Amazon Web Services CEO Reveals How He's Seeing AI Used | WSJ

TL;DR

AWS CEO Matt Garman describes how AI drives 3-10x productivity gains internally and why he rejects predictions of depression-level job losses.

Key Points

• 1.AWS sees AI parallels to the early cloud transition. Like convincing companies to leave data centers 20 years ago, Garman says AI adoption requires the same cultural shift, but is already delivering 3x–10x productivity improvements in Amazon's software teams.
• 2.Amazon deploys AI to every employee via Amazon Q. The internal tool lets all staff build agentic workflows and query data; Garman himself used it to prepare for this interview, though he warns against sending unreviewed AI-generated emails.
• 3.OpenAI models are now available on AWS Bedrock alongside Anthropic. Despite Anthropic being a major AWS investor and partner, Garman says AWS supports competing providers because customers want access to the best models, reflecting a multi-model strategy.
• 4.Project Reineer is a massive Indiana data center running Amazon's Trainium AI chips, built with Anthropic. Trainium powers AI training and inference; Amazon's custom chip strategy — including Graviton, which delivers 40% better performance at 20% lower cost — underpins its $200B 2025 capex plan.
• 5.Garman rejects 'Great Depression'-level AI job loss predictions. He argues fewer people are needed for the same task but that AI enables more work overall, and that software developers who understand systems thinking will remain in high demand.