US Treasury Threatens Sanctions as White House Accuses Moonshot of Distilling Anthropic's Fable
· 2 min read ·
The United States Treasury Department has escalated its warnings to Chinese artificial intelligence companies, with Secretary Scott Bessent declaring that sanctions remain a viable option following allegations that the China-based firm Moonshot improperly distilled Anthropic's Fable model.
White House Alleges Large-Scale Distillation
Bessent's remarks on Wednesday came just hours after Michael Kratsios, the White House's science and technology policy chief, publicly accused Moonshot of conducting large-scale distillation against American AI models. Distillation is a widely used training technique in which a smaller model learns from the outputs of a larger one. While it serves as a legitimate optimization method in many contexts, it can also infringe on intellectual property rights when performed without authorization.
"Open source is not open season on American IP," Bessent wrote on X. He warned that when Chinese firms carry out what he described as covert, industrial-scale distillation attacks that cross the line into intellectual property theft, both sanctions and Entity List designations would be on the table.
Earlier in the week, Bessent had already signaled that the U.S. government intended to scrutinize open source models originating from China for evidence of intellectual property theft, with sanctions to follow if violations were confirmed.
Export Control Concerns Over Nvidia GB300 Servers
Kratsios raised additional alarm by alleging that Moonshot had acquired Nvidia's GB300-equipped servers and accessed GB300 units in Thailand, presumably to train its AI models. These servers are part of Nvidia's Blackwell generation of chips, which are explicitly banned from being sold to companies based in China. The allegations have prompted questions about whether Moonshot violated U.S. export-control regulations to obtain the restricted hardware.
Experts Question the Distillation Timeline
Despite the accusations, some experts have pushed back against the notion that Moonshot's Kimi K3 model could have been developed primarily through distillation from Anthropic's Fable. Fable has only been publicly available since July 1, while Moonshot released K3 as an open-weight model the following week. The compressed timeline has led some to question whether distillation from Fable alone could account for K3's reported capabilities.
The release of K3 has sent ripples through the AI industry. Its advanced performance has called into question the foundational business models of leading American AI laboratories, fueling skepticism about whether these companies can continue to justify the enormous capital expenditures underpinning the frontier AI race.
Broader Debate Over Chinese Open-Weight Models
The controversy has also intensified an ongoing debate within Washington about the growing presence of Chinese open-weight models in the U.S. market. Dean Ball, a former White House AI adviser who now serves as OpenAI's Head of Strategic Futures, has been among those arguing that the United States should restrict or effectively ban the use of Chinese open-weight models. Proponents of such measures cite the need to preserve America's technological advantage and mitigate potential national security risks.
TechCrunch has reached out to both Moonshot and the Treasury Department for comment. As the intersection of open-source AI development, intellectual property enforcement, and export controls becomes an increasingly contentious frontier in U.S.-China relations, the outcome of this dispute could shape policy for the entire industry. What is your take — should the U.S. impose sanctions over AI model distillation, or would that risk stifling open-source innovation? Share this article and let us know where you stand.
