Jensen Huang, chief executive officer of Nvidia, told Axios in a "Behind the Curtain" video interview that he believes the United States has nothing to fear from Chinese open-source AI models and that efforts to ban them could do more harm than good.
"These Chinese models are excellent," Huang said. "Open-source models that are excellent should be used."
Why this matters
As the world's leading AI chipmaker, Huang argued that American companies should "absolutely" be allowed to use Chinese models — a position that runs counter to some former Trump administration officials and U.S. labs pressing Washington to block those models.
Among the concerned parties are Nvidia customers OpenAI and Anthropic, which have accused Chinese rivals of siphoning capabilities from their models and warned that Beijing's open-source surge threatens U.S. AI leadership.
Huang rejected the premise that OpenAI and Anthropic should fear open models. He said open models broaden the market by giving more people a first exposure to AI, while many users will still pay for the convenience, reliability and higher performance of closed services.
"There's no scenario where China runs U.S. companies off the road," Huang told Axios. "Zero possibility."
Where and when he spoke
Huang spoke with Axios in Fort Worth, Texas, at the opening of a new phase of an advanced manufacturing plant where Taiwan-based Wistron produces AI infrastructure for Nvidia. The visit highlighted U.S. capabilities in sophisticated manufacturing.
The big picture: Kimi K3 and market reaction
The Beijing-based Moonshot AI release of Kimi K3 triggered the sharpest wave of AI-related market concern since DeepSeek unsettled financial markets in January 2025. Kimi combines three features that rarely appear together: near-frontier performance, substantially lower prices, and open weights developers can download and customize.
That mix helped prompt a sell-off in Nvidia and other chip stocks, reviving worries that cheaper, more efficient models could undermine the rationale for the industry's massive AI infrastructure buildout.
Huang's view: the market misreads the impact
Huang said Wall Street and Washington have misunderstood the impact of Kimi — as they did with DeepSeek. His reasoning: cheaper, open models will make AI available to more people and businesses, increasing demand for the chips, data centers and compute power Nvidia sells.
"Free AI should be great for hardware," Huang said. "Free AI should be great for chips. Free AI should be great for data centers."
Policy and security: target misconduct, not models
On policy, Huang argued that restricting open models in the name of national security could actually leave the United States more vulnerable. He dismissed the notion that downloaded Chinese models create a "backdoor" to Beijing, saying companies can customize them and control access inside secure "sandboxes."
Huang suggested openness can improve security because external researchers can inspect models, uncover weaknesses and build defenses. "If everything just becomes one single model, one single point of attack, one single source of failure, I think the world is much, much more vulnerable," he said.
He applied the same logic to American models and urged Anthropic to make its restricted cyber model, Claude Mythos, available to "everyone" rather than limiting access. Huang said companies should harden powerful models through testing and rapid fixes instead of restricting access from the start. "Mythos should be available as a service," he said. "Remember: Just because Mythos is not available, open models are available anyhow. So I think: Let Anthropic run."
"Holding Anthropic back is not in the benefit of the United States," he added.
The race has no finish line
Huang views the AI competition differently than many in Washington who portray China's advances as an acute threat to U.S. dominance. He rejected the idea that it is "a race with an endpoint": "We're going to continue to use AI forever. The United States is going to be here for a long time. China's going to be here for a long time."
Responses and enforcement
Huang's remarks came hours after Treasury Secretary Scott Bessent told Fox Business that the administration is examining Chinese AI models for stolen U.S. intellectual property and considering sanctions. Bessent suggested the government could sanction overseas models if they are found to have "stolen" from American companies, citing "watermarks" of U.S. models inside Chinese ones.
Huang offered a different takeaway: "Distillation, learning from AI, learning from other sources of knowledge, is fundamental to intelligence," he told Axios. He said companies should face consequences for privacy violations or contract breaches, but his broader message was clear: focus enforcement on misconduct, not on banning models outright.
