U.S. Treasury Secretary Scott Bessent on Wednesday doubled down on his warning against Chinese AI companies and said sanctions remained on the table after White House officials accused Moonshot of unfairly extracting Anthropic’s Fables model.
Model distillation is a common AI training technique where a smaller model learns from the output of a larger model. Although this process may violate intellectual property rights, it is also widely used as a legal optimization method.
“It’s not open season for open source in US intellectual property,” Bessent wrote on X. “If a[Chinese]company secretly conducts an industrial-scale distillation attack that crosses the line of IP theft, sanctions and Entity List designation will be on the table.”
Earlier this week, Bessent said the U.S. government would investigate China’s open source models for signs of intellectual property theft and impose sanctions if found.
Bessent’s latest remarks came hours after White House science and technology policy chief Michael Crassios accused China-based Moonshot of conducting large-scale distillation, contrary to the U.S. model. He claimed that Moonshot “may acquire Nvidia’s GB300-based servers and access Thailand’s GB300 to train its own AI models,” raising questions about whether the company is violating U.S. export control rules.
The GB300 servers are part of Nvidia’s Blackwell generation and are prohibited from sale to Chinese companies.
Some experts dispute the idea that Kimi K3 could have been developed primarily through distillation from Fable, which has just been released to the public since July 1st. Moonshot released K3 as an open-weight model last week, but its advanced capabilities call into question the underlying business model of America’s leading AI labs and whether they can continue to justify the massive capital requirements supporting frontier AI competition.
The episode also intensified a broader debate in Washington over the influx of Chinese open models. Some, including Dean Boal, a former White House AI advisor and current director of strategic futures at OpenAI, have argued that the United States should limit or effectively ban the use of China’s open weight model in order to maintain U.S. technological superiority and reduce potential national security risks.
TechCrunch has reached out to Moonshot and the Treasury Department for comment.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
