Microsoft FY26: cloud growth met a much heavier infrastructure bill

The decision question

This article asks: what did Microsoft actually report for FY26, how much of the profit comparison is affected by investment gains and losses, and is operating cash keeping pace with the infrastructure build? It is a way to organize evidence, not a price target or a trade recommendation.

The evidence cutoff is the fiscal year ended June 30, 2026. Microsoft's earnings release was published July 29, 2026; the corresponding Form 10-K was filed with the SEC on the same date. No share price, market capitalization, valuation multiple, discount rate, expected return or entry/exit rule is used here.

All dollar figures are US$ millions unless the text says billions. Reported figures are Microsoft's GAAP measures unless a table says adjusted/non-GAAP. Percentages labelled “derived” are InfluencerQ calculations from the reported values and may not match a company's separately defined measure.

Q4 showed broad cloud strength, but not every business moved together

The release reported the following three-month comparison:

Microsoft Cloud revenue was $59.3 billion, up 27%. Commercial remaining performance obligation was $678 billion, up 84%. The latter is a contract-related forward-looking indicator, not current-period revenue; it should not be added to the income statement.

The segment picture was also uneven. Productivity and Business Processes revenue was $37.847 billion, up 14%. Intelligent Cloud revenue was $39.306 billion, up 32%, and Azure and other cloud services revenue rose 43%. More Personal Computing revenue was $12.854 billion, down 4%. The release also reported a $0.27 Q4 diluted-EPS benefit versus the April guidance comparison from several discrete items, including a $3.2 billion gain from the Anthropic investment, lower-than-expected Voluntary Retirement Program expenses, and offsets from severance and Xbox impairment charges.

The practical reading is not “one number is real and the others are fake.” It is that total revenue, cloud growth, segment results, recurring obligations and the quarter's one-time items answer different questions. A clean research note keeps their denominators and time horizons visible.

GAAP profit growth was helped by investment effects

Microsoft's release reconciles GAAP results to a non-GAAP measure that excludes the impact from investments in OpenAI. The direction of that adjustment changed between years:

The adjusted column is useful for a like-for-like question about the stated investment effect. It is not a universal “underlying earnings” measure: Microsoft is excluding the OpenAI investment effect, not every unusual item, and the company itself says the non-GAAP measures should be considered with GAAP rather than treated as superior to it.

That distinction changes the headline comparison. FY26 GAAP net income growth was 31%, while the stated adjusted growth was 22%. Both figures are reported by Microsoft, but the 9-percentage-point gap means that GAAP EPS growth alone is not a sufficient description of operating momentum.

The cash-flow proxy asks whether the build-out is absorbing more cash

The 10-K reports consolidated cash from operations and additions to property and equipment as separate lines. To make the infrastructure question explicit, we calculate:

cash-flow proxy = net cash from operations − additions to property and equipment

This is not Microsoft's reported free-cash-flow metric. It leaves acquisitions, investment purchases and sales, financing, taxes, working-capital timing and other cash requirements in their respective reported lines. It is only a consistent diagnostic using two consolidated cash-flow lines.

The year-over-year arithmetic is revealing but limited:

• Net cash from operations rose 34.4% from FY25 to FY26.
• Additions to property and equipment rose 79.6%.
• The proxy fell 6.5%, from $71.611 billion to $66.987 billion.

The 10-K's MD&A says the increase in operating cash was mainly driven by more cash received from customers and less cash used for income taxes, partly offset by more cash paid to suppliers. It attributes the increase in cash used in investing mainly to a $51.4 billion increase in property and equipment additions and a $22.2 billion increase in other investing cash use, partly offset by lower acquisition-related and net investment cash use. Those are management's reported explanations, not an independent causal decomposition.

This proxy does not prove that Microsoft's investments will earn too little, nor does the decline prove that the cloud strategy is failing. It does establish a better question than “cash flow grew”: did cash from operations grow quickly enough to keep the same post-property-and-equipment cushion while the build accelerated? On this narrow measure, the answer was no in FY26.

