New nuclear plants don't make financial sense.
The nuclear industries own modelling shows that it costs about 5x as much to build compared to wind power on a per kw basis.
Recent nuclear construction projects have all blown out their projected costs, the South Carolina plant blew out from $10B to $25B before it was completed. We will never know if it would have blown out further as they cancelled the project, Westinghouse went into bankruptcy, South Carolina still had to pay $9B but only got a fancy concrete slab to show for it.
The Vogtle plant in Georgia had a budget of $14.5B, which has been revised to $27.3B. Construction started in 2013 and was expected to finish in 2017, this is now targeted at 2022. So costs are up 88%, time is up 125%, and we are only two thirds of the way through construction.
This isn't a US problem, the Olkiluoto plant in France and the Hinkley plant in the UK have both costed twice as much and taken twice as long as projected. Neither is yet operational.
So it would be more reasonable to say that modern nuclear systems cost about 10x as much to build compared to wind power.
Admittedly these are all big, centralised, "old" designs. However despite considerable investment small modular reactors have never been certified and the first ones aren't projected until mid 2020s. Until proven otherwise, I believe it is prudent to expect a similar cost and time blowout on these systems, so the first is unlikely to be operational before 2030. Advocating for an unproven product ten years before it is deliverable is a bit too wishy washy for me, especially when fusion is only 20 years away.
Once operational nuclear is one of the cheaper systems to run, but operators have been seeking subsidies recently, and this in no way offsets the interest costs of having to pay off the upfront capital expense.
Nuclear power plants have never existed without considerable government support. Advocating for them is great, but I feel advocates should be upfront as to why it is worth spending billions extra for a fungible product.