The US delays with the Vogtle AP-1000s (the only recently completed US build) were extremely atypical.
First, it was a FOAK design. Which always takes longer to build, it is a prototype.
Second, the nuclear build know how in the nuclear engineers, construction workers, and supply chain was not really there any longer.
Third, they used a new permitting system, which in theory should have been better and probably will be better in the future: instead of ongoing individual checks and modifications, which made every nuclear power plant in the US a unique unicorn, you are now allowed to submit a master design and once approved you can build that over and over. Without changes.
Alas, Westinghouse wasn't actually done with the design when they submitted. So when they started building, they noticed that they had submitted plans that could not actually be built. Oops. That cause massive delays. And delays = cost.
And the suppliers fought each other, one went bankrupt etc. COVID also didn't help.
So how can we guarantee that the same won't happen in the future and that NOAK builds will be better? Well, for one they now have plans that are obviously buildable, because a bunch of AP-1000s have been built. So that exact thing absolutely can't happen.
Also, we can look to China. Turns out, China also built 2 FOAK AP-1000s. These also took about 10 years, despite China usually building in 5. And it turns out, China built some more AP-1000s after that. NOAK builds. And these took 5 years to build with buildable plans, experience building that reactors and a mature nuclear industry to back them.
So there is good reason to believe that future NOAK builds of the AP-1000 and of comparable reactors will be much faster and much cheaper than what we've seen so far.