If the recommissioning requires 20% of the work of new-build (in time and money), I'd rather have the "normal" 50-100% overrun in costs and time happen on recommissioning vs new-build.

The delay is usually heavily on the permitting/legal challenges/etc. stage, not actual concrete-pouring and equipment installation, so an already-once-permitted site will probably do better relatively, too.

Extra reactors at exising operating sites would make even more sense, though. Putting 2-4 extra AP-1000 at ~1GW/ea at each of the 54 facilities operating today would be huge and minimal incremental risk, and if there were a shortcut one time permitting process for it...