Perhaps orthogonal to your post, the WSJ today has an article of interest; I have provided a gift link for those without a subscription:

The AI Boom Is Transforming the American Economy Beyond Recognition
AI has rapidly become a huge economic factor while affecting everything from capital investment to how much you pay for an iPhone

The U.S. economy keeps putting more eggs in the artificial-intelligence basket.

Tech companies are spending hundreds of billions of dollars to meet AI computing needs and issuing billions of dollars of debt to help make those purchases.

The rapid data-center build-out is powering construction spending, hiring and municipal revenues. Meanwhile, a stock-market rally fueled by the rise in shares of chip makers and other companies benefiting from the AI boom has led to a massive increase in household wealth. That is helping to bolster consumer spending.

Combined, those factors are likely responsible for roughly one-third of the nation’s recent economic growth, according to Michael Pearce, an economist at Oxford Economics.

But the strength that the economy is drawing from the AI boom is also a vulnerability: What would happen if the boom fizzled out?

“It is very much an AI-driven economy right now,” said Barclays economist Jonathan Millar. “It’s hard to imagine that we would be anywhere near as resilient without that impetus.”

wsj.com - The AI boom is transforming the american economy beyond recognition

I have been one (among many) calling it a bubble, and I still believe it is, but you know: suppose it’s not. Suppose AI can grow to the sky and we just can’t imagine all the uses it will be put to (eventually.) The “internet bubble” surely was, and yet after the initial enthusiasm wore off and everybody went ‘Oh, yeah, bubble” it has kept growing: I now have three vendors for wired internet to my house, not to mention several sky-directed methods of connection. There are a bazillion thousand websites and vendors, and while Amazon gets all the notice, it still commands only 38% of on-line sales and 8% of total retail. (Walmart has more of the latter at almost 12%.) I can survey my property from a satellite for a dime, or plot a driving course to Montana - while moving and ordering a burger from Shake Shack as I chat with my brother. I can videoconference with someone in Croatia or get a pizza from the Dominos down the street because I’m too lazy to drive there.

Is “the internet” slowing down as a feature of society or commerce? I think not. Does anyone see a threat to it ahead? (Again, I don’t think so.)

There is much unknown about AI (just as there is about iPhones, the internet, etc.) in terms of its eventual effect on society - but here we are. [I have in my mind a New Yorker-ish cartoon with a guy buying some AI in a box saying “I have no idea what this is really for, but I gotta have some.]

I’d add to this already overlong post by calling out the major players: OpenAI, Anthropic, Grok, Oracle, Microsoft, Google, Meta, xAI, Amazon, DeepSeek, Baidu, Moonshot and others, but except for a few tangential thoughts I don’t really know enough about the internals. (I would group them into those which can use the function more-or-less self contained: Google, Facebook, Amazon, Baidu, and the others which will be selling the service to unaffiliated clients: Oracle, OpenAI, and so on. Perhaps both models will be successful, perhaps only one or two will survive in the “buy my service” category; dunno.)

Anyway, if you read down into the Journal article, it ascribes fully 1/3 of the US economic growth to AI buildout in hardware, siting, selling, using, power generation, etc. I would have thought more, given the relative flatline of other business, but what do I know? Still, if there comes a time when it’s “Whoa; stop building”, we all know we’re in for a world of hurt. Meanwhile the merry-go-round is spinning; I think we’re all along for the ride.