Chinese Regulators Probe DeepSeek and Moonshot Over Alleged Data Routing to Anthropic

The Cyberspace Administration of China is investigating artificial intelligence developers DeepSeek and Moonshot AI over user data that may have reached U.S.-based Anthropic, according to a report by The Information covered by The Next Web (https://thenextweb.com/news/china-cac-probe-deepseek-moonshot-anthropic-data). Regulators have questioned company personnel as part of an active inquiry, though no formal penalties have been announced.

The inquiry follows a threat intelligence report published by Anthropic on September 10, which stated that seven Chinese artificial intelligence labs had been relaying user requests to Claude. In that report, Anthropic publicly named six companies: Moonshot AI, DeepSeek, Zhipu, Xiaomi, SenseTime, and MiniMax. Chinese authorities reportedly summoned all seven organizations before narrowing their focus to DeepSeek and Moonshot.

Anthropic's documentation alleged that Moonshot routed more than 23 million interactions to Claude between May and July, including nearly 300,000 requests across a single 10-day span using 5,380 fraudulent accounts. The requests were allegedly forwarded silently and returned to users as if generated by Moonshot's Kimi model.

The investigation reportedly intensified around specific data transfers detailed in the reporting. According to The Information, one example involved a user assessed as likely affiliated with the People's Liberation Army who asked Kimi to analyze surveillance footage across police cameras in Chengdu, including feeds near defense-linked institutes and military facilities. Moonshot allegedly forwarded the query and footage to Claude without the user's knowledge.

While Anthropic's initial complaint focused on unauthorized extraction of model outputs, Chinese regulators are examining cross-border data transfer rules and whether sensitive domestic data was moved abroad. The regulatory scrutiny comes despite China's foreign ministry publicly dismissing a separate U.S. intelligence advisory on AI distillation on September 10.

The regulatory review coincides with significant milestones for both companies. DeepSeek is scheduled to brief the United Nations Security Council this week on international security risks from artificial intelligence, while Moonshot is preparing for an initial public offering in Hong Kong. Neither company has commented publicly on the ongoing investigation.