Azure Accelerates to 39% on a Bigger Base — and the FY26 Setup Just Got Cleaner
Key Takeaways
- Revenue of $76.4B (+18% reported, +17% cc) and EPS of $3.65 (+24% reported, +22% cc) cleared consensus by 3.6% on the top line and 8.3% on the bottom — but the headline number is Azure at +39% (constant currency, with no separate reported figure provided), accelerating off Q3's +35% on a much larger base and beating the StreetAccount/CNBC consensus of ~34-35% by ~400-450bps.
- Commercial bookings cleared $100B for the first time on a full-year basis (+30% cc), and commercial RPO grew to $368B (+35% cc) with 49% growth in the beyond-12-month tranche. Annuity mix held at 98%. The contracted backlog is now ~5x the trailing-12-month Microsoft Cloud revenue, which is the clearest demand anchor any hyperscaler has ever printed.
- FY26 guided to double-digit revenue and operating income growth with operating margins "relatively unchanged" YoY, capex growth moderating versus FY25, and a higher mix of short-lived assets — exactly the discipline framework the bear camp had been demanding three months ago. The Q1 capex guide of $30B+ is high but anchored to capacity-delivery timing on finance leases, not a new spending cycle.
- Hood now expects to remain "capacity constrained through the first half of FY26" — pushing the AI supply/demand crossover from her January "by Q4" framing to her July "hope by December" framing. That is the second consecutive call where the demand-runway extension has been the structurally bullish disclosure.
- Rating: Maintaining Outperform. This is the cleanest hyperscaler print of the cycle. Azure reaccelerated on a larger base, the FY26 capex framing is more disciplined than Q3's, the bookings/RPO acceleration confirms demand is getting harder to fulfill, and operating margins are guided flat-to-slightly-better despite the largest single-year capex absolute step-up in company history. Stock at all-time highs is the only friction; the fundamental trajectory does not warrant a downgrade.
Results vs. Consensus
| Metric | Actual | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $76.44B | $73.81B | Beat | +3.6% |
| EPS (Diluted GAAP) | $3.65 | $3.37 | Beat | +8.3% |
| Operating Income | $34.3B | ~$32.0B | Beat | +7% |
| Operating Margin | 45.0% | ~43.5% | Beat | +150bps |
| Microsoft Cloud Revenue | $46.7B | ~$45.5B | Beat | +2.6% |
| Azure Growth (cc) | +39% | +34.4-35.3% | Beat | +400-450bps |
| Commercial RPO | $368B | — | Step-change | +35% cc YoY |
| Capex (incl. fin leases) | $24.2B | ~$22-23B | Above guide | +~5-8% |
| Free Cash Flow | $25.6B | — | — | — |
Quality of the Beat
- Revenue: Broad-based across all three segments. Reported growth of 18% versus +17% cc means FX was a slight tailwind in the quarter. The $2.6B revenue beat versus consensus is concentrated in Intelligent Cloud (Azure) and Productivity & Business Processes (M365 commercial) — two of the three highest-quality lines in the P&L.
- Azure decomposition (qualitative): Hood explicitly said Azure was driven by "accelerated growth in our core infrastructure business, primarily from our largest customers" — i.e., enterprise migrations. Azure AI services were "generally in line with expectations" with capacity remaining short of demand. This is the same pattern as Q3: AI services constrained, non-AI services accelerating — and the beat is on non-AI.
- Margins: Operating margin of 45.0% (+2pts YoY) was driven by revenue out-running opex (+5% cc) plus continued discipline on layers/managers. Microsoft Cloud GM of 68% (-2pts YoY) on AI infra scaling was modestly better than expected. Headcount essentially flat YoY despite the largest year of capex spending in company history — this is the operating leverage story.
- EPS: $3.65 GAAP (+24% reported, +22% cc) versus $3.37 consensus on a 17% effective tax rate. The tax rate was lower than the guided 19%, which contributed roughly $0.10 of upside to the EPS print. Operationally, the beat is closer to ~$0.18, which is still wide.
- Cash flow: Operating cash flow of $42.6B funded $17.1B of cash capex with $25.6B left over for FCF. The cash-flow math comfortably absorbs the H1 capex acceleration that is coming.
