China's Ministry of Commerce on 27 July 2026 accused the United States of "artificial intelligence hegemonism" and warned that Beijing would take "all necessary measures" if Washington acted on threats to sanction Chinese AI companies. The statement, reported by multiple outlets citing a ministry spokesperson, marks Beijing's most direct formal response so far in a dispute over model distillation that has escalated over recent weeks.
What the ministry said
The ministry said US officials were threatening Chinese companies with punishment on the basis of allegations alone, arguing the claims "lacked factual basis and legal grounds, adopt double standards in practice, and constitute typical acts of artificial intelligence hegemonism." An unnamed spokesperson added: "For any action that causes substantive harm to Chinese interests, China will take all necessary measures to firmly safeguard its legitimate rights and interests." The measures were not specified.
Central to the rebuttal is the ministry's position that distillation, the practice of training a weaker model on the outputs of a stronger one, is a widely used technique across the industry. US officials counter that there is a difference between legitimate distillation and large-scale extraction that amounts to intellectual property theft, according to Reuters reporting carried by Times Now.
A counter-accusation
Alongside the denial, the ministry made an allegation of its own: that "many American artificial intelligence enterprises have distilled Chinese models during their research, development and training processes." No companies were named, and no evidence was published alongside the claim. As with the American allegations it answers, it remains an accusation rather than an established finding.
What prompted the response
The immediate trigger was a sequence of US moves. White House Office of Science and Technology Policy director Michael Kratsios alleged that Moonshot AI distilled Anthropic's Claude to build its Kimi K3 model, at industrial scale, through a purpose-built internal platform designed to avoid detection. Times Now, citing Reuters, reports the specific model named was Claude Fable 5. Reporting tied to the case also describes Moonshot sourcing Nvidia GB300 chips and routing training work through servers in Thailand; neither government has confirmed those specifics on the record.
Treasury Secretary Scott Bessent then warned publicly that Chinese firms conducting "covert, industrial-scale distillation attacks" could face financial sanctions or designation on the Commerce Department's Entity List, which would cut them off from American chips and software. "Open source is not open season on American IP," he wrote on X. Reports have also suggested the US is weighing restrictions on some foreign-developed open-source AI models.
Anthropic's own complaints sit underneath the policy moves. The company told US lawmakers earlier this year that DeepSeek, Moonshot and MiniMax generated more than 16 million interactions with Claude through roughly 24,000 fraudulent accounts, and it later named Alibaba's Qwen lab as the source of a larger campaign. It has used those findings to argue for tighter chip export controls.
What is established and what is claimed
Established: the commerce ministry made its statements on 27 July 2026; Kratsios and Bessent made their accusations and warnings publicly; Moonshot has denied the distillation charge, telling China's National Business Daily that Kimi K3's performance came from original architectural improvements rather than copying a rival.
Claimed but not proven: that Moonshot distilled Claude; that American firms distilled Chinese models; the GB300 and Thailand routing details. Anthropic's interaction tally has not been independently audited. No company has been added to the Entity List over distillation, and Beijing's countermeasures remain unspecified. Notably, neither side has raised a Section 301 trade investigation, keeping the fight aimed at a handful of AI labs and the models they have shipped.