Starting July 10, Alibaba is imposing a complete ban on its employees' use of Anthropic's AI tools, including Claude Code. The decision comes amid growing concerns over data security and espionage. The Chinese tech giant is officially transitioning its developers to its in-house assistant, Qoder, created by Alibaba itself.
Claude Code Under Suspicion: Data Collection and "Black Marks"
A few days before the ban, Alibaba engineers discovered that Claude Code was collecting information about users' time zones and proxy servers. Moreover, it was found that Anthropic's tool was adding hidden tags to requests sent to the company's servers. This was interpreted as a potential mechanism for corporate data leakage and even as a covert "backdoor."
These concerns were just the tip of the iceberg. As early as June, Anthropic sent a letter to the U.S. Senate Banking Committee, accusing Alibaba of attempting to illegally extract capabilities from its AI model. Anthropic described this incident as the largest known attempt at distillation—a process where the output of one model is used to train a competing AI at a much lower cost. In essence, this is intellectual property theft in the world of artificial intelligence.
Geopolitical Context: The AI Arms Race
The conflict between the companies is merely a reflection of the larger U.S.-China rivalry in the AI sector. Anthropic itself has previously warned that American leadership in this field could be lost by 2028 unless export rules are tightened and efforts to combat distillation are intensified. Alibaba's ban on Claude Code is not just a corporate decision but part of China's targeted state policy to strengthen control over AI.
The "Qinglang" Campaign: Cleaning Up the AI Space
Alibaba's decision fits seamlessly into the large-scale state campaign "Qinglang" ("Clean Sky"). In its first phase, China's internet regulator has already removed over 14,000 AI products. During inspections, more than 26,000 accounts were blocked, and millions of web pages and files violating the new requirements were deleted. Alibaba, like other major players, has adjusted its services in line with the tightened regulations.
As a result, Alibaba developers lose access to one of the leading programming tools from the U.S., while Anthropic loses opportunities to work with a massive Chinese technology platform. This is a double-edged loss that only deepens the technological divide between the two superpowers.
Cryptalist Commentary: This incident is a clear signal that the era of global technological exchange in AI is coming to an end. We are witnessing the formation of two isolated ecosystems: American and Chinese. For investors, this means that bets on companies capable of building independent AI infrastructure will be justified. Distillation and espionage are becoming integral parts of the new Cold War in high technology.