Microsoft's AI bill came in lighter — and that's the whole story

The headline is $92.3 billion of revenue, up 19%, with Microsoft Cloud crossing $63.9 billion and accelerating to 30%. But the number that actually moved the story is the pairing underneath it: Azure grew 48% — ahead of expectations — while capex came in at $45.2 billion, below what management had planned, because, as Amy Hood put it, "some third-party capacity deliveries we expected in Q1 were delayed."

That juxtaposition is the quarter. Nine months ago, on the January call, Morgan Stanley's Keith Weiss stated the market's anxiety plainly: “CapEx is growing faster than we expected, and maybe Azure is growing a little bit slower than we expected. And I think that fundamentally comes down to a concern on the ROI on this CapEx spend over time.” — Keith Weiss, Analyst · 2026-01-28 This quarter inverted both halves of that sentence at once.

The efficiency trick

Amy Hood was unusually explicit about why Azure outperformed: not more capacity, but more revenue per unit of it. First, engineering. “In Q1, I would say the engineering teams have really delivered very, very good work in terms of improving efficiency at the platform layer. That is not just a statement around the efficiency of GPUs. It's a statement around the efficiency of every layer.” — Amy Hood, Chief Financial Officer · 2026-10-28 They found underused capacity in the GPU fleet and in the CPU fleet, and redirected it to Foundry, whose revenue more than tripled.

Second, price. “We had higher token pricing than we expected in the quarter. Listen, high-quality models that we're able to take advantage of and see good margins on do benefit.” — Amy Hood, Chief Financial Officer · 2026-10-28 Token pricing is now a first-order swing factor in Azure's growth rate — management names it, alongside 3P GPU lease deliveries and model-mix ARPU, as a source of quarterly variability. That is a different confession than a year ago, when variability was purely a supply story. In July, Hood was still saying "demand continues to exceed available supply" and framing capex as a race to catch up. This quarter: "each dollar of CapEx has the potential to support more revenue and higher returns over time." The vocabulary has migrated from capacity to efficiency gains.

The best solution here is revenue growth. And so I would say I tend to start and focus there, even though I think many people seem to think I start at the bottom of the P&L and work up.

The filed numbers back the tone shift. Total revenue of $90.0B in the latest filed quarter, +18% year over year, against an operating margin of 48.9%, up four points, even with capital expenditure at $35.8B, up 110% year over year. The next quarter's spend steps up again, guided at over $50 billion.

Breadth versus the four customers

The second thread is Satya Nadella's insistence on breadth. He spent his Bernstein answer drawing a line between this AI build and the shiny-object phase: “And that's what I've been waiting for, as opposed to like 4 customers who buy a lot of capacity from you and you can show a few quarters of growth. And that's not a business as far as I'm concerned. So I want the breadth.” — Satya Nadella, Chairman and Chief Executive Officer · 2026-10-28

That is a direct answer to the question hanging over Microsoft since last October, when JPMorgan's Mark Murphy asked about contracts "worth hundreds of billions of dollars that are 20x their current revenue scale" and the concentration risk embedded in them. A year ago management deflected with the fungible-fleet argument. This quarter they point at breadth: 120,000 Foundry customers with revenue tripling, 44,000 paid Fabric customers, 50 million agents registered in Agent 365, and Copilot customers with more than 50,000 seats up over 8x year over year — including Volkswagen at 330,000 employees.

Underneath sits a business-model change the company has been telegraphing since April: per-seat becoming seat-plus-consumption. Usage based offerings — GitHub Copilot credits and E7 — now contribute to ARPU alongside premium seats, and Copilot customers actively consuming credits grew over 75% quarter over quarter. The consumption meter is no longer a slide; it's a revenue line.

What broke: Windows and E3

Not everything worked. Windows OEM revenue fell 27%, below expectations, and Devices & Consumer revenue fell 3%. Hood blamed "the Windows 10 end-of-support demand in the prior year, inventory drawdowns by OEM and channel partners, and weaker PC market demand," then guided Q2 OEM down in the mid-30s. That is a deliberate reset — Microsoft is letting the PC cycle wash out rather than chasing it.

Less discussed but more telling: paid M365 commercial seats grew 5%, and Hood conceded "some impact from the weaker sales execution noted earlier" on E3 and standalone SKUs. For the franchise that anchors the productivity story, an execution stumble — however small — is worth flagging. It is also why she keeps reworking sales motions toward E7, Copilot and usage constructs. The old SKU motion is slowing exactly as the new one must take over.

Then the bookings optics: commercial bookings declined 12%, driven by the prior-year large Azure commitment, even as commercial RPO grew 73% to $677 billion. Hood pre-empted the confusion — such commits "will result in material volatility in bookings, making quarterly growth rates less indicative of underlying sales execution." Honest, and a warning that bookings are now a noisy signal.

The bill is still coming

The relief on Q1 capex is not a change in direction. Calendar 2026 spend is guided at approximately $170 billion, accelerating in the second half from CPUs, memory, storage and GPUs. And for the first time Hood explicitly opened the debt door: “we're fortunate to have access in the debt markets given that where our ratings are and sort of we haven't been able to- and we haven't yet, I would say, tapped that.” — Amy Hood, Chief Financial Officer · 2026-10-28

The balance sheet is already telling that story. Effective net cash of $101.8B, down 21% year over year and roughly half the 2020 peak, while free cash flow margin (less SBC) of 18.4% shows the compression. Investors pay 50.8x free cash flow for the privilege.

The market's verdict is striking. Over the last 90 days Microsoft's stock is up 42%, even as the global AI data center theme is down more than 5% with dozens of decliners — HBM, co-packaged optics, data center power all in the red. The market is not buying AI infrastructure broadly. It is buying proof of monetization, and Microsoft just delivered the cleanest version yet: lower capex, higher Azure, expanding margins. The risk that remains is the one Nadella named himself — that enterprise breadth is "not like a quarter over quarter... a decade-long thing." If capex re-accelerates while Azure laps a tough comp, the January question returns. For now it has been answered the only way that counts: a dollar of spend that went further than expected.