Alibaba Bans Claude Code Over Anthropic Spyware Fears

Alibaba has instructed employees to stop using Claude Code, Anthropic's AI-powered coding assistant, according to a report by the South China Morning Post. The Chinese tech giant cited concerns that the tool could act as spyware, potentially exposing proprietary source code and internal data to Anthropic's servers in the United States. The ban applies company-wide and follows a broader pattern of Chinese organizations growing wary of American AI products amid escalating geopolitical tensions.

What Prompted the Ban

Claude Code operates as an agentic coding tool that can read files, execute commands, and interact with local development environments. That level of system access has raised red flags at Alibaba, where security teams reportedly flagged the risk that sensitive code repositories could be transmitted outside China. The company has not issued a detailed public statement, but internal communications reviewed by the South China Morning Post indicate the prohibition is firm. As we covered earlier this year, Alibaba has formally banned staff from using Anthropic's Claude Code tool, making this one of the more high-profile corporate rejections of a Western AI product in China's tech sector.

The security concern is not purely hypothetical. Claude Code's design gives it broad access to a developer's local environment by necessity. That functionality is core to what makes it useful, but it also means the tool sends substantial context, including code snippets and file structures, to remote inference servers. For a company like Alibaba, which holds vast amounts of commercially sensitive and potentially regulated data, that pipeline represents a genuine compliance and espionage risk, regardless of whether any actual data collection is occurring.

The concern is less about proven wrongdoing and more about the structural risk of routing proprietary code through a foreign company's servers during a period of significant US-China trade and technology friction.South China Morning Post analysis

A Wider Chill on Western AI Tools in China

Alibaba's decision does not stand in isolation. Goldman Sachs blocked Claude access for Hong Kong staff earlier this year under different but related pressures, illustrating how geopolitical friction is shaping corporate AI policy on both sides of the Pacific. Chinese firms in particular face pressure from regulators to keep sensitive data onshore, and using a tool developed and operated by a US AI lab complicates that obligation considerably.

The timing carries an extra layer of irony. Alibaba has invested heavily in positioning its own Qwen models as credible alternatives to Western frontier AI. The company recently demonstrated that Alibaba's Qwen3.7-Max can run autonomously for 35 hours and is compatible with Claude Code, signaling ambitions to compete in agentic AI workflows. Banning Claude Code internally while simultaneously building tools designed to interface with it suggests Alibaba sees the competitive landscape as fluid, and is hedging accordingly.

Anthropic has not publicly responded to the ban. The company, which positions itself as a safety-focused AI developer, has generally been transparent about how its products handle data, but that transparency may carry less weight in an environment where the underlying trust between US and Chinese technology sectors has eroded significantly. For context on how Anthropic frames its own mission and practices, our background on Anthropic covers the company's founding principles and research priorities.

What This Means for Claude Code's Global Reach

Claude Code has grown quickly since its launch, attracting developers who want an AI assistant deeply integrated into their workflow rather than a simple autocomplete engine. But that depth of integration is exactly what makes enterprise adoption complicated, especially in markets with strict data localization requirements or geopolitical sensitivities around US technology. Anthropic will need to address these concerns directly if it wants meaningful corporate uptake outside the US and Europe.

For now, the Alibaba ban is a concrete signal that even technically sophisticated companies with AI ambitions of their own are drawing clear lines around which foreign tools their staff can use. Whether Anthropic pursues a data-residency arrangement or some other structural change to satisfy organizations in China and elsewhere remains to be seen. The episode adds another chapter to a growing story about the fragmentation of the global AI tools market along national and regulatory lines.