Under the AI CapEx cycle, why is Microsoft entering a phase of rebalancing growth and valuation?

In June 2026, Microsoft experienced its most severe monthly sell-off since the dot-com bubble in 2000. On June 30 (Beijing time), Microsoft (MSFT) shares closed at $368.57, down 1.18% for the day, and briefly touched a 52-week low of $359.90 during the session. The cumulative decline for the month reached 18%, with market capitalization evaporating over $530 billion, pushing the stock to its lowest closing level since 2023.

This sell-off was not driven by a business contraction. Microsoft's fiscal 2026 second quarter (ending December 31, 2025) revenue reached $81.3 billion, up 17% year-over-year, exceeding market expectations of $80.27 billion; adjusted earnings per share were $4.14, also above the expected $3.97. Intelligent Cloud revenue was $32.9 billion, up 29% year-over-year, with Azure and other cloud services revenue growing 39%. Entering the third quarter (ending March 31, 2026), Azure's growth rate further accelerated to 40%.

Strong performance but a plummeting stock price — the market is repricing AI profitability. The core issue is not whether Microsoft's growth persists, but whether the return cycle and capital efficiency of AI investments are sufficient to support the current valuation framework.

Strong Azure Growth Fails to Soothe the Market

Azure is Microsoft's most valuable growth engine. In fiscal Q2 2026, Microsoft's total cloud business revenue reached $51.5 billion, up 26% year-over-year. The Intelligent Cloud segment contributed $32.9 billion in revenue in Q2, exceeding analysts' expectations of $32.39 billion. Azure grew 38% on a constant currency basis in Q2, further accelerating to 40% in Q3. By comparison, Amazon AWS grew at roughly 19% over a similar period, with Azure maintaining pressure on its core competitor in terms of market share expansion.

But the market's focus has shifted away from growth rates themselves. After the earnings release, Microsoft shares fell 7% in after-hours trading. CNBC cited analysts noting that while Azure growth met expectations, the market was more sensitive to capital expenditure intensity and operating margin guidance. Microsoft guided for an operating margin of 45.1% in fiscal Q3 2026, below the market expectation of 45.5%. Meanwhile, the company's Q2 gross margin narrowed to its lowest level in three years, just above 68%.

This is the core contradiction in the current valuation restructuring: Azure is still expanding, but the marginal cost of expansion is rising at a faster rate. The market is willing to pay a premium for growth, but only if that growth translates into predictable profit expansion. When capital expenditure growth consistently outpaces revenue growth, investors naturally ask a fundamental question: How much return does each dollar of AI investment actually generate?

Surging Capital Expenditure: From "Growth Narrative" to "Profit Validation"

Microsoft's capital expenditure (including finance leases) in fiscal Q2 2026 reached $37.5 billion, far exceeding Wall Street's expectation of $34.3 billion. In Q3, capital expenditure was $30.88 billion, up 84.39% year-over-year. The company raised its calendar 2026 capital expenditure guidance to approximately $190 billion, primarily for data centers, GPUs, and the underlying compute infrastructure for AI services like Azure and Copilot. It is estimated that about $25 billion of this figure comes from the impact of component price increases.

From an industry-wide perspective, Microsoft is not alone. Alphabet, Amazon, Meta, and Microsoft — the four major tech giants — plan to invest a combined approximately $725 billion in capital expenditure in 2026, up 77% from $410 billion in 2025. Goldman Sachs estimates that the combined capital expenditure of these four hyper-scalers from fiscal 2025 to 2030 will reach $5.3 trillion.

This level of spending is approaching the limits of cash flow. Bernstein estimates that in 2026, the combined operating cash flow of the four hyper-scalers will be approximately $635 billion, while capital expenditure is expected to reach $623 billion, nearly offsetting each other. Aside from Microsoft, the other hyper-scalers have already needed external financing to sustain expansion. As of early June 2026, Alphabet, Amazon, Meta, Microsoft, and Oracle had issued $159 billion in bonds for AI infrastructure construction, up 47% from 2025.

The rapid expansion of capital expenditure directly impacts profit margins. Microsoft guided for cloud business gross margin of 64% in fiscal Q4 2026, down 4 percentage points year-over-year, as early-stage costs of new capacity continue to outpace the revenue they generate. Stifel analyst Brad Reback noted in a June 25 research report that given "Azure gross margins are under pressure from accelerated capex," valuation expectations "appear significantly elevated," and lowered Microsoft's target price from $415 to $400.

AI Infrastructure Investment Cycle: GPUs, Data Centers, and Profit Redistribution

Understanding Microsoft's current market position requires viewing it within the macro cycle of AI infrastructure investment. This cycle involves three core layers: the compute layer (GPUs/chips), the physical layer (data centers), and the application layer (cloud services/AI products). The redistribution of profits among these three layers is reshaping the valuation framework of the entire tech industry.

The compute layer is currently the biggest beneficiary. Nvidia leads with an enterprise value (EV) of $4.8 trillion and a forward price-to-sales ratio of 10.4x. The massive capital expenditure from hyper-scalers directly translates into chip purchase orders, creating profit concentration in the compute layer.

The physical layer is becoming a new bottleneck. Industry analysis indicates that the bottleneck constraining cloud service provider expansion has shifted from GPU supply to the energized capacity of physical data centers. Who can bring compute capacity online faster and at lower cost is becoming a new competitive dimension. Microsoft added nearly 1 gigawatt of AI infrastructure capacity in Q2 and expanded to seven countries.

