Nvidia CEO Jensen Huang told CNBC that artificial intelligence is "potentially dangerous" and that the United States and China "absolutely must" work together on monitoring its risks. The comments, reported on Thursday, Oct. 15, 2026, came from a media briefing Huang held Wednesday evening in London.
"We absolutely should work with each other so that we advance the technology safely," Huang told CNBC. Yonhap Infomax, citing Associated Press coverage of the same London event, quotes him saying: "I think this technology is powerful and potentially dangerous. We must work together to develop the technology safely." Asked how to work with Chinese AI companies, he answered "dialogue, dialogue," according to that report.
What he wants from governments
Huang said he already has "dialogues with many of the researchers in China" and that those talks should deepen. "We ought to enhance the dialogue," he told CNBC. "The United States should do it... I advocated that the U.K. do the same thing." He cast the U.K. as a natural "convener" for AI safety talks because of its "intellectual capital," CNBC reported.
The briefing came after Nvidia said it would donate $250 million over five years to support agentic AI safety and security research in the U.K., Yonhap Infomax reported.
CNBC notes that AI safety has become a U.S.–China government topic. It points to September talks between U.S. Treasury Secretary Scott Bessent and China's He Lifeng on a possible "U.S.-China AI dialogue," including a channel for incidents "up to a national security level," and to a Chinese readout in which President Xi Jinping said the two sides can "exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI."
Regulation vs. company responsibility
CNBC frames Huang as a prominent opponent of heavier AI regulation, in contrast to OpenAI's Sam Altman and Anthropic's Dario Amodei, who have called for more of it. On security incidents, Yonhap Infomax reports Huang argued that responsible corporate fixes should come before stronger rules, saying problems like a recent cyberattack on Hugging Face should be solved through technology and engineering: "Before demanding greater authority, we must take responsibility and solve problems ourselves."
What it means for Tesla
Nvidia's chips power much of the AI training stack that Tesla, xAI and rivals rely on, and Huang's safety comments sit in the same public debate Elon Musk has pressed for years. Hours before CNBC's write-up, Musk posted on X that "Truth-seeking is the most important aspect of AI/SI safety." Tesla itself is pitching unsupervised FSD, Robotaxi and Optimus as AI products, so how chipmakers and labs talk about cross-border safety and regulation is not abstract for Tesla owners or investors. See our coverage of eight states suing OpenAI over ChatGPT child safety for the latest U.S. regulatory pressure on frontier labs.