A year might be enough time for the bubble to start leaking faster or the pop to become outright undeniable. And from what I’ve picked up from Ed Zitron’s reporting, a lot of the bigger data center deals are kind of tenuous to begin with (sketchy financing, no ability to actually meet overly aggressive timelines even with optimal money and no regulation, etc.). So anything helps.

Going on a bit of a tangent… Eliezer (and lots of rationalists) are so certain only international deals driven by their ideology have any hope of really stopping AGI (Eliezer’s Only Law Can Prevent Extinction is mostly about not supporting direct extralegal action against AI, but it also implies this, and of course AI 2040 is fixated on some cartoonishly impossible international deal), it would be funny if the surge data center opposition for purely pragmatic reasons prove to be a major cause of the bubble finally popping.