Damodaran: an AI crash could hurt more than dot-com

nyu finance professor aswath damodaran warns a potential ai market correction could inflict more damage than the dot-com bust because today's infrastructure investments are debt-financed and physical, unlike the capital-light dot-com era. he challenges the standard ai business model, arguing that compute-heavy usage prevents costs from dropping to zero, meaning growth with thin margins could destroy value, especially with low-price competitors like deepseek. damodaran also highlights a societal dilemma: if ai successfully replaces half of white-collar jobs to justify its valuation, it will impose 'insane costs' on society, while big tech companies risk building depreciating infrastructure that may become obsolete within five years.