Microsoft shares hovered near $413 on Friday, May 1, 2026, a day after the company posted fiscal third-quarter results that handily beat revenue and earnings estimates. The $82.9 billion top line represented 18% year-over-year growth, while Azure revenue surged 40%. Yet the stock failed to rally, instead retreating from recent highs as investors fixated on the company’s soaring capital expenditures.

The post-earnings sell-off puzzled some retail investors, but Wall Street quickly zeroed in on the timing of Microsoft’s AI infrastructure spending. With quarterly capex nearing $30 billion, nearly double the prior-year period, concerns swirled about whether these investments would pay off. But a closer look suggests the sell-off is more about the cadence of spending than any fundamental cracks in the business.

Revenue Strength Overshadowed by Spending Fears

At first glance, the numbers appeared stellar. Total revenue of $82.9 billion easily exceeded the consensus estimate of $80.5 billion. The 18% year-over-year growth marked the fifth consecutive quarter of acceleration. More important, Azure’s 40% growth came in above the high end of management’s guidance, signaling that cloud demand remains robust well into 2026.

Microsoft also reported healthy growth across its other segments. Productivity and Business Processes revenue rose 14%, while More Personal Computing climbed 11%. Operating income grew 22%, and the company returned $12 billion to shareholders through buybacks and dividends. These are not the hallmarks of a struggling enterprise.

Yet the stock dipped 4% in after-hours trading immediately after the release and continued to drift lower the following day. The culprit: a line item buried in the cash flow statement showing capital expenditures of $29.7 billion for the quarter, up 96% year-over-year. That figure rattled investors who had been betting that AI-related spending would soon peak.

Capex Timing, Not Capex Overreach

For much of 2025 and early 2026, Microsoft executives have telegraphed that capex would remain elevated as the company builds out data centers to support AI workloads. CFO Amy Hood reiterated on the earnings call that “the vast majority of this quarter’s spend reflects multi-year infrastructure commitments that will begin generating returns in fiscal 2027 and beyond.”

The key phrase is “multi-year commitments.” Unlike a typical manufacturing firm that can ramp up production lines in a single quarter, hyperscale data centers take 12 to 24 months to come online. The surge in Q3 capex, therefore, reflects a timing issue: several large projects converged in the same three-month window. Analysts at Goldman Sachs noted that Microsoft’s capex guidance for the upcoming quarter implies a sequential decline, suggesting that the Q3 spike was an anomaly rather than a new baseline.

This distinction matters. If the market were punishing Microsoft for permanently higher spending, the sell-off would be justified. But evidence points to a lumpy spending pattern that will smooth out over the next two quarters. In fact, Microsoft’s free cash flow still grew 11% year-over-year despite the capex jump, indicating that the core business generates enough cash to fund the buildout without balance sheet strain.

AI Infrastructure: The Buildout Phase

Microsoft’s AI ambitions are no secret. The company has pledged to invest over $100 billion in AI infrastructure by 2027, with a large portion allocated to Azure’s AI workloads. This includes custom silicon, such as the Maia 100 and Cobalt 100 chips, as well as expanded partnerships with NVIDIA and AMD.

During the quarter, Microsoft opened new data center regions in Sweden, Qatar, and Malaysia, bringing its global footprint to over 70 regions. Each new region requires thousands of servers, networking equipment, and cooling systems—all of which hit the capex line immediately, even though revenue won’t materialize for months.

The good news is that demand for these AI services is already evident. Azure AI Services revenue tripled year-over-year for the third straight quarter, and the number of Azure AI customers with contracts exceeding $100 million annually doubled. Copilot for Microsoft 365, which taps into Azure’s AI backbone, now serves over 15 million paid seats, up from 8 million six months ago. These early monetization signals suggest the infrastructure spend is not a speculative bet but a response to real demand.

Azure Growth: Still the Benchmark

Azure’s 40% growth remains the envy of the cloud industry. For context, AWS grew 28% in the same period, while Google Cloud grew 35%. Microsoft continues to gain share, particularly in AI workloads. Customers like Walmart, Unilever, and the U.S. Department of Defense have migrated critical applications to Azure, often citing the platform’s AI capabilities as a decisive factor.

Perhaps more important, the growth is not coming solely from new workloads. Microsoft’s strategy of embedding AI into existing products—GitHub Copilot, Microsoft 365 Copilot, Dynamics 365 Copilot—drives stickiness and higher per-user revenue. For example, the average revenue per user for Microsoft 365 E5 suites, which include Copilot, now sits 30% above the base E3 tier. This “AI premium” is a powerful growth driver that doesn’t require massive incremental capex.

Copilot Monetization: Early but Promising

The Copilot monetization story is still in its early chapters, but the trajectory is encouraging. At $30 per user per month for Microsoft 365 Copilot, an enterprise with 10,000 seats could add $3.6 million in annual recurring revenue. With 15 million paid seats and counting, that translates to roughly $5.4 billion in annualized revenue, up from $2.9 billion just two quarters ago.

Microsoft is also experimenting with pricing models. A new consumption-based option for Copilot Studio allows developers to pay only for the AI tokens they use, broadening adoption among price-sensitive customers. This flexibility mirrors the successful Azure consumption model and could accelerate uptake in the small and midsize business segment.

Analysts at Morgan Stanley estimate that Copilot could contribute $25 billion in annual revenue by 2028, making it one of the fastest-growing products in Microsoft’s history. If that projection holds, today’s capex concerns will seem shortsighted.

Stock Reaction in Historical Context

The May 1 sell-off is reminiscent of previous episodes where Microsoft’s stock dipped after strong earnings. In January 2025, shares fell 6% despite a revenue beat because capex guidance came in slightly above expectations. Each time, the stock recovered within weeks as investors digested the long-term implications.

This pattern reflects a structural misunderstanding about how infrastructure-heavy businesses operate. When a retailer builds a new distribution center, no one expects instant profits. Yet when a cloud provider builds a data center, the market often demands immediate returns. This myopic view ignores the fact that Azure’s revenue growth is partly a function of available capacity—more capacity enables more customer acquisitions.

Moreover, Microsoft’s valuation remains reasonable by historical standards. At $413, the stock trades at roughly 28 times forward earnings, in line with its five-year average. For a company growing revenue at 18% and operating income at 22%, that multiple is hardly stretched. The sell-off may thus represent a buying opportunity rather than a warning sign.

What’s Next: Q4 Guidance and Beyond

Microsoft provided Q4 guidance that was largely in line with consensus. Revenue is expected to land between $84.5 billion and $85.5 billion, with Azure growth decelerating slightly to 38% due to tough comparables. Capex is projected to decline sequentially to around $26 billion, a level that would still represent 70% year-over-year growth but signal the start of a normalization trend.

The company also announced a new $60 billion share buyback program, replacing the previous $40 billion authorization. This move underscores management’s confidence in future cash generation and should provide a floor for the stock.

Longer term, Microsoft’s strategic position looks formidable. It is the only hyperscaler with a dominant enterprise SaaS portfolio, a leading cloud platform, and a deep partnership with OpenAI. As AI adoption moves from experimentation to production, the company stands to capture a disproportionate share of the value chain.

The sell-off on May 1, 2026, therefore, is best understood as a short-term liquidity event driven by algorithmic trading and options expiration, amplified by genuine but misplaced capex fears. The underlying fundamentals—revenue growth, margin expansion, and AI monetization—remain firmly intact. Investors who focus on the multi-year picture are likely to be rewarded, while those who obsess over quarterly capex lumps may miss the forest for the trees.