The Cyberspace Administration of China has launched an investigation into DeepSeek and Moonshot AI following allegations that both companies secretly routed sensitive user data to Anthropic's servers. The probe, reported by The Information, stems from Anthropic's September 10 threat intelligence report which accused seven Chinese labs of illicit distillation through fraudulent accounts. Regulators have visited the offices of both firms to interview executives and determine if police, military, or state-linked corporate data was transferred to U.S. infrastructure. While the CAC initially summoned all seven named entities, it has narrowed its focus specifically to DeepSeek and Moonshot without announcing penalties.
Anthropic alleges Moonshot processed more than 23 million exchanges via 5,380 fraudulent accounts, while DeepSeek generated over 12.1 million exchanges in a 14-day period in July. The timing presents significant challenges for both companies: DeepSeek is scheduled to brief the UN Security Council on AI risks this week alongside Anthropic CEO Dario Amodei, just ahead of a summit between President Donald Trump and Xi Jinping. Meanwhile, Moonshot recently filed confidentially for a $3 billion Hong Kong IPO at a $50 billion valuation, a process that requires disclosure of any regulatory investigations in its prospectus.
This development marks a critical escalation in the technological decoupling between the United States and China, transforming what was previously a commercial dispute over model training practices into a matter of national security oversight. By investigating whether sensitive state-linked data was exposed to U.S. servers, Beijing is signaling that cross-border AI interactions are now subject to strict sovereignty controls. The narrowing of the probe to DeepSeek and Moonshot suggests regulators are targeting high-profile entities with significant international visibility, potentially using these cases to establish precedent for how Chinese AI labs interact with foreign infrastructure. The involvement of the Cyberspace Administration indicates that compliance frameworks for data localization and export controls are being actively enforced against leading domestic innovators.
The immediate institutional risk lies in the collision of regulatory scrutiny with high-stakes diplomatic and financial events. For DeepSeek, the investigation complicates its role as a representative of Chinese AI capabilities at the UN Security Council, potentially undermining its credibility in global governance discussions. For Moonshot, the requirement to disclose ongoing regulatory probes in its IPO prospectus introduces material uncertainty into its capital raising efforts, possibly affecting investor sentiment and valuation stability. Market participants should watch for formal penalty announcements or restrictions on API access, as these would indicate a broader tightening of rules governing the use of foreign large language models by Chinese enterprises.
