After all, can we really estimate the cost of securing spent fuel for over 100'000 years?
Do we need to? The point of modern reactor designs and nuclear reprocessing is to not only minimise the waste but the resulting waste does not need to be secured for 100000 years, several orders of magnitude less.
More importantly, the "tool" seems to cover only construction costs. Nowhere are decommissioning costs included, which are order of magnitude over the construction costs. Experience has shown both in the US and elsewhere, that these costs have been (willingly or not) underestimated by order of magnitude by the industry.
The costs haven't been under-estimated. The required scope was. To which I again ask, do we need to? Providing the site is not required for reclaiming why not just build a cement dome over it and let it sit. Decommissioning costs are very small if you're not under the delusion (in my opinion) that you will be using that land for anything other than a place to store a radioactive shit-ton of cement.
Now decommissioning a damaged reactor, that is a different and very expensive story, as is demolishing and removing reactors.