Alphabet reported Q4 2025 revenue of $113.83B, up 18% year-over-year and ahead of the $111.43B analyst consensus. Google Cloud grew 48% to $17.66B (versus $16.18B expected), while net income rose 30% to $34.46B.
The headline for enterprise tech leaders: Alphabet plans to nearly double capital spending in 2026. While the company didn't disclose the exact figure, analysts had expected roughly $75B for 2025, putting the 2026 number potentially north of $140B. That's infrastructure spending on the scale of building a small country's power grid.
This matters because it's a signal of how seriously Google views the AI compute race. When hyperscalers talk about "supply constraints" (as Alphabet did on the earnings call), they mean chips, power, and data center capacity are the bottleneck, not demand. The spending spike suggests Google sees a window to establish AI infrastructure dominance while competitors scramble.
What this means in practice: enterprises relying on Google Cloud should expect continued capacity expansion, but also potentially tighter allocation of high-end GPU instances as Google prioritizes its own AI workloads. The 48% Cloud growth rate indicates strong enterprise adoption, but the infrastructure investment signals Google believes the real revenue inflection is still ahead.
Worth noting: Alphabet executives declined to discuss the Google-Apple AI partnership during the earnings call, even when pressed by investors. That silence is telling. Whatever deal exists around Gemini integration into iOS (as reported by Bloomberg last month), Google isn't ready to detail the commercial terms.
The trade-off for Google: this level of capex commitment limits financial flexibility elsewhere. Alphabet crossed $400B in annual revenue for the first time, but the infrastructure bet means less cash for acquisitions, buybacks, or experimental projects. They're making a binary bet: win the AI infrastructure race or accept second-tier status in the next computing era.
History suggests this playbook works for Google. They made similar all-in bets on search infrastructure in the 2000s and mobile/cloud in the 2010s. The question is whether the AI market will reward infrastructure leaders the way previous cycles did, or if model efficiency improvements will commoditize compute before Google can recoup the investment.