Josh Brown

Low asset firms versus high asset firms.

So they're showing the tech bubble from 96 to 2003.

And what they're showing is the firms with high assets, actually the stock prices fell way more than the companies that kept their CapEx low and were considered low asset firms.

What do you think of this concept of an AI industrial bubble?

And is this the kind of thing that ultimately a lot of people might start thinking, which could shift the psychology and the stocks that are actually working?

If you think the profits of all of this AI activity are going to accrue to the platform, to the platforms themselves, then you have to own the platform.

Which they have.

And if you want to own the platform, that requires millions of machines

being plugged in in data centers all over the world.

And that's what these companies are investing in.

Now, that might turn out not to be true.

It may turn out the platforms are not where the profitability of this accrues.

It may not be about servers and GPUs and electricity.

The profits may ultimately accrue to a software level that we can't yet imagine.

We don't think they're going to accrue at the LLM level completely.

But there is a world in which all of these data centers and all this profitability gets driven down into a commoditized state.

And the software layer becomes the most valuable part of AI.

And if you believe that, you're probably calling people trying to get shares of OpenAI and Anthropic before they come public.

But we don't know for sure.

Is it the platforms?