Moonshot, DeepSeek secretly routed user requests to Claude, Anthropic claims

American AI giant Anthropic has accused Chinese rivals DeepSeek and Moonshot of covertly routing customer requests to Anthropic's Claude models, presenting the responses as if they originated from their own models. This latest claim adds to growing US criticism of China's AI industry practices.

Earlier this week, three US government agencies, including the NSA and FBI, alleged that six Chinese companies were engaging in an industrial campaign to steal US model outputs, likely with Chinese government involvement. Beijing dismissed these claims, accusing Washington of suppressing China's AI industry and threatening retaliation.

While US companies OpenAI and Anthropic have previously accused their Chinese counterparts of "stealing" their outputs to enhance their models, Anthropic's latest allegations are the first to specifically accuse Chinese firms of routing user requests directly to Claude. This would expose Chinese users' information to Anthropic.

Moonshot, a Kimi model developer, allegedly showed users Claude model outputs and passed them off as its own Kimi outputs. In a 10-day period, Moonshot allegedly routed 300,000 requests to Claude using 5,380 fraudulent accounts, primarily from Singapore and Japan. The company, which has become a leading AI player in China following the success of its K3 model release in July, expects its annual recurring revenue to reach $1 billion by the end of the year.

Anthropic's report also accuses DeepSeek of diverting user requests to Claude, although it does not specify the scale of the operation. Five other Chinese companies, including Alibaba, Z.ai, Xiaomi, and MiniMax, are accused of distilling their Claude models since February 2026. Anthropic claims these practices are illicit due to their industrial scale and the company's prohibition on using its models in China through fake accounts and stolen credit cards.

Legal and AI policy expert Joe Khawam noted that the report raises new data privacy concerns and potential legal implications under both US and Chinese laws. The timing of the reports, he suggested, could signal the Trump administration's intent to raise the issue of Chinese companies' distillation practices at upcoming US-China AI talks.