Alphabet, Amazon, Meta, Microsoft spend $400B

Data centers are the new product launch. That’s basically what changed. The biggest tech companies aren’t just shipping AI apps anymore — they’re pouring staggering sums into the physical layer underneath them, and the latest earnings cycle made that impossible to ignore. Alphabet, Amazon, Meta, and Microsoft are all telling investors the same thing in slightly different words: demand is here, capacity is tight, and the spending is going up anyway. (abc.xyz) ### Why is capex the story now? Because the numbers stopped looking like normal tech-company spending and started looking like industrial buildouts. Alphabet said first-quarter 2026 capex was $35.7 billion and raised full-year guidance to $180 billion to $190 billion. Amazon said it expects roughly $200 billion of capit(abc.xyz)ex target, but its annual report says it will keep investing in datacenters and AI infrastructure, and recent results keep pointing to the same buildout. (abc.xyz) ### What are they actually buying? Mostly compute, power, networking, and buildings. That means GPU clusters, custom chips, servers, fiber, cooling systems, and the giant facilities needed to run them. Meta has broken ground on ten AI-optimized data centers in the last twenty-four months. It’s also expanding its MTIA cu(abc.xyz) spend” in the old sense — it’s concrete, steel, transformers, and semiconductors. (about.fb.com) ### Why are they spending so fast? Because each company says demand is outrunning supply. Alphabet tied the spending to AI opportunities across Search, Cloud, and Gemini, while noting Google Cloud revenue topped $20 billion and backlog jumped above $460 billion. Amazon is making the same case through AWS — the idea is that if customers wa(about.fb.com)wed Azure above $75 billion in annual revenue, with margins pressured by AI infrastructure and usage growth. (abc.xyz) ### Where is the payoff showing up first? Not evenly. Meta probably has the clearest near-term “show me the money” story because it can point to AI improving engagement and ads inside an already massive business. It said feed and video ranking improvements lifted organic views, Threads time spent, and ad performance, wi(abc.xyz)o revenue. The cloud players have the same ambition, but their monetization path depends more on enterprise uptake and capacity delivery. (about.fb.com) ### So why are investors still uneasy? Because a lot of this spend lands before the returns do. The companies can show growth, but they still have to prove that today’s infrastructure binge won’t turn into tomorrow’s oversupply, margin pressure, or stranded hardware. Microsoft already flagged gross-margin pressure from AI infrastructure and product usage(about.fb.com)you the market still wants discipline, not just ambition. (microsoft.com) ### Is this really an AI race, or just cloud again? It’s both. The cloud buildout taught these companies that owning scarce infrastructure can lock in years of advantage. But AI adds a twist — the spending is tied not just to storage and compute demand, but to training frontier models, running inference at scale, and embedding AI into search, (microsoft.com)zier. (microsoft.com) ### What should people watch next? Two things. First, whether revenue tied directly to AI keeps rising fast enough to justify the capex ramp. Second, whether management teams get more precise about returns — not just “demand is strong,” but which products are monetizing, how quickly, and with what margins. The spending story is no longer about belief. It’s about proof. (abc.xyz) ### Bottom line? Big Tech’s AI boom has moved out of the demo phase and into the construction phase. The next test is simple — can four companies spending like utilities still earn like software?