Claude too expensive? Report says Microsoft is cancelling licenses for internal use after rising cost

Microsoft is scaling back internal use of Anthropic’s Claude Code AI tool, with most licenses set to be removed by June 30, the end of its current financial year. The decision comes as the company seeks to cut operating expenses, redirecting developers toward its own GitHub Copilot CLI instead. The Experiences + Devices team, responsible for Windows, Microsoft 365, and Surface, will phase out Claude Code usage entirely. The shift underscores the hidden cost of AI tools, where token-based pricing leads to rapid budget exhaustion as companies scale adoption. Uber previously faced a similar issue, depleting its 2026 AI budget by April after widespread use of Claude Code among 5,000 engineers. Despite the cost concerns, Microsoft remains engaged with Anthropic, reportedly negotiating to supply its custom Maia AI chips, which offer improved token efficiency. Microsoft first granted access to Claude Code in December, allowing thousands of developers to use the tool daily. However, rising costs have forced a reevaluation, even as the company continues strategic partnerships with Anthropic. The broader AI industry faces a paradox: tools designed to enhance productivity are now becoming significant financial liabilities for early adopters. Anthropic’s reliance on cloud services from Microsoft, Amazon, and Google further complicates the economics, with Microsoft’s $5 billion investment in Anthropic and its commitment to Azure cloud infrastructure. Meanwhile, Microsoft’s Maia 200 AI chip, introduced in January, promises 30% better token efficiency, potentially addressing some cost concerns in future deployments. The situation reflects the evolving challenges of AI adoption, where scalability clashes with budget constraints. Companies must balance innovation with financial sustainability, even as they deepen interdependent relationships in the AI ecosystem.