Bessent threatens China AI sanctions as Huang defends models

Free Instant SEO Audit

By NewSunSEO July 23, 2026

US Treasury Secretary Scott Bessent warned on Tuesday that the United States will scrutinise Chinese open-weight AI models for stolen intellectual property and could sanction their makers, as Nvidia chief executive Jensen Huang pushed the opposite message: open-weight Chinese models are excellent and American firms should be allowed to run them.

What Bessent said

Speaking on television, Bessent framed the issue as a line the US will enforce. "This administration supports open source models, but what we do not support is IP theft," he said. His evidence is technical: "We are finding watermarks of our US large language models on many Chinese models," he told Bloomberg, which first reported the comments. The practice he is pointing to is called distillation, where one model is trained on the outputs of another. Bessent called it theft and said action would come within weeks.

What triggered the warning

The trigger is Kimi K3, an open-weight model released by Moonshot AI, a Chinese lab. Kimi K3 has matched or beaten leading OpenAI and Anthropic models on some benchmarks while costing far less. Its release triggered a chip-stock selloff and renewed concern in Washington. US labs have accused Chinese competitors of siphoning their models for months.

Why Nvidia’s Jensen Huang pushed back

Speaking to Axios at a new chip plant in Texas, Huang said the panic around Chinese models is misplaced. "These Chinese models are excellent," he said. "Open-source models that are excellent should be used." American firms, he argued, should "absolutely" be allowed to run them. His argument is commercial: cheaper or free models put AI into more hands, and more AI means more demand for the chips, data centres, and power Nvidia sells. "Free AI should be great for chips," he said.

Huang also dismissed the idea that downloading a Chinese model gives Beijing a "backdoor". Firms can wall them off in secure sandboxes, he said.

The awkward comparisons Huang drew

Not everyone accepts the theft framing. Distillation is widespread, and US labs do it too. Clem Delangue, chief executive of Hugging Face, told TechCrunch: "We know distillation to be a very small factor," adding that it is "a practice that everyone is doing, including companies in the US."

There is also an irony in who is doing the accusing. Anthropic, one of the loudest voices warning about Chinese distillation, recently had a $1.5bn settlement approved over pirated books used to train its Claude model. Huang went further in his Axios remarks, urging the US government to stop restricting Anthropic’s own Mythos model. "Let Anthropic run," he said.

What happens next

Bessent will lead the US delegation at AI talks with China scheduled for September. The distillation dispute is expected to be on the table. For now, the split is sharp: the US government wants to fence Chinese models out, while the world’s most valuable chipmaker wants them let in.

FAQ

What did Scott Bessent say about Chinese AI models?

Treasury Secretary Scott Bessent said the US will scrutinise Chinese open-weight AI models for stolen intellectual property and could sanction their makers within weeks. He pointed to watermarks of US large language models appearing in Chinese models as evidence of distillation.

Why did Jensen Huang defend Chinese open-weight models?

Nvidia chief executive Jensen Huang, speaking at a Texas chip plant, said open-weight Chinese models are excellent and that American firms should be allowed to run them. He argued that cheaper, free models expand AI use, which drives demand for the chips, data centres, and power Nvidia sells.

Which Chinese model sparked the new US concern?

Kimi K3 from Moonshot AI matched or beat top OpenAI and Anthropic models on some benchmarks at lower cost. Its release is what triggered the chip-stock selloff and the renewed warnings from Washington about Chinese distillation practices.