Alibaba bans Claude for staff – Anthropic didn't want them using it anyway
Summary
Chinese tech giant Alibaba has ordered staff to stop using Anthropic's Claude Code, after it was found to be flagging users connecting from China. But Anthropic is already trying its best to stop Chinese firms from using Claude at all, and accuses Alibaba of running large "distillation" campaigns against it, saying it deployed around 25,000 fake accounts to train its own models on Claude.
Discussion on
Trending posts from X.
Anthropic thought the story was:
— 𝘊𝘰𝘳𝘳𝘪𝘯𝘦 (@OopsGuess) July 3, 2026
“We ban Chinese users.”
Then Alibaba replied:
“No. We ban you.”
That is the real twist.
Users did not find a hidden government-monitoring backdoor inside “authoritarian AI.”
They found it inside the priesthood of “democratic AI safety.”
Claude… https://t.co/5qYww9ezMB pic.twitter.com/0za1JyE1We
i think part of it is just that there is so much demand for frontier tokens from openai and anthropic that it strains supply
— FleetingBits (@fleetingbits) July 5, 2026
in addition, a lot of businesses are not comfortable using open source chinese models for customer perception reasons / want to buy from aws, gcp, azure
事件升级！阿里巴巴被曝，内部宣布全面禁用 Claude，全体员工被要求卸载 Anthropic 旗下所有产品。
— AB Kuai.Dong (@_FORAB) July 3, 2026
涵盖 Sonnet、Opus、Fable 等多个模型，以及 Claude Code 在内的 Agent 产品，7 月 10 日正式生效。 https://t.co/3Mj1ArgpZL pic.twitter.com/zIfgwQrjM5