The real economy without AI...

I have been pondering this, but also the reverse...AI without the real economy. Or at least as separate from the real economy.

Imagine you did the following
* List all of the firms that are pretty much "pure AI". Look up each one's trailing-four-quarter revenue, expenses, pretax profits (or operating profit or gross margin), and capex rate.
* List the big players that are "part AI", and find the numbers for their purely AI related division(s). e.g., the part of Amazon that is AI hosting within AWS, etc.
* Add 'em up.

Then, look at this agglomeration as if it were a single company: Cancel out all revenues and expenses that are flowing from subset to another, so you're looking only at the numbers crossing the "border" around this entire sum.
(You'd probably have to do this as trailing-four-quarters for each recent quarter end, to see how things are changing). At that point, the revenues are purely revenues from users of AI services, not other AI players.

Then: how does the future of that company look? Eliminating the circularity and seeing that the capex rate is around two orders of magnitude higher than revenues, I think the issues would be much more evident.

The issue here is that you can think of this "All AI" company as a newcomer to the world. If its profits are to support its capex and ultimately market cap, then at some point it has to have commensurate revenues and ultimately profits. How big is that number in 5-10 years? This composite company's revenue is someone else's expenditure. Where the heck are those going to come from in the real economy?

Hmm, maybe my Fable 5 subscription could do this for me...

Jim