Everyone talks about AI models, but the real gains may be elsewhere

I came across a post on Blossom social about a talk from Jensen Huang where he described AI as a “five-layer cake.”

Kew takeaway: AI isn’t just about models and chips ($NVDA, $GOOGL, $MSFT, $AMZN). To train and run large-scale AI, you also need:

\- Infrastructure (data centers, networking, cooling)

\- Energy (significant power to run the systems)

This means that the economic upside isn’t limited to the big tech names. Mid and smaller cap companies in infrastructure and energy could see significant growth if AI adoption continues.

AI is more then digital, it’s physical, capital-intensive, and energy-dependant.

Do you think the real AI winners are the mega-caps everyone knows, or the broader stack of infrastructure and energy companies?