Big Tech Is Spending $700 Billion on AI. Here’s How the Math Works

Four companies are on track to spend about $700 billion in 2026 if you add up their own guidance. That’s a hard number to picture, so here’s a comparison: in just the first quarter, their combined capital spending hit $129.8 billion, up 80% from a year earlier. Silicon AnalystsSilicon Analysts

Amazon is pointing to roughly $200 billion, and Meta has nudged its range upward, citing pricier memory chips and higher data center costs. Most of it goes to GPU clusters, custom silicon and data center construction. ValueaddvcValueaddvc

You’ll see slightly different totals depending on the source. That’s mostly bookkeeping. One analysis notes that definitions vary, including whether finance leases count and how Microsoft’s June fiscal year maps to calendar quarters. Silicon Analysts

So is it a bubble? Two facts pull in opposite directions. UBS estimates Amazon, Alphabet and Microsoft will spend about 102% of their cloud revenue on capex this year, though that doesn’t mean they’re burning more cash than they generate. Demand looks real too, with Google Cloud’s contract backlog around $460 billion, roughly double the prior year. Yahoo FinanceValueaddvc

The catch is timing. Right now, spending is running at nearly three times the pace income statements recognize, because depreciation is spread over years. To keep building, the companies are borrowing. Meta priced a $30 billion bond, Alphabet raised roughly $25 billion and then about $31 billion more, and Amazon added $24.9 billion. Goldman forecasts about $1.01 trillion of capex in 2027. Hyperscaler AI Capex 2026: $434B and the D&A Lag +2

The thing to watch is whether AI revenue catches up before the depreciation bill lands, because that wave arrives regardless of what AI revenue does. Silicon Analysts

Sources: Silicon Analysts (siliconanalysts.com), I/O Fund (io-fund.com), Yahoo Finance/UBS (finance.yahoo.com), ValueAdd VC (valueaddvc.com)