Well also that and that valuation i predicated on growth that' physically impossible.
AI companies have started talking about "GW" of installations which makes it simpler to analyze. The electricity capacity doesn't exist to meet the growth predictions. The manufacturing capacity doesn't exist to manufacture the generation capacity required to meet growth predictions. The capacity to build manufacturing facilities to match predictions doesn't exit.
The market has gone irrational because there's no way growth can match predictions. I ain't shorting though because Keynes isn't wrong.
I expect any other irrationality will also bite hard. Claude is pretty good, really. Not, like, perfect and AI will destroy the world and programmers, but the genie is out of the bottle here, something like it is here to stay. It's too useful. But let's say OpenAI stumble and suddenly their terrifying outgoings slightly tip over and they go very broke very fast, well, the market will panic and go omgomgomgomgAIisbadactually!!!111onelelevne and the entire thing, Anthropic included, will crash through the floor, even though people will, for example, still be spending a lot on Anthropic.