[renox]

In France the new nuclear reactors have been plagued by hardware issue, delays.. So no, it isn't 'easy' to predict the building costs, it's easy only when you're replicating an already existing reactor and only if it was built recently!

[JohnHaugeland]

> In France the new nuclear reactors have been plagued by hardware issue, delays

That's odd, can you show me some evidence of that?

[natmaka]

The new type of nuclear reactor is the "EPR".

The first project, in Finland (Olkiluoto-3) was a disaster: massive overcosts, 12 years late, customer lost, penalties paid, builder (Areva) saved thanks to public money. https://en.wikipedia.org/wiki/Olkiluoto_Nuclear_Power_Plant#...

The second project, in France (Flamanville-3) is an ongoing disaster: massive overcosts (from 3,3 to 19,4 billions €, and counting), aleady 10 years late and counting https://en.wikipedia.org/wiki/Flamanville_Nuclear_Power_Plan...

The third project, in China (Taishan-1 & -2) was officially 5 years late and 60% overcosts (and very probably not break even for the builder). After 1 year one of the 2 reactors was stopped mid-2021 after an incident.

The fourth project, in the U.-K. (Hinkley-Point C), started in 2018 and already announced overcosts and late delivery.

Old reactors: https://www.reuters.com/business/energy/frances-chinon-3-rea...

Decommission costs: https://news.ycombinator.com/item?id=18686703