Microsoft Earnings Call: AI Spend Surges, Cloud Soars

Microsoft ((MSFT)) has held its Q2 earnings call. Read on for the main highlights of the call.

Microsoft Earnings Call Signals Powerful AI Tailwinds Amid Margin Strain

Microsoft’s latest earnings call painted a picture of a company in the middle of a powerful AI-driven upcycle, with record cloud revenue, surging demand for Copilot and data platforms, and exceptionally strong cash generation. Management was candid about the trade-offs: heavy, near-term capital spending on AI infrastructure is pressuring margins and free cash flow, and there are pockets of weakness in gaming and parts of the personal computing portfolio. Still, the tone was clearly optimistic, with leadership emphasizing durable demand, accelerating AI adoption across enterprises and industries, and confidence that current investments will translate into stronger long-term earnings power.

Microsoft Cloud Crosses $50 Billion as AI Fuels Growth

Microsoft Cloud revenue surpassed the $50 billion milestone for the first time, reaching $51.5 billion and growing 26% year over year in constant currency. This acceleration underscores how central cloud and AI workloads have become to Microsoft’s story. Cloud gross margin came in around 67%, slightly better than expected but lower than a year ago, reflecting the cost of building out massive AI infrastructure. The message to investors is that AI is driving both volume and mix—more compute-intensive workloads and higher-value services—but requires sustained investment that temporarily compresses margins even as revenue scales.

Broad-Based Financial Strength with Expanding Operating Margins

At the company level, performance was robust: total revenue reached $81.3 billion, up 17% in constant currency, while operating income rose 21%, pushing operating margin up to a notably high 47%. GAAP gross margin was 68%, only slightly down year over year despite the AI spend. Adjusted EPS came in at $4.14, up 24% in constant currency after adjusting for the impact of OpenAI-related items. The combination of double-digit top-line growth and faster-growing operating income positions Microsoft as one of the most profitable and operationally disciplined mega-cap tech names, even while it leans aggressively into AI.

Cash Engine Remains Powerful Despite CapEx Drag

Microsoft’s cash generation remains a key pillar of the investment case. Operating cash flow surged 60% year over year to $35.8 billion, underscoring the cash-rich nature of its cloud and software franchises. The company returned $12.7 billion to shareholders in the quarter, up 32% year over year, through dividends and buybacks. Free cash flow, however, was only $5.9 billion and declined sequentially, a direct consequence of elevated cash capital expenditures. Investors are effectively seeing the trade-off in real time: Microsoft is choosing to reinvest aggressively in AI capacity, temporarily depressing free cash flow while maintaining generous shareholder returns.

AI Infrastructure and Custom Silicon Take Center Stage

Management highlighted major strides in AI infrastructure, adding nearly 1 gigawatt of capacity in the quarter alone. Microsoft showcased its custom silicon strategy with the Maya 200 AI accelerator—delivering more than 10 FLOPs at FP4 and over 30% better total cost of ownership versus the prior generation—and the Cobalt 200 CPU, which offers over 50% higher performance than its previous custom CPU. Roughly two-thirds of Q2 capital expenditures went toward short-lived assets like GPUs and CPUs to support AI workloads. The company is clearly positioning itself not just as a cloud provider but as an integrated AI computing platform with differentiated hardware and software.

Fabric and Foundry: Data and Model Platforms Gain Traction

Microsoft’s data and model platforms, Fabric and Foundry, are emerging as important growth drivers. Fabric’s annual revenue run rate topped $2 billion, with revenue up 60% year over year and more than 31,000 customers already on board, suggesting rapid adoption of its integrated analytics and data stack. Foundry is also scaling quickly, with the number of customers spending more than $1 million per quarter growing nearly 80%. Over 250 Foundry customers are on track to process more than 1 trillion tokens this year, a signal that large, production-grade AI workloads are increasingly being built on Microsoft’s infrastructure. Together, these platforms deepen customer lock-in and expand the monetization surface around AI.

Copilot and Agents See Rapid, Broad-Based Adoption

The company is seeing strong proof that AI assistants are becoming mainstream across its ecosystem. Microsoft 365 Copilot reached 15 million paid seats, with seat additions up more than 160% year over year. Daily active users are up 10x, and average conversations per user have doubled over the same period, indicating rising engagement and utility. GitHub Copilot hit 4.7 million paid subscribers, up 75% year over year, with Copilot Pro+ subscriptions rising 77% quarter over quarter. These metrics suggest Microsoft is successfully monetizing AI at the application and developer layers, not just in raw infrastructure, which could translate into higher ARPU and stickier enterprise relationships.

Bookings Strength and RPO Underscore Long-Term Visibility

Commercial demand indicators were robust. Commercial bookings grew 23% in constant currency, and commercial remaining performance obligation (RPO) climbed to $625 billion, up 10% year over year. The weighted average duration of RPO is roughly 2.5 years, and about 25% of commercial RPO is expected to be recognized in the next 12 months—a 39% increase versus last year. For investors, this growing backlog provides a line of sight into future revenue and underscores that customers are signing multi-year commitments for cloud and AI services, bolstering confidence in the durability of Microsoft’s growth trajectory.

