Microsoft's own earnings material gives the clearest read on that story, and it adds some caveats the stock note skips. Nothing below is investment advice. It separates what Microsoft reported from what a publisher modelled.
What Microsoft actually reported
For the quarter ended June 30, 2026, Microsoft's release gave these results:
- Annual revenue passed $331 billion, up 18%. Microsoft Cloud passed $214 billion, up 27%. Azure passed $100 billion, up 41%.
- Fourth-quarter revenue was $90.0 billion, and Azure and other cloud services grew 43%. Non-GAAP diluted EPS was $4.74, the same figure the 24/7 Wall St. piece uses.
- GAAP diluted EPS was higher, at $4.81. Microsoft said discrete items, including a $3.2 billion gain on its Anthropic investment, added $0.27 to EPS versus its prior guidance.
- Commercial remaining performance obligation (RPO) rose 84% to $678 billion.
The $4.74 figure excludes the impact of Microsoft's OpenAI investments, and the $4.81 figure does not. Neither figure is "clean" of the Anthropic gain. That gain sits inside the quarter's other income, so "beat the consensus" has an asterisk.
The $678 billion backlog, decoded
RPO is the headline bulls lean on. The stock note calls the commercial backlog the "deciding factor". Microsoft's call adds detail that changes how to read it:
- Management said RPO has a weighted average duration of 2.3 years. Roughly 30% will be recognized as revenue in the next 12 months, and that near-term slice is up 37% year over year. The portion beyond 12 months grew 112%.
- RPO excluding OpenAI grew 25%, compared with 84% including it.
- All sequential RPO growth came from customers outside frontier model companies.
- Nearly 90% of full-year Microsoft Cloud revenue came from customers outside frontier model companies.
So the 84% jump is real, but it is not a broad-based organic growth rate. A large share of the backlog is long-dated commitments, much of it tied to OpenAI contracts signed in the prior year. Microsoft itself warned that those contracts will cause "quarterly volatility" in bookings and RPO growth rates. RPO is also contracted revenue, not cash in hand.
The other side of that argument is that RPO growth outside OpenAI is healthy and was the sole source of sequential growth. That is a meaningful signal for enterprise buyers who commit to multi-year Azure and Microsoft 365 deals.
Demand beyond supply: the Azure capacity story
For IT admins, the most practical line is capacity. Azure grew 43% and customer demand continues to exceed available capacity. Growth beat expectations because of efficiency gains across the CPU and GPU fleet and process improvements that delivered new capacity earlier. That extra in-quarter capacity was quickly monetized.
For the next quarter, Microsoft guided Azure growth to roughly 45% in constant currency and still expects first-half growth to accelerate. It also cautioned that Azure's year-over-year growth can swing from quarter to quarter with capacity timing and contract mix. The stock note's "demand exceeds supply" quote is accurate, but the guidance is a forecast, not a result.
In practice, a supply-constrained cloud can mean tight regional GPU availability and quota approvals that take time. Pricing for scarce capacity can firm up too. Amy Hood pointed to spot-market pricing as evidence of the squeeze. That is general industry reasoning, not a Microsoft statement about any specific region or SKU.
The cost side: capex, cash flow and margins
This is where the stock note raises the right question and gets less precise on numbers. It says fiscal 2026 capex rose 79.62% to $115.95 billion and free cash flow fell 6.46%. I could not match those full-year figures to Microsoft's disclosures. Here is what the quarterly calls show:
| Quarter | Capex (incl. finance leases) | Free cash flow |
|---|---|---|
| Q1 FY26 | $34.9 billion | $25.7 billion |
| Q2 FY26 | $37.5 billion | $5.9 billion |
| Q3 FY26 | $31.9 billion | $15.8 billion |
| Q4 FY26 | $41 billion | $19.6 billion |
The sources for those figures are the first-quarter call (capex $34.9 billion, free cash flow up 33% to $25.7 billion), the second-quarter call (capex $37.5 billion, free cash flow $5.9 billion), the third-quarter call (capex $31.9 billion, free cash flow $15.8 billion), and the fourth-quarter call (capex $41 billion, free cash flow $19.6 billion).
