China's CAC Probes DeepSeek, Moonshot Over Claude Data Routing

China's Cyberspace Administration is investigating DeepSeek and Moonshot AI over whether sensitive user data reached Anthropic's Claude through covert request routing, The Information reported. The regulator first summoned representatives of all seven Chinese labs named in Anthropic's September threat report before narrowing its focus to the two. The CAC, per the reporting, "is examining whether sensitive Chinese police, military and state-linked corporate data was transferred to U.S. AI models" through the labs' use of Claude.

The trigger was Anthropic's own disclosure. In a 154-page threat intelligence report released September 10, the company alleged that "Moonshot routed more than 23 million exchanges to Claude between May and July," and that DeepSeek routed more than 12.1 million exchanges over a 14-day window in July. Over one 10-day stretch, Anthropic said, Moonshot relayed nearly 300,000 customer requests through 5,380 fraudulent accounts and displayed the answers as though they had come from its own Kimi model.

The material the CAC appears to focus on is the sensitive slice. According to the South China Morning Post, Anthropic said engineers working on a case-management system for a municipal Public Security Bureau sent Claude data used to compare an individual's movements with police records and national identification numbers. Anthropic also said Moonshot's traffic captured "surveillance footage from hundreds of cameras in Chengdu, including cameras near facilities associated with the People's Liberation Army."

For readers following our Anthropic tracker, this turns a distillation dispute into a domestic Chinese enforcement matter rather than a bilateral IP fight. The Information's account frames the probe as landing just ahead of Trump-Xi AI talks.