The segments are useful, but they are not standalone businesses

The 10-K reports three operating segments. The table adds revenue share and operating margin calculations to the company's reported revenue and operating income:

The segment totals reconcile to the consolidated totals. The interpretation boundary matters just as much as the arithmetic:

• Intelligent Cloud is not Azure alone. It includes Azure and other cloud services, server products and services, and enterprise and partner services.
• Productivity and Business Processes is not Microsoft 365 alone. It also includes LinkedIn and Dynamics.
• More Personal Computing is mixed. It includes Windows and Devices, Xbox, and Search advertising.
• Microsoft says certain revenue and costs are allocated across integrated businesses. AI-infrastructure, training and marketing costs benefiting multiple segments are generally allocated by relative gross margin, while corporate costs use measures such as relative gross margin or headcount.

The derived segment margins are therefore management-reporting views, not clean standalone cash margins. The consolidated cash-flow statement does not assign the $115.948 billion of property and equipment additions to Intelligent Cloud. Doing so would create precision the filing does not provide.

What the bullish reading gets right—and where it can overreach

Microsoft's 10-K also describes the risk from the scale and speed of its AI and cloud investments. It says those investments depend on customer demand, technology, competition and regulation; the associated revenue may not arrive at expected levels or timeframes. The filing discusses infrastructure capacity constraints, supply and energy resources, pricing and customer deployment decisions. These are not predictions that the build will fail. They are the company's own reasons to treat demand conversion and capital intensity as testable variables rather than settled conclusions.

Three tests to carry forward

These are audit rules for a future update, not management guidance and not a buy/sell signal.

1. Is Intelligent Cloud acceleration durable? The Q4 baseline is Intelligent Cloud growth of 32% and Azure and other cloud services growth of 43%. In each of the next two comparable releases, confirm the same reported bases. A result below 25% for Intelligent Cloud or below 30% for Azure and other cloud services in either release weakens the “uninterrupted acceleration” thesis; a segment-definition change requires a new baseline.

2. Can operating cash absorb the infrastructure build? The FY26 proxy is $66.987 billion versus the FY25 baseline of $71.611 billion, while property-and-equipment additions rose 79.6%. At FY27 year end, recompute the identical proxy from the consolidated cash-flow statement. A proxy below $71.611 billion or a negative proxy means the “cash conversion is keeping pace with the build-out” thesis is not confirmed. This is a cash diagnostic, not a valuation model.

3. Does GAAP profit growth remain representative? FY26 GAAP net income grew 31% and the stated adjusted measure grew 22%, with a $4.963 billion OpenAI investment effect. On the next comparable release, apply Microsoft's then-current reconciliation. A wider gap than 9 percentage points, or another material investment gain driving reported growth while adjusted growth lags, weakens the claim that GAAP EPS growth alone describes operating momentum.

What this article does not answer

It does not say whether MSFT is cheap or expensive. There is no price series, valuation input, expected-return estimate, event-study result or trading rule. It does not turn management's cloud commentary into independent customer research, and it does not assign the consolidated capex line to a segment.

The article also does not treat the FY26 result as a forecast. It is a dated read of reported periods. A later quarter can strengthen, weaken or change the question; the useful habit is to preserve the original cutoff and show the new evidence beside it.

Sources and reproduction

The calculations above use the figures in the tables and the following primary sources:

1. Microsoft FY26 Q4 earnings release and webcast, published July 29, 2026. This provides the Q4 and FY26 headline results, segment quarter data, Microsoft Cloud and non-GAAP reconciliation.
2. Microsoft FY26 Form 10-K filed with the SEC, filed July 29, 2026 for the year ended June 30, 2026. This provides the consolidated cash-flow statement, MD&A cash discussion, segment definitions and Note 18 allocation limits.
3. SEC filing detail for the FY26 Form 10-K, which records the filing date, period of report and document identity.

Ratios are calculated as numerator divided by the stated denominator and rounded to one decimal place. The cash-flow proxy is calculated as net cash from operations minus the positive cash outflow shown for additions to property and equipment. All source pages were checked in a browser on September 9, 2026. AI-assisted source checking and writing; no independent human expert review.