If AI bubble burst... the AI compute (where most of investment/capex going towards) depreciates 1-5 years, specialized data centers a little longer. For referen

If AI bubble burst... the AI compute (where most of investment/capex going towards) depreciates 1-5 years, specialized data centers a little longer. For reference most of the $$$$$ in capex of dotcom burst went into civil engineering / dumb fiber pipe layout and left US with a fuckload of fiber that lasts 30/40/50+ years. Still using capex from railroad bubble 100 years ago. Both bubbles at least left durable (agnostic networks for future to build off, i.e. upgrade switching and fiber + rail get more efficient utilization. VS specialized LLM data centers with bespoke cooling/power/networking needs - if LLM bubble fails there's a good chance they can't be repurposed economically because their OpEx is probably what broke the model in the first place. I feel like people are failing to grasp how big of a F US will eat if AI bursts relative to past bubbles. I mean it's better than tulip mania since fraction of capex will go towards electrification infra (assumign it gets built), but most of the $$$$$ is going towards fast obsolescent AI hardware also closer to tulips than fiber or rail in terms of stranded depreciated assets.