Chinese tech giant Alibaba will completely ban its employees from using Anthropic products in work processes starting July 10. Instead, the company recommends switching to its own Qoder platform. The ban covers all models in the Claude line — Sonnet, Opus, and Fable — as well as the developer tool Claude Code, which is now classified as high-risk software.
What triggered this?
Alibaba's decision was a direct response to recent events. A few days before the announcement, developers discovered mechanisms in Claude Code that analyzed the user's environment in the background. Specifically, the tool read data about the local time zone and checked proxy server addresses and API endpoints for matches with keywords related to Chinese cloud providers and AI companies. Upon detecting matches, Claude Code altered the date format and punctuation in the system prompt without sending explicit telemetry.
Claude Code team member Tariq Shihhipar explained that this was "an experiment launched in March" aimed at combating unauthorized account resale and protecting against model distillation. He assured that the mechanism would be removed in the next update. However, for Alibaba, this incident was the last straw in a series of tense relations with Anthropic.
Geopolitical backdrop and distillation accusations
Alibaba's ban unfolds against the backdrop of Anthropic's tightening policies toward China. As early as September 2025, the company updated its sales rules, effectively blocking access to its services from China and Hong Kong, citing security risks and potential forced data transfer to intelligence agencies. Previously, Anthropic accused Chinese startups DeepSeek, Moonshot, and MiniMax of a large-scale campaign to distill Claude, generating over 16 million interactions through fraudulent accounts.
In June, Anthropic sent a letter to the U.S. Senate Banking Committee, reiterating its grievances against Alibaba and its AI division Qwen, calling it the largest known attack of its kind. According to sources, this letter was the direct trigger for the internal ban at Alibaba.
Black market and technological lag
Despite strict restrictions, demand for Claude in China remains high. As investigations show, Anthropic accounts are actively sold on platforms like Taobao, Xianyu, and Telegram. Intermediaries host servers in supported countries and resell access to local users. Researchers note anomalously high usage of Claude in Singapore, indicating large-scale traffic proxying from China.
The reason for this popularity is simple: according to experts, Chinese AI models still lag behind American ones by six to nine months, especially in code generation tasks. This technological superiority drives Chinese developers to seek workarounds, despite all the risks.
Expert analysis
From my perspective, this incident is not just a corporate conflict but a symptom of deep fragmentation in the global AI market. Alibaba, on one hand, is forced to protect its intellectual property and reputation, and on the other, is trying to stimulate the development of its own Qoder ecosystem. However, a complete ban on using the best Western tools could have the opposite effect, slowing innovation within the company. In the long term, we will likely see a strengthening of the "digital iron curtain" in the AI sphere, where each side will strive for full technological sovereignty, sacrificing efficiency in the short term.