And I don't think the reason things aren't done isn't because they aren't economically viable, but because there's no interest in doing them. Could somebody start a megacity out in the middle of nowhere? I mean probably, and it might even be a good idea, but things like this require a visionary to get them going. Electric cars prior to Elon were glorified go-karts for hipsters. It's not that they weren't viable, but nobody had really shown the path forward yet.
And here there's already major interest and money behind it, so getting started won't be a problem. And from that point, the path forward will become ever more clear.
Trillions of dollars are being dumped into companies when frontier level open weight alternatives that are able to be run on ever more reasonable hardware already exist. It feels reminiscent of how everybody certainly knew that tulips could just be planted, but people operate under their own assumptions about the world and its future, and sometimes it can lead to quite odd outcomes.
I tend to frame this as beneficiary vs. investor perspective. I'm a beneficiary of AI, not an investor. I couldn't care less about whose name is on the model, it doesn't affect the way the technology transforms my life, or the world around me.