Argues there's a big misunderstanding about Google's market position. While OpenAI leads in current usage, Google is catching up quickly due to its efficient TPU compute architecture, ability to resell to Google Cloud enterprise customers, leading models in imaging/video, and its full-stack applications (Search, YouTube, Gmail, Maps) and advertising network.

Reveals they bought a lot starting in late spring (middle of the year) and have been adding recently on pullbacks. He finds Google compelling at 10x revenue and 29x forward earnings with 35% earnings growth, citing incredible capital allocation and upside from Waymo.

States Amazon has been slow/quiet on AI (partnering with Anthropic) but is starting to emerge as a winner, especially in advertising where AI is a documented growth driver. He likes their positioning, including AWS driven by enterprise AI traffic over the next five years.

Describes a huge period where AI compute demand outstrips supply, forcing hyperscalers (Mag 7) to fund massive capex—projected at ~73% of operating cash flow by 2026—a big shift from high free cash flow just years ago. The market also must digest OpenAI trying to borrow hundreds of billions from partners like Oracle and SoftBank.