Analysis
Microsoft closed out fiscal 2026 by beating Wall Street on every number that mattered. Fiscal fourth-quarter revenue came in at $90.01 billion, up 18% year-over-year and comfortably ahead of the $87.62 billion analysts had modeled, while adjusted earnings per share of $4.74 blew past the $4.24 consensus. The headline inside the headline: Azure and other cloud services revenue grew 43% for the quarter, and Microsoft confirmed that Azure's full fiscal-year revenue crossed $100 billion for the first time -- a threshold that puts the business behind only Amazon Web Services in absolute scale, and now ahead of Google Cloud.
The market's answer was immediate. Microsoft shares jumped roughly 7% in after-hours trading, a sharp break from the mood that has gripped the rest of the AI infrastructure trade over the past two sessions, where chip stocks have shed more than $1 trillion in value on fears that capital spending is outrunning demand. Microsoft essentially took the opposite side of that trade tonight: it guided FY2027 capital expenditures to $255-260 billion, up from roughly $190 billion in FY2026 -- a jump of 35-37% -- and investors rewarded rather than punished the number, at least in the first hour of trading.
Context matters here. Microsoft is the second of the four hyperscalers to report this earnings season, following Alphabet's July 22 print, where Google's own capex raise triggered a 7% stock drop on negative free cash flow. Meta followed hours before Microsoft with mixed results of its own: revenue beat, EPS missed, and its own capex guide ticked up to $130-145 billion for the year. The three prints together confirm that Big Tech's answer to the market's spending-discipline question is uniform -- more, not less -- but investors are pricing each company's answer differently depending on whether the revenue growth funding it shows up in the same quarter.
“Meta followed hours before Microsoft with mixed results of its own: revenue beat, EPS missed, and its own capex guide ticked up to $130-145 billion for the year.”
Microsoft's case for why investors bought its version: Azure's growth rate actually accelerated, and the $100 billion cloud milestone is revenue already booked, not a promise. Compare that to the chip sector's problem this week -- a selloff rooted in the idea that spending is running ahead of proven demand. Microsoft's report is effectively daylight between infrastructure spend and infrastructure revenue closing, not widening, which is exactly the reassurance chip investors have been missing.
For VCs and founders building on Azure, OpenAI's Azure exclusivity (even as it diversifies elsewhere), and Microsoft Copilot integrations, the read is straightforward: the hyperscaler funding the AI buildout just proved its own AI revenue is real and growing faster than its cloud business overall. That's a bullish signal for enterprise AI infrastructure spend continuing into 2027, and a green light for startups selling into Azure's ecosystem. It's a more cautious signal for anyone still betting the infrastructure buildout stalls before 2028 -- Microsoft just told the market it's accelerating, not pausing.
The bear case is the same one dogging every hyperscaler this earnings season: a $255-260 billion capex number for a single company is now larger than the entire venture industry's annual deployment, and Microsoft is one of four companies spending at this scale simultaneously. If AI-driven revenue growth decelerates even slightly in 2027, the depreciation math on that spending gets ugly fast -- Azure's operating margins are already absorbing heavier compute costs than any prior cloud cycle.
What to watch next: Amazon reports Thursday, July 30, the last of the big four this cycle, and will face the same question -- can revenue growth keep pace with a capex number now well north of $200 billion. If AWS also beats and reassures, expect the chip-sector selloff to stabilize on the theory that hyperscaler demand remains real even if near-term chip stock sentiment is jumpy. If Amazon stumbles, Microsoft's 7% pop tonight will look like an outlier rather than the start of a trend.
