Full Report

The numbers behind Microsoft Corporation: as-reported financial statements and company metrics for FY2021–FY2025, traced to the source filings, opened with the share-price history those statements have to justify. Every linked figure opens the exact page of the filing it was printed on, with the statement row highlighted. Amounts in US$ millions unless noted.

Reading notes: Microsoft's fiscal year ends June 30. Display unit is US$ millions except per-share (US$) and share counts (millions), exactly as printed in the 10-K. Core statements: FY2023–FY2025 income statement and cash flow are the three columns of the FY2025 10-K (pp.71, 74); FY2021–FY2022 (and FY2020 for the long-term record) are from the FY2022 10-K (pp.57, 60). Balance sheet: FY2024–FY2025 from FY2025 10-K p.73, FY2023 from FY2023 10-K p.62, FY2021–FY2022 from FY2022 10-K p.59. Revenue and operating income by segment: FY2023–FY2025 reflect the segment structure recast in the FY2025 10-K (p.118); FY2021–FY2022 reflect the structure as originally reported in the FY2022 10-K (p.95). Because of the FY2025 realignment the two ranges are not directly comparable (see discrepancies). FY2016–FY2019 figures in the long-term record come from the standardized data feed (SEC EDGAR XBRL) and are shown without page links; FY2020–FY2025 are cited to the filings.

Share Price — Full Available History — 37 Years

The stock closed at $385.10 on Jul 10, 2026 — up 8,997,418% over the window shown (+36.7% a year), trading between $0.00 and $542.07. At that close the stock trades at 28× FY2025 diluted EPS as reported below.

Source: market price feed, monthly closes, sampled from 9,197 source observations, Jan 1990–Jul 2026. Price return only, excludes dividends. Prices are split-adjusted (1:2 on Apr 16, 1990; ×1.5 on Jun 27, 1991; ×1.5 on Jun 15, 1992; 1:2 on May 23, 1994; 1:2 on Dec 09, 1996; 1:2 on Feb 23, 1998; 1:2 on Mar 29, 1999; 1:2 on Feb 18, 2003).

FY2025 at a Glance

Source: FY2025 consolidated statements [1] [2]. Click any linked figure to open the filing page with the row highlighted.

Revenue by Reportable Segment

Source: Note 19 (FY2025 10-K) / Note 20 (FY2022 10-K) — Segment Information; FY2023–FY2025 on the segment structure recast effective FY2025, FY2021–FY2022 as originally reported [3] [4]. Click any linked figure to open the filing page with the row highlighted.

Operating Income by Reportable Segment

Source: Note 19 (FY2025 10-K) / Note 20 (FY2022 10-K) — Segment Information [3] [4]. Click any linked figure to open the filing page with the row highlighted.

Income Statement

Source: Consolidated Income Statements [1] [2]. Click any linked figure to open the filing page with the row highlighted.

Columns marked E are consensus analyst estimates shown alongside reported results for direct comparison; they are not company guidance.

Estimate source: Yahoo Finance analyst consensus, as of 2026-07-11. Estimate figures link to the consensus source, not to filing pages.

Balance Sheet

Source: Consolidated Balance Sheets [5] [6] [7]. Click any linked figure to open the filing page with the row highlighted.

Cash Flow

Source: Consolidated Cash Flows Statements [8] [9]. Click any linked figure to open the filing page with the row highlighted.

Microsoft Cloud Commercial Backlog

Source: company filings [10] [11] [12] [13]. Click any linked figure to open the filing page with the row highlighted.

Revenue by Product Service

Source: company filings [10]. Click any linked figure to open the filing page with the row highlighted.

Consumer, Gaming Advertising

Source: company filings [14] [13] [15] [16]. Click any linked figure to open the filing page with the row highlighted.

Workforce Shareholder Returns

Source: company filings [17] [18] [19] [20]. Click any linked figure to open the filing page with the row highlighted.

Long-Term Record

Source: consolidated statements across filings; older years from the standardized feed [8] [1] [9] [2]. Click any linked figure to open the filing page with the row highlighted.

Operating KPIs

Source: company-reported operating metrics [10]. Click any linked figure to open the filing page with the row highlighted.

Analyst Consensus

Estimate source: Yahoo Finance analyst consensus, as of 2026-07-11. Estimate figures link to the consensus source, not to filing pages.

Traceability

359 of 427 figures on this page (84%) link to the filing page where they are printed — click a linked figure to open the source PDF at that page with the row highlighted. Unlinked figures come from standardized data feeds or pre-filing years.

• Microsoft's fiscal year ends June 30. Display unit is US$ millions except per-share (US$) and share counts (millions), exactly as printed in the 10-K.

• Core statements: FY2023–FY2025 income statement and cash flow are the three columns of the FY2025 10-K (pp.71, 74); FY2021–FY2022 (and FY2020 for the long-term record) are from the FY2022 10-K (pp.57, 60). Balance sheet: FY2024–FY2025 from FY2025 10-K p.73, FY2023 from FY2023 10-K p.62, FY2021–FY2022 from FY2022 10-K p.59.

• Revenue and operating income by segment: FY2023–FY2025 reflect the segment structure recast in the FY2025 10-K (p.118); FY2021–FY2022 reflect the structure as originally reported in the FY2022 10-K (p.95). Because of the FY2025 realignment the two ranges are not directly comparable (see discrepancies).

• FY2016–FY2019 figures in the long-term record come from the standardized data feed (SEC EDGAR XBRL) and are shown without page links; FY2020–FY2025 are cited to the filings.

• Quarterly income-statement figures are single-quarter (three-months) amounts from each Form 10-Q; Q4 FY2025 has no 10-Q (Microsoft files no fourth-quarter 10-Q) and is therefore omitted from the quarterly sequence. Quarterly cash flow is reported cumulatively year-to-date in the 10-Q and is not shown as a discrete-quarter statement.

• 2 figure(s) differed between the data feed and the filing; the filing value is shown (see the run's metrics/metrics_tab.json for the audit trail).

Franchise and Drawdown

Microsoft is a $281.7 billion-revenue software and cloud franchise that turns roughly 46 cents of every sales dollar into operating profit [1]. Its shares have fallen about 29% from a 2025 peak — not because profit stalled, but because free cash flow did. Capital spending on AI infrastructure now absorbs 47 cents of every operating-cash dollar, up from 27 cents four years earlier [2]. That divergence — profit rising while free cash flow fell — is the fact this report examines.

What Microsoft sells

For a reader meeting the company cold: Microsoft earns money in three reportable segments [3]. Productivity and Business Processes is the Office/Microsoft 365 subscription franchise plus LinkedIn and Dynamics — $120.8 billion of revenue at a 58% operating margin. Intelligent Cloud is Azure and the server business — $106.3 billion, growing fastest. More Personal Computing is Windows, search advertising, and Xbox gaming — $54.6 billion, the lowest-margin and slowest-growing of the three [4].

Source: FY2025 Annual Report (Form 10-K), Segment Results of Operations [5].

Two facts orient everything that follows. First, this is overwhelmingly a commercial software and cloud business: Productivity and Intelligent Cloud together are 81% of revenue and 89% of segment operating income. Second, the growth engine and the spending engine are the same segment — Intelligent Cloud, where Azure sits, is both the fastest-growing line and the reason capital expenditure has more than doubled in two years.

The scale of the machine

Source: FY2025 Annual Report (Form 10-K), Income Statements [6].

Microsoft earned $101.8 billion of net income on $281.7 billion of revenue in the fiscal year ended June 30, 2025 — a 36% net margin, on an operating margin of 46% [7]. Revenue has compounded at roughly 14% a year since FY2021, and net income at a similar pace [8]. Few companies of this size grow this steadily.

Source: derived from reported financials, FY2021–FY2025 10-Ks; FY2024–FY2025 per Segment Results [9].

What the stock has done

Source: market price data, as reported (close of 10 July 2026).

At $385 on 10 July 2026, Microsoft trades about 29% below its 52-week high of $542, a drawdown that erased roughly $1.1 trillion of market value from a company still worth about $2.9 trillion. The decline is unusual in one respect: every published sell-side price target sits above the current price — a range of $400 to $870, with a mean near $560 [consensus estimates]. That gap between a falling tape and unbroken analyst optimism — the price falling while every published target still sits above it — is the tension a value-minded reader should weigh.

Cash conversion and the capex turn

The reason the stock fell while profit rose is on the cash-flow statement. Operating cash flow climbed 15% in FY2025, to $136.2 billion. Free cash flow did not follow — it slipped about 3%, to $71.6 billion — because additions to property and equipment jumped 45%, from $44.5 billion to $64.6 billion [10].

Source: FY2021–FY2025 Cash Flows Statements; FY2023–FY2025 per FY2025 10-K [11].

