Though I'm more concerned about the effects of the current political climate, than the "AI" bubble popping. In the scenario of that going south, nothing will be normal for a long time.
This time around the ramifications might be larger, but it will still mostly be felt by those inside the bubble.
Personally, I would rather experience a slight discomfort from the crash followed by a normalization of hardware prices and the job market, than continue bearing the current insanity.
That's because the companies were largely public. In the dotcom era, the exit for startups was to IPO, as early as possible. When someone like pets.com was burning through cash with zero path to profitability, it was public knowledge, everyone could see it.
The AI built out is largely funded with private credit. It's a black box, we have no idea of true valuations. The mean time for a company to go public went from 4 years during dotcom to 14 years now. The rot can be hidden for a long time until the big funds go bust, with all the collateral damage that brings.
See Microsoft's recent "We don't understand how you all are not impressed by AI."
In the case of MS, you're right, Satya isn't going to fall on his own sword. They will just continue to bundle and raise prices, make it impossible to buy anything else (because you still need the other tools) and then pitch that to shareholders as success "Look how many people buy Copilot (even though it's forcefully bundled into every offering we sell("
There’s minimal risk to the decision makers. Meanwhile, every one of us peons is significantly more at risk of losing our jobs whether we could be effectively replaced with these AI tools or not because our own C-level execs decided to drink the snake oil that is the current bubble.