Microsoft shifts AI pricing to 'per seat plus consumption' as Copilot usage grows

Microsoft's shift to consumption-based pricing reflects a fundamental change in how the company monetizes artificial intelligence. Historically, software revenue has been tied to per-seat licenses—a customer buys access for a set number of users at a fixed cost. As AI assistants move from optional tools to continuous, integrated parts of daily work, that model no longer captures the full value customers derive from using them more intensely. By layering "consumption" on top of "per seat," Microsoft can capture both the baseline adoption cost and the upside from power users, frequent tasks, and broader enterprise deployment.

The company is already seeing this shift in practice: usage-based billing has been added to both Microsoft 365 Copilot and GitHub Copilot, and thousands of customers are already paying for and using these features, according to CEO Satya Nadella. Finance chief Amy Hood reinforced this, stating that Microsoft's expanding AI opportunity is "increasingly tied to usage and consumption growth" as AI experiences become more integrated across enterprise workflows. This is not a future bet—it is already generating meaningful revenue from GitHub Copilot's newly launched usage-based model.

The timing aligns with strong demand signals. Microsoft reported fourth-quarter revenue of $90.01 billion (up 18% year-over-year) and Azure cloud services revenue climbed 43% year-over-year, demonstrating that enterprises are leaning heavily into AI infrastructure. Nadella noted that the company continues to invest in AI infrastructure to meet demand that still exceeds available computing capacity, suggesting that adoption is outpacing supply. The consumption model positions Microsoft to capture more value as that demand translates into higher usage across Copilot's embedded features—Chat, CoWork, Autopilot, and Code—which Nadella called the "flagship super app" for different user roles.