Moonshot AI, the developer of the Kimi chatbot, faces allegations from Anthropic and US officials regarding the unauthorized distillation of AI models. Anthropic claims that on September 10, 2026, Moonshot routed approximately 300,000 user requests through 5,380 proxy accounts in Singapore and Japan to extract data from its Claude Opus model.
In July 2026, US officials, including OSTP Director Michael Kratsios and Treasury Secretary Scott Bessent, alleged that Moonshot used data from Anthropic’s Fable model to develop its Kimi K3 system. Authorities also claim the company accessed restricted Nvidia servers in Thailand, potentially leading to future sanctions.
OpenAI has not issued formal accusations but acknowledged the performance of Kimi K3, which launched on July 27, 2026, with 2.8 trillion parameters. The model's licensing terms require attribution for commercial products and include revenue-sharing agreements of up to 30% for large-scale users.
Founded in March 2023, Moonshot AI introduced its Kimi chatbot in November 2023. The company's latest Kimi K3 model is positioned as a frontier-level system designed to compete with major US-based AI technologies.
