How are organisations reckoning their AI addiction
Microsoft has begun rolling back employee access to Anthropic’s Claude Code in favor of its own GitHub Copilot CLI. The move is framed as toolchain unification, though reports suggest it is a cost-cutting measure following high AI usage.
Why it matters
It illustrates the tension between corporate AI experimentation and the financial reality of managing high-cost AI infrastructure.
In December 2025, Microsoft opened the doors for thousands of its engineers to use Anthropic’s Claude Code — an AI coding assistant — as part of an internal experiment to benchmark the best tools available. That test worked too well as some employees decisively chose it over Microsoft’s own AI coding assistance — GitHub Copilot CLI. Then, by May 2026, the software giant began cancelling most of those Claude Code licences, instructing teams behind Windows, Microsoft 365, Outlook, Teams, and Surface to migrate back to Copilot CLI before June 30. That day is apparently the last day of the company’s fiscal year. While it has called this rollback a “toolchain unification,” the financial reality, per multiple reports, suggests that engineers had used Claude Code a little too much.
The report balances corporate messaging with financial context.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in