If one would assume it's nearly all a bubble, How would you correct earnings for the US? I am interested in applying it to any investment that tracks AI heavy companies in the US.
One approach I've seen a few folks do is to fit a regression model of annualized real stock market returns over the next 10 years as some function of CAPE or 1/CAPE or log(CAPE).
It doesn't give a very good fit on training data, R^2 in the range of 0.2-0.3, i.e. it cannot "explain" most of the variation in 10 year returns.
CAPE based regression models like that have said the US stock market has been overpriced for the last decade! But investors in the US stock market have done pretty well over that period, with really good returns. Maybe these models are accurate but we've just gotten lucky? Maybe these models aren't very good. Hard to tell.
Elm capital publish estimates of expected returns of a few asset classes quarterly: https://elmwealth.com/capital-market-assumptions/
> This year the world’s four largest tech firms will spend $344 billion on AI
> Altogether, the four companies are expected to spend more than $344 billion for the year, with much of it going to the data centers necessarily to run AI models.
so both articles frame that $344B as estimates of capex within 1 year.
Apparently the total market capitalisation of the US stock market is $62.8 trillion. Shiller's CAPE ratio for the S & P index is currently about 38 -- CAPE is defined as current price / (earnings, averaged over the trailing 10 years)
That suggests that over the last 10 years, the average earnings of the US stock market is about $1.7 trillion annually.
So $344B of spending is about 1/5 of the average earnings of the total US stock market.
Still hard to interpret that, but 1/5 is an easier number to think about.