Segment Performance
| Segment | Revenue | YoY (cc) | Op. Income (cc) | vs. Estimate | Notable |
|---|---|---|---|---|---|
| Productivity & Business Processes | $33.1B | +14% | +19% | Beat | M365 Commercial Cloud +16% cc; Dynamics 365 +21% cc; Copilot record seat-add quarter |
| Intelligent Cloud | $29.9B | +25% | +23% | Beat | Azure +39% cc; on-prem server -3% cc (better than expected) |
| More Personal Computing | $13.5B | +9% | +33% | Beat | Search ex-TAC +20% cc (with 8pt third-party tailwind); Xbox content +12% cc |
| Total | $76.4B | +17% | +22% | Beat all three | RPO $368B (+35% cc); commercial bookings +30% cc on $100B+ annual base |
Productivity & Business Processes ($33.1B, +16% / +14% cc)
P&BP delivered a clean beat with M365 Commercial Cloud +16% cc (including 2pts of in-period revenue recognition benefit; underlying business "relatively stable" to Q3). Paid M365 commercial seats grew 6% YoY — slightly slower than Q3's +7% on a larger base — with installed-base expansion across all customer segments led by SMB and frontline workers. The decisive disclosure was on Copilot: "the largest quarter of seat ads since launch" with "a record number of customers returning to buy more seats." Specific deal references — Barclays expanding from 15K seats to 100K, UBS expanding from 55K to all employees, Adobe/KPMG/Pfizer/Wells Fargo each at 25K+ seats this quarter — tell you the enterprise inflection on Copilot is genuine.
"Customers continue to adopt Copilot at a faster rate than any other new Microsoft 365 suite with strong usage intensity as shown by our week-over-week retention. And we saw the largest quarter of seat ads since launch with a record number of customers returning to buy more seats." — Satya Nadella, CEO
Dynamics 365 +21% cc was a clean acceleration from Q3's +18% cc. LinkedIn +8% cc held steady, with Talent Solutions still soft on hiring-market weakness. M365 Consumer Cloud +20% reported on the January price increase plus +8% subscriber growth. P&BP operating income +19% cc with gross margin slightly higher YoY despite AI infra scaling — Hood credited "efficiency gains in Azure and Microsoft 365 commercial cloud" partially offsetting the AI drag.
Assessment: P&BP just delivered the cleanest Copilot data point of the cycle. Record seat-add quarter plus repeat-purchase customers is the missing piece in the Copilot-monetization story. The Q1 FY26 guide of M365 Commercial Cloud +13-14% cc reflects a deliberate moderation off this print's ~14% underlying rate plus seat growth that has been hard to forecast.
Intelligent Cloud ($29.9B, +26% / +25% cc)
This is the segment that defines the print. Azure and other cloud services +39% (cc reported as the primary metric — no separate non-cc figure given) versus the StreetAccount/CNBC consensus of +34-35%, on a base that was already 25% larger than the Q3 base. Hood's framing was deliberate: the upside came from "core infrastructure business, primarily from our largest customers," with Azure AI services "generally in line with expectations" but still capacity-constrained. This is now the third consecutive quarter where the non-AI side of Azure has surprised to the upside.
"In Azure and other cloud services, revenue grew 39%. Significantly ahead of expectations driven by accelerated growth in our core infrastructure business, primarily from our largest customers. As a reminder, new cloud and AI workloads are built and scaled using the breadth of our services. Revenue from Azure AI services was generally in line with expectations. And while we brought additional data center capacity online this quarter, demand remains higher than supply." — Amy Hood, CFO
Nadella's three-bucket migration framing — classic SQL/Windows/SAP/VMware migrations (Nestle as the marquee example: 200+ SAP instances, 10,000+ servers, 1.2PB of data with near-zero disruption); cloud-native scaling on Azure (e-commerce companies that came for AI but stayed for compute); and AI workloads — is the cleanest articulation of the Azure growth flywheel he has given. The "we're in the middle innings of migration at best" comment is the forward signal.
On-prem server revenue decreased 3% cc (better than expected) on transactional purchasing benefits. Enterprise & Partner Services +6% cc. Segment operating income +23% cc with gross margin down 4pts YoY on AI infra scaling — same pattern as Q3.
