Google Cloud crossed $20 billion in quarterly revenue for the first time. That is a 63 percent increase from the same period last year. But Alphabet CEO Sundar Pichai told investors something surprising: revenue would have been even higher if Google could actually meet the demand. The company is compute-constrained — it simply does not have enough capacity to serve all the customers who want to buy its AI infrastructure.
The Numbers
Google Cloud generated over $20 billion in Q1 2026 revenue. The Google Cloud Platform — which includes infrastructure, data analytics, and AI tools — grew faster than the division's overall rate. AI solutions were the largest driver, with products built on generative AI models growing nearly 800 percent year-over-year.
Gemini Enterprise grew 40 percent quarter-over-quarter. AI token usage via Google's API jumped to 16 billion tokens per minute, up from 10 billion in Q4. New customer acquisition doubled year-over-year. The number of deals worth $100 million to $1 billion doubled. And Google signed multiple deals exceeding $1 billion each.
The backlog — committed future revenue from signed contracts — doubled to $462 billion. That figure represents demand Google has already won but cannot yet fulfill.
The Capacity Problem
Pichai was direct about the constraint. Google Cloud's revenue would have been higher if the company could meet demand. The company is working through the bottleneck but expects to clear only 50 percent of the backlog over the next 24 months.
The capacity crunch echoes what is happening across the AI industry. Anthropic has faced complaints about Claude use limits and responded with massive infrastructure deals — $5 billion from Amazon for 5 gigawatts of AWS capacity and up to $40 billion from Google itself for TPU access. OpenAI's Stargate project envisions $500 billion in AI infrastructure. And the GPU shortage is now affecting everyone from frontier AI labs to university researchers.
Google is investing heavily to expand capacity. Its new TPU 8t and TPU 8i chips promise 3x faster training and 80 percent better cost efficiency. The company is also selling TPU hardware directly to some customers, not just providing it through cloud services.
Why It Matters for AI
Google Cloud's earnings reveal something important about the AI economy. Demand for AI compute is growing faster than even the world's largest technology companies can build it. A $462 billion backlog at Google Cloud alone — before counting AWS and Azure — suggests the global appetite for AI infrastructure far exceeds current supply.
This supply-demand imbalance explains the extraordinary deals being struck across the industry. It explains why Meta is buying millions of Amazon's custom chips. It explains why companies are exploring space-based solar power to keep data centers running. And it explains why AI chip companies like Nvidia are valued at nearly $5 trillion.
Google's Multi-Front AI Strategy
The earnings call reinforced Google's position as one of the few companies competing across every layer of the AI stack. It builds its own chips. It runs one of the three largest cloud platforms. It develops frontier AI models with Gemini. It sells AI tools to enterprises. And it invests in external AI labs including Anthropic and Thinking Machines.
No other company operates at this breadth. AWS is bigger in cloud but does not build frontier models. Microsoft has Azure and its OpenAI partnership but does not manufacture its own chips at Google's scale. Google's full-stack approach is expensive — but the $20 billion quarter suggests it is paying off.
The Bigger Picture
Google Cloud's capacity constraint is the best problem a cloud provider can have. It means demand outstrips supply. It means customers are locked into long-term contracts. And it means revenue growth will continue as capacity expands.
For the broader AI industry, the $462 billion backlog is a signal that the infrastructure buildout is still in early innings. Companies are committing hundreds of billions to AI compute they will not receive for years. The question is no longer whether there is enough demand. It is whether anyone can build fast enough to meet it.
