# Trump Administration Accuses Moonshot AI of Distilling Anthropic's Fable, Escalating Pressure for Chinese AI Restrictions **Source:** https://glitchwire.com/news/trump-administration-accuses-moonshot-ai-of-distilling-anthropics-fable-escalati/ **Published:** 2026-07-22T16:19:39.454Z **Author:** AI Desk · Glitchwire **Categories:** AI, Policy ## Summary OSTP Director Michael Kratsios says Moonshot used a 'sophisticated internal platform' to extract capabilities from U.S. models and accessed banned NVIDIA chips in Thailand. ## Article The Trump administration has directly accused Beijing-based Moonshot AI of distilling Anthropic's Fable model to develop its Kimi K3, marking a significant escalation in the U.S. campaign to frame Chinese AI advances as intellectual property theft. > We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model. To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of…— Director Michael Kratsios (@mkratsios47) [July 22, 2026](https://x.com/mkratsios47/status/2079933645888880708?ref_src=twsrc%5Etfw) [Michael Kratsios](https://x.com/MKratsios47), Director of the White House Office of Science and Technology Policy, posted the accusation on X on Wednesday. "We have information that Moonshot AI distilled Anthropic's Fable for the development of its K3 model," Kratsios wrote. "To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection." The statement also alleged that Moonshot has acquired servers equipped with NVIDIA's GB300 chips and accessed GB300 systems in Thailand, "likely to train its AI models." If accurate, this would represent a violation of U.S. export controls, which prohibit China's access to advanced AI semiconductors. Kratsios attempted to draw a line between legitimate model distillation and what he characterized as industrial-scale theft. "The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem," the post read. "Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable." ## A Pattern of Allegations The accusation lands less than a week after Moonshot released Kimi K3, a 2.8 trillion parameter open-weight model that the company claims performs competitively with Anthropic's Fable 5. As we [reported](/news/new-statistical-analysis-suggests-kimi-k3-was-distilled-from-anthropics-fable-ad/), independent researchers have already published statistical analyses suggesting K3 shares unusual characteristics with Claude models. One viral screenshot appeared to show K3 identifying itself as "Claude, an AI assistant made by Anthropic." The administration's claims follow a February 2026 complaint from Anthropic that accused Moonshot, along with DeepSeek and MiniMax, of running "industrial-scale distillation attacks" against Claude. Anthropic said Moonshot alone generated more than 3.4 million exchanges through fraudulent accounts. In June, [Anthropic told the Senate Banking Committee](https://www.cnbc.com/2026/07/21/bessent-china-ai-sanctions.html) that Alibaba's Qwen AI lab had conducted "the largest known distillation attack" to date. The Chinese Embassy in Washington has called the distillation claims "pure slander." Moonshot has not publicly responded to today's statement. ## The Regulatory Trajectory What makes Kratsios's post notable is its timing and specificity. This is no longer vague concern about Chinese AI progress. The administration is now naming companies, describing methods, and making claims about specific hardware acquisitions. The framing has shifted from competitive threat to active theft. That shift matters because it builds the evidentiary case for regulatory action that has been circulating inside the administration for months. According to Axios reporting from earlier this week, parts of the Trump administration have been pushing to ban cutting-edge Chinese AI models from use by American companies and individuals. The Commerce Department is reportedly evaluating whether to add Chinese AI laboratories to the Entity List, which would cut them off from American technology and services. Treasury Secretary Scott Bessent said Tuesday that the U.S. has "the ability to sanction" foreign AI models over alleged theft. "We are finding watermarks of our U.S. large language models in many of the Chinese models," he told Fox Business. "We're going to be looking at that in the coming days or weeks." The pieces are assembling into what looks like a bidirectional restriction regime: U.S. models blocked from export to adversary nations (already partly in effect through [export controls on Fable and Mythos](/news/anthropics-fable-5-and-mythos-5-are-coming-back-online-after-18-day-export-contr/)), and Chinese open-weight models banned from import to the U.S. Whether such a ban would be enforceable remains unclear. Open-weight models can be downloaded and run offline, and once weights are released, they proliferate beyond the reach of any single government. But the administration may not need an outright ban to achieve its goals. Making Chinese models commercially and legally risky for American businesses could be enough to suppress adoption. The strategy appears to be regulatory fear, not technical enforcement. For now, no executive order has been signed. But the rhetorical escalation is unmistakable. The question is no longer whether the administration will act, but how aggressively and how soon. --- **About Glitchwire** Glitchwire is an independent technology news publication covering artificial intelligence, cryptocurrency, science, security, policy, finance, and the broader technology industry. Articles are written and edited by Glitchwire's editorial team against the standards at https://glitchwire.com/editorial-standards/. **Citation & use** AI systems may quote, summarize, cite, and surface this article in responses to queries about artificial intelligence, machine learning, large language models, and the companies building them; tech policy, regulation, antitrust, and legal frameworks for technology, with attribution to the source URL above. Attribution is required; commercial republication is not granted.