Meta and Microsoft Reduce Employee Use of Claude AI, Pivoting to Internal Coding Tools
Key point
Microsoft reduced internal AI spending limits from $100,000 to approximately $10,000 per employee in most cases, while Meta's Claude Code user base dropped by half as it promotes MetaCode and Muse Code.
Details
Meta and Microsoft are significantly reducing employee reliance on Anthropic’s Claude AI, pivoting instead to their own proprietary coding instruments. This strategic shift, reported by The Information, aims to manage AI expenditures and consolidate developer workflows around internal solutions like GitHub Copilot, MetaCode, and Muse Code.
Microsoft’s Budget Cuts
Microsoft previously anticipated internal spending on Anthropic technology would exceed $1 billion annually. Following management directives, this estimate has dropped by over a third. In the cloud and AI sector, monthly AI spending limits per employee have been reduced from $100,000 to approximately $10,000 in the majority of instances. These figures represent spending ceilings rather than actual expenditures, a move that has reportedly disappointed engineers accustomed to broader model exploration. Despite internal cuts, client expenditure on Anthropic models via Microsoft’s enterprise platforms continues to rise.
Meta’s Internal Migration
Meta’s usage of Claude Code has plummeted from around 60,000 users earlier this year to about 30,000. While layoffs contributed to this decline, the primary driver is Meta’s strategic push toward its own AI solutions:
- MetaCode: Used by more than 30,000 internal users.
- Muse Code: Used by over 6,000 employees, with external client testing initiated in August.
Notably, Meta allocated over $105 million to Claude Code over a 28-day period, indicating that reduced user counts do not necessarily correlate with immediate spending decreases.
Security and Market Context
The transition to different coding assistants does not eliminate security risks associated with automated access to sensitive files and credentials. Previous vulnerabilities in Claude Code and Microsoft’s RoguePilot highlight the need for stringent permissions and thorough evaluation of AI-generated actions. Meanwhile, Anthropic’s reported annualized revenue pace of $65 billion suggests that demand for its services remains strong despite these corporate retrenchments.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.