Meta and Microsoft are Actively Cutting Employee Use of Claude AI

Meta and Microsoft are reducing employee use of Anthropic’s Claude AI while pushing their own coding tools, according to an October 5 report by The Information. The changes focus on internal spending and staff workflows, rather than ending customer access to Claude through Microsoft’s products.

The shift highlights growing pressure to control AI costs as coding assistants become part of daily software work. Both companies remain major Anthropic customers, but they also compete with it. Their decisions show how AI spending, product strategy, and control over developer tools are becoming closely linked.

Microsoft Cuts AI Spending Limits

Microsoft previously expected internal spending on Anthropic technology to reach at least $1 billion annually. That projection has reportedly fallen by more than a third after leaders asked employees to reduce Claude use and rely more on Microsoft tools, including GitHub Copilot and OpenAI models.

Within Microsoft’s cloud and AI division, monthly AI spending limits reportedly dropped from $100,000 per employee to around $10,000 in most cases. These figures describe spending caps, not what every employee actually spent. The lower limits have frustrated some engineers who previously had greater freedom to test models.

The distinction matters: lower internal budgets do not mean Microsoft has removed Claude from customer products. Reports indicate that customer spending on Anthropic models through Microsoft’s enterprise platforms continues to grow, even as the company tightens its own employees’ use.

At Meta, Claude Code users reportedly fell from roughly 60,000 earlier this year to about 30,000. Layoffs contributed to the decline, but the report identifies Meta’s push toward its own AI tools as the larger factor. These include MetaCode and Muse Code, which use Meta models.

MetaCode reportedly had more than 30,000 internal users, while Muse Code had over 6,000 employee users. Meta began testing Muse Code with outside customers in August. Despite the shift, Meta reportedly spent more than $105 million on Claude Code during one 28-day period, showing that fewer users do not necessarily mean low spending.

For security teams, changing coding assistants does not remove the risks created by automated access to files, commands, and credentials. Cybersecurity News previously covered Claude Code vulnerabilities involving repository settings that could enable unauthorized execution and API key theft. Anthropic fixed those issues before disclosure.

Microsoft’s tools have faced similar concerns. The patched RoguePilot vulnerability showed how malicious instructions inside a GitHub Issue could lead to repository takeover. These findings underline the need for limited permissions, trusted project settings, and careful review of AI actions.

Anthropic’s reported $65 billion annualized revenue pace illustrates continuing demand. The pullback reflects tighter budgets and competition, not a complete break with Claude.

Cut every SOC alert investigation by 21 min. Power your SOC with instant IOC context for immediate response: Integrate TI Lookup into your SOC