Microsoft Earnings Call: AI Growth Surges, Margins Squeezed

Microsoft ((MSFT)) has held its Q3 earnings call. Read on for the main highlights of the call.

Microsoft’s latest earnings call struck an upbeat tone, showcasing record cloud and AI revenue alongside rapid customer adoption. Executives framed the quarter as proof that the multiyear AI investment cycle is working, even as they candidly acknowledged margin compression, soaring capital spending, and pockets of weakness in legacy consumer businesses.

Record Top-Line Growth and Earnings Beat

Microsoft reported quarterly revenue of $82.9 billion and diluted EPS of $4.27, with performance exceeding internal expectations across revenue, operating income, and earnings. Management argued that strength was broad-based, underscoring resilience in core franchises even as the business shifts toward AI-driven, consumption-heavy models.

Microsoft Cloud Delivers Scale and Profit

Microsoft Cloud revenue climbed to roughly $54 billion, up 29% year over year, underscoring the company’s central role in enterprise digital transformation. Cloud gross margin hovered around 66%, with leadership pointing to robust demand for Azure and first-party AI services despite the drag from heavy infrastructure spending.

AI Revenue Run Rate Surges

AI annual revenue run rate surpassed $37 billion, growing 123% from a year earlier and confirming that AI is no longer a small experimental line item. Management emphasized that AI is now a broad business spanning infrastructure, models, and applications, with platform momentum that they believe will compound over multiple years.

Copilot and Agents Gain Traction

Microsoft 365 Copilot passed 20 million paid seats, with paid seat additions accelerating 250% year over year as customers scale pilots into production. Copilot queries per user rose about 20% quarter over quarter, while usage of first-party agents climbed sixfold year to date, suggesting deepening engagement rather than one-off trials.

Developer and GitHub Adoption Expands

GitHub Copilot in Enterprise is now used by nearly 140,000 organizations, nearly triple last year’s count, highlighting strong developer demand for AI coding tools. The company also noted GitHub Copilot CLI usage nearly doubled month over month and announced a shift toward usage-based pricing aimed at better aligning value with consumption.

Fabric and Foundry Power Data and Model Workloads

In data and AI infrastructure, Microsoft reported over 35,000 paid Fabric customers, up 60% year over year, and more than 10,000 Foundry customers using multiple models. Over 15,000 customers now use both Foundry and Fabric, while more than 300 customers are expected to process over 1 trillion tokens this year, with token usage growing 30% quarter over quarter.

New Silicon and Data Center Scale

Management highlighted a 40% improvement in inference throughput for its most-used models, reflecting ongoing efficiency gains in the stack. The company introduced its Maya 200 AI accelerator, said to deliver around 30% better tokens-per-dollar versus prior silicon, and ramped its Cobalt CPU while bringing the Fairwater data center online six weeks early and adding another gigawatt of capacity.

Cash Generation Supports Heavy Investment

Operating cash flow reached $46.7 billion, up 26% year over year, while free cash flow came in at $15.8 billion after substantial capital outlays. Microsoft returned $10.2 billion to shareholders through dividends and buybacks during the quarter, even as it devoted $31.9 billion of CapEx to scaling AI and cloud infrastructure.

Massive Consumer and Platform Reach

The company underscored the breadth of its ecosystem, with Windows monthly active devices surpassing 1.6 billion and Bing monthly active users reaching 1 billion. LinkedIn’s membership base has grown to 1.3 billion, Microsoft 365 Consumer is approaching 95 million subscribers, and Xbox set records for monthly users and game streaming hours.

AI Spending Weighs on Margins

Despite strong revenue, overall gross margin declined to 68%, with Microsoft Cloud margin settling near 66%, reflecting the cost of AI infrastructure build-out. Executives signaled that as AI usage grows and newer models roll out, near-term margin pressure will continue before scale efficiencies and pricing can fully catch up.

Operating Expenses Climb with AI Ambitions

Operating expenses rose sharply, driven mainly by research and development as well as compute and talent investments needed to support AI growth. Management framed this expense surge as a deliberate strategy to secure long-term leadership in AI, while indicating they will continue to balance hiring and efficiency measures.

CapEx Enters a Supercycle

Capital expenditures reached $31.9 billion in the quarter, and management expects CapEx to climb above $40 billion next quarter. They outlined plans for roughly $190 billion in capital investments through calendar 2026, which will suppress near-term free cash flow versus a normal cycle but is intended to build durable AI and cloud capacity.

Capacity Constraints Limit Azure Supply

Demand for Azure and AI services continues to outstrip available capacity, with management expecting supply constraints to persist at least through 2026. The company is carefully allocating scarce resources between its own first-party workloads and external customer demand, suggesting a seller’s market but also emphasizing execution risk.

Booking Volatility Around OpenAI and Contracts

Commercial bookings dropped significantly when including a large Azure commitment from OpenAI, with management pointing to a 46% decline tied to timing. Excluding OpenAI, bookings grew 7%, which executives argued better reflects underlying demand but also highlights how large partner deals and contract mix can distort reported bookings.

Headwinds in Consumer Hardware and Gaming

More Personal Computing revenue declined 1% overall and 3% in constant currency, with Windows OEM and devices revenue down materially, signaling ongoing PC and hardware softness. Gaming revenue also fell amid impairments and other gaming-related costs, underscoring that not all parts of the portfolio are participating in the AI-driven upswing.

Short-Term Profitability Tradeoffs

Guidance incorporates about $900 million of one-time costs tied to a voluntary retirement program, split between cost of goods sold and operating expenses. Management also expects Microsoft Cloud gross margin to compress further to around 64% near term as AI investments deepen, highlighting conscious tradeoffs between profitability and growth.

Shifting to Mixed Seat and Consumption Models

Executives noted that a growing mix of seat-plus-consumption pricing, especially for AI services, may introduce temporary volatility in bookings and renewal patterns. As customers adjust procurement and budgeting toward more usage-based buying, reported metrics may become bumpier even if underlying usage continues to rise.

Free Cash Flow Reflects CapEx Drag

Free cash flow of $15.8 billion was solid but meaningfully constrained by elevated capital spending as data centers and AI capacity come online. Management reiterated that investors should expect lower near-term cash generation than historical norms, with the expectation that these investments will yield higher returns over time.

Guidance Signals Growth with Ongoing Investment

Looking ahead, Microsoft guided Q4 revenue to $86.7–$87.8 billion, implying 13%–15% growth, with Intelligent Cloud expected to rise 27%–28% and Azure growing close to 40% in constant currency. They foresee continued strength in Microsoft 365 and LinkedIn, mid-teen declines in Windows OEM and devices, cloud gross margin around 64%, CapEx above $40 billion, persistent Azure capacity constraints, and a plan for operating margin expansion by fiscal 2026 as efficiencies build.

Management closed the call presenting a company leaning hard into AI at global scale, accepting margin and cash flow pressure today to secure durable growth tomorrow. For investors, the quarter underscored that Microsoft’s AI franchise is real and accelerating, but it also comes with higher capital intensity, greater complexity in metrics, and uneven performance in legacy consumer segments.