# Google Cloud customers spending 50% above commitments as AI demand drives record quarter > Source: > Published: 2026-07-23 15:24:58+00:00 # Google Cloud customers spending 50% above commitments as AI demand drives record quarter Alphabet's cloud division posted $24.8 billion in Q2 2026 revenue, crushing estimates by more than $2 billion as enterprise AI adoption accelerates Google Cloud CEO Thomas Kurian dropped a number on July 23 that should make every cloud competitor nervous. Existing customers are spending roughly 50% more than their committed amounts, a figure that reflects just how deeply enterprises are embedding AI into their operations. The spending surge isn’t theoretical. It showed up in Alphabet’s Q2 2026 earnings report, where Google Cloud revenue hit $24.8 billion, an 82% year-over-year jump from $13.6 billion. Analysts had penciled in roughly $22.4 billion, which means Google Cloud beat expectations by more than $2 billion in a single quarter. ## The numbers behind the momentum Google Cloud had already posted 63% year-over-year growth in Q1 2026, when revenue came in at $20 billion. So the growth is actually accelerating, which is the kind of trajectory that makes CFOs giddy and competitors anxious. Perhaps the most striking figure in the entire report is the backlog. Google Cloud’s total backlog reached $514 billion by the end of Q2 2026. In English: that’s the total value of contracted but not yet delivered services sitting on Google’s books. Kurian attributed the overspend trend to differentiated products and strong market execution. Translation: customers signed up for one thing, discovered they needed a lot more AI infrastructure than they originally planned, and kept swiping the corporate card. Under Kurian’s leadership since 2019, the division has transformed from a money pit into a sustained profit generator. By mid-2025, Google Cloud had already surpassed a $50 billion annual revenue run rate. ## Why AI is the engine The growth engine here is enterprise AI infrastructure and Google’s Gemini models. Companies aren’t just renting server space anymore. They’re building AI-powered applications, training custom models, and deploying inference workloads at scale, all of which consume significantly more compute resources than traditional cloud workloads. ## What this means for investors The $514 billion backlog deserves special attention from investors. Backlogs aren’t revenue, they’re promises. But a backlog of that magnitude provides extraordinary visibility into future earnings. **Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).