What Management Said
Read the full Q2 2026 transcript ↗More detailed outlook slides will be available on the Microsoft Investor Relations website when we provide outlook commentary on today's call. All growth comparisons we make on the call today relate to the corresponding period of last year, unless otherwise noted. We will also provide growth rates in constant currency when available, as a framework for assessing how our underlying businesses performed, excluding the effect of foreign currency rate fluctuations. Where growth rates are the same in constant currency, we will refer to the growth rate only.
This quarter, the Microsoft Cloud surpassed $50 billion in revenue for the first time, up 26% year-over-year, reflecting the strength of our platform and accelerating demand. We are building this infrastructure out for the heterogeneous and distributed nature of these workloads, ensuring the right fit with the geographic and segment-specific needs for all customers, including the long tail. The key metric we are optimizing for is tokens per watt per dollar, which comes down to increasing utilization and decreasing TCO using silicon systems and software. A good example of this is the 50% increase in throughput we were able to achieve in one of our highest volume workloads, OpenAI inferencing, powering our copilots.
We are seeing increasing demand for region-specific models, including Mistral and Cohere, as more customers look for sovereign AI choices. That means connecting their agents to systems of record and operational data, analytical data, as well as semi-structured and unstructured productivity and communications data. All up, the number of customers spending $1 million plus per quarter on Foundry grew nearly 80%, driven by strong growth in every industry. As agents proliferate, every customer will need new ways to deploy, manage, and protect them.
What went well
- The Microsoft Cloud surpassed $50 billion in revenue for the first time, reaching $51.5 billion, up 26% year-over-year (24% in constant currency).
- Q2 revenue was $81.3 billion, up 17% and 15% in constant currency, again exceeding expectations across revenue, operating income, and earnings per share.
- Earnings per share was $4.14, an increase of 24% and 21% in constant currency when adjusted for the impact from the OpenAI investment.
- Operating income increased 21% and 19% in constant currency, and operating margins increased year-over-year to 47%, ahead of expectations.
- Commercial bookings increased 230% and 228% in constant currency, driven by the large Azure commitment from OpenAI, the previously announced Anthropic commitment, and healthy core annuity growth.
- Commercial remaining performance obligation increased to $625 billion, up 110% year-over-year, with weighted average duration of approximately 2.5 years; roughly 45% is from OpenAI and the remaining balance still grew 28%.
- Intelligent Cloud revenue was $32.9 billion, up 29% and 28% in constant currency, with Azure and other cloud services up 39% and 38% in constant currency, slightly ahead of expectations.
- Microsoft 365 Copilot had a record quarter for seat adds, up over 160% year-over-year, reaching 15 million paid seats, with the number of customers over 35,000 seats tripling year-over-year and daily active users up 10x year-over-year.
- Microsoft Fabric's annual revenue run rate surpassed $2 billion with over 31,000 customers and revenue up 60% year-over-year, and the number of customers spending $1 million-plus per quarter on Foundry grew nearly 80%.
- Microsoft brought online its Maia 200 accelerator (10-plus petaflops at FP4, over 30% improved TCO) and Cobalt 200 CPU (over 50% higher performance), and added nearly 1 GW of total capacity in the quarter.
- Cash flow from operations was $35.8 billion, up 60%, and Microsoft returned $12.7 billion to shareholders, an increase of 32% year-over-year.
What went wrong
- More Personal Computing revenue was $14.3 billion and declined 3%.
- In gaming, revenue decreased 9% and 10% in constant currency, and Xbox content and services decreased 5% and 6% in constant currency, below expectations, driven by first-party content weakness with impact across the platform.
- The company recorded impairment charges in its gaming business, which increased operating expenses in the More Personal Computing segment.
- Search and news advertising ex-TAC increased only 10% and 9% in constant currency, slightly below expectations due to execution challenges as the third-party partnership benefit normalized.
- Free cash flow was $5.9 billion and decreased sequentially, reflecting higher cash capital expenditures from a lower mix of finance leases.
- Customer demand continued to exceed supply, leaving Azure capacity constrained across workloads, customer segments, and geographies.
- Company gross margin percentage was 68%, down slightly year-over-year, and Microsoft Cloud gross margin was 67%, down year-over-year on continued AI infrastructure investment.
