> Moreover, logistically, not building that accelerator means you quite likely never have it

That's exactly wrong. Who knows if you wait a bit the underlying technologies might cause a step change in price to build and make it so the project actually might get completed. For example waxahatchie vs cern