Sector Wins Highlight Real-World AI Use Cases

Microsoft called out specific industry deployments that show AI is moving beyond experimentation into scaled production. In healthcare, Dragon Copilot now supports more than 100,000 providers and documented 21 million patient encounters this quarter alone, up 3x year over year, illustrating how AI is being used to tackle administrative burdens in clinical settings. Enterprises such as Alaska Airlines, BMW, Land O’Lakes, and Unilever are using Foundry, Fabric, and agent-based systems to modernize operations and decision-making. These case studies help validate the practical value of Microsoft’s AI offerings and signal broad-based demand across verticals.

CapEx Surge and Short-Lived Asset Focus Raise Investor Questions

Capital expenditures jumped to $37.5 billion in Q2, with about two-thirds directed to short-lived hardware like GPUs and CPUs. Cash paid for property, plant, and equipment was $29.9 billion. Management acknowledged that the pace and mix of CapEx are key investor concerns, particularly since these assets depreciate quickly and free cash flow dropped to $5.9 billion. The company reiterated that it is balancing supply with rapidly rising demand, framing current spending as necessary to capture a generational AI opportunity. Still, the near-term optics are clear: Microsoft is front-loading investment to secure capacity, accepting short-term cash flow compression in exchange for future revenue and profit streams.

OpenAI Concentration Introduces Bookings and RPO Volatility

Approximately 45% of Microsoft’s commercial RPO is tied to OpenAI, creating notable concentration risk in its backlog. Management warned that this could introduce quarter-to-quarter volatility in bookings and RPO growth as large, multi-year AI contracts are signed and ramped. While the tight relationship with OpenAI strengthens Microsoft’s position in cutting-edge AI models and services, investors will need to monitor how this concentration evolves and whether demand from a broader base of AI customers offsets potential swings associated with a single strategic partner.

Gaming Weakness and Impairments Weigh on Personal Computing

The gaming segment was a clear soft spot. Gaming revenue declined 9% in constant currency, with Xbox content and services down 6%, falling short of expectations due to underperformance in first-party titles. The company also recorded impairment charges in the gaming business, which pushed up operating expenses in the More Personal Computing segment. These results highlight that, even as AI and cloud drive the broader story, Microsoft’s consumer-facing entertainment portfolio is facing cyclical and execution-related challenges that could continue to drag on the segment’s profitability.

AI Spend Pressures Gross Margins

Companywide, gross margin slipped slightly year over year to 68%, while Microsoft Cloud margins are feeling the weight of AI infrastructure build-out. Management guided to cloud gross margin of about 65% in the next quarter, down from the prior year. Investors should expect some ongoing margin pressure as the company scales AI capacity and usage, with management effectively arguing that this is a temporary phase: margins may compress now but should recover over time as utilization rises, pricing reflects higher value, and economies of scale kick in.

More Personal Computing Faces Mixed Trends and Ad Headwinds

The More Personal Computing segment saw revenue decline 3% in the quarter. Within that, Windows OEM and devices were largely stable, with Windows OEM growing 5%, though the company expects Windows OEM revenue to fall roughly 10% in the next quarter as the boost from Windows 10 end-of-support normalizes. Search and news advertising (excluding TAC) grew 9% in constant currency but landed slightly below expectations due to execution issues. Overall, the segment remains a smaller and more cyclical contributor compared with cloud, and is currently facing both normalization and operational challenges.

Supply Constraints Force Capacity Trade-Offs Across AI Businesses

Demand for AI capacity is outstripping supply, and management was explicit that it is having to make allocation trade-offs. Microsoft is prioritizing among its own Copilot and GitHub workloads, capacity commitments to OpenAI and Azure customers, and internal R&D teams. This constrained environment could create quarterly variability in Azure growth and revenue recognition, depending on the timing of new capacity coming online. For investors, this means AI demand is not the limiting factor—physical infrastructure is—and reported growth may be more lumpy than underlying demand trends would suggest.

Guidance Points to Continued High Growth with Ongoing AI Investment

For the upcoming quarter, Microsoft guided revenue to a range of $80.65–$81.75 billion, implying 15–17% growth, with foreign exchange expected to add about 3 points to overall revenue growth. Cost of goods sold is projected to grow 22–23%, and operating expenses 10–11%, reflecting continued AI-related spending. Segment guidance calls for mid-teens growth in Productivity & Business Processes, high-20s growth in Intelligent Cloud with Azure up 37–38% in constant currency, and a more subdued More Personal Computing segment, including a roughly 10% decline in Windows OEM and mid-single-digit declines in Xbox content and services. Microsoft Cloud gross margin is expected around 65%, down year over year, and operating margins are forecast to be slightly down in Q3 before improving over the longer term; management now expects operating margins in fiscal 2026 to be up slightly versus prior expectations, signaling confidence that the current investment cycle will ultimately be margin-accretive.

In summary, Microsoft’s earnings call underscored a company leaning hard into an AI super-cycle, with record cloud revenue, strong bookings, and rapid adoption of Copilot, Fabric, and Foundry more than offsetting pockets of weakness in gaming and personal computing. Near-term headwinds from heavy AI CapEx, margin pressure, and concentrated exposure to OpenAI were acknowledged, but management consistently framed them as the necessary cost of capturing a large and durable growth opportunity. For investors, the story remains one of strong execution, high visibility into future revenue, and a calculated bet that today’s intensive AI build-out will translate into even higher earnings power over the next several years.