US floats sanctions on Chinese AI models as Huang urges calm
Treasury Secretary Scott Bessent said Washington can sanction Chinese models over alleged IP theft, while Nvidia's Jensen Huang warned against calls for a ban.
The US government sharpened its rhetoric against Chinese AI this week, with Treasury Secretary Scott Bessent saying Washington will scrutinize Chinese open-source models for intellectual-property theft and has the authority to sanction their makers — even as Nvidia CEO Jensen Huang publicly argued the real danger lies in banning them.
The sanctions threat
Speaking on Fox Business on July 21, Bessent said that if overseas models are found to be stealing from American companies, the US 'has the ability to sanction them.' The administration's specific concern is distillation — training smaller models on the outputs of larger ones — which US labs have long complained Chinese developers use to absorb their capabilities. The remarks follow reports that the White House is weighing a broader ban on Chinese open-weight models, and they land amid the market shock from Moonshot's Kimi K3, whose release helped drag semiconductor stocks more than 20% off their June peaks and briefly cost Nvidia its spot as the world's most valuable company.
The industry pushback
The theft framing is contested inside the US industry itself. Microsoft CEO Satya Nadella noted American companies also rely on distillation, and Hugging Face CEO Clément Delangue argued China's progress reflects strong research teams rather than copying. Huang went furthest: in an Axios interview published July 22, he called Chinese models from DeepSeek, Alibaba, Tencent, MiniMax, and Baidu 'excellent,' and warned that restricting open models in the name of security would leave America more vulnerable while shrinking the global market that runs on US chips and infrastructure.
Why it matters
Export controls have so far targeted hardware; sanctioning models themselves would extend the tech war into software — and into open-source, where enforcement is close to impossible once weights circulate. The split between Washington's hawks and the industry's most powerful voices shows there is still no settled US theory of the case: whether Chinese open models are a security threat, a commercial challenge, or — as Huang contends — a customer base. Which view prevails will shape global AI supply chains for years.