The widening gap between the two bars is capital intensity. Capital expenditure has risen from 27% of operating cash flow in FY2021 to 47% in FY2025 — a structural shift in how much of the cash the business generates gets reinvested before any reaches shareholders [12].

Source: derived from FY2025 Cash Flows Statements (capex ÷ operating cash flow) [13].

The most recent quarter shows the pattern intact and intensifying. In the March 2026 quarter, operating cash flow rose 26% to $46.7 billion, but free cash flow was only $15.8 billion after $31.9 billion of capital spending [14]. Management guided to roughly $190 billion of capital expenditure for calendar 2026 — nearly three times FY2025's $64.6 billion outlay — and to another year of double-digit revenue and operating-income growth in FY2027 [15].

The bull case for that spending is demand that management says exceeds supply: Microsoft Cloud revenue reached $54.5 billion in the quarter, up 29%; Azure grew 40% in constant currency; the AI business surpassed a $37 billion annual run rate, up 123%; and commercial remaining performance obligations — contracted revenue not yet recognized — stood at $627 billion, up 99% year over year [16]. The skeptic's case was put to management on that same call by an analyst: "there is a bit of a disconnect that makes investors a bit nervous between how fast they are seeing CapEx growing and how fast they are seeing revenue growing" [17].

What the price implies

At $385, Microsoft trades at about 28 times trailing earnings and 23 times the consensus FY2026 estimate — rich against the market, unremarkable against its own 14% earnings growth. The metric that looks stretched is free cash flow: at roughly 40 times FY2025 free cash flow, the shares carry a free-cash-flow yield near 2.5%, because the AI capital cycle is holding that denominator down [18].

Source: derived from FY2025 10-K [19] and consensus estimates; price of $385.10 at 10 July 2026.

In FY2025 operating cash flow rose 15% to $136.2 billion while free cash flow fell about 3% to $71.6 billion as capex reached 47% of operating cash flow (up from 27% in FY2021), the board cut share repurchases from a $32.7 billion FY2022 peak to $18.4 billion in FY2025 while capex tripled from $23.9 billion to $64.6 billion (capex ~3.5x buybacks), and even normalizing capex to ~30% of operating cash flow lifts free cash flow only to ~$95 billion for a ~3.3% yield — so the 29% de-rate from ~40x to 28x trailing earnings removed froth rather than delivering a margin of safety [20].

Consensus expects the growth to continue: revenue near $329.5 billion in FY2026 and $384.4 billion in FY2027, both about 17% higher year over year, with EPS reaching $16.82 and then $19.36 [consensus estimates].

Source: consensus estimates (24–54 analysts, as of July 2026).

Whether 23 times forward earnings is a fair price for 15–23% earnings growth is most sensitive to whether the free-cash-flow drag is temporary or permanent — the reason the report does not stop here; Valuation and Scenarios takes the multiple apart in full.

The balance sheet

One question a downside-focused reader asks first is answered quickly. Microsoft held $94.6 billion in cash and short-term investments against $43.2 billion of total debt at fiscal year-end — roughly $51 billion of net cash — on stockholders' equity of $343.5 billion [21]. A business that generates $136 billion of operating cash a year against $43 billion of debt does not carry meaningful solvency risk. The debate over Microsoft is about the return on its reinvestment, not its survival; the full balance-sheet detail is laid out in Financials and Estimates.

The question this report answers

Microsoft is the rare case where the market's premium franchise has become, on a 29% drawdown, something closer to a fallen one — trading below every published target while its own cash generation is being reshaped by the largest capital-spending program in its history. The question the chapters that follow set out to resolve: is that drawdown a durable, high-return software and cloud franchise on temporary sale, or the early repricing of a capital-intensity shift that permanently lowers how much cash Microsoft returns for each dollar of profit it reports?

The numbers behind the franchise

Microsoft's ten-year record is one of widening scale and expanding margins: revenue roughly tripled to $281.7 billion since FY2016, operating margin climbed from 29% to 46%, and diluted EPS compounded near 19% a year. Reported profit is well-backed by operating cash, but three accounting choices — a longer depreciation schedule, one-off tax benefits, and a falling tax rate — flatter the trend at the edges. Consensus sees mid-teens revenue growth continuing into FY2027.

This chapter is the numeric spine for the rest of the report: what the income statement, segments, cash flows, and balance sheet actually show across the last decade, where reported earnings are and are not fully cash-backed, and what analysts expect next.

A decade of compounding

Over FY2016–FY2025, revenue grew from $91.2 billion to $281.7 billion and operating income from $26.1 billion to $128.5 billion [1]. The single interruption is FY2018, when net income fell to $16.6 billion on a one-time $13.7 billion charge for the U.S. Tax Cuts and Jobs Act — a tax event, not an operating one. Growth then reaccelerated: the FY2020–FY2025 revenue CAGR is 14.5% and the EPS CAGR 18.8%, the gap between the two explained by margin expansion, buybacks, and a declining tax rate.

Source: derived from reported financials, FY2016–FY2025 10-Ks; FY2023–FY2025 figures per the FY2025 Income Statements [2].

The margin story is as important as the growth story. Operating margin rose from 37.0% in FY2020 to 45.6% in FY2025, and gross margin held near 69% throughout despite the mix shift toward lower-margin cloud services [3]. That a business adding tens of billions in capital-intensive cloud revenue still lifted operating margin by nine points is the clearest evidence of the pricing power the through-line asks about.

Source: derived from reported financials, FY2020–FY2025 10-Ks [4].

Where the profit comes from

Microsoft reports in three segments, recast in FY2025 to move some products between them. On the current basis, Productivity and Business Processes (Microsoft 365, LinkedIn, Dynamics) generated $120.8 billion of revenue and $69.8 billion of operating income in FY2025; Intelligent Cloud (Azure, server products) $106.3 billion and $44.6 billion; and More Personal Computing (Windows, gaming, search, devices) $54.6 billion and $14.2 billion [5]. Productivity carries the fattest margin at 58%; Intelligent Cloud earns 42% and is the fastest grower.

Source: FY2025 Annual Report, Note 19 Segment Information [6].

The cleaner lens is "Microsoft Cloud," the cross-segment metric management steers to. It grew from $111.6 billion in FY2023 to $168.9 billion in FY2025 and is now the majority of revenue [7]. Forward visibility is unusually high for a company this size: commercial remaining performance obligations — contracted revenue not yet recognized — stood at $368 billion at June 30, 2025, against total RPO of $375 billion [8]. That backlog is the demand-side answer to the capex the report keeps returning to.

Earnings quality: cash-backed, with three flatterers

The first-order question a skeptic asks is whether reported profit is real cash. At the operating line, it is: FY2025 operating cash flow of $136.2 billion exceeded net income of $101.8 billion, a cash-conversion ratio of 1.34, held up by $34.2 billion of depreciation and amortization and $5.4 billion of growth in unearned revenue [9]. Accruals are not the problem. Where cash conversion has weakened is below operating cash flow — free cash flow fell to 70% of net income in FY2025 from roughly 90% four years earlier — and that gap is capital expenditure, an investment choice, not an accounting one. The mechanics of that capex turn are the subject of Franchise and Drawdown; here the point is narrower: the deterioration sits in capex, not in the quality of the earnings themselves.

Source: FY2025 Annual Report, Cash Flows Statements; free cash flow derived as operating cash flow less additions to property and equipment [10].

Three choices flatter the reported trend, and a reader comparing years should hold them in view.

Longer asset lives. In July 2022 Microsoft extended the estimated useful lives of server and network equipment from four years to six. The change lifted FY2023 operating income by $3.7 billion and net income by $3.0 billion — $0.40 of diluted EPS — by slowing depreciation [11]. The judgment is defensible, but it means part of FY2023's reported margin came from an estimate revision rather than operations, and it lowers the depreciation charged against each year of the AI build-out that followed.

One-time tax benefits. FY2022 net income included a $3.3 billion benefit ($0.44 EPS) from transferring intangible property out of Puerto Rico, and FY2021 a smaller benefit from an India Supreme Court ruling [12]. Those pushed the effective tax rate down to 13% in FY2022 and 14% in FY2021. The apparently flat net income from FY2022 ($72.7 billion) to FY2023 ($72.4 billion) is therefore misleading in both directions — FY2022 was lifted by the tax benefit, FY2023 by the depreciation change.

A drifting tax rate. The effective rate has since fallen from 19.0% in FY2023 to 18.2% in FY2024 to 17.6% in FY2025, on shifts in the geographic mix of earnings and foreign tax-credit timing [13]. Roughly 1.4 points of rate over two years is a modest tailwind to EPS that would reverse if the mix shifted back.

