U.S. Treasury Signals Sanctions Over Moonshot’s Alleged Distillation of Anthropic Model
Treasury Secretary Bessent threatened sanctions against Chinese AI firms after a White House official accused Moonshot of improperly distilling Anthropic's Fable model, escalating U.S. enforcement actions on AI intellectual property.
U.S. Treasury Secretary Scott Bessent doubled down on sanctions threats against Chinese AI companies on Wednesday, following a White House claim that Moonshot improperly distilled Anthropic's Fable model.
Bessent posted on X that "Open source is not open season on American IP," adding that "When [Chinese] firms conduct covert, industrial-scale distillation attacks that cross the line into IP theft, sanctions and Entity List designations will be on the table." His statement escalates enforcement language the Treasury introduced earlier in the week, when Bessent announced plans to examine open-source models from China for signs of intellectual property violations.
The accusation came from Michael Kratsios, the White House's science and technology policy chief, who alleged that Moonshot conducted large-scale distillation against U.S. models. Kratsios also raised export-control concerns, stating that Moonshot had acquired Nvidia GB300-equipped servers and accessed units in Thailand, potentially in violation of U.S. restrictions on advanced chip sales to Chinese firms. Nvidia's Blackwell-generation GB300 servers are banned from sale to Chinese companies.
Model distillation is a standard AI training technique where a smaller model learns from the outputs of a larger one. While the practice can breach intellectual property rights, it is also widely used as a legitimate optimization method. The line between legitimate optimization and IP theft remains contested among AI researchers and policymakers.
Moonshot released its K3 model last week as an open-weight system, drawing scrutiny over its advanced capabilities. Some experts question whether K3 could have been developed primarily through distillation of Fable, which became publicly available only on July 1. The model's performance has prompted debate in Washington about whether Chinese open-weight models pose a threat to U.S. technological competitiveness.
The dispute reflects broader tensions over open-source AI development and export controls. Former White House AI adviser Dean Ball, now Head of Strategic Futures at OpenAI, has argued for restricting or effectively banning Chinese open-weight models to protect American technological advantage and address national security concerns. TechCrunch reached out to Moonshot and the Treasury for additional comment.
