Microsoft hits $37B AI run rate

Microsoft just gave the market the cleanest AI revenue number any big tech company has offered so far. Not a vibe check. Not a “strong demand” line. A number — $37 billion in annualized AI revenue run rate for the quarter ended March 31, 2026. That matters because the whole AI trade has been stuck between two stories: huge spending now, vague payback later. Microsoft is trying to close that gap. (news.microsoft.com) ### What did Microsoft actually say? On its April 29 fiscal Q3 2026 earnings release and call, Microsoft said its AI business “surpassed an annual revenue run rate of $37 billion,” up 123% from a year earlier. The company also posted $82.9 billion in total revenue, up 18%, and said Microsoft Cloud revenue reached $54.5 billion, up 29%. (microsoft.com) ### What is an “annual revenue run rate”? Basically, it is a quarterly revenue pace turned into a yearly number. It does not mean Microsoft already booked $37 billion of AI revenue over the past 12 months. It means the latest quarter’s level, if you annualize it, points to roughly that scale. Companies use run rate t(microsoft.com) That said, it is still a much more concrete datapoint than the usual AI talk. (microsoft.com) ### Where is that money coming from? Mostly from Azure and Microsoft’s enterprise software stack. Azure and other cloud services grew 40% in the quarter, with AI services again contributing a meaningful chunk of that growth. Microsoft also said Microsoft 365 Copilot passed 20 million paid seats, up 250% year over year, which sh(microsoft.com)oftware too. (finance.yahoo.com) ### Why did this land so hard? Because investors have been asking a very simple question: are customers really paying for AI, or are vendors mostly subsidizing adoption while burning cash on chips and data centers? Microsoft’s number does not answer every part of that question, but it does answer the first one. Yes(finance.yahoo.com) traction, not just technical excitement. (msn.com) ### So why wasn’t the reaction only positive? Because the other side of the story is spending. Microsoft’s capex has surged as it races to build AI infrastructure, and investors are still trying to judge whether today’s demand justifies tomorrow’s bill. Managemen(msn.com)gs at once — AI is already big, and the buildout is still expensive. (msn.com) ### Why does Azure matter more than almost anything else here? Azure is where the AI thesis gets operational. Training models is flashy, but enterprise AI turns into money when companies run workloads, buy tools, and keep consuming cloud services every month. Think of Azure as the toll r(msn.com) growth number is what makes the $37 billion run rate feel less like a one-off headline and more like an installed business. (msn.com) ### What does this mean for everyone else? It raises the bar. Microsoft has now shown a concrete AI revenue figure at a scale that forces rivals to do more than talk about usage, pilots, or developer enthusiasm. Investors will want the same kind of evidence from (msn.com)es after the infrastructure bill arrives? That scrutiny is the real second-order effect here. (news.alphastreet.com) ### Bottom line? Microsoft did not end the AI monetization debate. But it moved the debate forward. The question is no longer whether generative AI can produce serious revenue. Microsoft just showed that it can. Now the fight shifts to efficiency — who can turn that revenue into lasting profit before the capex wave gets even bigger. (news.microsoft.com)