AI doesnt have to fail for the bubble to burst

The AI bubble could burst even if the technology doesn't fail, due to intense competition and pricing pressures. Companies are investing billions, but cheaper models could satisfy most users, reducing the incentive to pay more for slightly better performance. Competitors might also use distillation to train their own models from leading AI, and Chinese companies can acquire training data cheaply, potentially eroding the lead established by large investments. This competition could drive down prices, undermining the profitability anticipated from current spending.

This analysis uniquely highlights how distillation and cheap data acquisition by competitors, especially Chinese firms, could erode the lead of heavily funded AI companies, unlike other bubble theories focusing solely on overvaluation.