None of this is aggressive accounting; the disclosures are clear and the cash backing is strong. The takeaway is that a slice of the reported margin expansion and EPS growth is estimate- and tax-driven rather than purely operational — worth normalizing before extrapolating.

Stock-based compensation is a genuine, growing expense but a contained one: $12.0 billion in FY2025, about 4% of revenue, up from $9.6 billion in FY2023 [14]. It is expensed in GAAP earnings, and gross buybacks of $18.4 billion in FY2025 more than offset it — diluted share count fell from about 7.61 billion in FY2021 to 7.47 billion in FY2025 [15].

A fortress balance sheet

The balance sheet carries no solvency question. At June 30, 2025 Microsoft held $94.6 billion of cash and short-term investments against $43.2 billion of total debt — a net cash position of roughly $51 billion — with $343.5 billion of shareholders' equity [16]. What is changing is the composition of assets: property and equipment, net has grown from $59.7 billion in FY2021 to $205.0 billion in FY2025, now a third of the balance sheet, as the AI infrastructure moves from the cash-flow statement onto the balance sheet [17].

Source: FY2025 Annual Report, Balance Sheets (June 30, 2025), figures in $B [18].

Capital returns are steady rather than aggressive: dividends of $3.32 per share declared in FY2025 (a 21st consecutive annual increase) plus buybacks absorb a little over 40% of free cash flow, leaving the balance sheet to fund the capex internally so far [19].

What analysts expect

Consensus has growth decelerating only modestly. For FY2026, 52 analysts model revenue of about $329.5 billion (up 17%) and EPS near $16.82 (up 23%); for FY2027, revenue of about $384.4 billion (up 17%) and EPS of $19.36 (up 15%). The EPS-above-revenue growth in FY2026 assumes further margin expansion and the buyback, echoing the last decade's pattern rather than breaking from it.

Source: FY2025 actual per the FY2025 10-K Income Statements [20]; FY2026E–FY2027E are consensus estimates (52–54 revenue contributors), as reported.

The dispersion is the more revealing number. Published price targets run from $400 to $870, with a mean of $560 against a $385 share price, and the analyst tally is 12 strong-buy, 41 buy, 3 hold, and no sell ratings. A more than two-to-one spread between the highest and lowest target on a mega-cap is wide, and it maps directly onto the report's central question: the bulls capitalize the RPO and the AI run-rate, the cautious discount the $190-billion-a-year capex denominator.

Source: FY2025 actual per the FY2025 10-K [21]; FY2026E–FY2027E consensus range (24–36 EPS contributors), as reported.

The estimates carry a caveat the reader should hold: they are near-unanimously constructive, and the modest recent slippage in near-term quarterly EPS revisions is not yet visible in the full-year numbers. Consensus is describing continuation, not the scenario in which the capital-intensity shift the report is built around actually compresses returns.

Moat and Rivals

Microsoft does not have one moat; it has two very different ones stacked inside a 46% operating margin. In productivity software the advantage is wide and shows up cleanly in the numbers — a 58% segment margin on a locked, still-growing installed base. In cloud it is real but genuinely contested: Azure grew 40% last quarter, yet Google Cloud grew 63% and Amazon's AWS is larger than both. The AI tailwind behind all of it is quantified and strong, but it lifts every rival, and the capital it demands is the tension this report is built around.

The moat shows up in the margin

The first evidence for a durable franchise is that Microsoft keeps more of every revenue dollar than almost any company at its scale. In FY2025 it earned $128.5B of operating income on $281.7B of revenue, a 45.6% operating margin [1]. Against the mega-cap technology set that Microsoft's own 10-K names as competitors, that margin sits at the top — above Apple and Alphabet (both around 32%), Oracle (31%), Salesforce (20%), and well above Amazon, whose retail drag holds the group at 11% [2].

Source: operating income as a percent of revenue, each company's latest annual report (Form 10-K) — Microsoft FY2025 [3]; Alphabet FY2025 [4]; Apple, Oracle, Salesforce and Amazon as reported.

A blended margin, though, hides more than it shows. Microsoft reports in three segments, and their economics are not remotely alike: Productivity and Business Processes earned a 58% operating margin in FY2025, Intelligent Cloud 42%, and More Personal Computing 26% [5]. The blended margin hides two unlike moats. It is widest exactly where the capital intensity is lowest, and narrowest where the AI build is concentrated.

Source: FY2025 Annual Report (Form 10-K), MD&A — Reportable Segments; Productivity operating income $69.8B on $120.8B revenue, Intelligent Cloud $44.6B on $106.3B, More Personal Computing $14.2B on $54.6B [6].

The wide moat: the productivity franchise

The 58% margin in Productivity is the clearest moat Microsoft owns, and it is a switching-cost moat, not a technology one. Microsoft 365 sits inside the identity, document, and collaboration fabric of the enterprise — the directory people log in through, the files they have kept for a decade, the formats their partners expect. Leaving is not a software swap; it is a migration of the whole company. The number that reveals the lock-in is that paid Microsoft 365 commercial seats still grew 6% year over year off an installed base already numbered in the hundreds of millions, with growth led by average revenue per user rather than raw seat count as customers move up to the E5 tier and add Copilot [7]. A franchise that can raise price per seat on a base that large, year after year, is exercising pricing power a challenger cannot replicate by writing better code.

Copilot is the current test of that pricing power, and so far it is passing. Microsoft 365 Copilot paid seats surpassed 20 million, seat additions grew 250% year over year, and the count of customers with more than 50,000 Copilot seats quadrupled [8]. The significance is less the absolute number than the mechanism: Microsoft is monetizing AI by upselling a base it already owns — a per-seat premium layered on an existing subscription — rather than fighting to acquire new customers. That is the cheapest possible distribution for a new product, and it is available to almost no one else.

Sources: FY2025 Annual Report, segment KPIs (89M consumer subscribers, 6% commercial seat growth), as reported; Q3 FY2026 earnings call (over 20 million Copilot paid seats) [9] [10].

The read on this segment is that the moat is wide and has survived the two shifts most likely to break it — the move from packaged Office to cloud subscription, and the arrival of generative AI, which many assumed would let a startup route around the incumbent. Neither has happened. The main risk to that read is that Google Workspace and AI-native document tools slowly erode the low end; the seat-growth disclosure notes that gains came primarily from small-business and frontline offerings [11], which is where price competition bites first. What would change the read is seat growth turning negative or ARPU stalling — neither is visible yet.

The contested moat: a three-horse cloud race

Intelligent Cloud is where the investment case lives and where the moat is thinnest. Azure is a genuine number-two platform with real structural advantages — hybrid deployment, global scale, and the same enterprise identity relationships that anchor Office — and its growth is strong: Azure and other cloud services revenue grew 40% in constant currency in the March 2026 quarter [12]. But "strong" is not the same as "dominant," and the competitive tape has moved against the idea that Azure is pulling away.

In the same quarter, Amazon's AWS grew 28% year over year — its fastest in fifteen quarters — and crossed a $150B annualized run-rate, still the largest of the three [13]. Google Cloud grew 63% to $20B in the quarter, the fastest of the three by a wide margin, at a 32.9% operating margin, with backlog reaching $462B [14]. The growth-rate lead in cloud, in other words, no longer belongs to Microsoft — it belongs to the smallest of the three players.

Sources: Azure +40% constant currency, Q3 FY2026 call [15]; AWS +28% YoY, $150B run-rate [16]; Google Cloud +63%, ~$80B run-rate [17]. Microsoft does not disclose a standalone Azure revenue figure.

Microsoft's own filing is unusually candid about how contestable this is. The 10-K states plainly that "barriers to entry in many of our businesses are low" and that markets evolve with "changing and disruptive technologies" [18]. It also concedes that Azure's AI offerings compete with rivals "many of which are also current or potential partners" [19]. That last phrase describes a real erosion risk: to keep customers on Azure, Microsoft now hosts the models of its rivals. Management markets "the broadest selection of models of any hyperscaler," running OpenAI, Anthropic, and open-source models side by side [20]. That wins workloads, but it also commoditizes the model layer and turns Azure into a landlord competing on rent — scale, price, and reliability — rather than on a proprietary product. In cloud the read is a moderate moat, resting on scale and switching costs, narrowing rather than widening.

The tailwind is real, quantified, and shared

The industry force behind every one of these numbers is the migration of computing to the cloud, now compounded by AI. For Microsoft the tailwind is measurable, not rhetorical. Commercial remaining performance obligations — contracted revenue not yet recognized — reached $627B, up 99% year over year, with a weighted-average duration of about two and a half years [21]. The AI business alone surpassed a $37B annual run-rate, growing 123% [22]. This is genuine forward demand under contract, and it is the clearest evidence that the capex in the Return on AI Capital chapter is backed by contracted revenue.

