Anthropic, a leading AI developer, officially updated its sales policy on September 5, 2025, to restrict access to its services for companies that are more than 50% owned or controlled by entities from China. As reported by the Financial Times and Bloomberg, this decision extends to subsidiaries of Chinese firms, even if they are registered in other countries. Anthropic explains this move as a necessity to comply with the spirit of U.S. export controls and to mitigate national security risks. The company fears that its advanced models, such as Claude, could be used for military or intelligence purposes, and that Chinese laws could compel local companies to share data with security services. This step sets a significant precedent: a major private AI lab is voluntarily imposing strict restrictions, formalizing the technological divide and increasing pressure on other market players to adopt similar measures.
Anthropic Restricts AI Service Access for China-Controlled Companies
