Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effecti…

Microsoft canceled internal Claude Code licenses due to unsustainable token-based costs, while Uber exhausted its 2026 AI budget in four months and GitHub shifts to usage-based billing. The argument: AI vendor subsidies masking true inference costs are unwinding, forcing enterprises to confront unit economics. Practitioners building on third-party LLM APIs should model consumption costs carefully as flat-rate pricing disappears.