Assessment: Azure is now compounding at a rate that, if it holds, takes the line from $75B FY25 to ~$100-105B in FY26. The Q1 FY26 Azure guide of ~37% cc is essentially a flat-to-slight-deceleration off Q4's +39% on a much larger base — a confidence print. The "capacity-constrained through 1H FY26" disclosure is the structural anchor.
More Personal Computing ($13.5B, +9%)
MPC was the upside surprise. Windows OEM and Devices +3% on continued elevated channel inventory (the same dynamic Hood flagged in Q3). Search and news advertising ex-TAC +20% cc with ~8 points of favorable impact from third-party partnerships (the OpenAI/Bing arrangement, plus a low prior-year comparable). Xbox content and services +12% cc on first-party content and Game Pass — Game Pass annual revenue cleared $5B for the first time. Operating margin expanded 3pts YoY across all businesses; segment operating income +33% cc.
The Q1 FY26 guide is for MPC revenue of $12.4-12.9B — a mid-to-high single-digit YoY decline on Windows OEM channel correction and the lapping of the third-party search tailwind. The segment will look weaker on the surface in Q1, but the margin trajectory should hold.
Assessment: MPC continues to outpunch its weight on margin discipline. The +33% cc operating-income growth on +9% revenue is the kind of leverage that quietly compounds.
Key KPIs
| KPI | FY25 Q4 | YoY | Notable |
|---|---|---|---|
| Microsoft Cloud Revenue (Q) | $46.7B | +27% / +25% cc | FY25 annual: $168B+ (+23%) |
| Microsoft Cloud GM% | 68% | -2pts | AI infra scaling drag, partially offset by efficiency gains |
| Azure Annual Revenue (FY25) | $75B+ | +34% | First disclosure of explicit Azure annual figure |
| Commercial RPO | $368B | +35% cc | ~35% in next-12-months, +21% YoY; 49% growth beyond-12-months |
| Commercial Bookings (FY25) | $100B+ | +30% cc | First time clearing $100B annual |
| Annuity Mix | 98% | — | Stable |
| M365 Commercial Paid Seats | — | +6% | Slightly slower than Q3's +7% on a larger base |
| M365 Consumer Subs | — | +8% | January price increase still flowing through |
| Copilot Family MAUs | 100M+ | — | Across commercial + consumer |
| AI-feature MAUs (broad) | 800M+ | — | Across all Microsoft products |
| Foundry Customers | 14,000+ | — | 80% of Fortune 500 already on Foundry |
| Foundry Tokens (FY25 annual) | 500T+ | +7x | Platform diffusion signal |
| GitHub Copilot Users | 20M | +33% | From 15M in Q3; +75% QoQ Enterprise growth; 90% of Fortune 100 |
| Custom agents created (FY25) | 3M | — | Via SharePoint + Copilot Studio |
| Dragon Copilot encounters (Q) | 13M | +~7x | vs. 9.5M in Q3, broadening rapidly |
| Capex (incl. fin leases) | $24.2B | — | $5B from finance leases recognized at commencement |
| FY25 Total Capex | ~$88B | — | Implied from $17.1B Q4 cash-paid and FY framing |
| Headcount | — | ~Flat YoY | "Relatively unchanged" — operating leverage on cost |
The single most important number remains the $368B RPO. Up from $315B at Q3 — that is a $53B sequential increase. The 49% growth in the beyond-12-months tranche is the leading indicator that customers are committing to multi-year Microsoft Cloud spend at an accelerating pace. The crossover question — does capex growth ever catch revenue growth — is essentially answered by Hood: capex is anchored to delivering against the contracted backlog.
Key Topics & Management Commentary
Overall Management Tone: The most confident management has sounded in the entire AI cycle. Nadella opened with "I have never been more confident in my opportunity to drive long-term growth." Hood's tone shifted from Q3's "we'll be tight exiting June" hedging to Q4's "I hope I'm in better shape by December" framing — explicit acknowledgment that demand is outpacing the build. The capex discussion was unusually defensive in posture, which is a tell that management is willing to accept market pushback to keep funding the demand curve.
The Capacity Constraint Extends Into 1H FY26
Hood's most consequential disclosure was the explicit pushback of the supply/demand crossover: in January she expected balance by June; at Q3 she expected to be "tight" exiting June; at Q4 she now expects to remain capacity-constrained "through the first half of FY26." That is the third consecutive deferral of the crossover date in seven months, and it directly contradicts the AI-overbuild bear narrative that has periodically resurfaced through 2025.
