Anthropic has begun mass-blocking user accounts that access the Claude chatbot via VPN. Reports from users in Hong Kong indicate that accounts are being terminated without prior warning due to suspicious activity detected by the company's security systems. This move signals a stricter enforcement of existing regional restrictions, as the service is not officially supported in Hong Kong, Macau, or mainland China.
The crackdown follows recent tensions between Anthropic and Chinese AI developers. Last month, Anthropic accused seven Chinese laboratories of using industrial-scale model distillation to train their own systems by scraping Claude's responses. Companies such as Moonshot AI and DeepSeek allegedly bypassed geographical blocks using thousands of fake accounts and proxy networks to mask their origin.
Data shows that Moonshot AI alone generated over 23 million requests to Claude between May and July, masquerading as traffic from Singapore and Japan. While Beijing has denied these allegations, the tightening of security protocols by Anthropic raises concerns that users in other unsupported regions, including Russia, may face similar account terminations as the company enforces its access policies.
