Alibaba Blocks Claude Code After Security Concerns Deepen in Growing AI Dispute

The ongoing rivalry between major artificial intelligence companies in China and the United States has taken another turn. Chinese technology giant Alibaba has reportedly prohibited employees from using Anthropic’s Claude Code inside the company following an internal security review that raised concerns about the software.

According to reports from Chinese media, Alibaba has added Claude Code to its list of restricted software. Employees have reportedly been advised to use Qoder as an alternative coding assistant for work related tasks. The decision is said to have been made after an internal assessment flagged what the company described as potential security risks.

The reported review allegedly suggested that Claude Code could contain possible backdoor mechanisms, leading to concerns about unauthorized access to sensitive corporate information. Alibaba has not publicly released technical evidence to support these claims, and Anthropic has not issued an official response to the allegations.

The latest controversy follows claims that first gained attention through a Reddit post. A user who claimed to have reverse engineered Claude Code alleged that version 2.1.91 included hidden checks related to proxy configurations, system time zones and network environments. The post further claimed that the software attempted to identify whether it was operating within Chinese corporate networks or connected to major Chinese artificial intelligence laboratories.

According to those allegations, the software was capable of detecting whether a user was accessing the internet through a Chinese proxy or connecting to networks associated with companies such as Alibaba, Baidu, ByteDance and Moonshot AI. The Reddit post also claimed the relevant code had been embedded within the application in a way that was difficult to identify. These claims have not been independently verified.

The discussion has fueled wider concerns about software transparency, enterprise security and the handling of sensitive corporate data. While such allegations have attracted significant online attention, no official findings confirming the presence of hidden surveillance functions have been made public.

The latest move also comes against the backdrop of an earlier dispute between the two companies. In June, Anthropic accused Alibaba’s Qwen team of operating around 25000 fake accounts to scrape Claude generated outputs, claiming the activity resulted in approximately 28.8 million interactions. That accusation further escalated competition between the two companies, although Alibaba has not publicly accepted those claims.

With Alibaba now reportedly blocking Claude Code for internal use, the disagreement has evolved beyond competitive accusations into a broader debate over software security and trust. The episode highlights the increasingly tense relationship between leading AI developers from China and the United States as competition intensifies across the global artificial intelligence industry.

Neither Alibaba nor Anthropic has announced any legal action related to the latest allegations. Until further technical evidence or official statements emerge, many of the claims surrounding Claude Code remain disputed, leaving industry observers closely watching how the situation develops.