I read this opinion in that blog of Austin Vernon that made the HN rounds about a year ago [1]. In the end his argument is basically "if some energy generation method has moving parts, it has to be more expensive that one that doesn't".
But let's look at the numbers.
Here's a study by the EIA that looks at the capital costs and also at the Operating and Maintenance costs of various types of power plants [2].
The costs that are interesting for us are (all costs per GW alternative current of capacity):
- solar without storage: $1.3 BN (page 175)
- solar with 4 hours of storage: $1.8 BN (p. 180)
- coal without CO2 sequestration: $3.7 BN (p. 46)
- nuclear AP1000 (similar to Vogtle): $6.0 BN (p. 107)
- natural gas power plant: $1.0 BN (p. 83)
First of all, notice that the natural gas power plant comes to be the cheapest, despite the fact that it has turbines, including steam turbines. So the general assertion that things with moving parts are more expensive than those without is not quite correct.
You can counter that the estimates are from 2019, and by now solar is probably already cheaper and it will continue to get so, and I suppose you are right. But the panels constitute only 18% of the cost of the solar panel plant, the rest being inverters, transformers, something called BOP (balance of plant), etc. So, if the panels go to zero, the cost of the solar power plant gets reduced by 20% only.
On the other hand, the same is true for nuclear: only about 20% of the cost of a nuclear power plant is attributed to the nuclear part, the rest to the turbines, BOP, etc. So, even if by a miracle the nuclear part were to cost zero, the cost per GW would still be about $5 BN. This is higher than the coal power plant, because the steam generated by coal is much hotter than the one generated by the current generation of nuclear reactors.
Does this prove Austin Vernon's point?
Not necessarily.
First, the fact that gas power plants are very cheap is reason to hope. If we make nuclear reactors that don't use steam (which is corrosive) but some other gas, then it's possible the cost could go down significantly.
Well, less than one year ago China hooked to the grid exactly such a plant [3]. It uses Helium as a coolant, which means that the turbines can be cheaper (they should be similar to the ones used in a gas-firing plant). It also runs much hotter than a regular pressurized water reactor (about 700 Celsius vs 300 Celsius), which means the efficiency is higher.
Can the US build such reactors? I don't see why not, see for example Xe-100 [4]. But the first step is to get back to the nuclear technology learning curve.
[1] https://austinvernon.site/blog/nuclear.html
[2] https://www.eia.gov/analysis/studies/powerplants/capitalcost...
[3] https://en.wikipedia.org/wiki/HTR-PM
[4] https://www.energy.gov/ne/articles/x-energy-developing-pebbl...