Microsoft (MSFT) Q3 2026 Earnings Transcript

Microsoft reported Q3 revenue of $82.9 billion, operating income up 2016% in constant currency, and EPS of $4.27, driven by surging cloud and AI demand. AI business ARR surpassed $37 billion, Copilot paid seats topped 20 million, and Azure growth remained strong, though management flagged over $40 billion of Q4 capex and continued supply constraints. Guidance for Q4 calls for $86.7 billion to $87.8 billion in revenue and roughly $190 billion of calendar 2026 capex, reinforcing a heavy investment cycle but also sustained top-line momentum.

Analysis

The market’s instinct will be to treat this as another “great quarter, spend more” MSFT print, but the more important signal is that Microsoft is converting AI from a margin-dilutive story into an operating leverage story faster than expected. The key second-order effect is that its model mix is shifting from discrete software licenses to metered usage layered on top of seats, which should extend monetization life far beyond the initial Copilot install cycle. That makes the current capex intensity less like a cost spike and more like an arms race to lock in the highest-velocity enterprise workloads before competitors can match context, identity, and workflow integration.

The biggest winners are the companies that can ride Microsoft’s distribution rather than fight it: infrastructure vendors tied to accelerated compute, selected data/agent platform adjacencies, and services firms that implement these workflows. The losers are point-solution vendors whose value prop is “AI assistant” without embedded data, governance, or workflow ownership; they risk being commoditized as Microsoft packages baseline intelligence into existing bundles. The most underappreciated competitive dynamic is that usage-based pricing should make procurement easier for buyers while simultaneously widening Microsoft’s share of wallet, because customers can fund AI out of productivity gains and process savings rather than new headcount budgets.

The main risk is not demand; it is execution on supply and unit economics over the next 2-4 quarters. If component inflation persists or capacity comes online too late, gross margin expansion could stall even with strong top-line acceleration, and investors may punish the stock for capex visibility before revenue catches up. A second risk is customer pushback on consumption bills, but the call strongly suggests Microsoft is preempting that with entitlement-plus-overage packaging, which should blunt churn and reduce budget shock.

Contrarian view: the Street may still be underestimating how quickly this becomes a platform monetization cycle rather than a product cycle. If Copilot usage is already habitual, then the right framing is not seat penetration but token intensity per employee; that implies Microsoft can keep expanding revenue per installed base even with modest seat growth. The tradeable implication is that MSFT’s multiple may deserve to stay structurally above legacy software peers as long as Azure supply remains tight and Copilot usage keeps inflecting.

Key Decisions for Investors

• Long MSFT 3-6 months: buy on any post-earnings digestion; thesis is that metered AI monetization and booked demand will offset capex overhang. Risk/reward favors 10-15% upside if the market starts capitalizing FY27 operating leverage, with downside limited unless Azure growth decelerates meaningfully.
• Long ACN vs short a basket of low-context SaaS names (e.g., SNOW, SSTK) for 3-9 months: ACN should benefit from implementation spend as enterprises operationalize Microsoft’s agent stack, while point solutions face pricing pressure. Target a 20-30% relative move if AI workflow adoption broadens.
• Buy NVDA on pullbacks, but size smaller than MSFT: the supply-tight capex plan implies sustained accelerator demand over the next 4-6 quarters. Use a 6-12 month horizon; upside is tied to allocation discipline, but risk is that Microsoft’s custom silicon gains take some share.
• Short a basket of enterprise software names with weak workflow ownership for 6 months: focus on vendors where AI is a feature, not distribution. The risk/reward is favorable if Microsoft continues bundling enough capability to slow new-logo conversion and compress renewal power.
• For higher convexity, consider MSFT call spreads 3-6 months out: this expresses the view that the market still underprices AI margin durability while limiting exposure to capex and supply-chain headlines.