Back to News
PolicyAI Understanding briefing

China probes DeepSeek and Moonshot over alleged data leaks to Anthropic

The Cyberspace Administration of China has opened an investigation into DeepSeek and Moonshot AI, alleging they routed sensitive user data to Anthropic's servers, according to The Information.

4 min readRead the linked source
Source-provided image accompanying China probes DeepSeek and Moonshot over alleged data leaks to Anthropic
Source referenceSource recorded
Publisher
decrypt.co
Source link
decrypt.cohttps://decrypt.co/379120/china-probes-deepseek-moonshot-data-leaks-anthropic-claude
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Start here

Key terms

API (Application Programming Interface)
A structured way for one software system to send requests to and receive responses from another system.
AI Governance
Policies, standards, and oversight mechanisms that guide how AI is developed and used in society.
Distillation
Compressing knowledge from a large teacher model into a smaller student model.
Test yourselfAI Ethics Quiz

What happened

Chinese regulators have launched a formal investigation into DeepSeek and Moonshot AI, alleging that both companies secretly routed sensitive user data, including police and military information, to Anthropic's servers. This action follows Anthropic's recent threat intelligence report accusing several Chinese labs of illicit .

The Cyberspace Administration of China (CAC) has opened an investigation into DeepSeek and Moonshot AI, according to a report by The Information. The investigation stems from claims that both companies secretly routed sensitive user data to Anthropic's servers. Regulators reportedly visited the offices of both companies to interview executives and staff, seeking to determine whether sensitive police, military, and state-linked corporate data ended up on U.S. servers.

The probe follows the publication of Anthropic's fourth threat intelligence report on September 10, a 154-page document that accused seven China-based labs, including DeepSeek and Moonshot, of 'illicit .' Anthropic alleged that these labs were using fraudulent accounts to reroute queries to Claude, effectively using Anthropic's infrastructure to train cheaper rival models. While the CAC initially summoned representatives from all seven named companies, it has since narrowed its focus specifically to DeepSeek and Moonshot. No penalties have been announced as of the report's publication.

The timing of this investigation is particularly sensitive for both companies. DeepSeek is scheduled to brief the United Nations Security Council on AI risks this week, sharing a stage with Anthropic CEO Dario Amodei. This event occurs just one day before President Donald Trump's September 24 summit with Xi Jinping, where AI is reportedly on the agenda. Additionally, Moonshot AI confidentially filed for a Hong Kong IPO on September 3, targeting a $50 billion valuation. Any disclosed regulatory investigation would need to be included in its prospectus before the listing can proceed, potentially impacting its financial trajectory.

Source details: decrypt.co

Why it matters

This investigation marks a significant escalation in the regulatory and geopolitical tensions surrounding AI data sovereignty and cross-border data flows. It directly impacts the operational security and public relations of major Chinese AI firms, particularly Moonshot AI, which is currently preparing for a high-profile IPO. The timing coincides with high-level diplomatic meetings between the U.S. and China, potentially influencing broader discussions.

This development represents a concrete regulatory response to allegations of cross-border data leakage involving major AI models. It highlights the growing friction between U.S. AI companies and Chinese regulators regarding data sovereignty and the integrity of AI training data. The involvement of sensitive state-linked data elevates the issue from a commercial dispute to a national security concern.

For the AI industry, this probe underscores the risks associated with relying on foreign AI infrastructure for model training or inference. It may lead to stricter data localization laws in China or increased scrutiny of API usage by domestic firms. The potential impact on Moonshot's IPO and DeepSeek's diplomatic engagements demonstrates how regulatory actions can have immediate financial and geopolitical consequences for AI companies.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Interactive Concept Check+10 Points
AI Ethics Quiz

Impossibility results in algorithmic fairness (e.g. Kleinberg et al., Chouldechova) show what?

What to watch next

Monitor for official statements from the Cyberspace Administration of China regarding the investigation's findings, any penalties imposed on DeepSeek or Moonshot, and how this regulatory action affects Moonshot's upcoming Hong Kong IPO prospectus and DeepSeek's upcoming UN Security Council briefing.

Watch for official announcements from the CAC regarding the outcome of the investigation, including any fines, operational restrictions, or public statements clarifying the nature of the data flows. The absence of penalties so far suggests the investigation is still in its early stages.

Monitor the progress of Moonshot AI's IPO, specifically whether the regulatory investigation is disclosed in its prospectus and how investors react to this risk factor. Additionally, observe the dynamics of the upcoming U.S.-China summit and the UN Security Council briefing, as the investigation may influence the tone of diplomatic discussions on .

Related guides & quizzes

AI EthicsAI Models ExplainedFuture of AITest what you know — try a free AI quizLook up an AI term in our glossary
Found this useful?