Jose Najarro

And so the need to kind of replace it very often is very high.

And so I would say that Amazon right now, even though it's a great business, the returns on invested capital might be much lower than they were in the past.

And if you were to agree with me and you don't have to,

people out there were to agree with me, then you would have to reach the conclusion that it's likely overvalued because with low reinvestment returns and potentially negative reinvestment returns, it doesn't matter what happens in the long term, you're going to be toast.

Off profits, not of revenue.

Because some people think about the revenue opportunity, but the revenue is also associated with a lot of costs in the process, right?

To service the data centers, but also the electricity that goes in servicing those data centers.

So it's $220 billion of cash flows, which means for just what they're doing this year, which means that by need revenue of additional revenue of $300 billion or $350 billion or whatever it is.

But sorry to interrupt you, but I 100% agree with you.

And we won't invest in it, so we're not subject to that potentially lower negative reinvestment risk, right?

But Google, again, is also a great example.

And in fact, for most hyperscalers, it's a good example of thinking about return on invested capital.

Because like Google before, the reason why also they were able to compound is because

they needed relatively little capital to produce the earnings and cash flows of what they were doing into new projects or even just enhancing their search business.

For the search business was extremely high margins, extremely low capital intensiveness, but now the tides have shifted and

what they want to do now is super capital intensive.

And I don't want to deliver the point and repeat myself too much, but it makes it at least questionable.

And so something that I like to say to people is you got to be cautious because it's not a question about whether AI is going to

change the world.

I think we all know that in a way that already has.