Google is positioning Gemini and TPU infrastructure as its wedge to close a decade-long cloud market share deficit against AWS and Azure.
Google Cloud, currently third in global cloud market share behind AWS (~31%) and Azure (~22%), is explicitly betting that AI-native infrastructure — including Gemini models, TPUs, and Vertex AI — will be its primary lever to close the gap. Google's cloud revenue grew 28% YoY to $12.3B in Q1 2025, outpacing both AWS and Azure's growth rates. The strategic thesis is that enterprise AI workloads will reward first-mover model quality and proprietary hardware over incumbent cloud lock-in. Google is framing its historical disadvantage in enterprise sales as surmountable because AI buying decisions are being made by new stakeholders.
Google's growth narrative depends on winning AI-native workloads, which means Vertex AI will get aggressive pricing and capability investment to compete with SageMaker and Azure ML. TPU access and Gemini API pricing are likely to get more competitive as Google fights for developer mindshare. If you're currently locked into AWS Bedrock or Azure OpenAI Service, this competitive pressure benefits you — expect pricing moves in the next two quarters.
Run a cost comparison this week: take your highest-volume LLM API call, benchmark Gemini 1.5 Pro on Vertex AI against your current provider using identical prompts, and calculate cost-per-1M-tokens at your actual usage tier.
Go to console.cloud.google.com, navigate to Vertex AI > Model Garden > Gemini 1.5 Pro
Tags