Pop!(?)
So, the US midterms seem to have kicked off a rout of AI stocks. This is based on promises to introduce regulation (gasp!) and clamp down on datacentre building (shock!).
But, neither of those things feel like they should have kicked off a huge sell-off of any stock related to AI. Any reasonable company would already be behaving in a way that regulation doesn't impact (and, in fact, regulation would favour large incumbents as they would have the resources to manage them). And, slowing down building data centres that were already a long way from adding value doesn't feel like a panic moment.
What gives?
My theory: the hypothetical infinite growth machine that is AI has suddenly been brought out of fantasy economics land. All that had to happen was the implication of consequences and dealing with reality, and suddenly all the glitter came off. Ed Zitron will probably have a 3,000 word essay on this by Saturday, with far more research than I, but, I'd bet good money he'll be saying the same thing.
So, stocks are being analysed not as a potential gateway to owning some vaguely guessed at proportion of all human productivity, but as businesses with earnings and losses and debts and (maybe some) profits. In light of that, well...
Pop.