Guidance Changes
| Metric | Period | Current guidance |
|---|---|---|
| Total company revenue | Q3 FY2026 | $80.65B-$81.75B, or 15%-17% growth, with strong commercial growth partially offset by consumer businesses |
| FX impact on total revenue growth | Q3 FY2026 | +3 points to total revenue growth |
| FX impact by segment | Q3 FY2026 | +4 points in Productivity and Business Processes, +2 points in Intelligent Cloud and More Personal Computing |
| FX impact on COGS and operating expense growth | Q3 FY2026 | +2 points |
Performance Breakdown
| Metric | YoY | Note |
|---|---|---|
| Total revenue | +17% (+15% cc) to $81.3B | Growing demand with focused sales execution while investing for long-term growth. |
| Microsoft Cloud revenue | +26% (+24% cc) to $51.5B | Surpassed $50 billion for the first time on platform strength and accelerating demand. |
| Productivity and Business Processes revenue | +16% (+14% cc) to $34.1B | Consistent core execution plus increasing Copilot contribution; operating margin rose to 60%. |
| Intelligent Cloud revenue | +29% (+28% cc) to $32.9B | Driven by Azure and improved on-premises server results. |
| Azure and other cloud services | +39% (+38% cc) | Efficiency gains across the fungible fleet allowed reallocation of capacity to Azure; demand continued to exceed supply. |
| More Personal Computing revenue | -3% to $14.3B | Gaming decline and normalization of the search third-party partnership benefit, partly offset by Windows OEM strength. |
| M365 commercial cloud | +17% (+14% cc) | Consistent core execution and strong Copilot results; ARPU led by E5 and M365 Copilot, seats +6% to over 450 million. |
| M365 commercial products | +13% (+10% cc) | Higher-than-expected Office 2024 transactional purchasing. |
| M365 consumer cloud | +29% (+27% cc) | ARPU growth; subscriptions grew 6%. |
| LinkedIn revenue | +11% (+10% cc) | Driven by marketing solutions. |
| Dynamics 365 revenue | +19% (+17% cc) | Continued growth across all workloads. |
| On-premises server business | +2% (+1% cc) | Demand for hybrid solutions including SQL Server 2025 launch and higher transactional purchasing ahead of memory price increases. |
| Windows OEM and devices | +1% (unchanged cc) | Windows OEM grew 5% on strong execution and continued Windows 10 end-of-support benefit; inventory levels remained elevated. |
| Search and news advertising ex-TAC | +10% (+9% cc) | Below expectations on execution challenges as the third-party partnership benefit normalized. |
| Gaming revenue | -9% (-10% cc) | First-party content weakness; Xbox content and services down 5% and 6% in constant currency. |
| Earnings per share | +24% (+21% cc) to $4.14 (adjusted) | Revenue outperformance and operating margin expansion to 47%. |
Earnings Call Themes & Trends
| Topic | Previous mention | Current period | Trend |
|---|---|---|---|
| Commercial RPO scale | $392B in Q1 FY2026 (weighted duration ~2 years) | $625 billion, up 110% year-over-year, with weighted average duration extended to about 2.5 years; ~45% from OpenAI | — |
| OpenAI backlog concentration | Prior quarter OpenAI share of RPO | Approximately 45% of commercial RPO from OpenAI; the remaining ~55% (~$350B) grew 28% on broad demand | — |
| Custom silicon | Earlier Maia and Cobalt generations | Maia 200 accelerator (over 30% improved TCO) and Cobalt 200 CPU (over 50% higher performance) brought online, alongside NVIDIA and AMD | — |
| Fairwater AI super factory | Single Fairwater sites announced prior quarter | Atlanta and Wisconsin sites connected via an AI WAN into a first-of-its-kind AI super factory | — |
| Microsoft 365 Copilot seats | Prior-quarter seat base | Record seat adds up over 160% year-over-year to 15 million paid seats; daily active users up 10x year-over-year | — |
| Foundry customer scale | Prior-year $1M+/quarter customer base | Customers spending $1M+ per quarter on Foundry grew nearly 80%; over 250 customers on track to process 1 trillion-plus tokens this year | — |
| CapEx short-lived asset mix | ~half short-lived in Q1 FY2026 | Roughly two-thirds of CapEx on short-lived assets (primarily GPUs and CPUs) this quarter | — |
| OpenAI accounting | Equity method based on share of operating losses (Q1: $4.1B loss) | Post-recapitalization, gains/losses now based on share of change in net assets, driving a $10 billion GAAP gain in OI&E | — |
Q&A Summary
Keith Weiss (Morgan Stanley) noted the stock fell despite 24% EPS growth, citing investor concern that CapEx is growing faster than expected while Azure grew a bit slower, and asked how to think about ROI on the investment.
