Alphabet plans $175B–$185B annual AI spending, 6x increase in 4 years, as Google Cloud reports 75% of customers now using AI in production

Alphabet CEO Sundar Pichai announced a jump from $31 billion in 2024 CapEx to $175–$185 billion planned for this year — roughly a sixfold increase — to build data centers and chips for AI workloads. Google Cloud CEO Thomas Kurian stated that nearly 75% of Google Cloud customers have moved from testing AI to actually running it in their live business systems.

Google introduced the Gemini Enterprise Agent Platform, which acts as a control center to build, deploy, and monitor AI agents (autonomous AI systems that complete multi-step tasks without human intervention per step). The platform includes a new reasoning model called Gemini 3.1 Pro, plus new image, video, and audio AI models in preview, and now supports competitors' AI models like Anthropic's Claude Opus 4.7 — giving customers choice rather than lock-in.

Early customers show concrete speed gains: Merck is rolling this out to 75,000 employees; Citi Wealth built an AI team member called 'Citi Sky' that speaks multiple languages; and security teams using AI agents reduced threat investigation time from 30 minutes to 60 seconds. For business users, this means routine work like analyzing trends, finding dead inventory, or writing security reports can now happen in minutes instead of hours.

Google Cloud partnered with Apple as Apple's "preferred cloud provider" to build foundation models powering next-generation Apple Intelligence and a more personalized Siri coming later this year — signaling that enterprise AI infrastructure, not just consumer devices, is becoming a competitive battleground.