What Will a Nuclear Plant Cost, and How Long to Build?
Bottom line is we don't know. But here's an educated guess.
I am including both plants that are presently approved by the NRC and designs that should have approval in under 2 years. I researched this in some detail, following the citations from the different A.I. queries.1
First off, what do we need to look at?
Cost to Build: Covers construction, equipment, financing, and regulatory fees. Varies with design complexity and market conditions.
Construction Timeline: From groundbreaking to operation, affected by permitting, labor, and delays.
Advantages: Safety, efficiency, scalability, and environmental perks.
Disadvantages: Cost overruns, waste, and public perception risks.
Approval Status: For unapproved designs, NRC review timelines (typically 18–24 months) matter.
Renewable Comparison: Wind and solar are cheaper and faster but falter without massive storage.
Please note, the below is not so much how much and how long if Colorado goes nuclear. Rather this is to look at the various designs to see which one(s) Colorado should focus on.
NRC-Approved or Near-Approved Nuclear Designs
1. NuScale VOYGR (Approved: 50 MWe; 77 MWe Under Review)
NuScale’s VOYGR is a small modular reactor (SMR), certified in 2023, with an uprated 77 MWe version under review. A six-module plant yields 462 MWe.
2. TerraPower Natrium (Under Review)
Natrium is a 345 MWe sodium-cooled reactor with molten salt storage, targeting coal plant sites.
3. X-energy Xe-100 (Pre-Application)
The Xe-100 is an 80 MWe gas-cooled SMR, scalable to 320 MWe (four modules), targeting Texas.
4. Westinghouse AP1000 (Approved)
The AP1000, a 1,100 MWe PWR, was certified in 2006 and operates at Vogtle.
5. APR1400 (Approved)
The APR1400, a 1,400 MWe PWR from South Korea, was certified in 2019, with units in South Korea and the UAE.
Here’s the comparison
Which begs the question - why would we use anything other than the APR1400? We get 1.4GW sooner for less than the AP1000 1.1GW. And the price per GW is cheaper than any of the upcoming SMRs.
TerraPower which seems to be everybody’s favorite SMR is 3 - 4 times the cost per GW. We don’t get it any sooner. And while the first unit will be less than the APR1400, add the second and it’s more - and still not the same amount of power.
Final Thoughts on SMRs
First off, a lot of people tend to think of SMRs as something that comes off an assembly line, is put on a truck, dropped off at the site, plugged in, and is running.
Nope. A SMR needs to be constructed at each location. All the custom pieces are hopefully being manufactured ahead of time and are delivered to the job site as needed. But we’re still talking 6 - 10 years to build one.
Second we take the SMR estimates of cost and build time as writ. Nope, they’re estimates and as most estimates of upcoming technology, the estimates are likely below what it’ll end up being.
The SMRs have some interesting new technology. I’m confident we’ll see some achieve significant use. But they’re no game changer. And quite likely for the next 15+ years the APR1400 will remain the most bang for the buck.
Citations
Rather than put 23 citations here, please go to the link below and it will lay out the details, the numbers I have above, and the citations for each.
I went to a lot of the citations. Every citation from the different A.I.’s was to a credible source and then numbers it pulled were correct. Nice validation.
I’m glad that the XE-100 and Natrium reactors are being built but we need to get several years of operational experience before committing to build a lot more of them.
If we want to do something more than twiddle our thumbs in the next decade then the only sensible approach is to deploy more copies of the latest PWR (or BWR) designs - like the APR1400 or the AP1000. If we were smart, we’d ally ourselves with a few other states and put in an order for a dozen plants, agreeing to spread the cost among all participants.
Educated guesses on construction time in the United States should be based on the experiences of other building processes in the United States. I'd think variations from the actual experience of building would need to be explained.
Most recent experience: Vogtle 4. Looking at an already established site with previous construction and operation, an established vendor and an experienced operator -- and the build process took a bit over 11 years from when a license was issued until operation.
Combined License Issued: 02/10/2012
Date of 10 CFR 52.103(g) Finding: 07/28/2023
What part of the processes do you expect would be substantially shorter for (presumably) a new operator, a vendor with little experience in the United States, a new site, and in a state that hasn't had a nuclear plant operating since 1989?