A senior White House technology official on Wednesday accused China’s Moonshot AI of taking from Anthropic’s most advanced large language model, Fable, to support the development of its newest K3 model, while also alleging the company obtained cutting-edge Nvidia artificial intelligence chips.
“We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model,” Michael Kratsios said in a post on X. “Large-scale, covert industrial distillation aimed at stealing proprietary US technology and undermining American research is unacceptable.”
Kratsios also said Moonshot AI had secured servers equipped with Nvidia NVDA.O GB300 chips, among the company’s powerful AI processors, and used them in Thailand, “likely to train its AI models.” The allegation follows comments in February from a senior Trump administration official, who said Chinese AI startup DeepSeek’s latest AI model had been trained on Nvidia’s most advanced AI chip — a scenario that could point to a breach of U.S. export controls.
Moonshot AI and the Chinese Embassy in Washington did not immediately respond to requests for comment.
The accusations come amid a broader U.S. government campaign over alleged Chinese intellectual property theft in artificial intelligence. In April, the U.S. State Department ordered diplomats worldwide to spotlight what it described as extensive efforts by Chinese companies, including Moonshot AI, to take intellectual property from U.S. AI labs, according to a diplomatic cable reviewed by Reuters.
In artificial intelligence, distillation refers to training a smaller model on the outputs of a larger and more expensive system, a technique often used to reduce the cost of building a capable new AI tool.
OpenAI has also raised concerns with lawmakers, warning that DeepSeek was targeting the ChatGPT creator and other leading U.S. AI companies in an effort to copy their models and use them for its own training, Reuters reported in February.
The State Department’s April cable said, “AI models developed from surreptitious, unauthorized distillation campaigns enable foreign actors to release products that appear to perform comparably on select benchmarks at a fraction of the cost but do not replicate the full performance of the original system.”
Chinese AI startup Moonshot on Friday unveiled Kimi K3, a 2.8 trillion-parameter model that it said is the world’s largest open-weight AI system and delivers performance approaching US giant Anthropic’s frontier Fable model.
The launch, which comes a month after Anthropic’s Fable and Mythos models were abruptly withdrawn by the US government due to security concerns, underscores how quickly China’s open AI ecosystem is narrowing the gap with the most advanced US systems.
