Third generation reactor designs like the AP1000 and EPR were, prior to actual construction, promoted as increasing safety
and reducing construction costs/time via standardization and simpler passive safety features, compared to older Generation II/II+ reactors that currently make up most of the world's operating reactors. Now that they're actually under construction these reactor designs too are terribly over-budget and over-schedule, just like the 1980s reactor projects that burned US utilities bad enough to scare them away for a generation.
AFAICT "late and over budget" usually happens with megaprojects of all kinds, and unfortunately commercial power reactors are only available in "mega" sizes. The AP1000 and EPR especially were designed larger than median Gen II units, for economies of scale, but seem to be reaping anti-economies of scale. Small modular reactors with a faster learning curve and more units to learn across might finally tame schedule and cost problems... but it's hard to find investors optimistic and deep-pocketed enough to spend a decade iterating on the concept with real hardware. Most "nuclear startups" are stuck at nothing but simulations and PowerPoint.
Shouldn't government be making these sorts of high-risk, high-reward, long term investments? They should, but unfortunately political opposition to nuclear power is much more concentrated, if not more broad, than support.