Ways to think about token pricing in the AI supply crunch

Ways to think about token pricing

I argue that the current AI supply crunch is unstable and likely temporary. With massive data center investments and improving inference efficiency, foundation models risk becoming low-margin commodity infrastructure rather than retaining sustainable pricing power. The market will eventually settle into a new equilibrium, but predicting the exact outcome remains impossible due to unknown variables in future use cases and training costs.

"At one extreme, there are two or three giant minds that run half of everything and have massive pricing power, and at the other extreme LLMs look like databases - there'll be millions of them, some very big and some very small, and the value is in what you build on top."