Alibaba Bans Employees from Using Anthropic's Claude Code Over Security Concerns

Alibaba Group Holding bans its employees from using Anthropic's Claude Code starting July 10, 2026, citing security risks from embedded backdoors that track Chinese users. The decision reverses a policy allowing up to $1,400 in monthly reimbursements for external AI tools, directing staff to use its Qoder platform instead. This follows Anthropic's accusations that Alibaba-affiliated accounts conducted a massive model distillation campaign, highlighting growing US-China AI fragmentation.

Alibaba ban on Anthropic

• ▪Alibaba Group Holding banned its employees from using Anthropic's Claude Code and other Anthropic products in workplace environments starting July 10, 2026
• ▪Alibaba Group Holding issued an internal notice on July 2, 2026, adding Claude Code to its list of high-risk software with security vulnerabilities
• ▪Alibaba Group Holding instructed its employees to use the company's in-house AI coding platform, Qoder, instead of Anthropic's Claude Code

Alleged model distillation operation

• ▪Anthropic sent a letter to US Senators Tim Scott and Elizabeth Warren on June 10, 2026, accusing Alibaba-affiliated operators of generating 28.8 million interactions with Claude via 25,000 fraudulent accounts
• ▪Anthropic has previously accused other Chinese artificial intelligence firms, including DeepSeek, Moonshot AI, and MiniMax, of engaging in industrial-scale model distillation attacks
• ▪Anthropic characterized the Alibaba-affiliated activity between April 22 and June 5, 2026, as the largest known data extraction campaign aimed at model distillation

Claude Code backdoor concerns

• ▪Alibaba Group Holding's ban on Claude Code was triggered by an internal security audit that identified potential embedded backdoor risks in the software
• ▪An Anthropic employee stated on X that the environment-inspecting feature was an experiment launched in March 2026 to prevent unauthorized resellers and model distillation
• ▪Security researchers discovered that Claude Code versions released from April 2026 contained code inspecting local timezones and proxy configurations to identify China-based users

AI tool reimbursement reversal

• ▪Alibaba Group Holding reversed its flexible financing policy, prohibiting engineers from receiving reimbursements for using external artificial intelligence tools
• ▪Since early 2026, Alibaba Group Holding had allowed engineers to claim up to $1,400 monthly in reimbursement for using third-party AI tools like Claude Code, GPT, and Gemini

US-China AI fragmentation

• ▪Financial institutions JPMorgan Chase and Goldman Sachs restricted their staff members from accessing Anthropic models in Hong Kong
• ▪Alibaba Group Holding filed a lawsuit in the U.S. District Court for the Northern District of California on June 23, 2026, challenging its inclusion on the Pentagon's Section 1260H list
• ▪Anthropic suspended numerous Chinese user accounts under personal and team plans in late June and early July 2026 without offering refunds