Google Cloud crossed $20 billion in revenue for the first time in Q1 2026, up 63% from the same quarter a year earlier, and Alphabet CEO Sundar Pichai told analysts on Wednesday's earnings call that the number would have been higher if the company could deliver enough compute to meet demand. The unit's backlog doubled in the quarter to $462 billion. AI products were the largest single driver of growth.
Products built on Google's generative AI models grew nearly 800% year-over-year, while Gemini Enterprise grew 40% quarter-over-quarter. API token throughput climbed to 16 billion tokens per minute, up from 10 billion in Q4 2025. New customer acquisition doubled year-over-year, and the count of deals worth between $100 million and $1 billion also doubled, with Google signing multiple billion-dollar-plus contracts in the quarter.
Existing customers are spending faster than they planned. Pichai said cloud customers outpaced their initial commitments by 45% quarter-over-quarter, a figure that helps explain why the backlog has run so far ahead of revenue recognition. Google Cloud expects to work through 50% of that $462 billion backlog over the next 24 months, leaving a long tail of contracted revenue stretching well beyond.
Key facts
- 01Google Cloud revenue passed $20B in Q1 2026, a 63% jump from Q1 2025.
- 02Backlog doubled in the quarter to $462B, with 50% expected to clear over 24 months.
- 03Products built on Google's genAI models grew nearly 800% year-over-year.
- 04API token throughput hit 16B tokens per minute, up from 10B in Q4 2025.
- 05Gemini Enterprise grew 40% quarter-over-quarter; customers outpaced initial commitments by 45% QoQ.
The capacity story is the one investors fixated on. "Obviously, we are compute constrained in the in the near-term," Pichai said on the call. "And as an example, our cloud revenue would have been higher if we were able to meet that demand. So we are working through that moment, and we are investing, but we have a robust, long-range planning framework…we see extraordinary opportunities ahead."
“Google Cloud's backlog doubled in a single quarter to $462 billion, and Pichai said the company expects to work through only half of it over the next 24 months.”— Jaeden Schafer
Pichai pointed to demand for both TPU hardware and data center capacity as the bottleneck. Google sells cloud infrastructure through Google Cloud Platform and, with some customers, sells TPU hardware directly. He framed the constraint as a planning problem rather than a structural one, noting Google evaluates spending against return on invested capital before pushing further into new buildouts.
The mix shift inside Cloud is worth watching. Google Cloud Platform grew faster than the Cloud division as a whole, which also includes Google Workspace and a range of data analytics and machine-learning tools. Pichai attributed the GCP outperformance to demand for Gemini Enterprise and the company's broader AI stack, not to legacy workloads migrating from on-premises systems.
The Q1 print continues a pattern across the AI hyperscalers, where revenue is large and growing fast but trails what customers say they want to buy. We covered Microsoft's disclosure earlier this week that Copilot has crossed 20 million paid seats, and Meta's $4 billion quarterly Reality Labs loss alongside AI capex tracking past $125 billion. Google's $462 billion backlog now sits as the most concrete data point for unmet enterprise AI demand on any public balance sheet.
Skeptics will note that backlog math is not revenue math. Multi-year contracts with consumption-based pricing can be renegotiated, deferred, or stretched, and a backlog clearing only 50% over 24 months means the other half lands at an indeterminate future date. If model costs keep falling and customers re-architect workloads onto cheaper inference, the dollar value of that backlog could compress before it is recognized. Pichai did not break out how much of the $462 billion is locked-in minimum commitment versus expected consumption.
There is also the question of whether Google can actually build its way out of the constraint. TPU fabrication, power contracts, and data center construction operate on multi-year timelines that do not flex to a single quarter's demand spike. The company is investing aggressively, but so is every other hyperscaler chasing the same grid capacity, the same chip supply, and increasingly the same nuclear and gas baseload deals.
Google Cloud's Q1 makes the bull case for the AI infrastructure trade about as cleanly as any single data point this year. A division growing 63% at $20 billion in quarterly revenue, with a backlog larger than the GDP of most countries and customers consistently spending past their original commitments, is not a business with a demand problem. The risk has migrated to the supply side, and to whether Alphabet's capex discipline can stay coherent while every competitor races to pour concrete. For now, the constraint is a luxury problem, but it is still a problem, and the next four quarters will show whether Google can convert that backlog into recognized revenue or watch some of it slip to rivals with spare capacity.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
