Microsoft (MSFT) trades near $424 after Q3 beat as OpenAI revenue share is eliminated

Microsoft Corporation (NASDAQ: MSFT) closed its fiscal 2026 third quarter with revenue of $82.9 billion, up 18 percent year over year, and adjusted diluted earnings per share of $4.27, beating Wall Street consensus of roughly $81.4 billion in revenue and $4.07 in earnings. Azure and other cloud services grew 40 percent in constant currency, the AI business surged 123 percent to a $37 billion annualised run rate, and Microsoft 365 Copilot crossed 20 million paid commercial seats. Yet the stock entered the print already down roughly 12 percent year to date, with shares trading near $424 against a 52-week range of $356.28 to $555.45, and management guided fourth-quarter revenue at a midpoint of $87.25 billion, below the LSEG consensus of $87.53 billion. The deeper signal is calendar 2026 capital expenditure guided to roughly $190 billion, including a $25 billion lift from higher component pricing, a number that reframes the entire investment debate around Microsoft for the next eighteen months.

How does the Microsoft FY26 third-quarter beat reconcile with a Q4 revenue guide that fell short of Wall Street consensus?

The headline result was unambiguous. Revenue of $82.9 billion grew 18 percent on a reported basis and 15 percent in constant currency, operating income of $38.4 billion expanded 20 percent, and Microsoft Cloud revenue exceeded $54 billion, up 29 percent. Intelligent Cloud, the segment that houses Azure, generated $34.7 billion in revenue, up 30 percent. Azure itself, the closest single proxy for AI infrastructure demand, grew 40 percent in constant currency against a strong prior-year comparable, ahead of the 37 to 38 percent management had guided, with Microsoft attributing the upside to capacity coming online earlier than planned in the Fairwater datacenter in Wisconsin.

What complicated the picture was the forward look. The fourth-quarter revenue range of $86.7 billion to $87.8 billion implies year-over-year growth of 13 to 15 percent, but the midpoint sits below the consensus build that buy-side models had constructed around accelerating commercial cloud trends. The shortfall is not in the commercial business, where Microsoft 365 commercial cloud is guided to 13 to 14 percent constant-currency growth and Azure to 39 to 40 percent constant currency. It is concentrated in More Personal Computing, where Windows OEM revenue is expected to decline in the high teens against a prior-year comparable inflated by Windows 10 end of support, channel inventory normalisation, and a softer PC market dragged down by rising memory prices.

For executives modelling Microsoft as a pure cloud and AI story, the implication is that consumer-adjacent friction is masking the strength of the commercial trajectory. Investors who treat the segments as a single number will see a soft guide. Investors who decompose it will see a commercial business accelerating into the fiscal fourth quarter on a base of more than $600 billion in remaining performance obligations and a Copilot installed base that is finally beginning to translate seat additions into revenue per user expansion.

Why has Microsoft pushed calendar 2026 capital expenditure to $190 billion and what does it signal about AI infrastructure economics?

Capital expenditure is the line item carrying the entire investor narrative. Third-quarter capex of $31.9 billion came in below the $34.9 billion Visible Alpha consensus, but Chief Financial Officer Amy Hood guided the fourth quarter to over $40 billion and the full calendar year 2026 to roughly $190 billion, a figure approximately 61 percent above calendar 2025. Roughly $25 billion of the calendar 2026 number is attributable to higher component pricing rather than additional physical capacity, which provides some accounting comfort but does not eliminate the absolute scale of the spend.

The composition of that spend is what executives should focus on. Roughly two-thirds of third-quarter capex went to short-lived assets, primarily graphics processing units and central processing units, which Hood explicitly framed as the assets most directly correlated with near-term revenue. The remaining third went to long-lived datacenter shells and finance leases with monetisation horizons of fifteen years and beyond. This split tells executives that Microsoft is leaning harder into demand-driven, faster-payback silicon than into speculative real-estate buildouts, a discipline that distinguishes its capex profile from peers spending at similar absolute levels.

The market reaction to Microsoft capex has been instructive. The fiscal second-quarter print in January triggered a 10 percent single-day decline despite a beat, the worst single-session reaction in the stock since March 2020, because the company offered no upward revision to capex guidance and the buy-side concluded that supply constraints would persist longer than modelled. The fiscal third-quarter response reframes the question. Capex is now expected to grow faster, supply remains constrained at least through 2026, and yet management expressed firmer conviction that Azure growth will accelerate in the second half of calendar 2026. The risk for investors is that the lag between capex outlay and revenue recognition has not shortened, even as the absolute capital commitment has stepped higher.

What does the restructured Microsoft and OpenAI agreement actually change for shareholders and competitors?