Source: Q3 FY2026 earnings call — commercial RPO $627B (+99%), AI run-rate over $37B (+123%) [23] [24].

Three qualifications keep the tailwind from being a clean bull point. First, it lifts the rivals as much as Microsoft: Google Cloud's backlog of $462B [25] and AWS's re-acceleration [26] show the same wave, so a strong tailwind is not by itself a competitive edge. Second, the headline demand is concentrated: Microsoft's RPO grew only 26% excluding OpenAI, so the near-doubling leans heavily on one counterparty that Microsoft also funds [27]. Third, capturing the tailwind is exactly what compresses margins today — the 10-K notes that building AI infrastructure is "reducing operating margins" [28]. The tailwind is strong and durable; it is not free, and it is not Microsoft's alone.

Where the moat is, and what would move the read

The honest summary is a bifurcated one. In productivity the moat is wide, visible in a 58% margin and a base Microsoft can keep re-pricing, and it has survived two platform shifts; execution there compounds a real structural advantage. In cloud the moat is moderate and contested — scale and enterprise switching costs are genuine, but a well-funded rival is now growing faster, the model layer Microsoft sells is commoditizing, and the company's own filing calls the barriers to entry low. Execution is not a moat, and much of Azure's recent story is execution against opponents who execute too.

The read is most sensitive to the OpenAI relationship: OpenAI is simultaneously the source of the AI run-rate, the largest driver of the RPO backlog, and a company Microsoft both funds and competes with as it broadens its model roster. How that relationship — its accounting, its dilution, and its concentration — bears on the durability of the demand shown here is taken up in The OpenAI Stake.

Return on AI Capital

Microsoft's AI build-out is backed by real, contracted demand — a $627 billion backlog and an AI run-rate near $37 billion — but the return on it is not yet visible in reported profit. Capital spending of $64.6 billion in FY2025 ran at nearly three times the $22.0 billion depreciation charge, so the cost of this capital has only partly landed. And a six-year server life, set in 2022 just as a rival now shortens its own for AI hardware, flatters today's margin. The demand is contracted; the return on the capital is not yet demonstrable in reported profit.

The cash-conversion gap this report is built around lives almost entirely in this line item. Chapters on the numeric spine established that reported profit is cash-backed at the operating line; the open question is whether the cash Microsoft is now redirecting into infrastructure earns an adequate return, or whether reported margins are being held up by how slowly that infrastructure is being expensed. The forward walk of those margins against consensus is developed in Margin Bridge; this chapter stays on the return the capital earns.

The spend, and how fast it is expensed

Capital expenditure has more than tripled in four years — from $20.6 billion in FY2021 to $64.6 billion in FY2025 [1]. Depreciation, the charge that turns that spending into a cost of doing business, rose far more slowly, from $9.3 billion to $22.0 billion, and actually fell in FY2023 despite a larger asset base [2] [3].

Source: FY2025 Annual Report (Form 10-K), Consolidated Statements of Cash Flows [4] and Property and Equipment note [5].

The two lines diverge because the fleet is young. When a company adds assets far faster than it depreciates the ones it already owns, the depreciation charge is still climbing toward the run-rate of spending. Property and equipment, net of accumulated depreciation, reached $205.0 billion at June 2025, up from $135.6 billion a year earlier; accumulated depreciation itself grew by $17.2 billion in the year, to $93.7 billion [6]. Management guided to roughly $190 billion of capital expenditure in calendar 2026, and stated it remains "confident in the return on these investments given higher demand signals and increasing product usage" [7]. On that trajectory, the depreciation line has years of catch-up ahead of it, and the reported operating margin of 46% reflects an asset base still only partly charged.

The demand behind it is real, and increasingly long-dated

The build is not speculative. The commercial remaining performance obligation — contracted revenue not yet recognized — stood at $627 billion at March 2026, up 99% year over year, and the AI business reached a $37 billion annual revenue run-rate, up 123% [8]. Azure grew 40% in constant currency in the same quarter, with demand described as exceeding available capacity [9].

Source: Q3 FY2026 Earnings Call, CFO prepared remarks [10].

Two qualifications keep that backlog honest. First, its growth is concentrated: excluding Azure commitments from OpenAI, commercial RPO grew 26%, not 99% [11]. Much of the headline figure rests on a single counterparty that is itself a heavy user of Microsoft's own capital. Second, the backlog is getting longer-dated: Microsoft expected to recognize about 40% of total remaining performance obligation within twelve months as of June 2025 [12], but only about 25% as of March 2026, with a weighted-average duration near two and a half years [13]. The demand is contracted, but less of it converts to revenue soon.

The depreciation question — a six-year life against faster-aging hardware

Whether this capital earns its keep turns partly on an accounting estimate: how long the equipment lasts on the books. Microsoft has twice extended that estimate. In FY2021 it lengthened server life from three to four years [14]; effective FY2023 it extended server and network equipment from four to six years, which is why depreciation dipped that year even as the fleet grew [15]. Its FY2025 filing still depreciates computer equipment over "two to six years" [16].

That estimate is moving in the opposite direction from a direct peer. Amazon shortened the life of a subset of its servers and networking equipment from six years to five, effective January 2025, citing "the increased pace of technology development, particularly in the area of artificial intelligence and machine learning" [17]. Alphabet depreciates servers and network equipment over roughly six years [18].

Sources: Microsoft FY2025 10-K [19] and Q4 FY2022 call [20]; Amazon FY2025 10-K [21]; Alphabet FY2024 10-K [22].

The tension matters more now than a year ago because the composition of the spending has shifted toward exactly the assets that age fastest. In the June 2025 quarter, management said more than half of capital spending went to long-lived assets supporting monetization "over the next fifteen years and beyond," with the remainder on servers, "both CPUs and GPUs" [23]. By the March 2026 quarter the mix had flipped: "roughly two-thirds of our CapEx was for short-lived assets, primarily GPUs and CPUs," a proportion the company expected to hold through calendar 2026 [24]. GPUs depreciated over five years rather than six carry a 20% higher annual charge on the same asset — arithmetic that Amazon's disclosure applies to itself and Microsoft, so far, does not. Microsoft does not break out a GPU-specific life, so the effect cannot be sized precisely from the filings; the direction, however, is not in doubt, and the cloud gross-margin trend already shows it — Microsoft Cloud gross margin fell to 69% in FY2025 "driven by the impact of scaling our AI infrastructure" [25], and to 66% by the March 2026 quarter [26].

Management's answer, and what would change the read

Microsoft's CFO addressed the risk directly. As spending pivots toward short-lived GPUs and CPUs, she argued, those assets are "generally aligned with the duration of the contracts," so the depreciation on them tracks the revenue they earn; the long-lived portion is datacenter capacity leased over 15 to 20 years that she is "very confident" will be used [27]. It is a coherent defense: if a five-year GPU serves a contract of similar length, matching the two is exactly what accounting is meant to do, and the concern is not that the assets are mismarked but that the six-year disclosed life may be generous relative to how the newest, GPU-heavy cohort actually performs.

One further item sits alongside the capital question rather than inside it. Microsoft holds about 27% of OpenAI, accounted for under the equity method, with $13 billion of funding commitments of which $11.8 billion was funded by March 2026 [28]. That stake has swung reported results in both directions: OpenAI-related losses cut net income by about $2.0 billion ($0.28 of EPS) in the first nine months of FY2025, while the first nine months of FY2026 carried a one-time $4.5 billion gain ($0.60 of EPS) from OpenAI's recapitalization [29]. OpenAI is thus both the largest single driver of the RPO backlog and a volatile line below operating income — a reason to read the AI return through operating metrics, not headline EPS.

The measured read: the capital is backed by contracted, fast-growing demand, so this is an investment build rather than a speculative one, but the return is not yet demonstrable in the numbers, and two things make today's reported margin flatter than the likely steady state — a depreciation charge running well below the pace of spending, and a six-year server life that looks generous against a now-majority-GPU fleet and a peer moving the other way. What would change the read, in either direction: the short-lived asset mix normalizing rather than rising further; Microsoft Cloud gross margin stabilizing instead of stepping down each quarter (guided to about 64% for the June 2026 quarter [30]); depreciation converging toward capex without margins giving way; and commercial RPO excluding OpenAI continuing to compound in the mid-20s. Those are the lines to watch in each quarterly filing.

Margin Bridge

Consensus has Microsoft's revenue growing about 17% a year through FY2027 while EPS grows faster — margin expansion. Three cost waves the reported margin has so far outrun are still building: owned depreciation catching up to a capex line running near three times it, finance-lease amortization and interest compounding, and a $196.6 billion off-balance-sheet lease book that doubled in nine months and lands through FY2031. This chapter walks what that does to the forward margin.

