Amazon’s stock cratered 11.5% in after-hours trading on February 5, 2026, moments after the company announced a $200 billion capital expenditure plan for the year. The crash wasn’t because AWS is failing—it’s because AWS just posted $35.6 billion in Q4 2025 revenue, up 24% year-over-year, the fastest growth rate in 11 quarters. Wall Street punished Amazon anyway. That’s the new math: you can dominate cloud computing or satisfy quarterly earnings expectations, but not both.
The $200B isn’t optional spending. It’s what staying ahead costs when four tech giants will collectively spend $650 billion on AI infrastructure this year. Amazon’s commitment represents a more than 50% increase from 2025 levels, earmarked primarily for AWS data centers, chips, and satellite networks. And investors immediately decided that’s too much.
AWS is growing faster than ever—and that’s exactly the problem
AWS revenue hit an annualized run rate of $142 billion in Q4 2025. That 24% growth outpaced every quarter since early 2023, when the cloud market was still riding post-pandemic enterprise migration waves. CEO Andy Jassy told investors the company is “adding more incremental revenue and capacity than others”—a direct shot at Microsoft Azure and Google Cloud.
But here’s the tension: AWS can’t grow 24% without massive infrastructure investment, and that infrastructure takes years to generate returns. The infrastructure spending patterns across cloud providers reveal a market where capital deployment has become the primary competitive weapon. Amazon consumes much of its own infrastructure—AWS powers Amazon’s e-commerce logistics, Prime Video, Alexa—which theoretically shortens ROI timelines compared to Google or Microsoft. In practice, it means every dollar of capex hits profitability harder.
The market’s reaction was brutal and immediate. Shares plunged because traders can’t reconcile record AWS growth with an infrastructure bill that won’t pay off until 2028 or later. This is the cloud war’s structural absurdity: success requires spending more than shareholders will tolerate.
Google took market share while AWS spent billions defending its lead
AWS still commands the largest slice of global cloud infrastructure, but competitive pressure from Google isn’t just about market share—it’s about who controls the next decade of enterprise AI. Google Cloud posted 48% growth in the same quarter AWS grew 24%. That gap matters.
The $200B isn’t conquest spending. It’s defensive. AWS faces customer demand it can’t currently meet—enterprises migrating AI workloads, startups like Anthropic and OpenAI requiring massive compute clusters, Fortune 500 companies finally moving legacy systems to the cloud. Amazon’s choice is binary: build the capacity now and accept margin compression, or lose deals to competitors who already built it.
And Google isn’t slowing down. Neither is Microsoft. The cloud war has shifted from software differentiation to raw infrastructure capacity. Whoever blinks first—whoever decides quarterly profits matter more than 2030 market position—loses.
The margin squeeze nobody wants to talk about
AWS operating income rose to $12.5 billion in Q4 2025, up from $10.6 billion the prior year. Sounds healthy. But income growth lagged revenue growth, signaling margin compression—the inevitable result of pouring billions into data centers that won’t reach full utilization for 18-24 months.
Jassy framed the spending as existential: “It’s very different having 24% year-over-year growth on $142 billion annualized run rate than to have a higher percentage growth on a meaningfully smaller base.” Translation: AWS is so large that maintaining growth requires infrastructure investments no other cloud provider can match. That’s both AWS’s competitive moat and its profitability problem.
The honest trade-off: AWS can maintain leadership or maintain margins, but not both simultaneously. Amazon chose leadership. Investors chose to sell.
Analysts project AWS revenue could hit $600 billion by 2036, driven by AI workloads that don’t exist yet. That’s the bet—that today’s $200B capex unlocks a decade of dominance in a market still being invented. Against that: immediate margin pressure, investor skepticism, and competitors spending just as aggressively with fewer internal consumption demands.
Two opposing forces, colliding at full speed. The cloud war’s new rule: spend everything, or lose everything.