"I talked about my gosh, in January and said, I thought we'd be in better supply demand shape by June. And now I'm saying, hope I'm in better shape by December. And that's not because we slowed CapEx. Even with accelerating the spend and trying to pull leases in and get CPUs and GPUs in the system as quickly as we can we are still seeing demand improve." — Amy Hood, CFO
"I'm focused on building backlog, building business, and delivering capacity. Which we are seeing has a good ROI today. In terms of our ability to get that done. So I don't want people to get overly focused on a pivot point because when you're in sort of these expansive moments, picking a data point usually means you're gonna pick to be too conservative." — Amy Hood, CFO
Assessment: This is the most important disclosure on the call. Demand is still outpacing supply at the largest hyperscaler, with $368B of contracted backlog backstopping the spend. The "I'm not as focused on a pivot point" framing is Hood signaling that capex will keep flexing to demand for at least another 12 months. The cost of being wrong about this — losing share in a once-in-a-decade platform shift — vastly exceeds the cost of overspending and depreciating an extra $5-10B over 15 years.
The FY26 Capex Framing — More Disciplined Than Q3
The forward capex framing is the cleanest piece of Hood's outlook. FY26 capex growth will moderate versus FY25; the mix will shift further toward short-lived assets (servers, GPUs, CPUs, networking, storage) that are "really correlated to the backlog we see and the curve of demand"; H1 will be heavier than H2 due to finance-lease delivery timing. The Q1 capex guide of $30B+ is the highest absolute number Microsoft has ever guided to for a single quarter, and it is being framed as a capacity-delivery timing dynamic rather than a step-up in run-rate spending.
"Capital expenditure growth, as we shared last quarter, will moderate compared to FY '25 with a greater mix of short-lived assets. Due to the timing of delivery of additional capacity in H1, including large finance lease sites, we expect growth rates in H1 will be higher than in H2." — Amy Hood, CFO
The strategic logic is identical to what Hood said at Q3: short-lived assets are revenue-tied; long-lived assets (data center shells, networking, power) have 15+ year monetization arcs. The mix shift toward short-lived means the capex-to-revenue elasticity gets tighter. The Q1 $30B+ figure is high in absolute terms but $5B of Q4 capex was finance-lease commencements — which means a meaningful portion of any quarter's headline capex is timing rather than incremental spend.
Assessment: The FY26 capex framing is more disciplined than the spring narrative implied. "Operating margins relatively unchanged YoY" is the cleanest single-line summary — Microsoft is committing to absorbing the AI infra ramp without margin degradation, which is the fundamental investor-protection clause in this story.
The Compounding S-Curves and the Cost-per-Token Story
Nadella spent material airtime on the software-efficiency thesis, with the most concrete data point being that on the GPT-4o family of models (highest inference token volume on Azure), software optimization alone is delivering "90% more tokens for the same GPU compared to a year ago." Combined with Q3's "cost-per-token has more than halved" commentary, the message is unambiguous: Microsoft believes software efficiency compounds faster than AI demand scales, which is what will drive AI-side gross margin expansion over the next 18-24 months.
"Take, for example, the GPT-4o family of models, which have the highest volume of inference tokens. Through software optimization alone, we are delivering 90% more tokens for the same GPU compared to a year ago." — Satya Nadella, CEO
"The difference between a hoster and a hyperscaler is software. And the same is going to be true here." — Satya Nadella, CEO
Assessment: The 90% tokens-per-GPU efficiency gain is the single most important technical disclosure on the call. It directly addresses the AI-margin durability question and supports the "operating margins unchanged YoY" guide. If software efficiency continues to compound at this rate, the AI-side margin trajectory turns from a drag into a tailwind starting FY27.
OpenAI and the Foundry Multi-Model Strategy
The Foundry framing on this call is notable for what is no longer there: a single-OpenAI-anchored AI story. Nadella highlighted Foundry's multi-model breadth (OpenAI, DeepSeek, Meta, xAI's Grok, Black Forest Labs, Mistral coming soon), with the key disclosure being that Microsoft "sim-shipped 15 models from OpenAI alone on Foundry this year." 14,000 customers are using the Foundry agent service; 80% of the Fortune 500 are on Foundry; 500T tokens served via Foundry API in FY25 (+7x YoY). OpenAI's Cosmos DB and PostgreSQL footprints on Azure were also called out — "OpenAI uses Cosmos DB in the hot path of every ChatGPT" — confirming the deep multi-product integration.