The capital-intensity question the report is built around shows up in cash flow first — the Return on AI Capital chapter traced the capex-versus-depreciation gap. It reaches reported earnings through a slower channel: the depreciation and lease charge that the profit-and-loss statement has not yet fully absorbed.

What consensus is modeling

Consensus is not modeling a flat business. Across the analysts covering the stock, FY2026 revenue lands near $329.5 billion and FY2027 near $384.4 billion — roughly 17% growth in each year — while diluted EPS is put at $16.82 for FY2026 and $19.36 for FY2027, against $13.64 reported in FY2025.

Source: consensus analyst estimates, as of July 2026, against reported FY2025 results. EPS growth exceeding revenue growth implies margin expansion.

Two cautions sit under those numbers. First, the FY2026 EPS figure is flattered by a one-time item: the gain Microsoft booked when OpenAI recapitalized added about $1.02 to a single quarter's diluted EPS and roughly $5.9 billion over nine months, none of it operating or repeatable — the mechanics are in The OpenAI Stake. Excluding it, clean earnings growth is lower, closer to the revenue pace. Second, even holding the operating margin flat — not expanding it — requires revenue growth and cost discipline to fully absorb the depreciation and lease charge that is still landing.

That operating margin has, in fact, been rising. It went from 41.8% in FY2023 to 44.6% in FY2024 to 45.6% in FY2025 [1], and reached 46% in the March 2026 quarter [2]. Consensus continuation embeds that level holding or climbing. The question this chapter puts to that assumption is whether it can, given what is building beneath the line.

Wave one: depreciation catching up to capex

Microsoft's reported depreciation has lagged its spending. In FY2025 the property-and-equipment depreciation charge was $22.0 billion, up from $15.2 billion and $11.0 billion in the two prior years [3], while additions to property and equipment ran at $64.6 billion [4]. Depreciation of roughly a third of capex means most of the cost of the AI fleet has not yet reached the income statement. The comprehensive non-cash charge — depreciation, amortization, and other in the cash-flow reconciliation — grew 53% to $34.2 billion in FY2025, against 15% revenue growth [5].

Source: FY2025 Annual Report (Form 10-K), Cash Flows Statement [6]. Total D&A includes finance-lease right-of-use amortization; capex excludes leased assets.

Management guides calendar-2026 capital spending to roughly $190 billion, with the short-lived-asset mix — GPUs and CPUs — holding near the two-thirds share of the March quarter [7]. A single year of spending at that rate implies a steady-state depreciation run-rate well above today's charge — on the order of $35 billion just from one year's additions, against the $22.0 billion booked in FY2025 (workings below). Spending recurs and grows, so the depreciation line climbs for years before it stops catching up.

Wave two: the lease book, on and off the balance sheet

The second wave runs through leases, and it is moving faster. Total finance-lease cost — right-of-use amortization plus interest — rose from $1.9 billion in FY2023 to $4.8 billion in FY2025 [8], and in just the first nine months of FY2026 reached $5.7 billion, already above the full prior year, split $3.9 billion amortization and $1.8 billion interest [9]. The interest portion sits below operating income, so it pressures net income even in a quarter where the operating margin holds.

Source: FY2025 Annual Report, Leases note [10]; Q3 FY2026 Form 10-Q, Note 12 [11]. FY2026 shown at nine months; full-year run-rate is higher.

What is already on the balance sheet is only the leading edge. Finance-lease property and equipment stood at $77.6 billion at cost by March 2026, up from $53.9 billion nine months earlier, and the finance-lease liability of $62.9 billion carries $21.7 billion of imputed interest still to be expensed over a weighted-average thirteen-year term [12]. That $21.7 billion is interest expense the company is contractually committed to recognize, independent of how demand evolves.

Behind that sits the larger number. Datacenter leases signed but not yet commenced — off the balance sheet entirely until the space is ready — reached $196.6 billion at March 2026, from $92.7 billion at the June 2025 year-end, more than doubling in three quarters [13]. These commence between FY2026 and FY2031 with terms up to 21 years.

Source: FY2024 10-K [14]; FY2025 10-K [15]; Q1 FY2026 10-Q [16]; Q2 FY2026 10-Q [17]; Q3 FY2026 10-Q [18]. Jun 2024 combines operating and finance leases not yet commenced.

The FY2024-to-FY2025 dip is informative rather than reassuring: leases move off this line as they commence and onto the balance sheet, so a falling balance meant space was going live faster than new deals were signed. Through FY2026 the direction reversed hard as AI-datacenter commitments were booked. Amortized over its roughly thirteen-year weighted term, a $196.6 billion book implies on the order of $15 billion a year of right-of-use amortization at full commencement, plus interest — a charge that phases in over FY2026–FY2031 regardless of how the AI revenue lands.

The gross-margin anchor

The clearest place the wave already shows is Microsoft Cloud gross margin, the profitability of the segment absorbing most of the spend. It has stepped down each measurement: 71% in FY2021, 69% in FY2025 as AI-infrastructure scaling outweighed Azure efficiency gains, 66% in the March 2026 quarter, and guided to roughly 64% for the June quarter [19] [20].

Source: FY2021 and FY2025 figures per management KPIs; Q3 FY2026 and Q4 guide per the Q3 FY2026 earnings call [21] [22].

The sensitivity is straightforward. Microsoft Cloud revenue was $168.9 billion in FY2025 [23], so each percentage point of cloud gross margin is worth roughly $1.7 billion of gross profit — and more as the base grows. The five-point slide already visible, from 69% toward the guided 64%, is on the order of $8 to $9 billion of annual gross profit at that revenue base, before FY2027's higher revenue is applied.

The counter-case

None of this has cracked the reported margin yet, and the reasons it hasn't are real. Consolidated operating margin reached 46% in the March 2026 quarter, up slightly year over year, because Microsoft offset the cloud-gross-margin compression with cost discipline elsewhere — headcount fell year over year and operating expenses grew only 9% against 18% revenue growth [24]. Management's position is that the short-lived GPU and CPU assets are matched to the duration of the contracts they serve, so their depreciation tracks the revenue they generate rather than running ahead of it, and it remains confident in the return given an AI run-rate above $37 billion growing 123% [25]. Mix helps too: as higher-margin Copilot and Microsoft 365 revenue scales, it lifts the blended margin against the cloud-infrastructure drag.

The honest read is that reported margin expanded over the last three years while the depreciation charge sat at a third of capex and the largest leases were still off the balance sheet. Both of those conditions are now reversing at once: the property-depreciation run-rate is heading from $22 billion toward the mid-$30-billions from a single year's spending, finance-lease cost has already passed a $7-billion annual pace, and a $196.6 billion lease pipeline begins converting into amortization and interest through FY2031. Set against FY2027 revenue near $384 billion, an incremental $20-to-$25 billion of combined charge is six-odd points of margin that revenue growth and cost discipline must more than absorb for the expansion consensus models to hold.

This is not a call that the margin collapses — Microsoft has the growth and the mix to fight the drag, and it has done so far. It is that the cost wave now runs opposite to the tailwind that produced the last three years of expansion, which makes continued margin expansion a demanding base case rather than an extrapolation. The read would flip constructive if the combined depreciation-and-lease charge stabilizes as a share of revenue and Microsoft Cloud gross margin holds at or above the guided 64%; it would confirm the concern if the charge keeps climbing faster than revenue and the guide steps below 64% into FY2027.

The OpenAI Stake

Microsoft owns about 27% of OpenAI, funds it, and sells it Azure capacity — one relationship that now sits on both sides of the ledger, flattering reported earnings while backing much of the contracted demand behind the AI build. In the December 2025 quarter investments in OpenAI added $10.0 billion to other income and $1.02 to diluted EPS, lifting reported EPS to $5.16 (+60%) against adjusted EPS of $4.14 (+24%), while in the March 2026 quarter that same OpenAI relationship carried Microsoft's commercial RPO to +99% growth ($627 billion) versus only +26% excluding OpenAI and turned commercial bookings to -4% including OpenAI versus +7% without it — and Microsoft's equity-method loss from OpenAI was about $4.1 billion in the September 2025 quarter alone, nearly the entire $4.7 billion 'Other, net' drag recorded across all of fiscal 2025, even as Microsoft surrendered its right of first refusal to be OpenAI's compute provider as OpenAI's Azure commitment rose by an incremental $250 billion. One counterparty thus supplies about 20% of the December-quarter diluted EPS ($1.02 of $5.16) and roughly 73 points of the 99% commercial RPO growth: the same loss-making, part-owned customer does double duty as earnings flatterer and demand backstop. The counter-fact sits in the same disclosure — the funding is capped at a disclosed $13 billion, Microsoft voluntarily publishes an adjusted EPS that excludes OpenAI, and the equity-method losses are immaterial against net income above $100 billion, disclosure hygiene rather than concealment. Reported earnings have nonetheless become a weaker guide to run-rate profit than at any recent point.

