Wind and solar don't have fuel, but the conversion is bad and the resources is distributed and hard to capture.
> Engineering and regulatory costs won't necessarily come down over time. As use of nuclear power wanes, costs will likely go up, the positive feedback reinforcing the trend.
Why would they not come down? In most technology we use it does come down. Nuclear is really the exeption where normal development has not happened.
Regulatory costs are comming down right now already. The US has started a wide range of reforms to address this problem. The US is going from essentially impossible to regulate a new reactor to only very expensive. Interestingly nuclear went from both parties being against, to both parties being for and 3-4 nuclear reform bills have gone threw congress with little issue. Its suprising to say the least. Its still baby steps but at least its not continued political deadlock.
Canada is doing much better and there is actually now feasable to go threw the regulatory process again. This has lead many US and British companies to move to Canada.
This should allow now reactor types comming to market in the next 10 years and those be much smaller in scale then PWRs and thus have very much smaller capital costs, thus building them should be cheaper.
None no of that is a certainty but its also not crazy either.
> In 100 years time, fusion may be developed, but we may have seen peak fission already.
Fission is 10^12 more dense then chemical energy, Fusion is about 10^15. The difference there is not all that relevant specially not as a society already ignore the 10^12 difference.
Most Fusion don't really add much, other then maybe being better at nuclear proliferation but that is not really a hugly practical problem with fission specially in the western world.
Solar and wind have to replaced fairly often, and once first generation installations must get replaced, and cheap nuclear is available, the situation might reverse itself.
That said, in certain places it will be hard to beat these but there many places where they are not optimal and there is still the intermitency question. Many GenIV nuclear plants can be perfectly perfectly load following or can be connected to utility scale heat storages.