But would the cost of that structures not be pretty insignificant compared to the cost of the reactors? Or build the support structure for six reactors but initially only one reactor, sell five more when the first one works. In relative terms it makes no difference whether a cost overrun of 100 percent is for one or six reactors, in absolute term it of course makes a difference and it seems you might have a better chance of not getting your project cancelled if the absolute overrun in smaller. But this might of course, as you mentioned, also be a matter of scale, hard to tell without knowing where exactly the additional costs come from.
16.5% for the buildings and roughly 50% for construction services, engineerig services, owner's costs, structures and improvement according to the aricle I linked in my other post. I suspect that specifically those costs are the ones that cause most of the of overruns and delays. Also, the whole process is certified by the NRC and you can't get creative with it. There are 4, 6 and 12 reactor configurations for a given plant.