A stake that reversed the sign on earnings

For two years the OpenAI investment was a drag on reported profit. Microsoft accounts for it under the equity method, recognizing its share of OpenAI's operating losses through the "Other, net" line of other income. That line deepened from a $223M loss in FY2023 to $1,319M in FY2024 and $4,725M in FY2025, and management attributes it primarily to net losses on equity-method investments, including OpenAI [1]. In the nine months to March 2025 the OpenAI drag alone was $2.7B [2].

The recapitalization reversed that. On October 28, 2025 Microsoft signed a new agreement, OpenAI formed a public benefit corporation, and Microsoft's holding was set at approximately 27% on an as-converted diluted basis. The step-down in proportionate ownership produced a dilution gain [3]. In the December 2025 quarter, investments in OpenAI added $10.0 billion to other income, and $5.9 billion for the six months — figures the filing states relate primarily to that dilution gain [4].

Source: Q3 FY2026 Form 10-Q, Note — Other Income (Expense), Net [5].

Because the gain landed in one quarter, it dominates the year. The March 2026 quarter carried a negligible $14M OpenAI drag on net income [6], so essentially all of the nine-month swing is the single recapitalization event. Microsoft itself now publishes an adjusted diluted EPS that strips OpenAI out. On that measure, the December quarter's headline 60% EPS growth becomes 24%, and the OpenAI dilution gain accounts for $1.02 of the $5.16 reported [7].

Source: Q2 FY2026 Form 10-Q, Non-GAAP Financial Measures reconciliation [8].

The sign flipped. A year earlier OpenAI was a small drag, so adjusted EPS sat above reported; after the recapitalization it is a boost, so reported sits above adjusted. Neither direction reflects Microsoft's operating performance.

How the number is built

The gain is an accounting re-measurement of a private stake, not cash and not operations. OpenAI is a variable interest entity, but Microsoft has determined it is not the primary beneficiary, so the entity is not consolidated and is carried under the equity method [9]. Microsoft does not simply book 27% of OpenAI's results. It uses the hypothetical liquidation at book value method — recognizing income or loss based on the change in what it would receive if OpenAI's net assets were distributed at book value — because its liquidation rights and priorities differ from its ownership interest [10].

The recapitalization also changed the basis of that calculation. As CFO Amy Hood described it, Microsoft now records gains or losses "based on our share of the change in their net assets on their balance sheet as opposed to our share of their operating profit or losses from their income statement" [11]. A change in a book-value measurement drove other income to $10 billion; adjusted for OpenAI, the same line was slightly negative [12].

The stake is small on the balance sheet relative to its earnings effect. Equity-method investments — of which OpenAI is the dominant part — rose from $6.0 billion at June 2025 to $11.5 billion at December 2025, against total funding commitments of $13 billion, $11.7 billion of it funded [13]. An investment carried near $11.5 billion generated a $10 billion swing in reported other income in a single quarter. That mechanics — a private, unobservable book value driving a nine-figure earnings line — is why Microsoft reports an adjusted number that excludes it.

The circular demand

The same counterparty sits on both sides of Microsoft's ledger. Microsoft funds OpenAI and holds roughly 27% of it; OpenAI is at the same time one of Microsoft's largest cloud customers.

Sources: Q1 FY2026 Form 10-Q, Note 17 [14]; Q2 FY2026 Form 10-Q, Note 1 [15].

The link is visible in the bookings line. In the September 2025 quarter commercial bookings rose 112%, which management attributed to Azure commitments from OpenAI alongside growth in large contracts — and that figure predated the new deal [16]. The October agreement then added an incremental $250 billion of contracted Azure purchases by OpenAI [17]. That commitment flows into commercial remaining performance obligation — $392 billion and up 51% at the September quarter, on a weighted-average duration of about two years [18]. The contracted backlog that underwrites the capital program is therefore concentrated in part on a counterparty Microsoft part-owns and funds.

The exclusivity that once justified the investment has loosened as the dollars have grown. The FY2025 10-K described OpenAI's API as exclusive to Azure with Microsoft holding a right of first refusal on OpenAI's new capacity needs [19]. Under the October agreement, Microsoft will no longer have that right of first refusal to be OpenAI's compute provider [20]. The relationship is deeper in committed spend and looser in exclusivity than it was a year ago.

The capital that is not on the balance sheet

Much of the datacenter capacity that serves the OpenAI-anchored demand is contracted through leases not yet on the balance sheet — $196.6 billion of not-yet-commenced leases, primarily for datacenters, at March 2026 [21] — committed capital the reported capex and depreciation lines do not yet carry; the Margin Bridge chapter walks that lease book and its finance-lease conversion in full.

What would change the read

OpenAI has turned Microsoft's GAAP earnings into a poor guide to run-rate profit — a multi-billion drag that became a one-time boost — and has tightened the loop between Microsoft's own capital, its equity stake, and the cloud demand that justifies the build, while the not-yet-commenced lease book shows the capital commitment is larger than the reported capex implies.

The strongest fact against reading this as a concern is Microsoft's own transparency. It discloses the HLBV method, the $13 billion commitment, the not-yet-commenced leases, and — unusually — publishes an adjusted EPS that removes the OpenAI noise entirely [22]. The equity-method losses are immaterial against net income above $100 billion, and the $250 billion is contracted revenue from a counterparty whose demand, management says, exceeds available supply [23].

Three things would sharpen the read. First, disclosure of OpenAI-related revenue as a separate line would let an investor judge how much of the RPO backlog is genuinely third-party. Second, the not-yet-commenced lease book flattening or falling would signal demand catching up to committed supply — while continued growth signals the opposite. Third, the pace at which finance-lease interest and depreciation from commenced datacenters land in the income statement will show whether the committed capital compresses the margins consensus extrapolates. Until OpenAI's own funding position and the pricing of its Azure commitments are visible, the circularity can be described but not fully priced.

Demand Durability

The contracted demand that justifies Microsoft's AI build is unusually concentrated in one counterparty. Microsoft's own disclosures show it: commercial remaining performance obligations grew 99% to $627 billion including OpenAI, but only 26% excluding it [1]. That counterparty is loss-making, spending far ahead of revenue, and dependent on serial external financing. Whether the capex earns a return turns partly on a customer whose solvency the filings do not vouch for.

How much of the backlog is one customer

Microsoft frames nearly every contracted-demand figure twice — with OpenAI and without — which makes the concentration measurable rather than a matter of inference. In the March 2026 quarter, commercial bookings fell 4% including OpenAI's Azure commitments but rose 7% excluding them; the RPO backlog grew 99% including OpenAI and 26% without [2]. Two quarters earlier, bookings had jumped 112% — "driven by Azure commitments from OpenAI" — and even that surge, Microsoft noted, excluded "the incremental $250 billion Azure commitments from OpenAI announced yesterday" [3].

Sources: Q3 FY2026 earnings call [4]; Q1 FY2026 earnings call [5].

The gap between the two bars is the load a single customer carries.

Source: Q3 FY2026 earnings call, CFO remarks [6].

An analyst put the concern to management directly. On the October 2025 call, JPMorgan's Mark Murphy described "a new era where the contractual commitments from a small number of AI natives are just incredibly large… contracts worth hundreds of billions of dollars that are 20 times their current revenue scale," and asked how Microsoft evaluates the ability of those companies to follow through and "placing guardrails on customer concentration for any single entity" [7]. The $250 billion Azure commitment against OpenAI's revenue base is roughly that multiple.

In the December 2025 quarter investments in OpenAI added $10.0 billion to other income and $1.02 to diluted EPS, lifting reported EPS to $5.16 (+60%) against adjusted EPS of $4.14 (+24%), while in the March 2026 quarter that same OpenAI relationship carried Microsoft's commercial RPO to +99% growth ($627 billion) versus only +26% excluding OpenAI and turned commercial bookings to -4% including OpenAI versus +7% without it — and Microsoft's equity-method loss from OpenAI was about $4.1 billion in the September 2025 quarter alone, nearly the entire $4.7 billion 'Other, net' drag recorded across all of fiscal 2025, even as Microsoft surrendered its right of first refusal to be OpenAI's compute provider as OpenAI's Azure commitment rose by an incremental $250 billion [8] [9] [10].

What the counterparty's economics look like

Microsoft does not report OpenAI's financials, but its own equity-method accounting is a window onto them. Microsoft holds an approximately 27% as-converted interest in OpenAI, accounted for under the hypothetical-liquidation-at-book-value (HLBV) method, and has committed $13 billion of funding, of which $11.8 billion was funded by March 2026 [11]. Across all of fiscal 2025, the "Other, net" line — which Microsoft says "primarily reflects net recognized losses on equity method investments, including OpenAI" — was a $4.7 billion drag [12].