Microsoft announced on April 27 a material revision to its OpenAI partnership, ending the revenue-share payments that had defined the commercial relationship since 2019 and removing Azure’s exclusivity as the sole cloud provider for OpenAI workloads. Chief Executive Satya Nadella confirmed on the earnings call that Microsoft retains a royalty-free licence to OpenAI intellectual property through 2032, that OpenAI remains a large customer of Microsoft compute services beyond AI accelerators, and that the equity stake remains intact. Hood added that the revenue share OpenAI pays Microsoft now extends through 2030, providing a longer visibility window than the prior structure.

For competitors, the revision is consequential. Amazon Web Services and Google Cloud Platform can now pursue OpenAI workloads directly, which adds incremental cloud demand for the broader hyperscaler market but dilutes what had been a structurally exclusive Microsoft advantage. The strategic offset for Microsoft is that exclusivity had become a constraint rather than a moat. Microsoft customers have been demanding model diversity, and Azure Foundry already hosts models from Anthropic, open-source providers, and Microsoft’s own MAI family. Over 10,000 customers have used more than one model on Foundry, and the share of customers using both Anthropic and OpenAI models doubled quarter on quarter.

The shareholder calculus is more nuanced. Eliminating the revenue share to OpenAI is margin accretive at the gross-profit level. Retaining royalty-free IP rights through 2032 protects Microsoft’s ability to integrate frontier model capability into Copilot, GitHub, and security products without recurring licence cost. The dilution risk is that as OpenAI distribution broadens, the Azure consumption uplift specifically attributable to OpenAI workloads becomes harder to defend in competitive bids. Microsoft’s remaining performance obligation of $627 billion, up 99 percent year over year and weighted to a duration of approximately two and a half years, suggests the contracted demand base is sufficient to absorb that dilution, but the margin of safety has narrowed.

Is Microsoft 365 Copilot adoption translating into the revenue per user expansion the market has been waiting for?

This was arguably the most decisive datapoint in the quarter for the longer-term valuation case. Microsoft 365 Copilot paid seats crossed 20 million, with seat additions growing 250 percent year over year, the fastest pace since launch. Customers with deployments exceeding 50,000 seats quadrupled year over year, Accenture has reached 740,000 seats in what Microsoft described as its largest Copilot win to date, and Bayer, Johnson and Johnson, Mercedes, and Roche each committed to deployments of 90,000 seats or more.

The economic significance is in the average revenue per user trajectory. Microsoft 365 commercial cloud grew 19 percent on a reported basis and 15 percent in constant currency, with Hood attributing the acceleration to both E5 attach rates and Copilot uplift. The product itself is also evolving in ways that support pricing power. Agent Mode is now the default experience across Word, Excel, and PowerPoint as of the past week, Cowork enables delegation of long-running tasks rather than purely synchronous assistance, and weekly engagement intensity has reached parity with Outlook usage levels, a meaningful behavioural threshold for a productivity product less than three years old.

The risk for executives modelling Copilot is that the business model is migrating from pure per-seat pricing toward a hybrid of seats plus consumption, mirroring the transition GitHub Copilot announced this week with a usage-based pricing model effective June 1. Hood was explicit that bookings will look different in this model because consumption-based usage flows through revenue without first appearing in deferred revenue. The implication is that traditional bookings growth will become a less reliable leading indicator of cloud and AI revenue, which complicates how the buy-side calibrates forward expectations and may compress the multiple investors are willing to assign to remaining performance obligations.

How is the consumer business performing relative to the commercial cloud and AI growth engine?

More Personal Computing posted revenue of $13.2 billion, down 1 percent on a reported basis and down 3 percent in constant currency, the segment dragging on consolidated growth. Windows OEM and Devices declined 2 percent, Xbox content and services fell 5 percent against a strong prior-year comparable that benefited from first-party content launches, and search advertising revenue excluding traffic acquisition costs grew 12 percent on volume and revenue per search gains across Bing and Edge. Bing crossed one billion monthly active users for the first time, Edge gained share for the twentieth consecutive quarter, and Microsoft 365 consumer reached nearly 95 million subscribers.

The fourth-quarter guidance for More Personal Computing implies revenue of $11.75 billion to $12.25 billion, with Windows OEM expected to decline in the high teens against a prior-year comparable that benefited from Windows 10 end of support, elevated channel inventory, and PC market softness driven by rising memory costs. Hood was explicit that the range of potential outcomes is wider than normal, which executives should read as a signal that the consumer hardware ecosystem is absorbing the same memory pricing inflation that is feeding into Microsoft’s own cloud capex.

For competitors and downstream PC original equipment manufacturers, the read-through is that memory inflation is now a multi-quarter headwind rather than a transient supply-chain anomaly, and pricing actions across the PC ecosystem will continue through the second half of calendar 2026. For Xbox, the recent Game Pass pricing changes reflect a deliberate pivot away from subscriber growth at any cost toward unit economics, a discipline that mirrors the broader Microsoft posture on consumer business profitability.

What does the $627 billion in commercial remaining performance obligations tell executives about future revenue visibility?

