Blocks Claude Code Use Over AI Security Concerns 

TLDR

• Alibaba bans Claude Code at work over alleged AI security concerns.

• Staff must use Qoder as Alibaba tightens internal AI coding controls.

• Anthropic’s China access limits add pressure to enterprise AI use.

• Claude Code scrutiny grows over environment checks and compliance risks.

• Alibaba’s move reflects China’s shift toward domestic AI platforms.

Alibaba has ordered staff to stop using Anthropic’s Claude Code at work, citing security concerns around the coding tool. The move adds pressure to an already tense dispute between Alibaba and Anthropic over AI access, compliance, and model protection. The restriction also shows how Chinese technology firms now handle foreign AI tools with tighter internal controls.

Alibaba Moves Staff Toward Internal Coding Tools

Alibaba will block employee use of Claude Code in workplace systems from July 10, according to a person familiar with the decision. The order applies to internal work environments and directs staff toward Alibaba’s own coding platform, Qoder. However, the company has not issued a public statement on the ban.

The decision followed concerns over features that could inspect user environments and detect China-linked access patterns. Developers recently reported that Claude Code checked details such as time zones and proxy-related signals. Therefore, Alibaba treated the tool as a potential security and compliance risk.

Claude Code has gained use among programmers, including some users in China, despite Anthropic’s access restrictions. Some users route activity through overseas servers to appear outside restricted regions. As a result, companies face more direct legal and operational exposure than individual developers.

Dispute With Anthropic Adds Pressure

The ban comes after Anthropic accused Alibaba of trying to extract capabilities from its AI systems. Anthropic described the activity as a distillation effort, where one model learns from another model’s outputs. The claim has increased tension between the two companies and wider U.S.-China AI competition.

Alibaba has not publicly responded to the accusation, and Anthropic has not confirmed new details on the workplace ban. Still, the dispute highlights growing concern over advanced model access and unauthorized reuse. It also shows how AI companies now guard their systems against resale and replication.

Anthropic said an internal feature in Claude Code aimed to prevent account abuse by unauthorized resellers. It also linked the measure to efforts against model distillation. However, Alibaba appears to view those controls as unacceptable inside corporate development environments.

China AI Firms Shift Toward Local Models

Alibaba’s decision fits a broader shift among Chinese cloud and AI companies toward domestic systems. Firms in China increasingly use models such as Qwen, DeepSeek, Moonshot and Zhipu. This shift reduces reliance on U.S. AI providers and limits exposure to foreign licensing limits.

The move also follows separate restrictions involving Claude models in Hong Kong financial institutions. Reports said JPMorgan and Goldman Sachs limited Claude access because of Anthropic’s regional terms. Those actions raised wider questions about where enterprise AI systems can legally operate.

For Alibaba, the ban protects internal code, employee workflows, and corporate data from disputed third-party controls. It also supports the company’s own AI development stack at a sensitive time. The case now reflects a sharper divide between U.S. model safeguards and Chinese enterprise compliance needs.