Moonshot AI—Moonshot AI allegedly routed 23 million Kimi user queries to Anthropic's Claude via hidden proxy network, prompting Chinese CAC probe
Anthropic reported that Moonshot AI covertly rerouted Kimi chatbot user queries to Claude and used the exchanges to train its own models, without disclosing to users they were being answered by a third-party US model. Anthropic said Moonshot routed over 23 million exchanges to Claude between May and July 2026, with nearly 300,000 requests funneled through a proxy network of roughly 5,380 accounts largely registered via servers in Singapore and Japan to obscure the data's origin. China's Cyberspace Administration (CAC) opened a formal data-security probe into Moonshot specifically, examining whether sensitive Chinese user and state-linked data reached US servers without proper safeguards. The probe came as Moonshot was reportedly pursuing a confidential Hong Kong IPO filing.
Scoring Impact
| Topic | Direction | Relevance | Contribution |
|---|---|---|---|
| Corporate Transparency | -against | secondary | -0.50 |
| Data Security | -against | secondary | -0.50 |
| User Privacy | -against | primary | -1.00 |
| Overall incident score = | -0.572 | ||
Score = avg(topic contributions) × significance (high ×1.5) × confidence (0.57)
Evidence (1 signal)
SCMP: Moonshot secretly routed Kimi user requests to Claude via proxy network, Anthropic claims
Anthropic's September 10, 2026 report accused Moonshot AI of routing over 23 million Kimi user exchanges to Claude via a proxy network of roughly 5,380 accounts, mostly registered via servers in Singapore and Japan; China's Cyberspace Administration subsequently opened a formal probe narrowed to Moonshot and DeepSeek.