Assessment: The Foundry disclosure shows Microsoft converting OpenAI from an exclusive partnership into a portfolio anchor — same-day model access plus a multi-model platform underneath. The 80% Fortune 500 penetration is the platform-diffusion signal. OpenAI concentration is still material but is now structurally less dominant in the Foundry story.
Copilot Reaches Enterprise Inflection
Q4 was the cleanest Copilot quarter to date. The "largest quarter of seat ads since launch" framing plus the specific deal references (Barclays 100K, UBS all-employee, Adobe/KPMG/Pfizer/Wells Fargo each 25K+) translate Q3's "hundreds of thousands of customers" abstraction into concrete enterprise deployments. M365 Copilot Family of Copilot apps surpassed 100M MAUs (commercial + consumer combined) — first time disclosed. The 800M MAU figure for "AI features across products" is broader scope but signals the surface-area diffusion.
"Customers created 3 million agents using SharePoint and Copilot Studio." — Satya Nadella, CEO
The agent-creation metric is the most strategically important disclosure on the consumer/SMB monetization arc. 3 million agents created in FY25 (versus the 1 million Q3 number) implies the agent-creation flywheel is compounding QoQ. GitHub Copilot at 20 million users (up from 15M at Q3) with 90% Fortune 100 penetration is the developer-side equivalent.
Assessment: The Copilot story has crossed from "early adopter" to "enterprise standard." The 13M Dragon Copilot physician-patient encounters (up ~7x YoY versus Q3's 9.5M up 50% QoQ) is the leading indicator that vertical Copilots are also crossing the chasm.
Guidance & Outlook
| Metric | FY26 Q1 Guide | Implied Growth | Notes |
|---|---|---|---|
| Productivity & Business Processes | $32.2-32.5B | +14-15% | ~3pts FX tailwind; M365 Commercial Cloud +13-14% cc |
| Intelligent Cloud | $30.1-30.4B | +25-26% | Azure ~+37% cc; capacity-constrained through 1H FY26 |
| More Personal Computing | $12.4-12.9B | -mid/high SD | Windows OEM channel correction; lapping search tailwind |
| Total Implied Revenue | ~$74.7-75.8B | +12-13% | Above LSEG ~$73.6B consensus |
| COGS | $24.3-24.5B | +21-22% | AI infra scaling continues |
| Operating Expense | $15.7-15.8B | +5-6% | Disciplined |
| Microsoft Cloud GM% | ~67% | -2pts QoQ | AI infra scaling drag |
| Capex | $30B+ | — | Finance-lease delivery timing |
| FY26 Op Margins | "Relatively unchanged YoY" | — | Despite AI ramp |
| FY26 Effective Tax Rate | 19-20% | — | Up from 17% in Q4 |
| FY26 Revenue/OI Growth | "Double-digit" | — | Both lines |
The Q1 FY26 Azure cc guide of ~37% is the most aggressive forward Azure guide of the cycle. It implies essentially flat sequential constant-currency growth off the Q4 +39% print, on a base that is now ~25% larger than the Q3 base. The +37% guide is materially above the ~33% Street had been pricing pre-print and signals that the capacity-delivery timing of finance-lease sites in H1 will translate to revenue almost immediately.
Implied FY26 trajectory: "Double-digit revenue and OI growth" with "operating margins relatively unchanged" implies revenue of ~$310-315B (versus FY25's $281.7B) and operating income of ~$140-145B (versus $128B). The $30B+ Q1 capex is high but is anchored to capacity-delivery timing rather than a permanent step-up.
Street at: Pre-print LSEG consensus was tracking to ~$73.6B for Q1 FY26 and Azure ~33% cc; the guide is materially above on both lines.
Guidance style: The FY26 framing is unusually direct for Microsoft — "double-digit revenue AND operating income growth" is a structurally bullish statement that locks Hood into specific outcomes 12 months out. The "operating margins relatively unchanged" anchor is the most important investor-protection commitment.
