I like the example of dotcom bubble enabling fiber investments. Hyping up internet applications enabled setting up infrastructure, which in turn made the hype r

I like the example of dotcom bubble enabling fiber investments. Hyping up internet applications enabled setting up infrastructure, which in turn made the hype real, albeit years later. I think in this wave, GPUs will be largely outdated in few years and they are not probably going to be powering up the usage by the time hype will become real. The actual utility of this infra investment I think comes primarily from money spent on RDing the next gen AI chips, which wouldn't have been possible if there wasn't any hype spending money on today's chips.