Meta and Microsoft have begun reducing employee use of Anthropic’s Claude AI as they shift focus to their own proprietary coding tools, according to a report from The Information on October 5. This change aims to prioritize internal costs and workforce processes without eliminating customer access to Claude through Microsoft’s services.
The move highlights the growing need to manage AI-related expenses as coding assistants become essential in software development. Despite being major clients of Anthropic, both companies are also competitors, reflecting a blend of AI investment strategies and control over developer resources.
Microsoft had previously projected that its internal spending on Anthropic technology would exceed $1 billion annually, but this estimate has decreased by over a third following management's directive for employees to limit their use of Claude and to increasingly utilize Microsoft tools like GitHub Copilot and OpenAI frameworks.
In Microsoft’s cloud and AI division, monthly AI spending limits have reportedly been reduced from $100,000 per employee to around $10,000 in most cases. It is important to clarify that these figures represent spending caps rather than actual employee expenditures. This budget reduction has reportedly disappointed some engineers who previously had more freedom to explore various models.
However, the decrease in internal funding does not mean Microsoft is withdrawing Claude from its client-facing products. Reports indicate that client spending on Anthropic models via Microsoft’s enterprise platforms continues to grow, even as the company restricts its employees' usage.
At Meta, the number of Claude Code users has reportedly dropped from approximately 60,000 earlier this year to about 30,000. While layoffs have contributed to this decline, the primary reason is attributed to Meta’s strategic pivot towards its own AI solutions, including MetaCode and Muse Code, which are developed using Meta’s proprietary models. MetaCode is reported to have over 30,000 internal users, while Muse Code has more than 6,000 users. Meta began external testing of Muse Code with clients in August.
Despite the reduction in users, Meta reportedly spent over $105 million on Claude Code over a 28-day period, indicating that fewer users do not necessarily mean lower spending. For cybersecurity teams, the transition to alternative coding assistants does not eliminate risks associated with automated access to sensitive files, commands, and credentials.
Previous analyses by Cybersecurity News have identified vulnerabilities in Claude Code that could allow unauthorized execution and API key theft, which Anthropic addressed before public disclosure. Microsoft’s products have faced similar issues, with the resolved RoguePilot vulnerability demonstrating how malicious commands in a GitHub Issue could lead to repository takeovers. These incidents highlight the importance of enforcing strict permissions, using trusted project settings, and thoroughly evaluating AI-generated actions.
Anthropic’s reported annualized revenue pace of $65 billion indicates continued demand for its services. This reduction in employee usage reflects tighter budgets and increased competition rather than a complete withdrawal from engagement with Claude.