Analyst Q&A Highlights
Quality and Durability of the Quarter
- Keith Weiss, Morgan Stanley: Opened with "I don't think I've ever seen a quarter where, like, everything came together this well." Asked about head app vs. tier-2 application diffusion as a platform-maturity indicator. Nadella framed it as "head apps shaping the platform" with "broad diffusion" following — the key signal of platform durability.
Assessment: The framing of "tier-2 applications" being tracked is significant — it says Microsoft is now monitoring breadth of platform usage, not just headline customer wins.
SaaS AI Monetization & Long-Term Margin Trajectory
- Mark Moerdler, Bernstein: Asked how software companies will monetize AI for SaaS, and how horizontal vs. agentic monetization differs, and what the long-term SaaS AI margin trajectory looks like. Nadella drew the analogy to server-to-cloud: cloud expanded the addressable base by orders of magnitude due to flexibility/burst/spin-up. AI will follow the same arc — "intelligence is basically log of compute" so compute usage compounds. Hood added that pricing models will blend per-user, tiered, and consumption.
Assessment: The "log of compute" framing is Nadella's clearest articulation of the AI demand thesis. It is the rebuttal to anyone modeling AI demand as bounded by current GPU capacity.
Migration Durability
- Karl Keirstead, UBS: Asked about the durability of the migration-driven Azure acceleration. Nadella cited three drivers — classic migrations (Nestle: 200+ SAP instances, 10K+ servers, 1.2PB), cloud-native scaling, and AI workloads — and said Microsoft is "in the middle innings of migration at best."
Assessment: "Middle innings" is the strongest forward signal Nadella has given on migration runway. The Nestle disclosure is also a competitive signal — that scale of migration would have historically gone to AWS.
What Surprised in the Quarter
- Brent Thill, Jefferies: Asked Nadella what surprised him in the quarter. Nadella said nothing in particular surprised him but emphasized the maturation of the AI app stack — "very stateful app patterns that are emerging that require quite a bit of rethinking of even the app stack."
Assessment: Nadella's framing of "rethinking the app stack" is the strategic positioning for the next 5 years — Microsoft is selling not just compute but the app-architecture pattern.
Copilot Beyond Chat
- Raimo Lenschow, Barclays: Asked about Copilot evolving beyond chat into data-and-agents architectures. Nadella confirmed: starts with chat UI but quickly evolves to spawn applications, monitor asynchronous work, and integrate with non-Copilot code generation tools (citing GitHub Copilot's code-review agent being used even when developers used Claude Code or Cursor to write the code).
Assessment: The disclosure that GitHub's code-review agent is used even with non-Microsoft AI coding tools is strategically important — it says Microsoft's developer-tools moat is the platform, not the model.
Capex Pivot Point & ROI
- Kash Rangan, Goldman Sachs: Asked when capex growth and Azure growth crossover. Hood explicitly declined to pick a date — "I'm not as focused on trying to pick a date at which revenue growth and CapEx growth will meet and cross. I'm focused on building backlog, building business, and delivering capacity." Both Hood and Nadella tied capex to the $368B contracted backlog.
Assessment: Hood's refusal to pick a pivot point is the cleanest signal yet that capex is demand-anchored, not budget-anchored. The implicit message is that capex will keep flexing to demand for as long as the backlog grows faster than capacity comes online.
Margin Trajectory & Internal AI Productivity
- Michael Turrin, Wells Fargo: Asked how Microsoft holds operating margin flat in FY26 while absorbing the Azure mix shift. Hood emphasized three drivers: (1) revenue out-running opex (the durable margin driver), (2) S-curve compounding efficiencies (software optimization, hardware optimization), (3) talent concentration on highest-opportunity areas.
Assessment: Hood's framing — "starts and ends with product" — is the cleanest single line on the margin algorithm. It says Microsoft will not cost-cut its way to margin protection; it will revenue-grow its way there.
What They're NOT Saying
- FY27+ capex framing: Hood reaffirmed FY26 capex growth will moderate but declined to discuss FY27. Given that FY26 capex still growing on top of an ~$88B FY25 base implies ~$95-100B+ for FY26, the FY27 number is the next debate.
- OpenAI partnership economics: Foundry was framed as a multi-model platform with OpenAI as one of many anchors. The economics of the OpenAI relationship — revenue share, capacity reservation, equity stake mechanics — were not discussed. This is appropriate while the relationship is being repositioned but means the concentration risk is still opaque.
