[DigitalToday reporter Jinju Hong (홍진주)] China demanded that the United States stop reviewing sanctions on its artificial intelligence companies and publicly warned of possible countermeasures.

Cryptopolitan, a blockchain media outlet, reported on July 27 that China’s Ministry of Commerce said the United States was threatening sanctions on Chinese AI firms. It said China would take all necessary steps to protect its interests if Washington moved to take action.

The dispute centers on the United States investigating whether Chinese laboratories engaged in so-called model “distillation” and considering sanctions over alleged intellectual property violations. U.S. senior officials are reported to be examining whether Chinese labs replicated technology based on high-performance U.S. AI models.

China’s Ministry of Commerce rejected the claims head-on. A ministry spokesperson said the repeated threats of sanctions by the United States lacked facts and a legal basis and amounted to “AI hegemony.” The spokesperson also said some Chinese models were released around the same time as leading U.S. systems and that in areas such as front-end coding they had already reached the top level. The spokesperson also stressed that innovation does not belong exclusively to any one country.

China countered that the United States had also used Chinese models in research and training. The ministry said several U.S. AI companies had used Chinese models through distillation. It also said about 200 U.S. startups had asked their government not to block access to Chinese open-source models, arguing that doing so could weaken competitiveness.

China’s stance on U.S. claims of intellectual property violations is also hardline. A Chinese government spokesperson previously described the allegations as baseless “smearing” and rebutted them as criticism born of prejudice. China said the two sides should respect the common understanding reached by their leaders and warned against an expansion of conflict.

Behind the clash is the rapid spread of Chinese AI models. On OpenRouter, Chinese models accounted for 46.4 percent of total routing token traffic as of July 2026. U.S.-made models accounted for 35.7 percent. DeepSeek alone accounted for 17.6 percent. A Hugging Face study released on March 16, 2026 found Chinese open-source models made up 41 percent of total open-source model downloads.

Cost competitiveness is cited as a driver of the spread of Chinese models. The explanation is that Chinese technology has lower computing costs, and that in early 2026 the United States imposed export restrictions on frontier models such as Anthropic’s Claude Mythos 5 and Fable 5, creating more room for overseas models to fill gaps. After Moonshot AI released Kimi K3, developers said its performance was close to Anthropic’s Fable and OpenAI’s ChatGPT while its price was much lower.

The United States is also reviewing additional responses. There is a possibility it will introduce new procurement rules on Chinese technology and again raise the option of adding Chinese labs to the Commerce Department’s Entity List, a trade restriction roster. Steps to intensify public pressure on U.S. companies’ use of Chinese models are also being discussed. The Entity List restricts purchases of U.S.-made products without permission.

The gap is also large in market terms. OpenAI’s recent valuation was estimated at $852 billion and Anthropic at $965 billion. Moonshot’s valuation was reported to be about $30 billion. Still, as Chinese models continue to spread on the back of price and accessibility, a key point to watch is how tighter U.S. regulation will affect the technology rivalry and corporate adoption trends.