said Britain was well positioned to convene international dialogue because of its “intellectual capital”, suggesting that its research institutions and technical expertise could help facilitate discussions among countries and companies that might otherwise struggle to cooperate directly.
Such a role could give the UK an opportunity to contribute to international AI governance even as the United States and China remain the central competitors in advanced AI development. Convening discussions, however, would be only an initial step; industry experts have noted that meaningful cooperation would depend on whether participants could agree on shared priorities and establish practical mechanisms for handling risks.
Huang’s comments also underline the importance of distinguishing AI safety from the broader question of how quickly the technology should advance. Calls for greater oversight do not necessarily imply opposition to AI development, just as support for continued development does not eliminate the need for safeguards. The disagreement concerns how to balance potential benefits against risks and what responsibilities companies and governments should assume.
For Nvidia, the debate has particular relevance because the company supplies advanced computing chips used to train and operate AI systems. Its business is closely tied to demand for powerful computing infrastructure, placing Huang at the intersection of the commercial expansion of AI and discussions about its potential consequences.
His call for cooperation does not resolve the disagreements over regulation, development speed, or national security. Nor does dialogue alone guarantee that companies will prevent harmful outcomes. Effective safeguards would still require technical testing, monitoring, accountability and clear procedures for addressing failures.
Nevertheless, Huang’s position is that AI safety should not be treated exclusively as a domestic policy issue or as another front in the US-China technology rivalry. He is urging the countries and companies developing the technology to maintain communication even as they compete.