
Microsoft's AI Revenue Just Hit $37 Billion. The Memory Crunch Is About to Make Everything More Expensive.
Azure +40%. AI revenue +123% to $37B annualized. 20M Copilot seats. And $190B capex. Why the strongest AI print in Big Tech history barely moved the stock.
Microsoft just put up the strongest AI revenue number in Big Tech history. Annualized AI revenue of $37 billion, up 123 percent year over year. Twenty million paying Microsoft 365 Copilot seats. Azure growing 40 percent. The stock barely moved.
That gap, between extraordinary numbers and a flat reaction, is the single most important data point in technology investing right now. Here is what is going on under the hood.
The Top Line Was Excellent
Microsoft delivered $82.89 billion in fiscal Q1 revenue, up 18 percent year over year, beating the $81.39 billion consensus. EPS came in at $4.27 against a $4.06 estimate. Azure and other cloud services grew 40 percent, well past the 37 percent the sell side expected.
The AI line is what everyone wanted to see. Annualized AI revenue of $37 billion is up from roughly $16.5 billion the same time last year. That growth rate, 123 percent, is not normal at this scale. It includes Azure AI services, AI model builders, and Copilot products across the stack.
Twenty million paid Copilot seats is the consumer relevant proof point. CFO Amy Hood said there is more growth to come and emphasized that enterprise adoption is still in early innings.
And Yet The Stock Did Not Move
The reason was the capital spending number. Microsoft guided to roughly $190 billion in capex for fiscal 2026. The Visible Alpha consensus going into the print was $154.6 billion. That is a $35 billion miss to the upside. Wall Street was not expecting that much.
Inside that gap, Hood pointed to about $25 billion in higher component prices. Component prices means memory, specifically the DRAM and high bandwidth memory that goes into AI accelerators. The global memory market is in the tightest squeeze since the 2017 to 2018 cycle, and every hyperscaler is bidding against every other hyperscaler for the same constrained supply.
Memory Is The Hidden Tax
This is the part most retail investors miss. The story for the last two years has been about Nvidia and the GPU shortage. The story going into 2026 is about everything around the GPU. HBM3 and HBM3e modules. DDR5 server memory. The packaging and substrate that ties it all together.
Samsung, SK Hynix and Micron are sold out for 2026 across high end memory. Lead times are stretching. Spot prices are climbing. None of that shows up in a consumer earnings call as a line item, but it shows up in capex guidance and gross margins.
Microsoft gross margin printed at 67.6 percent, the narrowest since 2022. That is the canary. Depreciation from the hyperscaler buildout is mounting and will keep pressuring margins for at least the next four quarters.
Q2 Guidance Did Not Help
Microsoft guided fiscal Q2 revenue to a range of $86.7 to $87.8 billion. The midpoint sits slightly below the $87.53 billion consensus. Operating margin is expected to tick down to around 44 percent from 46.3 percent.
Fiscal Q3 capex came in at $31.9 billion, up 49 percent year over year but actually below the $34.9 billion consensus. That undershoot on near term spending plus the giant overshoot on full year spending tells you the curve is steepening into the back half. Microsoft is front loading the bill.
What This Means For The Stack
If Microsoft, the most disciplined hyperscaler in the group, is calling out $25 billion in component price pressure, every other AI buyer is feeling the same wind. That includes Meta, Amazon, Alphabet, Oracle, and the wave of neoclouds raising debt to build capacity.
It also means the memory makers are about to have a very good year, and the AI customers are about to have a very expensive one. The revenue is real. The infrastructure cost to deliver it is also real. The gap between those two numbers is why the market is no longer throwing confetti at every Big Tech earnings beat.
Sources: CNBC reporting by Jordan Novet, Microsoft Investor Relations press release, Visible Alpha consensus data, FactSet, Yahoo Finance, Quartz.