Alibaba to Ban Claude Code Over Alleged Backdoor Risk, Source Says
July 3, 2026 – 7:30 am
Alibaba will bar its employees from using Anthropic’s Claude Code inside workplace environments starting July 10, according to a person familiar with the matter cited by Reuters. The stated reason is an alleged backdoor built into the coding tool, though Alibaba has not confirmed the move publicly and did not immediately respond to a request for comment.
The ban was initially reported by Chinese financial outlet Yicai before Reuters corroborated it through its source. It comes at a pivotal moment between Anthropic and Alibaba, already embroiled in accusations of model theft and, now, an alleged spying mechanism within Claude’s own tooling.
Claude Code is Anthropic’s command-line coding agent, used by developers to write and debug software from a terminal. Its workplace ban at a company the size of Alibaba is significant.
The alleged backdoor, detailed in a Reddit post published on June 30 by user LegitMichel777, claims to have discovered a covert mechanism within Claude Code designed to detect and track users linked to Chinese corporate networks, cloud regions, or AI labs, including Alibaba, Baidu, ByteDance, and Moonshot AI.
Anthropic has not publicly addressed these allegations. A member of its Claude Code team, Thariq, responded on social media that the mechanism was intended to prevent account reselling and model distillation, and it would be removed in the next release. This fix was reportedly underway by July 1.
This incident is part of a broader dispute. In a letter to US senators dated June 10, Anthropic accused operators linked to Alibaba’s Qwen AI lab of employing nearly 25,000 fraudulent accounts to extract Claude’s capabilities, generating over 28.8 million exchanges between April 22 and June 5.