By my own arithmetic, those capex quarters sum to about $145 billion. That is well above $115.95 billion. The gap could come from a different definition, such as cash paid for property and equipment rather than total capex including finance leases. Microsoft's Q4 cash paid for PP&E was $35.8 billion against $41 billion of total capex. I can't confirm which definition the note used, so treat the $115.95 billion figure with caution. Readers should check the cash flow statement rather than rely on a derived percentage.
The direction is not in dispute. Capex is large and rising, and Microsoft expects fiscal 2027 capex to grow again. Management also expects fiscal 2027 free cash flow to stay positive. Q1 capex is guided to more than $50 billion.
Three more facts matter here:
- Margins: Microsoft Cloud gross margin was 65% in Q4. Microsoft attributed the decline to the mix shifting toward Azure and to AI infrastructure investment, partly offset by efficiency gains. It guided Cloud gross margin to be "relatively stable" quarter over quarter.
- Short-lived assets: Roughly two-thirds of Q4 capex went to CPUs and GPUs. Hood argued that this makes spending flexible, because if demand changes the company can slow the largest cost component. Skeptics would counter that chips depreciate quickly and can become obsolete, so flexibility is not the same as safety.
- Accounting changes: From the start of fiscal 2027, Microsoft is extending the estimated useful life of datacenters and office buildings from 15 to 25 years. It said the depreciation effect on fiscal 2027 operating income is minimal. The bigger effect is that more datacenter leases will be operating leases, which don't count as capex. That moves its calendar-2026 capex expectation to about $175 billion. Headline capex comparisons across periods are therefore not like-for-like.
Copilot and the rest of the portfolio
The 30 million paid Microsoft 365 Copilot seats are confirmed. Microsoft said net paid seat adds more than doubled sequentially. It also said paid Microsoft 365 commercial seats grew 6%, so Copilot's seat count has no meaningful denominator without a comparable definition. Dynamics 365 grew 13%, with ERP bookings healthy and CRM seeing longer sales cycles.
The picture for Windows readers is less cheerful. Windows OEM and Devices revenue fell 7% in the quarter. Microsoft guided that segment to a high-teens revenue decline for fiscal 2027. It cited higher component costs that push up device prices, a prior-year comparison that benefited from Windows 10 end of support, and elevated inventory. Cloud is carrying the company while PC demand softens, which matters if you are budgeting for hardware refreshes.
How to read the 24/7 Wall St. valuation
The stock note's numbers fall into two groups:
- Publisher model outputs: the $636.14 target, the $738.76 bull case, the $539.40 bear case, the yearly targets up to $945.64 in 2030, and the Buy rating with 90% confidence. These come from the publisher's own weighting model. They are not Microsoft guidance, and the 2030 figure in particular is long-range extrapolation.
- Market data: the $534.55 price, the 14 Strong Buy and 39 Buy ratings, the $587.63 consensus target, and the forward P/E comparison (Microsoft 25x, Alphabet 23x, Amazon 24x). These change daily and aren't supported by Microsoft's filings.
The note's operating-margin comparison (45.1% for Microsoft, 34% for Alphabet, 13.7% for Amazon) is consistent in spirit with Microsoft's reported 45% operating margin. Amazon's lower figure reflects its retail business, which makes the comparison less direct than it looks.
Its own caveat is fair: the targets rest on current strategy execution, and either faster Copilot adoption or an AI demand slowdown could move them sharply. Its bear case of $539.40 sits barely above the quoted price, which suggests the model's downside is narrow. That is a modelling choice, not a market fact.
Bottom line
Microsoft's primary-source numbers support the core of the story: Azure growth accelerated, and demand still outruns supply. Guidance points to more acceleration, and the backlog is at a record. The costs are real too, with Cloud gross margin down, capex rising, and free cash flow lumpy from quarter to quarter.
The caveats are what the stock note leaves out. RPO growth excluding OpenAI is 25%, not 84%. Accounting changes complicate capex comparisons. The full-year capex figure in the note doesn't match the quarterly disclosures. Whether the cloud and AI businesses justify a higher valuation depends on whether this capacity converts to revenue and cash. For admins and IT buyers, the practical takeaway is to plan for tight Azure capacity, especially for GPU workloads. Expect Microsoft to keep pushing usage-based billing alongside per-seat licensing.