Google just crossed a milestone that cements its position in the AI arms race. The company's Q4 2025 earnings call revealed Alphabet's annual revenue surpassed $400 billion for the first time, driven by explosive 48% growth in Cloud and the rapid adoption of Gemini 3. CEO Sundar Pichai announced the company will pour between $175 billion and $185 billion into CapEx for 2026—a staggering increase that signals Google is doubling down on AI infrastructure as customer demand outpaces even its own projections.
Google isn't just winning the AI race—it's lapping the competition. The company's Q4 2025 earnings call laid bare the scale of its AI transformation, with CEO Sundar Pichai revealing that Alphabet's revenue broke through the $400 billion barrier for the first time. But the real story isn't what happened last quarter. It's what Google's planning to spend next year.
The tech giant announced it will invest between $175 billion and $185 billion in capital expenditures for 2026, according to Pichai's prepared remarks. That's an eye-watering sum that dwarfs most companies' entire market caps, and it signals Google sees no slowdown in enterprise AI demand. "We're seeing our AI investments and infrastructure drive revenue and growth across the board," Pichai told investors.
The spending spree comes as Google Cloud hit escape velocity. Cloud revenues grew 48% year-over-year to reach a $70 billion annual run rate, while backlog—the total value of committed customer contracts—exploded 55% quarter-over-quarter to $240 billion. That backlog figure represents "a wide breadth of customers, driven by demand for AI products," Pichai said, and it's become the clearest indicator yet that enterprise AI spending isn't slowing down.
Google exited 2025 with double the new Cloud customer velocity it had in Q1, and the number of deals exceeding $1 billion "surpassed the previous three years combined." Even more telling: existing customers are outpacing their initial commitments by over 30%, suggesting enterprises are finding more use cases for AI than they originally anticipated.
