From a different perspective: the fact US businesses can assemble massive power plants on demand, in weeks, is quite an illustration of the economic dominance of the country.
If you mean the financial power to pay for such power plants, then I agree. But the technical capabilities to quickly provide stationary or mobile power plants can be purchased from a few large, international companies; noteworthy suppliers can be found in the USA, Europe and Asia. And they will build it for you in almost any country in the world, as I said, provided you pay accordingly. Of course, the delivery time of turbines is often a bottleneck, but this applies more to the very large ones than the smaller ones that should actually be installed here (and it also applies equally to all manufacturers).
They're big, by the standard of a domestic source such as a car, but they're not what I'd call "massive" in an economic dominance sense. About the size of a large shipping container, give or take, eyeballing from other photos that tell me which objects the turbines even are: https://www.capacitymedia.com/article/musks-xai-data-centre-...
It's a large power plant (400+ MW and rising), rapidly assembled from small, modular components. That's a remarkable thing from any point of view.
The other threads on this topic (powering AI hyperscalers) are usually about nuclear fission, which most of the large players are investing in. Those power plants often take 10+ years to build. One could imagine small reactor modules akin to these 16 MW shipping containers, built in factories and shipped on demand like these, to be assembled into a full power plant in weeks. If someone were to get that business model working, they'd dominate the industry. (Just how big of a premium did xAI pay their vendor, to have all this shipped halfway around the world on a priority schedule?)
There’s a reason this isn’t generally done for real power plants; these little turbines are pretty inefficient. This sort of thing is not uncommon for data centers and other things that need either a lot of backup power or a lot of temporary power, but it would not make for an economically viable power plant.
I mean 400 MW is tiling a 4km square with cheap PV, even after accounting for capacity factor. And 400 MW is also what you get from 1052 Tesla Model 3's at max acceleration, according to their website (though obviously draining a 74kWh battery at 510hp empties it rather a lot faster than "overnight"), and ICEs are also tiny power plants so you get the same power out from order-of-magnitude same number of those.
In fact, this is why I'm putting the money I can set aside for investments, into China rather than the USA: they're speed-running cheap PV and batteries, which are cheaper than most everything else right now.
