Tech giants Amazon, Alphabet, Microsoft and Meta all released first-quarter earnings results after Wednesday market close, projecting spending this year will reach as much as $725 billion, with the vast majority earmarked for AI chips, servers and data centers.

Add in Tesla's revised $25 billion 2026 budget reported last week—up from $20 billion three months ago—reported last week, and the five companies are on track to spend up as much as $750 billion this year, more than double last year’s spend.

Alphabet increased its capital expenditure (capex) guidance to between $180 billion and $190 billion, up $5 billion from its prior $175 billion to $185 billion guidance.

Meta also increased guidance for its spend to $125 to $145 billion, up $10 billion from its prior $115 billion to $135 billion range.

Microsoft, which previously did not guide on this year’s spend, projected $190 billion in spend, up from analyst expectations of around $150 billion and nearly triple last year’s number.

Amazon did not increase guidance for its spend during its first quarter earnings, but its $200 billion guidance is the largest single-year capex commitment from any U.S. company over the past decade and tops the list (the firm said a majority of spend will be going toward AI, but also its satellite business.).

On its April 22 earnings call, Tesla raised its guidance from $20 billion to $25 billion as the company shifts more heavily into AI, including self-driving software, humanoid robots and in-house chip manufacturing.

Alphabet stock is up 7%, while Amazon’s stock is down 2%, Microsoft’s is down 5% and Meta’s has plummeted 10%.

First quarter earnings results from the four AI hyperscalers—the giant tech companies that own the infrastructure most of the AI industry runs on—made clear that good earnings isn't enough. Despite beating Wall Street expectations across metrics including revenue and growth, Amazon, Microsoft and Meta’s stocks are all down, with Meta facing the steepest loss. Alphabet’s results were the only to be embraced by the market despite increasing projections for this year’s spending to as much as $190 billion. Alphabet CEO Sundar Pichai told analysts Wednesday the company is "compute constrained" and that Google's cloud business has more than $460 billion in signed contracts waiting to be filled, justifying the increase in spending. Demand for AI computing power is outrunning what these companies can build: customers are waiting in line to rent space in their data centers, and the specialized chips that power AI—mostly made by Nvidia—are in short supply. In the AI race, none of these companies wants to be the one that runs out of capacity while a competitor lands a major customer, fueling the increase in spending. Still, Wall Street is nervous because no one yet knows whether the revenue from AI will eventually justify the staggering bills.

Though Tesla isn't a hyperscaler, Musk compared Tesla's $25 billion to Amazon's $200 billion and Alphabet's $175 billion to 185 billion guidances during its April 22 earnings call, framing Tesla's spend as part of the same race. That money will cover AI infrastructure and a $3 billion AI chip research facility in partnership with Intel, as well as the start of Tesla's Optimus humanoid robot production this summer. The $5 billion guidance hike came just three months after Tesla's last update and helped erase the stock's post-earnings gains.

“We are seeing unprecedented internal and external demand for AI compute resources,” Alphabet CFO Anat Ashkenazi said on Wednesday's earnings call, adding that around 50% of the company’s capex would go to servers that handle data used in AI models and around 40% would go to data centers and networking equipment. She added that 2027 spending is expected to “increase significantly” to meet demand, as Alphabet indicated the clearest payoff from its AI spend with its Google Cloud revenue at $20 billion in its first quarter, more than $2 billion above analyst estimates and up 63% from the last year.

The AI buildout doesn't just rest on the hyperscalers' own businesses, but leans heavily on customers like OpenAI being able to pay the bills. The Wall Street Journal reported on Monday that OpenAI missed a number of internal revenue and user targets, and that CFO Sarah Friar privately told colleagues the company might not be able to afford its massive spend if revenue doesn't grow fast enough. The market reaction was swift: Oracle, which signed a $300 billion five-year deal to supply OpenAI with compute, dropped 7.7%, CoreWeave fell 7.4%, SoftBank sank nearly 10%, and chip stocks Nvidia, Broadcom, AMD and Arm all slid between 2% and 6%. The hyperscalers themselves were partly insulated from the selloff on the logic that Microsoft, Amazon, Alphabet and Meta have diversified AI customer bases and their own consumer products to fall back on. Still, investors will continue to closely watch whether that gap between AI revenue and AI spending closes with a focus on key customers as these companies continue to dramatically increase spending.