Alibaba has instructed employees to stop using Anthropic’s Claude Code at work, according to a person familiar with the matter, as per a recent report by Reuters.
As per the report, this is part of an escalating dispute between the two firms as the US and China jostle for dominance in artificial intelligence.
The ban, first reported by Chinese media, follows Anthropic’s accusation last month that Alibaba had been illicitly extracting the capabilities of its Claude models through “distillation,” a technique that trains a smaller system on a more powerful one’s outputs. In a letter seen by Reuters, Anthropic told two US senators that Alibaba’s effort was designed to accelerate its own progress toward matching the company’s advanced Mythos Preview capabilities.
The latest trigger came after developers discovered that Claude Code inspects user environments, checking details such as timezone and proxy settings, and embeds subtle markers into prompts sent to Anthropic’s servers. An Anthropic employee confirmed on X on Tuesday that the functionality was “an experiment we launched in March” aimed at preventing account abuse by unauthorised resellers and protecting against model distillation.
A person familiar with Alibaba’s order told Reuters that employees are being directed to use the company’s own coding platform, Qoder, instead. The person noted that while Anthropic’s China-related restrictions are hard to enforce on individuals who can route traffic through US-based servers, enterprises face greater legal and compliance exposure.
Alibaba has not publicly addressed Anthropic’s allegations, and both companies declined to comment to Reuters. The row comes as Chinese cloud and AI firms increasingly turn to domestic and open-source models, including DeepSeek, Alibaba’s own Qwen, Moonshot, and Zhipu, even as Chinese AI models gain ground in the US market.