500B of compute infrastructure with order of magnitude greater deprecation / need to return on capital. Compute isn't concrete infra with 50+ years of value, more like 5 years, i.e. need to produce 50-100B worth value per year to break even. On top of the “$125B hole that needs to be filled for each year of CapEx at today’s levels” according to Sequoia. All which with could be wiped by PRC pacing SOTA model, unless 500B of more compute will lead to qualititive model differences. Right now, I don't know where that value is coming from, so either a lot of investors are getting fleeced, or this is a Manhattan tier strategic project... privately funded, which makes even less strategic sense.