In this graph you have the marginal cost of traditional generation compared to total cost of new built renewables.
https://www.lazard.com/media/451885/grphx_lcoe-07.png
From Lazard Levelized Cost of Energy 2021.
https://www.lazard.com/perspective/levelized-cost-of-energy-...
I didn't find all that much to go on. Aside from a French Court of Audit report in 2012, this information seems to be mostly non-public? Anyway, it looks like the answer is staffing and maintenance. It seems to be common for nuclear power plants to employ 500 to 1000 people. Fessenheim in France employed 700 people at the time of its closure, plus an additional 600 to 2000 during maintenance (per Wikipedia).
Meanwhile, nuclear seems to be very technically complicated with a bunch of weird Advanced Physics-y shit that can go wrong. Like, water pipes exploding because the nuclear radiation caused water molecules to split up by radiolysis and a design flaw allowed the hydrogen to collect somewhere. Or borated water eating through steel.
And now that I type this, it occurs to me that you can't just turn off the plant for safe maintenance work -- inside the dome is always going to be irradiated and probably require hazmat suits.
And, as you alluded to, nuclear plants use the steam based thermodynamic cycle but run cold because they have to maintain water in a liquid state while absorbing heat from the reactor. The power achievable from the Carnot cycle depends on the magnitude of the temperature difference: nuclear plants run at 300° C, coal plants about 570° C, and CCGT at 1000° C up to 1500° C.
"Study of Equipment Prices in the Power Sector"
https://esmap.org/sites/default/files/esmap-files/TR122-09_G...
From table ES3 we can see that steam based plants are expensive in pure equipment costs. Which makes sense comparing a tiny gas turbine giving direct mechanical force. Then you can use the waste heat for steam generation if that makes sense in your location.
Comparing this to coal and nuclear plants huge boiler and turbine setup with about the same output. Nuclear adding even more complexity of different loops and safety requirements.
Haven't dug deeper into that report than the executive summary though.
https://en.wikipedia.org/wiki/Base_load#Description
It is a straight LCOE shoot out. Storage may help at the limits but is a pure loss from a system perspective.
Admittedly you need much less storage (until a tsunami comes along). But pumped storage only costs around $1/watt. (The unit of $/watt rather than $/kWh because the dam is only a minor fraction of the cost of pumped storage, and it drops quadratically with size.) Storage + renewable generation still costs less than 1/2 of nuclear.