Meta and Microsoft Restrict Employees’ Use of Claude AI Tools

Summary

Meta and Microsoft are reducing employees’ internal use of Anthropic’s Claude as they shift more coding work toward proprietary tools and tighten AI spending. The move does not represent a withdrawal of Claude from customer-facing Microsoft products, and both companies reportedly remain substantial Anthropic customers. Microsoft had expected internal spending on Anthropic technology to exceed $1 billion annually, but that estimate fell by more than one-third after employees were directed toward GitHub Copilot and OpenAI frameworks. In Microsoft’s cloud and AI division, monthly internal spending ceilings reportedly dropped from $100,000 per employee to about $10,000 in most cases; these are limits rather than measures of actual consumption, and some engineers reportedly objected to the reduced flexibility. At Meta, Claude Code users reportedly fell from roughly 60,000 earlier in the year to about 30,000, with layoffs contributing but the company’s own MetaCode and Muse Code also driving the change. MetaCode reportedly has more than 30,000 internal users, while Muse Code has over 6,000, and Meta began external testing of Muse Code with clients in August. Despite the smaller Claude Code user base, Meta reportedly spent more than $105 million on Claude Code during a 28-day period, showing that user counts and spending can move differently. The article also notes that changing coding assistants does not remove security risks: past Claude Code issues could enable unauthorized execution and API-key theft, while Microsoft’s remedied RoguePilot vulnerability showed how malicious commands in a GitHub Issue could help take over repositories. The reported shift therefore reflects tighter budgets and stronger competition over developer tools, rather than an end to commercial engagement with Claude; the article also cites Anthropic’s reported annualized revenue pace of $65 billion as evidence of continuing demand.