Moonshot AI Hit by US Scrutiny Over Alleged Data Siphoning

Moonshot AI, the Chinese company behind the Kimi chatbot, is facing mounting scrutiny from US authorities over allegations of siphoning data from rival models. The company's latest system, Kimi K3, has been accused of distilling data from Anthropic's Claude and Fable models.

The allegations began with Anthropic's claim in September 2026 that Moonshot used around 5,380 proxy accounts to pull data out of its flagship model, Claude Opus. The company alleged that the goal was to feed this data back into Moonshot's own systems.

The US government has also made similar claims, with officials including OSTP Director Michael Kratsios and Treasury Secretary Scott Bessent accusing Moonshot of distilling data from Anthropic's Fable model in July 2026. They alleged that the company used sophisticated methods to avoid detection and accessed banned Nvidia servers in Thailand.

Moonshot AI was founded in March 2023 and launched its Kimi chatbot in November 2023. The company's latest system, Kimi K3, carries 2.8 trillion parameters and is considered a frontier-level model, competitive with leading US systems.