United States Treasury Secretary Scott Bessent reinforced his stern warnings to Chinese artificial intelligence firms on Wednesday, indicating that potential sanctions remain under consideration. This comes in the wake of an accusation from a White House official, who alleged that Moonshot had improperly distilled Anthropic’s Fable model.
Model distillation, a prevalent AI training technique, involves a smaller model learning from the outputs of a more extensive one. Although this methodology can potentially infringe upon intellectual property rights, it is also broadly recognized and utilized as a valid optimization strategy.
Secretary Bessent articulated his stance on X, stating, “Open source is not open season on American IP.” He further elaborated that if Chinese firms engage in “covert, industrial-scale distillation attacks that cross the line into IP theft,” then sanctions and designations to the Entity List would be actively considered.
Earlier in the week, Secretary Bessent had already indicated that the U.S. government intended to scrutinize open-source models originating from China for any evidence of intellectual property theft, with sanctions to be imposed should such theft be confirmed.
These recent comments from Bessent closely followed accusations made by Michael Kratsios, the White House’s science and technology policy chief. Kratsios specifically accused the Chinese company Moonshot of undertaking large-scale distillation efforts targeting U.S. models. He further alleged that Moonshot had either acquired Nvidia’s “GB300-equipped servers” or accessed “GB300s in Thailand,” presumably for the purpose of training its AI models, thereby raising concerns about potential infringements of U.S. export control regulations.
The GB300 servers belong to Nvidia’s advanced Blackwell generation, a product line subject to prohibitions against sales to Chinese entities.
However, some industry experts contest the premise that Moonshot’s Kimi K3 model could have been developed predominantly via distillation from Fable, given that Fable only became publicly accessible on July 1. Moonshot’s release of K3 last week as an open-weight model, coupled with its sophisticated capabilities, has prompted questions regarding the fundamental business strategies of prominent U.S. AI laboratories. This has, in turn, cast uncertainty on their ability to continually justify the substantial capital investment required to compete in the cutting-edge AI race.
This incident has also amplified a wider discussion within Washington concerning the increasing prevalence of Chinese open models. Noteworthy figures, such as Dean Ball – a former White House AI advisor and current Head of Strategic Futures at OpenAI – advocate for the U.S. to either restrict or outright ban the deployment of Chinese open-weight models. Their rationale centers on safeguarding America’s technological supremacy and mitigating potential national security vulnerabilities.
TechCrunch has sought comments from both Moonshot and the Treasury Department regarding these developments.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.