Did Moonshot Steal Technology from Anthropic? U.S. Claims Kimi K3 Was Distilled from Claude Fable

The United States is accusing the Chinese AI startup Moonshot AI of intellectual property theft. The administration claims that the company’s latest AI model, Kimi K3, was created using technology distilled from Anthropic’s AI models. These allegations have reignited calls in Washington for tougher measures against Chinese AI companies. Nevertheless, many prominent figures in the U.S. tech sector—including Nvidia CEO Jensen Huang and Elon Musk—oppose this stance, arguing that the United States should not fear Chinese “open-weight” AI models.

On Wednesday, Michael Kratsios, head of the White House Office of Science and Technology Policy, accused Moonshot AI of using Anthropic’s Fable AI model to develop Kimi K3. He noted that the administration possessed information suggesting the company had created a “sophisticated internal platform” to carry out large-scale distillation of U.S. AI models while attempting to evade detection.

“We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model,” Kratsios wrote in a post on X.

“To do this, they developed a sophisticated internal platform designed to perform large-scale distillation from U.S. models, allowing them to rapidly switch between multiple access methods to avoid detection.” https://x.com/mkratsios47/status/2079933645888880708?s=20

Kratsios stated that Moonshot AI had acquired servers equipped with Nvidia GB300 chips and accessed additional GB300 systems in Thailand to train its models. While acknowledging that AI distillation is a common technique for creating smaller, more efficient models, he argued that there is a clear distinction between legitimate distillation and what he described as “large-scale, covert industrial distillation” aimed at stealing protected U.S. technology.

Could the United States ban Kimi K3?

Kratsios’s remarks came shortly after U.S. Treasury Secretary Scott Bessent indicated that the administration was investigating whether Chinese AI companies had improperly extracted knowledge from U.S. AI models. He added that companies found to have stolen U.S. intellectual property could face sanctions.

However, the United States is not currently planning a blanket ban on Chinese AI models or companies. Instead, the Trump administration is considering targeted restrictions aimed at reducing U.S. companies’ reliance on Chinese AI platforms, without imposing sweeping federal controls.

Why is the United States concerned about Kimi K3?

At the centre of the controversy is Kimi K3, Moonshot AI’s latest open-weights large language model. Unlike closed AI models from companies like OpenAI and Anthropic, open-weights models allow developers to download, customize, and run them on their own infrastructure.

As the model gains popularity, U.S. officials are questioning whether it was developed using technology improperly extracted from U.S. AI models, while also expressing concern over the speed at which the Chinese AI industry is closing the gap with the United States.

U.S. Tech Industry Calls for No Ban

Meanwhile, several major U.S. tech companies and startup founders are urging the Trump administration not to restrict access to Chinese open-weight AI models.

In an interview with Axios, Nvidia CEO Jensen Huang stated that U.S. companies should “without a doubt” be allowed to use Chinese AI models, arguing that banning them would do more harm than good.

Huang described Chinese open-source AI as “excellent” and downplayed concerns that downloading such models would automatically expose companies to security risks or hidden backdoors. Instead, he noted that organizations can customize the models and deploy them in secure environments. He also maintained that open models enhance security, as researchers can examine them, identify vulnerabilities, and develop more robust safeguards.

According to Huang, the market has misunderstood the impact of models like Kimi K3, just as it did earlier this year with DeepSeek. Rather than reducing demand for AI hardware, he believes that more affordable and accessible AI models will foster wider adoption, thereby driving greater demand for Nvidia’s chips, data centres, and computing infrastructure.

Tesla CEO Elon Musk also appeared to back Huang’s stance, replying “True” to a post highlighting the Nvidia leader’s argument that the United States should not fear Chinese AI models.

Silicon Valley Says a Ban Would Hurt U.S. Innovation

Separately, nearly 200 Silicon Valley companies—including startup accelerator Y Combinator and encrypted email provider Proton—have urged the Trump administration not to restrict access to Chinese open-weight AI models.

In letters sent by the newly formed Little Tech Association, the companies argued that U.S. startups rely on affordable, open AI models; Many of them cannot easily be replaced by more expensive proprietary alternatives from OpenAI or Anthropic, a famous publication reports. They warned that sweeping restrictions would harm U.S. innovation more than China’s. Instead of a total ban, they advocated for targeted safeguards against intellectual property theft, while allowing developers to continue accessing AI models that are already available worldwide.