Last week’s earnings calls confirmed what many suspected: AI demand is reshaping cloud infrastructure economics. AWS, Azure, and Google Cloud all beat analyst estimates, but Google Cloud’s 63% revenue jump to $20 billion stood out—the highest growth since the segment was broken out in 2020. Synergy Research pegged total cloud infrastructure spend at $129 billion for the quarter, with analyst John Dinsdale noting, “That was some quarter.”

Google’s growth is driven by enterprise AI solutions, which CEO Sundar Pichai said became the primary cloud growth driver for the first time. Revenue from Gemini-powered products grew 800% year-over-year. Notably, Google’s custom TPUs are gaining traction as an alternative to Nvidia GPUs, reducing dependency on a single hardware vendor—a key architectural consideration for teams designing cost-efficient training pipelines.

AWS grew 28% to $37.6 billion, topping StreetAccount estimates by nearly $1 billion. Usage of Bedrock, their managed AI agent platform, surged 170% quarter-over-quarter. AWS also announced OpenAI models will land on Bedrock, a strategic move to reduce vendor lock-in for customers building multi-model architectures. CEO Andy Jassy reported “unprecedented demand” for the new agent-building service.

Azure posted 40% growth, slightly above expectations. CEO Satya Nadella noted the number of customers using Anthropic and OpenAI models through Azure doubled sequentially. Microsoft guided Q2 Azure growth at 39%, above consensus.

All three are investing heavily: combined CapEx guidance for this year approaches $600 billion. Smaller neocloud providers like CoreWeave now hold 5% of the market, offering specialized GPU clusters that challenge the big three on cost and latency for specific workloads. For engineering teams, the takeaway is clear: AI is driving architectural decisions from hardware selection to multi-cloud agent orchestration, and the infrastructure layer is evolving faster than ever.

Source: CNBC