“Folks were pricing in something like 70-plus gigawatts for next year outside of China of new AI data center capacity built. Now it looks more like 40 to 50 gigawatts. … And at a deeper level, when you get down to it, the revenues from the AI ecosystem just aren't big enough at this point. And they're not growing fast enough at this point to justify the amount of capex that's being contemplated for the next two years.”

…there, folks were pricing in something like 70 plus gigawatts for next year outside of China of new AI data center capacity built. Now it looks more like 40 to 50 gigawatts of data center capacity outside of China. And so you bring down that estimate. That's 40 to 70, that's like a 40% decrease on the low end of the new estimate. And that is a huge haircut to all the numbers in the AI infotrade. And at a deeper level, when you get down to it, the revenues from the AI ecosystem just aren't big enough at this point. And they're not growing fast enough at this point to justify the amount of capex that's being contemplated for the next two years. So that does two things. One, it brings down the overall capex numbers and estimates, and midterms, which we'll talk about, is also a part of this. But it also makes people doubt the ROI on all the capex that's happening. And so then for the players that are in maybe a more tenuous position relative to Microsoft and Meta and Google, who have billions and billions of dollars of free cash flow and…