The AI Buildout and Who Actually Gets Paid

The short version, though the end is where it pays off: the AI buildout may be tracking the railroads of the late 1800s and the fiber-optic buildout of twenty-five years ago. Both technologies were every bit as transformative as advertised. Neither made money for the people who financed the track and the cable. A fifth of American railroad mileage ended up in receivership, and WorldCom and Global Crossing went bankrupt laying fiber. The profits went instead to the businesses that rode on top of the network after somebody else had already paid for it — Pullman and Adams Express then, Google and Netflix later.

I am not endorsing every claim in the piece,  but it is a useful counterweight whenever the conversation turns toward tilting a portfolio into the AI buildout itself rather than owning the broad market and the businesses that stand to benefit from cheaper intelligence everywhere. Being right about a technology and wrong about who captures the profit from it can be an expensive mistake. The dot-com era is full of people who got the first part exactly right.

Read the whole thing. The last few paragraphs are the most important:

Read Porter Stansberry’s full post on X →

‍