Microsoft’s rally is the market voting that AI demand still matters more than AI sticker shock. The cleanest reason is that the business is still converting heavy infrastructure spend into real growth: Azure grew 40% in the latest quarter, Microsoft Cloud revenue rose 29% to $54.5 billion, and management is still talking like a company chasing constrained supply rather than apologizing for weak uptake. At $372.97, this does not look like a broken AI story getting a sympathy bounce. It looks like investors deciding the capex backlash against MSFT is missing the one thing that matters most: demand is still there.
The strongest bull point is simple: the growth engine is still firing at hyperscale. In Q3 FY26, Microsoft posted $82.9 billion in revenue, up 18% year over year, while Azure grew 40%. That is not the profile of a platform spending ahead of reality. It is the profile of a platform that is still monetizing AI and cloud demand fast enough for the market to tolerate massive investment.
The second point is visibility. Microsoft disclosed $627 billion in commercial remaining performance obligations, which is exactly the kind of number that supports the market’s willingness to look through giant capex. Customers are not dabbling here; they are signing long-dated commitments into Microsoft’s stack. That backlog matters more than the hand-wringing around spending because it suggests Microsoft is not building capacity into a vacuum.
The third point is that profitability gives Microsoft room to spend from a position of strength. A 46.8% operating margin and 39.3% net margin are elite by any standard, and they help explain why the TickerSpark Score stays strong at 74 overall, including a perfect 100 on Profitability and 85 on Growth. Even valuation is less stretched than the popular narrative suggests: MSFT trades at 22.13 times trailing earnings with a 0.74 PEG, cheaper on P/E than Apple at 34.27 and Alphabet at 25.74 while delivering faster growth than Apple and comparable margins to Alphabet. If the market were staring at runaway spending with no earnings power behind it, those numbers would not hold up this well.
The pushback is real, and it is not hard to understand. Management has pointed to roughly $190 billion of 2026 capex, and that is large enough to make any investor question whether returns can keep pace. The OpenAI relationship also looks less exclusive after the April rewrite, which weakens the easy version of the “Microsoft owns the AI stack” argument. Add in the fact that the stock is still below its 50-day and 200-day moving averages and down 21.1% year to date while the technology sector is up 25.5%, and the skepticism has a real foundation.
That still does not beat the bull case because the operating numbers are stronger than the narrative of capex excess. Microsoft has beaten earnings estimates in seven straight reported quarters, consensus still sits at Buy with 66 buys against 16 holds, and recent sentiment has stayed strongly positive. If Azure were decelerating hard or margins were cracking, the spending backlash would land. Instead, investors are looking at 40% Azure growth and a $627 billion backlog and concluding the spend is moat-building, not value-destructive.
That leaves MSFT looking more like a high-quality AI platform in a temporary credibility fight than a hyperscaler whose economics are rolling over. We would respect the volatility because momentum is still weak, with the Momentum component of the TickerSpark Score at 30 and technicals not yet repaired, but the fundamental case is stronger than the tape has implied. For investors who want AI exposure without paying NVDA-style multiples, Microsoft still looks like one of the cleaner ways to own the demand side of the buildout.
What would change our mind is not another scary capex headline by itself. The real trigger would be a quarter where capex stays enormous while Azure growth and commercial commitments lose altitude at the same time. Until that happens, the market’s read looks right: Microsoft is spending big because demand is still bigger.