- Azure AI vs. non-AI explicit decomposition: Q3's "16 points from AI services" disclosure was not repeated in the Q4 prepared remarks. Hood instead said AI services were "generally in line" while non-AI drove the upside. This continues the Q3 trend of moving away from explicit AI/non-AI segmentation, which reduces visibility into a key growth driver.
- Copilot revenue: Massive seat-add disclosure but no direct revenue figure. The ARPU contribution to M365 Commercial Cloud +16% cc is opaque. Management is clearly pricing Copilot but not yet ready to disclose monetization mechanics — appropriate but means modeling requires assumptions.
- GPU useful-life extension: Hood was asked about software efficiency and tokens-per-GPU but did not touch the depreciable-life question. The 90% efficiency gain on GPT-4o tokens is a candidate for a future useful-life extension, which would be a multi-billion-dollar EPS tailwind. Management is keeping it as discretionary upside.
- Specific FY26 segment growth rates: Hood guided "double-digit revenue and operating income growth" at the company level but did not break it down by segment. The segment-level FY26 growth assumptions are inferable from the Q1 FY26 guide but not confirmed.
- Tariff-driven Windows OEM normalization quantification: Hood acknowledged the inventory headwind is bigger than usual but did not quantify it. The Q1 MPC guide ($12.4-12.9B) is wider than normal as a hedge.
Market Reaction
- After-hours move (July 30, 2025): +9% in extended trading, briefly clearing $550 — first time above that level.
- Next-day close (July 31, 2025): All-time high of $555.45.
- Day-of-publication move (StockTitan Argus): +3.95% regular session, peak intraday +4.2%, ~$157B added to market cap (~$4.13T total).
- Volume: ~1.6x daily average — elevated.
- Drivers cited: Azure +39% (~400-450bps ahead), $368B RPO, $100B+ FY25 bookings, Q1 Azure guide ~37% cc, FY26 framing of double-digit revenue + OI growth with margins flat.
The +9% AH move on top of a stock already at all-time highs is unusual. The market is digesting (1) the FY26 guide being meaningfully above Street expectations, (2) the capacity-constraint extension into 1H FY26, and (3) the bookings/RPO surge confirming demand is getting harder to fulfill. The next-day all-time high reflects index-buyer rebalancing more than incremental marginal-buyer enthusiasm — but the price action is unambiguously confirmatory of the bull thesis.
Street Perspective
Debate: Is Azure Growth Sustainable at ~35%+ Through FY26?
Bull view: The Q1 FY26 guide of +37% cc is forward-looking, not just trailing. RPO grew $53B sequentially. Hood is signaling capacity-constrained through 1H FY26 — meaning the ceiling on Azure growth in 1H is supply, not demand. Migration runway is "middle innings" per Nadella, with marquee enterprise wins (Nestle SAP) confirming Microsoft is taking AWS share at scale.
Bear view: +39% on the Q4 base is a high-water mark by definition. The base effect alone makes 1H FY26 the easiest YoY compares; H2 should naturally decelerate. AI-services contribution was deliberately not quantified — if AI is decelerating relative to non-AI, the headline number is masking a mix-shift problem.
Our take: The bull case has cleaner near-term evidence (the +37% cc Q1 guide on a much larger base, plus the 1H capacity constraint). H2 deceleration to ~32-34% would still be a reaccelerating trajectory versus FY25 — well within "double-digit" company-level guidance. We side with the bull case.
Debate: Is the Capex Trajectory Disciplined or Out of Control?
Bull view: $30B+ Q1 capex is high in absolute terms but is anchored to finance-lease delivery timing for capacity that backs $368B of contracted RPO. FY26 capex growth is moderating versus FY25; mix is shifting toward short-lived (revenue-tied) assets; FY26 operating margins are guided "relatively unchanged" — exactly the discipline framework the bear camp was demanding three months ago. Free cash flow of $25.6B in Q4 means the cash-flow math still works comfortably.
Bear view: Q1 capex is the largest single-quarter capex print of any company in history. Hood explicitly refuses to discuss when capex growth and revenue growth crossover. FY27+ capex is unbounded.
Our take: "Operating margins relatively unchanged YoY" is the bull anchor. As long as Microsoft can guide operating margin flat through the AI ramp, the capex magnitude is a multiple-compression risk, not a fundamental risk. The bear case is structural but does not bite until FY27 is guided.
