Big Tech Is Now Planning $725 Billion of AI Spending, and Investors Are Picking Sides

The four biggest US cloud companies reported first-quarter results at the end of April, and the headline was not profits but plans. Together Alphabet, Amazon, Microsoft and Meta are on track to spend roughly $725 billion on capital expenditure in 2026, according to Financial Times figures cited by Tom's Hardware, a 77% jump from the record $410 billion of last year. Investors rewarded some of them and punished others.

Four budgets, one direction

Microsoft set calendar-2026 capex at $190 billion, well above the $152 billion average analyst estimate. Its finance chief attributed about $25 billion of the total to rising prices for memory chips and other components, a reminder that part of the bill is inflation rather than new capacity. Alphabet raised its full-year range to $180 billion to $190 billion and said spending would increase significantly again in 2027. Meta lifted its range by $10 billion at both ends to $125 billion to $145 billion, also citing higher component pricing. Amazon held to its roughly $200 billion plan, which management described as largely unchanged.

The market splits the field

The reaction depended less on the size of the check than on the evidence of payoff. Alphabet jumped about 7% as Google Cloud grew 63%. AWS accelerated to its fastest pace in 15 quarters, at 28% growth, yet Amazon shares slipped about 1% because the market had expected more. Microsoft fell roughly 3%, and Meta was the clear loser: its stock dropped about 6% in one account and closer to 10% in another, with investors unconvinced that its spending is tied to revenue the way a cloud provider's is.

That is a key distinction. Microsoft, Amazon and Google rent capacity to others, so spending flows into a backlog of contracts. Meta is building for its own products and ads, where returns are indirect and harder to model. An analyst quoted in coverage called the bear case on this spending "garbage," but the Meta move shows investors are not applying that confidence to every name.

Why the number keeps rising

Three forces explain the upward drift. First, demand: all three clouds say they are capacity-constrained, so they build as fast as chips, land and power allow. Second, price: memory and component costs are climbing, so the same number of servers costs more than planned. Third, competition: no executive wants to be the one who underbuilt while a rival locked in the customers.

The result is a collective bet of unusual scale. At $725 billion, the four companies would be spending more in a single year than the annual economic output of many mid-sized countries. For suppliers such as Nvidia, memory makers, networking firms and power-equipment companies, that is a revenue forecast. For skeptics, it is the sum that has to be earned back.

What it means for the bubble debate

The optimistic reading is that revenue is following spending: Google Cloud at 63% growth and AWS at 28% look like real demand. The cautious reading is that capex is growing far faster than the revenue it supports, and that part of the 2026 increase buys higher prices rather than more computing power. Both can be true at once. The test comes when depreciation from this wave of buildout starts to hit income statements, squeezing margins unless revenue keeps pace.

The supply-chain angle

Every dollar of capex is someone's revenue. Chip designers, memory makers, networking companies, server assemblers and power and cooling specialists all price in the hyperscalers' plans. That is why the capex guidance, not the earnings per share, is the number that moves the broader market on these days. A trim from any of the four would ripple through dozens of stocks, while a raise buys the whole supplier base another quarter of confidence.

What to watch

• Whether guidance rises again in July, or the first signs of restraint appear.
• Depreciation growth and operating margins at each company as new capacity comes online.
• Memory and component price trends, which determine how much of capex is price versus volume.
• Meta's ability to link its spending to measurable ad or product revenue.

Sources

1. Big Tech's AI spending plans reach $725 billion — Tom's Hardware
2. Alphabet and Amazon outpace Meta in AI as Big Tech's Q1 2026 earnings confirm cloud dominance — The Next Web
3. Big Tech's AI spending problem — Tech Brew
4. Big Tech Q1 2026 Earnings Power $700B AI Capex Spree — HeyGoTrade