AI revenue reaches $25B, justifying data-center spending boom

Global AI sales excluding China reached $25 billion in the first quarter of 2026, exceeding the industry's estimated $21 billion in depreciation costs tied to data-center and chip investments for the second consecutive quarter, according to a report from research firm Exponential View. The milestone suggests the hundreds of billions of dollars that tech companies are pouring into artificial intelligence infrastructure may be economically sustainable — though margins remain thin.

"For now, the economics are holding. But the margin for error is narrow," the report states. Depreciation charges still consume more than two-thirds of revenue, leaving a small buffer to cover other costs such as power, labor and financing. Azeem Azhar, founder of Exponential View and an investor in dozens of startups, said the data shows the industry is "just about" clearing the depreciation hurdle and "roughly speaking, it's improving over time."

The findings address one of the central questions hanging over the AI boom: whether customer demand is large enough to justify the unprecedented capital spending. The biggest US technology companies — Meta Platforms Inc., Alphabet Inc., Microsoft Corp. and Amazon.com Inc. — plan to spend as much as $725 billion this year on capital expenditures, much of it on AI infrastructure, in one of history's largest corporate spending sprees. Much of the AI boom has been measured from the supply side through disclosures from Nvidia Corp. and the hyperscalers, while demand has been harder to quantify because many of the most important AI labs, including OpenAI and Anthropic, remain private.

The Funding Machine Behind the Buildout

The capital flowing into AI extends well beyond the hyperscalers. More than 63 so-called neolabs — startups that raise billions of dollars out of the gate to focus on frontier research rather than products — are collectively valued at more than $300 billion and have raised about $48 billion, according to data compiled by Menlo Ventures partner Deedy Das. That accounts for 16% of the roughly $283 billion invested in startups other than OpenAI or Anthropic over the past year.

Many of these rounds are structured in tranches, allowing lead investors to buy in at a lower valuation before a second, higher-priced tranche is marketed to other firms. Ineffable Intelligence, the reinforcement-learning startup founded by former Google DeepMind scientist David Silver, raised $1.1 billion in what was touted as Europe's largest seed round. In reality, the company raised $11 million from Sequoia at about a $55 million pre-money valuation, then an additional $1.1 billion at a $4 billion pre-money valuation weeks later — a markup of more than 70 times for the same company.

"In a market where fundraising runs on vibes, a billion-dollar headline is worth a lot more than an accurate one," said Jaya Gupta, a partner at Foundation Capital. The practice has become common enough that Mercor CEO Brendan Foody called it the "Sequoia Scam" in a social media post, though he later acknowledged it is "common practice in the industry across all top firms."

Thin Margins, Narrow Room for Error

The Exponential View report flags that financing risk is shifting into capital markets through leases, debt and equity, particularly among the neoclouds — smaller cloud providers that lease Nvidia GPUs to AI startups. If demand softens or interest rates rise, these companies face the most pressure.

For investors, the data provides the first fundamental validation that AI spending may be self-sustaining. Nvidia, whose data-center revenue has grown more than fivefold since the start of the AI boom, trades at about 30 times forward earnings. The risk of a "capex bubble burst" narrative — where hyperscaler spending collapses because AI applications fail to generate sufficient returns — has receded with two consecutive quarters of revenue covering depreciation. But the narrow margin means any slowdown in AI adoption or a shift in enterprise buying patterns could quickly reverse the math.

"At this stage of an investment in any kind of capital expenditure, you wouldn't expect to have dramatically jumped over that hurdle because if you had, you were probably leaving something on the table," Azhar said.

This article is for informational purposes only and does not constitute investment advice.