No, you don't know what you are talking about. The budget, that's EDF. EDF builds things, the ASN ensures the safety is up to standards. Hell, the ASN _is_ responsible for these costs because they are the ones ensuring everything is up to par. There are many EPRs running, in Finland, in China, but none of these have safety requirements as drastic as the Flamanville EPR.
Additionally, EDF offered multiple estimates at the time of construction, from "everything goes well" to "oh dear god so many delays". Many of these estimates were perfectly in line with the current costs. It is, once again, political decisions that only allocated the minimal budget, while knowing full well that it would go over (but that EDF would take the blame, and not the government).
And once again, this is also a result of letting our nuclear industry decay for 40 years, losing every single person that has the knowledge on how to build nuclear plants either to retirement or other countries.
> The ones who underestimated Flamanville's cost by 80%? (and it's still not in production)
It's an entirely new design of a complicated piece of technology, of course it will be over budget and delayed. Grand Paris Express and the Paris Philharmonic were also over budget and delayed, does that invalidate the opinion of all experts involved in both projects?
Being over budget and off schedule is most of the purpose. Once a plant is delivered, the gravy train stops. No one actually involved wants that.
The Finns recognized the gravy train would stop soon, anyway, and elected to deliver a (more-or-less) working reactor, after consuming 5B euros beyond the 3B quoted cost, more than a decade late. In the US, corruption is secure enough to go far, far beyond that price, with no demand to deliver.
Because the most important thing about a project as complex as a big acoustic-friendly building or multiple hundreds of kms of automated metros or a nuclear power plant is that the predictions about budget and time are perfect?
I wonder what that means for software developers who notoriously struggle to predict basically the same stuff, only on much less drastic scales (if an engineer underestimates how much time and effort a feature will take, usually people don't die. If a building crumbles or a nuclear power plant has an accident or a train crashes people do die).
The ones who underestimated Flamanville's cost by 80%? (and it's still not in production)
Let's maybe not trust them completely, and have our eggs in other baskets as well.