Then the pace jumped. In the September 2025 quarter alone, the OpenAI investment sat in other income at roughly $4.1 billion of loss — nearly the entire prior full year in three months — and CFO Amy Hood confirmed "that increased loss was all due to our percentage of losses in OpenAI due to the equity method" [13]. HLBV recognition does not map linearly to OpenAI's total losses — Microsoft absorbs a share weighted by its liquidation priority, not a flat 27% — so the figure understates as much as it reveals. But the direction is unambiguous: the losses Microsoft records from OpenAI are large and accelerating.

External reporting fills in the counterparty's own scale. OpenAI was valued near $500 billion in its late-2025 recapitalization, with reported annualized revenue around $20 billion at the end of 2025 and cash losses on the order of $10 billion or more a year; its own circulated projections show losses continuing to build through 2028 before a targeted turn to positive cash flow near the end of the decade. Against that revenue base, OpenAI has signed roughly $1.1 trillion of multi-year compute commitments across a handful of suppliers.

Source: external reporting on OpenAI's disclosed infrastructure agreements, 2025–2026; figures are company announcements and press accounts, not Microsoft filings, and are estimates.

The arithmetic is stark: about $1.1 trillion of contracted spending against roughly $20 billion of revenue, funded by a plan that depends on both scaling and continued fundraising. Microsoft's $250 billion is the third-largest single line in that stack — and it is the RPO that Microsoft's investors are, in part, capitalizing when they underwrite the AI build.

The loop, and the backstop Microsoft gave up

The relationship is also circular in a way the filings make explicit. Microsoft funds and part-owns OpenAI, and OpenAI's Azure purchases flow back into the RPO backlog that underwrites Microsoft's capex — the same dollars appearing as Microsoft's investment on one side and Microsoft's contracted revenue on the other. The wider version of this loop, in which chip vendors, cloud providers, and model labs finance each other's purchases, has drawn open comparison to late-1990s vendor financing.

The October 2025 agreement sharpens the demand-durability question in a way that belongs to this chapter's lens. As OpenAI's dollar commitments rose, Microsoft gave up a protection: it "will no longer have a right of first refusal to be OpenAI's compute provider" [14]. The commitments grew even as exclusivity loosened. The same recapitalization also produced a one-time equity-method gain and reset the carrying value of the stake, but those accounting mechanics bear on reported earnings rather than demand durability and are taken up in The OpenAI Stake.

What makes the risk manageable — and what would change the read

Management's answer to the concentration question is the strongest counter-case, and it is a substantive one. Nadella argued that the "risk of concentrating on a single customer can be mitigated" because Microsoft builds "a highly fungible fleet" usable for first-party products, enterprises, and other digital natives alike [15]. Hood added the practical point that capacity commitments convert to cost on a delivery schedule: the "CPU and GPU and the storage gear doesn't come into play until the contracts start happening," giving Microsoft "a lot of lead time" to reassess as each tranche approaches [16]. If OpenAI faltered, the datacenters it would have filled could in principle serve the broader Azure book — where demand still "exceeds available capacity" and the AI run-rate reached $37 billion, up 123% [17]. Fungibility is the reason a counterparty shock need not become a stranded-asset event.

The bounding facts on the other side: RPO growth outside OpenAI is a healthy-but-ordinary 26%, so the 99% headline that most visibly "backs" the capex is not repeatable without the same customer; the right-of-first-refusal Microsoft surrendered was exactly the mechanism that would have redirected OpenAI's compute back to Azure; and re-tenanting purpose-built AI capacity at the same price and utilization is an assertion, not yet a demonstrated fact. The read that fits the evidence: the demand is real and partly fungible, but its growth is concentrated in a counterparty that is years from self-funding, so the return on the AI build is more exposed to one customer's trajectory than the consolidated backlog implies.

What would move that read, in Microsoft's own disclosures:

• RPO growth excluding OpenAI — if the ex-OpenAI line re-accelerates toward the headline rate, the concentration is diversifying away; if it stalls near or below the low-20s while the headline stays inflated, the reliance is deepening [18].
• The equity-method loss run-rate — a quarterly OpenAI drag that keeps climbing from the ~$4.1 billion September-quarter mark signals the counterparty's burn is widening faster than its revenue [19].
• Azure ex-AI utilization — Microsoft's fungibility defense only holds if non-OpenAI, non-AI demand can absorb capacity; continued "exceeds available capacity" commentary supports it, a swing to excess capacity would not [20].

Ownership and Pay

Microsoft is a professionally run mega-cap, not a founder company. Directors and officers together own under 1% of the stock — about three-hundredths of a percent — and the largest holders are index funds. Alignment comes from pay design, not proprietary ownership: over 95% of the CEO's target compensation is performance-based, and hedging and pledging are banned. Pay is large and rising, and the board has steadily redirected cash from buybacks toward AI capital spending.

For an investor who prizes founder control and heavy insider skin in the game, this is the chapter where Microsoft does not fit the template — and where it is worth being precise about what replaces founder alignment, and what that costs.

Who owns Microsoft

There is no founder block on the register. Bill Gates left the board in 2020 and no longer appears among the beneficial owners; the two largest holders are passive index managers — The Vanguard Group at 8.95% and BlackRock at 7.30% as of the September 30, 2025 record date [1]. Every director, nominee, and named executive combined — 18 people — held 2,279,620 shares, which the proxy marks with an asterisk meaning less than 1% of the company [2]. Against roughly 7.4 billion shares outstanding, that group holds about 0.03%.

Satya Nadella himself owned 900,572 shares [3] — worth roughly $347 million at the July 10, 2026 close of $385. That is real money in absolute terms, but it is about 0.01% of the company he chairs and runs. The distinction matters for this kind of investor: Nadella is a steward paid in equity, not an owner-operator whose personal balance sheet rises and falls with the enterprise in any material proportion.

Sources: 2025 Proxy Statement, 5%-owner table [4]; director and officer ownership table [5]. Percentages of ~7.4B shares outstanding.

The board that oversees this is conventional large-cap: 12 director nominees at the December 2025 annual meeting, 11 of them independent, with Nadella the sole insider [6]. Nadella holds the combined Chairman and CEO role [7], a concentration of authority that the board offsets with a lead independent director rather than by splitting the seats.

Alignment by pay, not by stake

Because the ownership stake is small, alignment is engineered through how executives are paid rather than what they already hold. The mechanics are on the stronger end of large-cap practice. Over 95% of the CEO's annual target compensation is performance-based, and at least 70% of every named executive's target pay is delivered in equity [8]. Performance stock awards carry a relative total-shareholder-return modifier measured against the S&P 500; for the fiscal 2023–2025 period Microsoft placed in the 80th percentile, triggering a 150% payout on that portion [9].

Two governance details reinforce the alignment. Directors and executive officers are prohibited from hedging their Microsoft stock, from buying it on margin, and from pledging it as loan collateral [10] — so the equity they do hold cannot be quietly monetized or levered away from shareholders' interests. And shareholders vote on pay every year: at the 2024 annual meeting, over 91% of votes cast supported the say-on-pay resolution [11]. That is solid support, though below the mid-90s that the most uncontroversial pay programs draw.

What management is paid

The pay is large. Nadella's fiscal 2025 total compensation was $96.5 million — a $2.5 million salary and $84.2 million in stock awards, the rest in incentive and other pay [12]. The next four named executives clustered between $12 million and $29 million.

Source: 2025 Proxy Statement, Summary Compensation Table (fiscal 2025) [13].

Scale gives the number its context. Nadella's pay was 480 times the $200,972 total compensation of Microsoft's median employee [14]. And the trajectory is upward: compensation actually paid to the CEO — the SEC's mark-to-market measure that captures how outstanding equity revalues each year — rose from $56.9 million in fiscal 2022 to $131.1 million in fiscal 2025 [15].

The proxy frames pay against a favorable long-run record: over Nadella's tenure since 2014, Microsoft's cumulative total shareholder return exceeded 1,500% and its market capitalization grew by almost $3.4 trillion, as revenue tripled and net income quadrupled [16]. That case is real. The nearer-term picture is more even-handed: on the SEC's own pay-versus-performance table, a $100 investment at the start of the five-year window was worth $255 in Microsoft stock by fiscal 2025 against $276 for the peer group index [17]. Over that specific stretch, the equity-heavy pay grew while the stock modestly trailed its own technology peers.

