Microsoft reported fiscal third-quarter 2026 results on April 29 that demonstrated the company's AI monetization story is accelerating at a remarkable pace. Azure AI revenue is now running at $37 billion annualized, up 123% from a year ago, while Azure and other cloud services grew 40% -- exceeding analyst estimates of 38.8%. The Intelligent Cloud segment posted $34.68 billion in revenue, and total company revenue beat on the top and bottom lines.
The results validate Microsoft's massive bet on AI infrastructure, though they came with a significant cost. Capital expenditures for fiscal 2026 are now projected to reach $190 billion, up from earlier estimates, partly due to soaring memory component prices. CEO Satya Nadella said the company is seeing AI transform every layer of its technology stack, from infrastructure to applications, and that Copilot enterprise subscriptions continue to scale rapidly.
What happened
Microsoft's Q3 FY26 earnings showed broad-based strength across its business segments. Azure AI services -- which include revenue from model builders, API access, and companies running AI workloads on Azure -- grew 123% year over year and now contribute approximately 18% of total Azure revenue, up from 12% a year ago. The company reported that Copilot enterprise subscriptions have surpassed 5 million seats, driven by adoption in Microsoft 365, GitHub, and Dynamics 365.
The $190 billion capex forecast for fiscal 2026 represents a staggering investment in AI infrastructure, including GPU clusters, data center construction, and memory procurement. Microsoft cited a $25 billion impact from higher component prices, primarily HBM (high-bandwidth memory) used in AI accelerators. Despite the massive spending, management reiterated that AI investments are generating returns and that capacity constraints, not demand, are the binding factor.
Why it matters
Microsoft is the first hyperscaler to demonstrate AI revenue at meaningful scale. The $37 billion annualized run rate from AI services alone would make it larger than most enterprise software companies. This validates the thesis that AI workloads will drive a multiyear cycle of cloud infrastructure spending, benefiting not just Microsoft but the entire AI supply chain including Nvidia, AMD, and cloud infrastructure providers.
The Copilot adoption trajectory is equally significant. With 5 million enterprise seats, Microsoft is proving that AI productivity tools have real demand at enterprise scale. Each Copilot subscription generates $30 per user per month in additional revenue, representing a massive upsell opportunity across Microsoft's installed base of over 400 million commercial Office users.