Alibaba has instructed its employees to discontinue use of Anthropic’s Claude Code programming assistant, effective July 10, classifying the external software as high-risk and mandating a transition to its proprietary internal development platform, Qoder. The directive coincides with Anthropic’s ongoing enforcement of geographic restrictions that block Chinese companies and their foreign-owned subsidiaries from accessing its AI models. Prior to this internal compliance measure, Anthropic had been actively closing access loopholes, including a March deployment of a detection mechanism designed to identify Chinese users attempting to bypass regional limits. Anthropic representative Thariq Shihipar clarified that the tool was initially deployed to prevent unauthorized reselling and protect against model distillation, though the company confirmed it had since implemented stronger mitigations and intended to retire the feature. Alibaba’s decision to restrict the developer tool reflects heightened corporate scrutiny over foreign AI dependencies, particularly within software engineering workflows. By consolidating development operations around Qoder, the tech giant aims to maintain regulatory compliance while safeguarding internal code and intellectual property from cross-border access policies. The ban highlights the growing operational friction between global AI model providers and major Chinese technology firms, as both navigate tightening geopolitical controls, security protocols, and the strategic push toward domestic AI infrastructure.