Debate: Is Copilot Real Monetization or Adoption Theater?
Bull view: Q4 was the largest seat-add quarter since launch with a record number of repeat-purchase customers. Specific enterprise references (Barclays 100K, UBS all-employee, Adobe/KPMG/Pfizer/Wells Fargo 25K+ each) are fully concrete. 3M custom agents created via SharePoint/Copilot Studio in FY25. M365 Commercial Cloud +16% cc with ARPU growth driven by E5 + Copilot.
Bear view: Microsoft still won't break out Copilot revenue. Seat counts and customer counts are adoption metrics, not monetization metrics. The ARPU contribution from Copilot is opaque and likely smaller than the bull case implies.
Our take: The bull case has the cleaner near-term evidence; the bear case is correct that disclosure is incomplete. We trust the M365 Commercial Cloud +16% cc print as the holistic signal that ARPU is moving, even without explicit Copilot decomposition.
Debate: Is the Stock at All-Time Highs Already Pricing Perfection?
Bull view: EPS is growing 22-24% on the latest print, FY26 implied EPS growth is double-digit, and the multiple has not expanded materially against earnings power. The stock is at all-time highs because the fundamentals are at all-time highs, not because the multiple has stretched.
Bear view: Multiple is at the upper end of the 5-year range. Any execution misstep — capex magnitude exceeding margin discipline, Azure deceleration before H2 base effects work through — could trigger a derate.
Our take: Multiple is full but earnings trajectory supports the price. We do not see a re-rating catalyst from here, but we see continued earnings power that supports the valuation.
Model Update Needed
| Item | Prior Assumption | Suggested Update | Reason |
|---|---|---|---|
| FY25 Revenue | ~$278B | $281.7B (actual) | Q4 print |
| FY26 Revenue Growth | +12% | +12-13% | Q1 guide above Street + double-digit FY framing |
| FY26 Op Margin | ~45% | ~45% ("relatively unchanged") | Hood explicit guide |
| FY26 Capex | $95B | $100-105B | Q1 $30B+ guide + finance-lease timing |
| FY26 Azure Growth (cc) | +30% | +33-36% | Q1 guide +37%; 1H capacity constraint |
| FY26 EPS | ~$15.00 | $15.30-15.70 | Q4 print + FY framing; offset by 19-20% effective tax rate |
| FY26 Microsoft Cloud GM% | ~67% | ~67% trough | Q1 guide; software efficiency offsets |
| FY27 Capex Growth | +10% YoY | +5-10% YoY | Implied "moderation continues" |
Valuation impact: Modest upward revision to NTM revenue and EPS, with a more meaningful upward revision to FY26 Azure growth and a tighter capex framing. Net intrinsic-value impact is positive; multiple is full but trajectory supports.
Thesis Scorecard Post-Earnings
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: Azure compounds at 25%+ on broadening workload mix | Strongly Confirmed | +39% cc this Q vs. +35% Q3; Q1 FY26 guide ~+37% cc |
| Bull #2: Copilot ARPU drives M365 Commercial Cloud reacceleration | Confirmed | "Largest seat-add quarter since launch"; Barclays/UBS/Adobe/KPMG/Pfizer/Wells Fargo deals |
| Bull #3: AI-side unit economics hold through capex ramp | Confirmed | FY26 op margins guided unchanged; +90% tokens/GPU on GPT-4o |
| Bear #1: Capex magnitude breaks margin trajectory | Strongly Challenged | FY26 OI growth double-digit; margins flat; capex growth moderating |
| Bear #2: AI demand will overshoot supply | Strongly Challenged | Capacity-constrained through 1H FY26 — third consecutive deferral |
| Bear #3: OpenAI concentration risk | Improving | Foundry multi-model strategy diversifying anchor; 80% Fortune 500 on Foundry |
Overall: Thesis materially strengthened. Two of three bear points moved from "challenged" to "strongly challenged"; the third is improving. This is the cleanest hyperscaler print of the cycle.
Action: Maintaining Outperform. Monitor (1) Q1 FY26 print for Azure deceleration vs. the +37% cc guide, (2) FY27 capex framing on the January call, (3) any explicit Copilot revenue disclosure as the seat base scales.