According to a report published on October 5 by The Information, Meta and Microsoft have begun to significantly lower employee utilization of Anthropic’s Claude AI as they pivot toward their own proprietary coding assistants. The shift is described as a strategic change that focuses on internal expenditure and workforce processes, rather than removing customer access to Claude through Microsoft’s offerings. Both companies remain substantial clients of Anthropic while also competing with the firm, reflecting a convergence of AI investment, product strategies, and control over developer resources.
Microsoft had previously projected that internal spending on Anthropic technology would exceed $1 billion annually. Following a management directive for employees to curtail Claude use and increasingly adopt Microsoft tools such as GitHub Copilot and OpenAI frameworks, that estimate has fallen by more than one‑third. Within Microsoft’s cloud and AI division, monthly AI spending limits have reportedly been cut from $100,000 per employee to roughly $10,000 in most cases. The report notes these figures are spending ceilings, not actual expenditures, and that the tighter budgeting has disappointed some engineers who previously enjoyed broader freedom to experiment with models. Despite the internal cutbacks, Microsoft has not withdrawn Claude from its client‑facing products; client spending on Anthropic models via Microsoft’s enterprise platforms continues to rise.
At Meta, the number of Claude Code users reportedly dropped from about 60,000 earlier this year to approximately 30,000. While layoffs have contributed to the decline, the report attributes the primary cause to Meta’s strategic move toward its own AI solutions—MetaCode and Muse Code—built on Meta’s proprietary models. MetaCode now has more than 30,000 internal users, and Muse Code has over 6,000 employee users. Meta began external testing of Muse Code with clients in August. Notably, during this transition Meta allegedly allocated over $105 million to Claude Code over a 28‑day period, indicating that a reduction in users does not automatically translate to lower spending.
The report also addresses cybersecurity considerations. It notes that switching to different coding assistants does not eliminate risks associated with automated access to sensitive files, commands, and credentials. Prior analyses by Cybersecurity News identified vulnerabilities in Claude Code that could permit unauthorized execution and API key theft; Anthropic patched those flaws before public disclosure. Microsoft’s tools have faced similar issues, exemplified by the remedied RoguePilot vulnerability, which showed how malicious commands embedded in a GitHub Issue could lead to repository takeover. The article stresses the importance of stringent permissions, trusted project settings, and thorough review of AI‑generated actions.
Finally, Anthropic’s reported annualized revenue pace of $65 billion underscores sustained demand for its services. The retrenchment by Meta and Microsoft is characterized as a tightening of budgets and intensifying competition, rather than an outright abandonment of Claude.