Anthropic Takes Action Against Covert Access of Claude by Chinese Firms
AI safety company Anthropic has initiated a stringent crackdown to prevent Chinese technology companies from clandestinely accessing its premier Claude AI models. These firms have been using intricate offshore workarounds, marking a significant intensification in the US-China AI technology rivalry.
As reported by the Financial Times on July 3, 2026, major Chinese conglomerates such as Ant Financial (an Alibaba affiliate) and ByteDance (TikTok’s parent company) have been bypassing Anthropic’s rigid geographic restrictions. Ant Financial allegedly provided its employees with corporate Claude accounts linked to its Singapore-based subsidiary, while ByteDance compensated engineers for personal Claude subscriptions accessed via VPNs.
Other companies have been routing access through overseas subsidiaries using cloud platforms like Microsoft Azure. Although these practices do not contravene U.S. or Chinese law, they infringe upon Anthropic’s terms of service. These terms explicitly prohibit Chinese companies and foreign entities under their control from utilizing Claude.
The extent of the violation is astounding. In a letter submitted to the U.S. Senate on June 10, 2026, Anthropic implicated operatives associated with Alibaba’s Qwen AI lab in the largest-ever “distillation attack” against its models. Approximately 25,000 fraudulent accounts generated over 28.8 million interactions with Claude between April 22 and June 5, 2026. These accounts systematically extracted its capabilities to train competing AI systems. Alibaba has refuted these allegations.
Anthropic is currently scrutinizing user accounts for indications such as computer time zones and usage patterns. Simultaneously, it is advocating for stronger legal penalties against AI model distillation attacks in Congress.
Source: BanklessTimes – Anthropic Moves to Block Chinese Firms Using Claude via Offshore Workarounds