Earlier quoted context omitted.
There are thousands of items on the agenda in the present day, everything from better open source LLMs to moon-base railway designs, so even if 99% of these turn out to be duds that can’t be budged, there are still sufficiently many remaining to absorb all bonafide super-geniuses on Earth several times over.
In this particular item on the agenda, chip fabs, it is not clear spending more money one time will produce a result, much less a faster result by attracting "bonafide super-geniouses." One might respond that "you can't win if you don't play" which would lay bare the gamble.
Of course if they were somehow disproportionately forced to only ever work on chip fabs for the rest of their lives, then that would create huge waste, but this doesn’t seem likely.