Microsoft: This Sell-Off Doesn't Make Any Sense

Microsoft delivered Q3 FY26 results above revenue and EPS estimates, with Azure growth at 40% and AI revenue up 123% year over year. However, ballooning CapEx is now projected at $190B for 2026 and capacity constraints are limiting Azure acceleration despite strong backlog and Copilot user growth. Shares remain 27% below all-time highs, though the stock trades below its 15-year average P/E while FY26-FY28 EPS is projected to grow 18.7%.

Analysis

The key market setup is not “MSFT beats” but a temporary mismatch between demand visibility and supply monetization. When capex rises faster than near-term capacity, the equity can de-rate even while the underlying franchise is strengthening, because investors are paying today for cash flows that may not arrive until the next 4-8 quarters. That creates a classic second-order winner set: power, networking, server OEMs, and semiconductor suppliers benefit immediately, while competitors without comparable balance-sheet firepower face a widening AI infrastructure gap.

The most important competitive implication is that MSFT is using balance-sheet scale to lock in future share, not just defend current share. If Azure backlog is strong but constrained, the bottleneck shifts negotiating leverage toward customers willing to pre-commit capacity and toward suppliers able to deliver scarce components; that tends to compress margins for everyone else in the stack before it shows up in reported cloud growth. The flip side is that persistent shortages can slow monetization of AI demand, giving nimbler hyperscalers and specialized cloud providers a window to take incremental workloads if they can source capacity faster.

From a risk perspective, the next 1-2 quarters matter more for sentiment than fundamentals: investors will punish any sign that incremental capex is yielding lower-than-expected near-term revenue conversion. The bigger upside catalyst is not another earnings beat, but evidence that capacity additions are unlocking a step-up in Azure growth without a proportional increase in capex intensity; that would re-rate the stock over a 6-12 month horizon. A contrarian read is that the selloff may already be pricing in a slowdown that is mechanically self-inflicted by supply constraints, which is more fixable than a demand problem.

The asymmetry is favorable if one believes AI revenue is still in an early adoption curve and the market is over-indexing on near-term free-cash-flow optics. If the capex cycle proves front-loaded, EPS growth can re-accelerate into FY27-FY28 as utilization improves and depreciation lags the revenue ramp. In that case, the current drawdown looks more like a financing of future growth than a deterioration in franchise quality.

Key Decisions for Investors

• Long MSFT on a 6-12 month horizon; add on weakness if the market continues to punish capex because the setup is for delayed, not broken, monetization. Risk/reward improves if shares remain below the prior high while revenue visibility stays intact.
• Pair trade: long MSFT / short a weaker cloud platform with less capex flexibility over 3-6 months. The thesis is that the scarcity regime rewards balance-sheet scale and supply access, not just software demand.
• Long NVDA and/or ANET into any MSFT post-earnings consolidation over the next 1-3 months. MSFT’s capex intensity is effectively a demand transfer to AI infrastructure suppliers with cleaner near-term earnings acceleration.
• Buy MSFT call spreads 6-9 months out, funded by selling upside further out. This expresses the view that the stock can re-rate once the market sees capacity conversion, while capping premium if capex stays punitive.
• Monitor for a reversal trigger: if Azure growth decelerates without a corresponding capex step-down over the next two quarters, reduce exposure. That would imply the market is right to question incremental ROI rather than just timing.