According to a person familiar with the matter, the internal directive followed growing scrutiny over features within Claude Code that can reportedly help identify users with links to China. The individual, who was not authorised to speak publicly, said employees have instead been directed to use Alibaba's in-house AI coding platform, Qoder.
Neither Alibaba nor Anthropic immediately responded to requests for comment, while Alibaba has yet to publicly address allegations recently made by the U.S.-based AI company.
The latest development comes amid an increasingly intense rivalry between Chinese and American artificial intelligence firms as competition accelerates over advanced AI technologies.
Claude Code, Anthropic's AI-powered programming assistant, has gained popularity among software developers, including users in China, despite the company's restrictions on access by individuals and organisations based in the country.
Anthropic alleges model distillation
The dispute intensified after Anthropic alleged that Alibaba attempted to extract the capabilities of its Claude AI models through a process known as "distillation"—a machine learning technique in which a smaller model is trained using the outputs of a more advanced system.
According to a letter reportedly sent by Anthropic to two U.S. senators and seen by Reuters, the company claimed the alleged activity was intended to accelerate the development of AI systems capable of matching the performance of its advanced Mythos Preview model.
Anthropic also disclosed that Claude Code recently included experimental mechanisms designed to inspect aspects of users' computing environments, including timezone settings and proxy-related information, while embedding subtle markers into prompts transmitted to the company's servers.
Responding to concerns raised by developers, an Anthropic employee explained the purpose of the feature in a post on X.
"An experiment we launched in March"
The employee said the mechanism was intended to prevent account abuse by unauthorised resellers and help protect the company's AI models from distillation efforts.
Compliance concerns drive policy
The source familiar with Alibaba's decision told Reuters that while Anthropic's restrictions on users in China can often be bypassed by individuals using servers located in the United States, companies face far greater legal and regulatory risks when using restricted AI services.
According to the source, those compliance considerations played a significant role in Alibaba's decision to prohibit employees from using Claude Code internally.
The development reflects the broader technological competition between the United States and China, where access to advanced AI models has become increasingly sensitive amid tightening export controls and national security concerns.
As U.S. AI companies strengthen safeguards against unauthorised access, resale and model distillation, many Chinese technology firms have accelerated investment in domestic and open-source alternatives, including Alibaba's Qwen, DeepSeek, Moonshot AI and Zhipu AI.
At the same time, Chinese-developed AI models have continued expanding into international markets, including the United States, a trend that has attracted growing attention from policymakers and industry analysts concerned about the future balance of global AI leadership.
Alibaba's reported ban on Claude Code was first disclosed by Chinese media outlets and has since drawn wider attention as competition between leading AI developers continues to intensify.