Microsoft puts a $37 billion annual run-rate number on its AI push, reframing what “profit” can mean

Microsoft is turning a recurring debate about generative AI economics into a single operational benchmark: an annualized revenue run rate. CEO Satya Nadella pointed to the metric on Microsoft’s April earnings call, saying the company’s AI business has surpassed an annual revenue run rate of $37 billion, up 123% year over year.

The comments are being repeated in coverage of Microsoft’s latest quarterly update and tie the AI growth rate to a concrete number the company said it is tracking as of its fiscal third quarter of 2026. The coverage also states that Microsoft disclosed the figure in a company filing dated April 29, 2026, describing it as a run-rate measure rather than a reported GAAP revenue line item.

Microsoft is also linking the AI growth to paid usage. The same coverage says Microsoft has 20 million paid Copilot seats, framing Copilot, Microsoft’s generative AI assistant included across productivity tools and developer workflows, as a major unit of adoption that converts interest in AI into recurring customer spending.

The $37 billion run-rate number is being positioned as a sign that Microsoft’s AI ramp is moving from pilot projects toward scaled revenue generation. The coverage characterizes the growth rate as still early in the cycle, but the headline is unambiguous: Microsoft is reporting AI as large enough to be discussed in annualized revenue terms.

Alongside the AI metric, the coverage highlights additional cloud figures that it says help explain the demand environment. It reports Intelligent Cloud revenue of $34.681 billion, up 30%, and says Azure growth was 40% in constant currency. It also reports that Microsoft Cloud revenue reached $54.5 billion for the quarter, up 29%, tying AI monetization to broader enterprise cloud momentum.

The company’s AI push, according to the reporting, sits inside this larger cloud engine rather than as a standalone business that rises or falls independently. That matters because cloud revenue is a key driver of Microsoft’s overall operating leverage, and scaled AI features are increasingly delivered through the same infrastructure layers that underpin Azure services.

There is an important limitation to the $37 billion framing. A run-rate metric is an annualized estimate based on current-quarter performance and management’s definition, so it is not the same as a formally reported full-year revenue number. Microsoft, in the coverage, does not appear to provide further breakdowns there on margin contribution, cost structure, or how much incremental operating profit the AI business is generating at this stage.

Investors and analysts, meanwhile, will likely focus next on whether the AI run-rate can maintain its growth as seat growth continues, whether Microsoft can sustain customer demand without outpacing underlying capacity, and how management will describe AI profitability over time. The company’s next earnings cycle should clarify how the run-rate metric evolves alongside Copilot seat growth and Azure results.