China’s Ministry of Commerce has told the United States to stop threatening sanctions against Chinese artificial intelligence companies.

The ministry warned that it will take “all necessary measures” to defend its interests if Washington moves ahead, underscoring how AI has become another pressure point in the broader U.S.-China technology rivalry.

China questions the basis for U.S. action

A spokesperson for China’s Ministry of Commerce said the U.S. threats have no factual or legal basis and amount to what Beijing calls “AI hegemony.”

According to the spokesperson, senior U.S. officials reportedly want to investigate Chinese labs over the alleged “distillation” of advanced American models and may impose sanctions on claims of intellectual-property theft.

The spokesperson said several Chinese models were released at roughly the same time as leading U.S. systems and already rank among the best in some areas, including front-end coding. Innovation, the spokesperson added, belongs to no single country.

China’s strongest criticism is that many U.S. AI companies have distilled Chinese models during research and training, according to the Ministry of Commerce. The ministry also said close to 200 U.S. startups have asked their own government not to cut off access to Chinese open-source models, arguing that such a move would weaken their competitiveness, as Cryptopolitan reported.

A government spokesperson had also previously dismissed the theft allegations, calling them an unfounded “smear” rooted in prejudice.

Beijing ended by urging both countries to honor a consensus reached by their leaders.

Adoption of Chinese models continues to rise

Cryptopolitan reported that Chinese models now account for 46.4% of routed token traffic on OpenRouter, a platform that allows developers to switch between systems, compared with 35.7% for U.S.-built models. DeepSeek alone makes up 17.6% of that total as of July 2026. A Hugging Face study published on March 16, 2026, found that Chinese open-source models accounted for 41% of all open-source model downloads.

The reported reasons for the wider adoption of Chinese technology include lower compute costs and U.S. export limits imposed earlier in 2026 on frontier models such as Anthropic’s Claude Mythos 5 and Fable 5, which created openings that foreign systems filled. That backdrop helps explain why any new sanctions or procurement restrictions would matter beyond the companies directly targeted, since access to widely used models can affect developers’ tool choices and enterprise workflows.

After Moonshot AI launched Kimi K3, developers said the model’s performance was close to Anthropic’s Fable and OpenAI’s ChatGPT, while being available at a much lower price. That development has prompted the U.S. to consider further measures against Chinese technology, even as Chinese providers continue to gain share in open-source and routed usage.

OpenAI’s most recent valuation was $852 billion, Anthropic’s was $965 billion, and Moonshot’s was reported at roughly $30 billion.

Cryptopolitan reported on July 20 that the U.S. is likely to introduce new procurement rules, renew threats to add Chinese labs to the Commerce Department’s Entity List, and increase public pressure on American companies using Chinese models. The Entity List already restricts hundreds of Chinese firms from buying U.S. goods without a license.