Hood said investors should view Azure guidance as an allocated-capacity guide, since capital (especially short-lived GPU/CPU spend) first serves rising M365 Copilot and GitHub Copilot usage and first-party apps, then long-term R&D and product innovation, with the remainder to Azure; she noted that if all newly online Q1-Q2 GPUs had gone to Azure the growth KPI would have been over 40%. Nadella said to consider the gross-margin profile of the whole portfolio (Azure plus M365, GitHub, Dragon, and Security Copilot), each an incremental TAM, and that Microsoft optimizes the best lifetime-value portfolio while supply-constrained.
Mark Moerdler (Bernstein) asked how investors get comfortable capturing enough revenue over the 6-year server useful life when RPO duration is only 2.5 years, given the AI-centric CapEx.
Hood said average duration blends short-dated M365/BizApps contracts with longer Azure contracts, that the majority of GPUs being bought are already contracted for most of their useful life (including for the largest customers), so much of the perceived risk is not present, and that margins actually improve as assets age. Nadella added that software keeps running the latest models on the aging fleet, which is why Microsoft continually 'ages the fleet' by riding Moore's Law each year.
Brent Thill (Jefferies) asked about the 45% of backlog tied to OpenAI and its durability.
Hood reframed it to the 55% (~$350 billion) that reflects the breadth of Microsoft's portfolio across customers, solutions, industries, and geographies, growing 28% with high confidence and a consistent adoption curve. On OpenAI, she said it is a great partnership, Microsoft remains its provider of scale, and it keeps Microsoft on the cutting edge of app innovation.
Karl Keirstead (UBS) asked, regardless of allocation, about the magnitude of capacity coming online, referencing the 1 GW added and the Fairwater Atlanta/Wisconsin sites.
Hood said Microsoft is adding capacity as fast as possible globally, that specific sites like Atlanta and Wisconsin are multi-year deliveries so investors should abstract away from individual locations, and that the focus is securing power, land, and facilities and filling them with GPUs/CPUs as quickly and efficiently as possible to meet demand.
Mark Murphy (JPMorgan) asked Satya to put the Maia 200 inference accelerator's achievements in perspective versus TPUs, Trainium, and Blackwell, and asked Amy about gross-margin ramifications.
Nadella said Microsoft has built its own silicon for a long time and is thrilled with Maia 200's performance (e.g., FP4 GEMMs running GPT-5), enabled by end-to-end model-silicon-system co-optimization and close round-tripping with its superintelligence team, but stressed Microsoft will not be locked into any one option and values its NVIDIA and AMD partnerships to keep the best TCO across generations, since leadership is a continuous, multi-generation game.
Brad Zelnick (Deutsche Bank) asked about the momentum of enterprises becoming 'frontier firms' and how much their Microsoft spend can expand.
Nadella described compounding adoption across the three major suites (M365, security, GitHub), highlighted Work IQ making the Microsoft 365 data layer the most important enterprise database of tacit knowledge, and said the agentic plane, together with Fabric, Foundry, GitHub tooling, and low-code/no-code tools, is transforming customer service, marketing, and finance workflows.
Raimo Lenschow (Barclays) asked about the CPU side of Azure and whether customers realizing they must be in the cloud to do AI is driving cloud transitions.
Nadella said AI workloads are not just AI-accelerator compute; agents spawn containers running on general compute and storage, so Microsoft builds the fleet in ratios of compute and storage to accelerators. He added that cloud migrations continue (e.g., SQL Server growing as an IaaS service in Azure), which is why Microsoft keeps its commercial cloud balanced with its AI cloud so customers have all infrastructure elements in their region.
More on Microsoft Corp
Reported 2026-01-28 · figures from the Microsoft Corp Q2 2026 earnings call.
See how Top Bucket AI works for your firm
Request DemoTop Bucket AI Research
Stay ahead of private markets
Research and market intelligence for private-markets professionals.
✓
You're subscribed.
Thanks for signing up.