SK Hynix debuting today on the Nasdaq. The first thing that comes to mind for me was Micron’s gross margin expansion as reported in its most recent quarterly earnings: from 75% to 85%. I have a feeling a lot of folks still need to fundamentally adapt their mental model on compute. Most of us still reason CPU style but the future is GPU and HBM and pushing the boundaries on running them hot if you will. Not enough people realize that they’ll have to get used to the sound of the fan working on the max setting around the clock.
Productivity gains are often the result of technological advances that create bearing points. These often lead to the lions share of the returns accruing to a handful of companies that scale the technology. In this chart you also see McAFee’s point of the relatively absent productivity gains of the web 1.0 and 2.0 era.
The technology that were scaled in that era essentially serve as the birthing pains of the current bearing point. AI is going to be a major source of bifurcation and a decent chunk of the web 2.0 era incumbents are gonna have a really hard time making it through the birth canal. We needed the path dependency of 140 characters to show us the potential of compute at scale to ultimately arrive at flying cars. The subtleties around inflation really matter in our current era. The opportunity cost is simply wildly different
