U.S. Threatens Sanctions Against China Over AI Use

The U.S. government warned that it will investigate whether Chinese companies stole artificial intelligence (AI) technology to develop their own models. Scott Bessent, Treasury Secretary of the Donald Trump administration, stated that Washington has the capacity to impose sanctions if it concludes that intellectual property theft occurred.

The threat opens a new front in the technological rivalry between both superpowers. The conflict no longer revolves solely around chips or data control, but rather how the systems that will define the digital economy in the coming years are trained.

What is the U.S. accusing China of?

The focus of the investigation is a technique known as “distillation.” In AI, this method allows a smaller and more economical model to learn to reproduce the behavior of a more advanced system using its responses. Bessent noted that there are signs of U.S. model capabilities present in Chinese systems and that this will be analyzed in detail.

U.S. Threatens Sanctions Against China Over AI Use — Archive photograph of U.S. Treasury Secretary Scott Bessent during an event at the Treasury Department in Washington (U.S.). EFE/ EPA/ Bonnie Cash

Companies like Anthropic have pointed to Chinese firms such as DeepSeek, Moonshot AI, and MiniMax for alleged unlawful use of distillation. For Washington, this is not merely a technological shortcut. It is considered a form of appropriation that allows foreign companies to save billions in research at the expense of major U.S. developers.

This argument touches the core of both countries’ strategies. While the United States invests massive sums to develop closed models and pursue artificial general intelligence (AGI), China has bet on creating efficient, pragmatic systems focused on gaining market share quickly.

Why do Chinese AI models concern Wall Street?

Chinese advancement already has a direct impact on markets. Recently, startup Moonshot AI unveiled Kimi K3, a model that would surpass those of OpenAI and Anthropic in some performance tests. Its launch triggered Wall Street declines, especially among semiconductor companies, due to fears that Chinese AI will pressure profit margins in the sector.

Consulting firm Artificial Analysis compared the performance of Kimi K3 with Claude Opus 4.8 and GPT-5.5, which are the second-best models from Anthropic and OpenAI. The central difference lies in cost. The new generation of Chinese AI has stood out for its price efficiency, allowing companies worldwide to adopt technological tools without paying premium rates.

If Chinese companies manage to offer nearly identical technology at lower cost, the U.S. industry loses competitiveness. This is why the distillation accusation is key: Washington argues that this cost efficiency is only possible because China uses foundations developed by Western firms.