Commercial remaining performance obligation reached $627 billion, up 99 percent year over year, with a weighted average duration of approximately two and a half years. Roughly 25 percent will be recognised as revenue in the next twelve months, up 39 percent year over year, and the portion recognised beyond twelve months grew 138 percent. Excluding OpenAI commitments, commercial bookings grew 7 percent against what management characterised as a significant prior-year comparable.

The interpretive question for executives is whether the remaining performance obligation acceleration represents genuine multi-year demand commitment or whether it reflects accounting treatment of a small number of very large contracts, including the OpenAI commitment itself. Hood pushed back on concentration concerns by noting that the non-OpenAI portion alone exceeds $250 billion, which suggests the underlying enterprise contract base is broader than headline concentration metrics imply. Even with that caveat, the duration and scale of the obligations create a credible visibility runway through fiscal 2028 that few large-cap technology peers can match.

The competitive read-through is that customers are willing to commit to multi-year Azure consumption at scale, which is the strongest possible market validation of the AI infrastructure thesis. The risk is that as the seats-plus-consumption model proliferates, contracted commitments may shift toward shorter-duration consumption credits rather than multi-year licence obligations, which would gradually reduce the visibility advantage Microsoft has historically enjoyed.

How should institutional investors interpret the Microsoft stock reaction in the context of the broader AI capex debate?

Microsoft entered the print with shares around $424, down approximately 12 percent year to date and roughly 23 percent below the all-time high of $555 reached in late October 2025. The fiscal second-quarter reaction in January, a 10 percent single-day decline, established the playbook for how the buy-side reacts to capex acceleration without commensurate revenue visibility. The fiscal third-quarter print contains both the trigger for that pattern and the partial counter-argument.

The trigger is the calendar 2026 capex guide of $190 billion, which is well above consensus estimates of approximately $135 billion to $160 billion that prevailed entering the print. The counter-argument is the combination of Azure acceleration, Copilot seat momentum, and the OpenAI revenue-share elimination, all of which improve the unit economics of incremental capex deployment. Microsoft’s forward price-to-earnings multiple has compressed to roughly 22 to 25 times trailing twelve-month earnings, the lowest range in three years, against an S&P 500 forward multiple of approximately 21 times. For executives benchmarking the valuation against peers, the multiple reflects skepticism that may be partially priced in, but the next leg of re-rating depends on Azure growth visibly accelerating into the first half of fiscal 2027 as guided.

Key takeaways on what this development means for Microsoft, its competitors, and the AI infrastructure industry

• Microsoft delivered a clean operational beat on revenue, operating income, and earnings per share, but Q4 revenue guidance came in below consensus and calendar 2026 capex of $190 billion is materially above buy-side expectations, setting up a fresh round of debate on the lag between AI infrastructure spend and revenue conversion.
• Azure growth of 40 percent in constant currency, ahead of the 37 to 38 percent guide, signals that the supply constraint has eased modestly, and the Fairwater datacenter coming online six weeks ahead of schedule provides a tangible execution datapoint that capacity additions are accelerating.
• The OpenAI partnership restructuring eliminates the revenue-share payment to OpenAI, retains royalty-free intellectual property rights through 2032, and removes Azure exclusivity, creating margin accretion for Microsoft while opening competitive opportunity for AWS and Google Cloud Platform.
• Microsoft 365 Copilot crossed 20 million paid seats with year-over-year seat additions growing 250 percent, and the quadrupling of customers with deployments exceeding 50,000 seats validates that enterprise AI productivity adoption is reaching scale beyond pilot deployments.
• The migration from per-seat to seats-plus-consumption pricing, mirrored in the GitHub Copilot business model change effective June 1, will reshape how investors interpret bookings and remaining performance obligations as leading indicators of cloud and AI revenue.
• Commercial remaining performance obligation of $627 billion, up 99 percent year over year with 25 percent recognising in the next twelve months, provides the strongest revenue visibility runway among large-cap hyperscaler peers and partially offsets capex concerns.
• Capex composition matters as much as the absolute number, with two-thirds of third-quarter spend allocated to short-lived GPU and CPU assets that correlate directly with near-term revenue rather than long-dated real estate, a discipline distinguishing Microsoft from purely speculative peer buildouts.
• Memory pricing inflation is now a structural cost driver, contributing $25 billion of the calendar 2026 capex increase and dragging the consumer Windows OEM business through higher PC pricing, with multi-quarter implications across the entire technology hardware supply chain.
• Forward price-to-earnings of approximately 22 to 25 times places Microsoft at the low end of its three-year valuation range, suggesting that capex risk is at least partially priced in, with the next re-rating contingent on visible Azure acceleration in the first half of fiscal 2027.
• The voluntary retirement program adding roughly $900 million in one-time costs to fourth-quarter operations reflects a deliberate posture on headcount discipline that is expected to drive year-over-year headcount declines and operating expense growth in the mid-to-high single digits in fiscal 2027.