The US model for a forced private capital AI data center buildout is in trouble, while Asian countries do much the same thing, but with governments planning and setting guardrails. This all is quite apart from serious doubts about the return on investment for those data centers, which makes the future outlook for them uncertain if not bleak ...

There is no A.I. race - it is a worldwide phenom of tech builders in many countries each using the generally same building blocks delivering some level of pattern matching - called A.I. Everyone will converge in pretty much the same place - except the big A.I. companies will not make it to the end of the race - their financials are desperate and they simply cannot outrun their invested capital and show any level of ROI. The future is A.I. everywhere - on your watch, phone, in your car, embedded in your apps - and none of that needs a centralized data center - and these guys know it.

This article asks an important question. Why should AI be given all the knowledge that humanity has acquired over millions of years without also having to learn what humans learned as we acquired that knowledge...