Chinese technology giant Alibaba has imposed a strict ban on employees using the Claude Code tool from the American company Anthropic. This decision is a direct response to identified mechanisms that, according to developers, allowed Anthropic to track and identify users associated with China.
The Core of the Conflict: From Distillation to a "Backdoor"
Claude Code is an AI assistant for developers that, despite Anthropic's official restrictions on access from China, has gained popularity among Chinese programmers. Alibaba's management has instructed its employees to switch to their own programming platform, Qoder, effectively recognizing Claude Code as a threat to corporate security.
The situation escalated after Anthropic accused Alibaba of "distillation"—a method where a weaker model is trained on the responses of a more powerful one. Anthropic claims this process helped the Chinese side gain accelerated access to the advanced capabilities of their Mythos Preview model. These accusations were outlined in a letter to two U.S. senators; Alibaba has not publicly commented on them.
The ban by Alibaba came a few days after developers reported built-in mechanisms in Claude Code for checking the user's environment. The tool analyzed the time zone and proxy data, inserting hidden markers into requests to Anthropic's servers. An Anthropic representative called this feature an "experiment launched in March" to protect against unlicensed resellers and model distillation.
Anthropic Closes Loopholes, China Goes Offline
Anthropic is tightening restrictions, closing loopholes that allowed users associated with China to access Claude via VPNs, cloud platforms, and proxies. The company prohibits access to Claude both within China and for Chinese companies abroad, citing national security concerns.
Against the backdrop of American developers' attempts to prevent unauthorized access and distillation, Chinese cloud and AI companies are actively transitioning to domestic and open-source models—such as DeepSeek, Alibaba Qwen, Moonshot, and Zhipu. At the same time, Chinese AI models are simultaneously penetrating the U.S. market, causing concern among some industry experts.
Expert Opinion: This incident is not just a corporate dispute but a clear marker of the fragmentation of the global AI industry. We are witnessing the formation of two isolated technology blocs, where development tools become instruments of geopolitical pressure. For the crypto and Web3 sectors, which historically value decentralization and openness, this is an alarming signal: the future may be far less global than we are accustomed to thinking.