The application layer faces pressure to validate return on investment. Microsoft's AI business annualized revenue run rate has reached $37 billion, up 123% year-over-year. This scale is quite significant in absolute terms — as Microsoft CEO Satya Nadella said during the Q2 earnings call, "We are only in the early stages of AI diffusion, but Microsoft has already built an AI business that is larger than some of our largest business lines."

However, there remains a significant gap between the $37 billion annualized revenue and the $190 billion annual capital expenditure. Wedbush analyst Dan Ives believes that tech giants are in an "AI arms race" and will not cut spending due to stock price declines. He points out that the next 6 to 12 months will enter the commercialization phase for AI. But there are also views that as stock prices continue to fall, some hyper-scalers may have to lower capital expenditure commitments in their Q2 earnings reports.

Valuation Restructuring: From 27x to 19x

The sharp price adjustment has pushed Microsoft's valuation to its lowest level in a decade. As of June 29, 2026 (Beijing time June 30), Microsoft's P/E ratio (TTM) was 21.87, with a market cap of approximately $2.74 trillion. Based on expected profits over the next 12 months, the P/E ratio is about 19x. This valuation is not only below Microsoft's 10-year average of 27x but also below the S&P 500's valuation of approximately 20x.

From a broader perspective, Microsoft's enterprise value is about $2.7 trillion, with a forward price-to-sales ratio of about 7x, lower than Alphabet's 7.9x. Among the "Magnificent Seven," Microsoft's P/E ratio is only higher than Meta's.

The logic of valuation compression is not complicated: when the market shifts its assessment of a company from "growth-driven" to "profit-validation," valuation multiples naturally adjust downward. Investors are no longer simply paying for revenue growth; they demand verifiable returns for every dollar of capital expenditure. This shift in valuation framework is not uncommon in tech history — from early Amazon to Tesla in recent years, similar "narrative shift" cycles have occurred.

But a low valuation alone does not constitute a buy signal. The key is to determine whether the current valuation compression is an overreaction or a reasonable reset. Jack Ablin, chief investment strategist at Cresset Wealth Advisors, which holds Microsoft shares, said: "Although the valuation looks cheap and seems like a good trade, my sense is that investors are acting first and asking questions later." Michael Burry, the inspiration for the movie "The Big Short," recently bought call options on Microsoft with a strike price just above $700, expiring in 2028. This news pushed Microsoft shares up 5.7% on Friday, June 26, to $372.97.

Conclusion

What Microsoft is experiencing is not a growth crisis but a restructuring of valuation logic. Azure is still growing at 40%, the AI business has an annualized revenue of $37 billion, and commercial remaining performance obligations (RPO) reached $625 billion in Q2, up about 110% year-over-year. These fundamental indicators have not collapsed.

What has truly changed is the market's assessment framework for the return cycle of AI investment. When a company invests nearly $200 billion annually in infrastructure and the market demands verifiable returns for every dollar, compressing the valuation from a 27x P/E to 19x is a mathematical inevitability.

AI infrastructure represents the largest capital expenditure cycle in this generation of the tech industry. Goldman Sachs expects total capital expenditure from 2026 to 2031 to reach $7.6 trillion. In this cycle, profits will be redistributed among the compute layer, physical layer, and application layer. Microsoft, as a player that encompasses all three — chip buyer, data center operator, and provider of cloud services and AI products — will have its valuation trajectory serve as a bellwether for the entire industry.

Key variables to watch in the coming quarters include: whether Azure gross margins can stabilize and recover after the peak of capital expenditure, whether AI business revenue can accelerate to approach the scale of capital expenditure, and whether a coordination mechanism for capital expenditure will emerge among the hyper-scalers. The answers to these questions will determine whether Microsoft's current valuation represents "deep value" or a "value trap."

FAQ

Q: What was the main reason for the sharp drop in Microsoft's stock price in June 2026?

Market concerns over the imbalance between Microsoft's AI capital expenditure scale and its return cycle. Microsoft's calendar 2026 capital expenditure guidance is approximately $190 billion, while AI business annualized revenue is $37 billion, a gap that has raised investor doubts about near-term profitability. Additionally, Azure gross margins are under pressure from accelerated capex, and operating margin guidance fell short of market expectations.

Q: What is the actual performance of Azure cloud business?

Azure growth remains strong. Azure grew 39% in fiscal Q2 2026, further accelerating to 40% in Q3. Microsoft's cloud business quarterly revenue reached $51.5 billion, up 26% year-over-year. The issue is not growth itself, but the unprecedented speed at which capital expenditure required to support this growth is expanding, squeezing near-term profit margins.

Q: Will Microsoft's AI investments ultimately generate sufficient returns?

Microsoft's AI business annualized revenue has reached $37 billion, up 123% year-over-year. Commercial remaining performance obligations stood at $625 billion, up about 110% year-over-year, indicating strong demand-side support. However, validating returns takes time — the AI infrastructure investment cycle is long, with high upfront costs and delayed payoff. The market is currently in a transition period from a "growth narrative" to "profit validation."

Q: Is Microsoft's current valuation attractive?

Microsoft's current P/E ratio (TTM) is about 21.87x, and about 19x based on expected profits over the next 12 months. This is below its 10-year average of 27x and also below the S&P 500's roughly 20x. The valuation is indeed low, but whether it constitutes a buying opportunity depends on whether AI investments can translate into predictable profit growth in the coming quarters.