When people talk about the AI bubble bursting, I keep wondering where the investors who still want growth will put their money.
So many good-looking software opportunities look like LLM wrappers to me. If the interesting new businesses depend on the same models, why not invest in the providers? And if picking a model provider is too uncertain, why not invest in building the infrastructure they need?
You can move through several businesses and still end up betting on AI.
That does not mean a bubble needs a replacement industry before it can burst. Investors can simply decide the current price is too high. Falling appetite for risk can be enough to bring prices down, as the Federal Reserve's valuation analysis explains.
Take an imaginary company with 100 shares at $10 each. It is valued at $1,000. If its market price falls to $6, it is now valued at $600. The missing $400 does not have to turn up in another investment. It was a valuation, not cash sitting inside the shares.
But that still leaves my question. For investors looking for the next big source of technology growth, what looks more promising?
My suspicion is that disappointment with one AI business will often send them to another part of the same industry. From the wrapper to the provider. From the provider to the infrastructure.
None of those is automatically a better investment. A useful application might have customers the provider cannot reach. A provider needs to cover its costs. Infrastructure needs enough paying demand to justify building it.
AI can keep attracting investment while investors lose money. I can see how the prices fall. I am less sure what replaces the attraction.