Alphabet Plans $190B AI Infrastructure Spending After 63% Cloud Sales Jump
Alphabet is focusing its AI infrastructure on custom tensor processing units to optimize costs, performance and cloud economics as AI demand rises. Its cloud division delivered 63% year-over-year sales growth with a $460 billion backlog, while the company forecasts $180 billion to $190 billion in 2026 capital expenditures for AI buildout.
1. TPU-Centric AI Infrastructure
Alphabet is prioritizing its tensor processing units to gain control over hardware costs, boost processing performance and improve cloud economics as enterprise AI workloads increase. This shift toward in-house AI chip development aims to reduce reliance on third-party GPUs and secure long-term cost advantages.
2. Cloud Sales and Backlog Growth
The cloud segment reported 63% year-over-year revenue growth and now carries a purchase order backlog near $460 billion, reflecting strong demand from large enterprises. Turning this backlog into booked revenue will be critical for sustaining growth and demonstrating cloud service profitability.
3. 2026 Capital Expenditure Outlook
Management has guided capital spending of $180 billion to $190 billion in 2026, earmarked primarily for data centers, AI accelerators and TPU deployment. This represents a significant increase in infrastructure investment as Alphabet seeks to scale its AI capabilities globally.
4. Investor Implications and Risks
While aggressive infrastructure spending could underpin future AI revenue growth, it also raises concerns about margin pressure and return on investment. Investors will closely monitor upcoming TPU utilization rates, cloud profitability metrics and AI-driven revenue contributions.