Anthropic: For national security reasons, the company does not currently offer commercial access to Claude in China

Anthropic: For national security reasons, the company does not currently offer commercial access to Claude in China

Anthropic Restricts Chinese Entities from Claude AI Amid National Security Concerns

Anthropic has implemented a policy prohibiting Chinese-controlled companies from accessing its Claude AI models, citing legal, regulatory, and security risks. The update to its terms of service explicitly bars firms majority-owned or controlled by Chinese entities—regardless of geographic location—from using Claude models, including Claude 3.5 Sonnet, and associated developer tools according to the company's announcement. This restriction extends to subsidiaries, joint ventures, and foreign-incorporated divisions, effectively cutting off access for major Chinese tech firms like ByteDance, Tencent, and Alibaba as reported by industry sources.

The company's decision emphasizes preventing adversarial use of its technology by authoritarian regimes, stating that Chinese entities "could use our capabilities to develop applications and services that ultimately serve adversarial military and intelligence services" according to Anthropic's public statement. While Anthropic acknowledges the policy will reduce revenue by "low hundreds of millions of dollars," it maintains that safeguarding sensitive AI capabilities is critical to mitigating strategic risks as detailed in their terms update.

This move marks a shift from geographic-based restrictions to ownership-focused policies, distinguishing it from prior actions that blocked Chinese users based on location according to industry analysis. In response, Chinese AI firms have accelerated efforts to capture displaced clients. For example, Zhipu AI launched a migration toolkit offering its GLM-4.5 API at a lower cost, with incentives such as 20 million free tokens and higher throughput compared to Claude as reported in tech coverage. Alibaba also previously introduced migration programs for developers affected by OpenAI's restrictions, signaling competitive dynamics in the AI sector according to industry reports.

The policy underscores growing tensions between U.S. tech firms and foreign entities over AI governance. While Anthropic's stance aligns with broader U.S. national security priorities, it raises questions about balancing innovation, market access, and geopolitical risks. As enterprises adapt to these restrictions, the financial and operational implications for both providers and users will likely remain a key focus for investors and policymakers.