Google's cloud AI budget may rise as capacity runs out
Google is pushing its AI spending higher because cloud demand already exceeds the computing capacity it can deliver. Alphabet raised full-year capital expenditure to $180–$190 billion after Google Cloud’s $462 billion backlog nearly doubled in one quarter, and CEO Sundar Pichai said revenue would have been higher with more capacity. That backlog is more than ten times Google Cloud’s 2025 revenue of $43.2 billion, according to 24/7 Wall St.
Key Takeaways
- Alphabet lifted 2026 capex guidance to $180–$190 billion after Q1 spending hit $35.7 billion.
- Google Cloud’s $462 billion backlog nearly doubled in one quarter; over half may convert within 24 months.
- Pichai said capacity constraints kept Google Cloud revenue below its potential.
- Internal AI coding mandates and a reported Gemini 3.5 Pro delay add more pressure on shared GPUs.
Why is Google raising its AI capital budget again?
Investors once worried Alphabet was spending too aggressively on AI infrastructure. In Q4 2025, Google guided to $175–$185 billion in capital expenditures for the year—nearly double 2025 spending.
One quarter later, the company booked $35.7 billion in capex in Q1 alone and raised the full-year outlook to $180–$190 billion. Management’s explanation was straightforward: demand outran supply, and waiting enterprise customers already lined up for cloud AI services.
For more coverage of infrastructure and AI arms races, see BlasterPost’s Future Tech & AI Wonders section.
How large is the Google Cloud backlog—and why does it matter?
Google Cloud’s order book now stands at $462 billion and nearly doubled in a single quarter. Management expects more than 50% of that backlog to turn into revenue within 24 months.
By comparison, Google Cloud generated $43.2 billion in revenue in 2025, so the backlog is more than ten times that annual base. The company also said billion-dollar-plus cloud deals signed in 2025 exceeded the combined total from the prior three years.
Reporting ahead of earnings has put that same $462 billion figure at the center of Alphabet’s growth story, even as near-term model and regulatory headlines rattled the stock.
What else is straining Google’s AI capacity?
Capacity pressure is not only external. Bloomberg reported Google delayed Gemini 3.5 Pro after engineers struggled to hit internal performance goals, with a late-June retraining run failing to meet benchmarks.
Google also required engineers to use AI tools to help generate code—an initiative that boosts productivity but consumes the same GPUs sold to enterprise cloud customers. In effect, Google is competing with itself for scarce compute.
The upshot for investors and enterprise buyers is a capacity story, not a demand drought: customers are waiting, internal AI use is rising, and Alphabet’s answer so far is to spend more to catch up.