What the board does with the cash

The largest capital-allocation decision the board has made is not on the pay tables — it is the steady rotation of shareholder cash from buybacks into AI infrastructure, the governance leg of the cash-return question that Valuation and Scenarios takes up in full.

Source: fiscal 2023–2025 figures from FY2025 Annual Report (Form 10-K), Cash Flows Statements [18]; earlier years derived from reported filings.

The reversal is clean. Share repurchases peaked at $32.7 billion in fiscal 2022, then fell to $18.4 billion in fiscal 2025 — nearly halved — while capital expenditure rose from $23.9 billion to $64.6 billion over the same three years [19]. In fiscal 2022 buybacks exceeded capex; by fiscal 2025 capex was roughly three-and-a-half times buybacks. The dividend was the untouched leg, growing each year from $18.1 billion to $24.1 billion [20].

That is a deliberate, reversible choice rather than a forced one — Microsoft carries net cash and generated $136 billion of operating cash flow in fiscal 2025 [21]. But it is the point at which management's incentives and the shareholder's return question meet: the same board that ties 95% of the CEO's pay to performance has bet a rising share of distributable cash on infrastructure whose return is still unproven, a return examined in Return on AI Capital.

Alignment and the cash-return choice

For a value investor who favors founder-owners with a large personal stake, Microsoft is a deliberate mismatch on ownership and a strong fit on alignment mechanics. Insiders hold a rounding error of the equity, so there is no owner-operator whose fortune moves with the business; what stands in for that is an unusually disciplined pay structure — overwhelmingly performance-based, equity-delivered, hedging- and pledging-banned, and voted on annually. The strongest fact against a benign read is that pay has climbed toward $131 million of compensation actually paid while the stock modestly trailed its technology peers over the measured five-year window, and that the board is steering an increasing share of shareholder cash into AI capex rather than buybacks. What would sharpen the picture in either direction: buybacks recovering toward their fiscal 2022 peak, which would signal management believes the capex wave has crested, or say-on-pay support slipping from the low-90s, which would signal owners are pushing back on the pay-versus-performance gap.

Valuation and Scenarios

At $385.10, Microsoft trades at 28x trailing and roughly 23x forward earnings, about 40x free cash flow, and a 2.5% free-cash-flow yield. In FY2025 operating cash flow rose 15% to $136.2 billion while free cash flow fell about 3% to $71.6 billion as capex reached 47% of operating cash flow (up from 27% in FY2021), the board cut share repurchases from a $32.7 billion FY2022 peak to $18.4 billion in FY2025 while capex tripled from $23.9 billion to $64.6 billion (capex ~3.5x buybacks), and even normalizing capex to ~30% of operating cash flow lifts free cash flow only to ~$95 billion for a ~3.3% yield — so the 29% de-rate from ~40x to 28x trailing earnings removed froth rather than delivering a margin of safety. That is a de-rate, not a discount: the fall from October's $542 high compressed a peak-cycle multiple to a still-premium one, and the return case rests on growth and on capex normalizing, not on a low entry price.

What the price is

Sources: share price and 52-week high from market data (close of $385.10 on July 10, 2026; high of $542.07 on October 28, 2025); P/E on FY2025 diluted EPS of $13.64 [1]; FCF yield on FY2025 free cash flow of $71.6B [2].

Microsoft carries a market capitalization near $2.87 trillion (7,465 million diluted shares at $385.10) [3]. Against $94.6B of cash and investments and $43.2B of debt, roughly $51B of net cash sits inside that figure — about $6.89 per share — so on an enterprise-value basis the business is valued near $2.82 trillion [4]. The table below triangulates the price across the four lenses that matter for this model — earnings, cash, and enterprise value — and shows each at the current price and, where applicable, at October's peak.

Sources: EPS, share count and operating income from the FY2025 10-K income statement [5]; free cash flow (operating cash flow $136.2B less capex $64.6B) from the cash-flow statement [6]; forward EPS are consensus estimates; multiples derived at market prices.

The lenses agree on direction. Microsoft is priced as a growth compounder: even the FY2027 consensus earnings put the multiple near 20x, and the cash-based measures — 40x free cash flow, a 2.5% yield — are richer still because capex now absorbs a far larger share of operating cash than it did five years ago (Return on AI Capital).

How much the drawdown discounted

Source: market prices, monthly closing values, May 2024 – July 2026.

The reader's interest here is a fallen star, held to a strict margin-of-safety test. Microsoft fits the first half: the stock is down 29% from its October high and has round-tripped roughly two years of gains, and sentiment has clearly cooled. It fits the second half far less well. The drawdown moved the trailing multiple from about 40x earnings to 28x and the free-cash-flow yield from 1.8% to 2.5%. That is a repricing from expensive to less expensive — not to cheap. On the cash the business generates today, the yield remains below what a risk-free instrument offers, and the entire equity return must come from growth.

Where Microsoft does clear one of the reader's bars is solvency. With ~$51B of net cash and $136B of annual operating cash flow against $43.2B of debt, the chance of financial distress is immaterial [7]. The risk in the stock is not that the business fails; it is that the price already embeds a return the business may not deliver if AI capital earns less than hoped. And the buyback pullback that holds down today's free cash flow is a deliberate, reversible allocation choice, not distress: with $136B of operating cash flow the board can restore repurchases once capex crests, so the depressed cash return is a governance decision rather than a sign of strain.

What the current price implies

Taken as arithmetic rather than narrative, the multiple sets a hurdle. A 2.5% free-cash-flow yield means that, at a 9% required return, the price is consistent with free cash flow compounding at roughly 6-7% in perpetuity — a demanding assumption to hold forever, though not an implausible one for a franchise that has grown revenue at a mid-teens rate. The forward earnings yield of 4.4% frames the same point from the income statement: the equity offers little current yield, so nearly all of the expected return is deferred into growth that has not yet been earned.

The counter-argument is that current free cash flow is artificially depressed. Capex ran at 47% of operating cash flow in FY2025, against 27% four years earlier (Financials and Estimates). Normalize capex back toward 30% of operating cash flow and free cash flow would be near $95B rather than $71.6B, lifting the yield to about 3.3% [8]. That reframes the cash multiple — but even on normalized free cash flow the yield sits in the low-3% range, still short of a margin-of-safety threshold. The valuation question is therefore not whether reported free cash flow is understated (it is), but whether even the normalized figure justifies the price.

Three scenarios

The report's earlier chapters isolate the variables that decide the answer. Whether the AI build is investment or a permanent step-up in intensity (Return on AI Capital); whether reported margins survive depreciation catching up to capex and the removal of a one-time OpenAI gain (The OpenAI Stake); and whether the cloud franchise holds share and price as rivals scale (Moat and Rivals). The three paths below hold those variables at low, middle, and high settings. Prices are illustrative — earnings times an exit multiple, not forecasts — and are anchored to a fiscal-2028 view.

Sources: driver mechanics from the FY2025 10-K and the Q3 FY2026 earnings call (cloud gross margin guided to ~64%, Intelligent Cloud revenue growth 27-28%) [9]; base EPS path from consensus estimates; prices illustrative (EPS × exit multiple), not forecasts.

The base case is close to what analysts already model, and the market's published targets sit almost entirely above the current price. Of 56 targets, the lowest is $400 — still 4% above spot — with a median of $550 and a mean of $560, implying 43-45% upside, and a high of $870. No covering analyst carries a target below the price, and the rating book holds zero sell ratings.

Source: consensus analyst price targets (56 analysts) and current price, market data.

That one-sidedness is itself informative for a pessimism-priced-in read: after a 29% drawdown, the sell side still models continuation and not one target implies downside. The bear scenario is therefore not the consensus — it is the case the price is not discounting. The evidence that would move the read between these paths is specific and observable, and it is where the next few quarters will settle the through-line.

What to watch

Each item below is a single line in a named filing with a threshold that tilts the read. Together they distinguish the bull's "capex is investment, FCF re-expands" from the bear's "intensity is permanent, margins compress."

Sources: capex, depreciation and buyback lines from the FY2025 10-K cash-flow statement [10]; cloud gross-margin guide, RPO and adjusted-EPS framing from the Q3 FY2026 earnings call [11].

The case is most sensitive to the first two lines. If capex plateaus and cloud margin holds near 64%, free cash flow re-expands and today's 40x multiple compresses on its own, without a re-rating — the durable-franchise reading. If capex keeps climbing while cloud margin steps down, the depreciation and finance-lease costs the AI build has deferred begin to land on the income statement, and a 23x forward multiple built on extrapolated margins becomes the vulnerability. The March 2026 quarter already carried the tension: 21% earnings growth and a $37B AI run-rate up 123%, set against a company gross margin down year-over-year "driven by continued investment in AI infrastructure" [12]. Both readings remain live; the filings above are where each is confirmed or refuted.