Google Will Fund Three Nuclear Projects with Elementl Power

Here's the best evidence that they're delusional:

  Elementl Power to date has not built any nuclear power projects. The 
  company has said it is technology agnostic, and has not determined what 
  type of reactor it would use to produce nuclear power, noting that it would   
  choose the technology furthest along in development when it is ready to start 
  construction.

The only reactor which anyone has built in the last 30 years is the AP1000. Construction started in 2009 and ended in 2024. That 15 years was particularly impacted by the problems of building an all-new design and the bankruptcy of the vendor. We should be able to do better if we could build reactors on an Nth-of-a-kind basis instead of a first-of-a-kind basis and the obvious path to that is standardize on a design and build a bunch of them despite high costs, something China can and will do, which could get to the place where Russia is where they can build VVER reactors very similar to our PWRs reliably in other countries.

4th generation reactors using molten salts or liquid metals or gaseous coolant might operate at higher temperatures and cut down the high capital costs of the powerset but they are not a bird in the hand -- here you are going to wait 10 years not to build from an existing plan, but 10 years for somebody to build a test reactor.

The times involved are so long that the only countermeasure is to "start early" and pick a ready design and start building now. "Technology agnostic" is often a polite way to say that you don't want to make any real investments at all.

It seems possible, if ambitious, to build 10 AP1000 reactors in the next 10 years and thus add 10GW of power by 2035. Any other LWR is going to add a few years to get regulatory approval. Any post-LWR reactor is going to take 10 years just to be ready to build.