Alibaba will block employee use of Anthropic’s Claude Code in workplace environments from July 10, citing alleged security risks linked to embedded backdoors, according to sources familiar with the matter.
The move, first reported by Chinese outlet Yicai and confirmed by Reuters, follows a recent escalation in tensions between the two companies. Last month, Anthropic accused operators associated with Alibaba of running a massive “distillation” campaign that used thousands of fake accounts to extract capabilities from its Claude AI models — an allegation Alibaba has not publicly addressed.
Alibaba internally directed staff to switch to its in-house coding platform, Qoder, and reportedly views Claude Code’s environment-checking features — such as inspecting time zones and proxy signals — as unacceptable security and compliance risks inside corporate development settings. The tool was designed to prevent account abuse by unauthorized resellers, but Alibaba perceives those controls as a direct threat to data sovereignty and internal workflows.
The ban also reflects a broader pivot among Chinese tech firms toward domestic AI models, including Qwen, DeepSeek, Moonshot, and Zhipu, reducing reliance on U.S. providers. This case intensifies the ongoing U.S.-China rift over AI model access, compliance, and intellectual property protection.
