Nuclear Power Correlations with Average Retail Electricity Rates: 2018-2024

6/29/2025

We’ve now arrived at our final, major contributor of electrons to the US grid: nuclear power.

Unlike earlier power generation schemas – nuclear represents a slightly different format in a variety of ways.  For one – nuclear energy outputs really didn’t vary much over this six-year examination.  It’s domestic footprint – as we’ll dissect more closely later on – has remained relatively stoic.  As the following chart demonstrates – US nuclear production only dwindled by roughly 22,500 GWh/year during this multiyear window; encapsulating not even a 3% drop. 

Market shares (in terms of total energy generation and cumulative sales) during this time frame barely budged southward beyond 1%.  Given this YoY consistency in domestic nuclear capacity – we didn’t feel inclined to run any single-factor ANOVA test(s) to measure between-annual data population variances: if those from more profoundly dynamic power generators failed to yield even a decent number of significant variations – then there is certainly no reason to believe that nuclear would behave any differently!

But the correlation data between nuclear power, and average, retail electricity prices tell a slightly different story…. Immediately below are the graphed trendlines of each annual group’s Pearson r value, as calculated using each state’s average grid costs and both its respective % of cumulative electricity sales and total energy output (in separate graphs).

Directly above you’ll find the commensurate r values for the Spearman Rank correlation procedure; since these state-by-state data sets didn’t appear to fit within the parametric classification, as shown in the following histograms:

It’s logical to assume that such parallel and consistent, YoY population sets would yield very similar correlation coefficients; regardless of whether the data was normally distributed, or converted into ordinal rankings.  But that doesn’t appear to be the case; especially as it relates to the % of total power generation: there is a progressive improvement in nuclear power, and its congruency with declining grid prices over time. 

A few further bullet points to note:

  • Due to continued reductions in  industry-wide fuel, operating, and capital expenses - nuclear energy generation costs fell from 2020 ($34.08/MWh) to 2021 ( $32.32/MWh), then again in 2022 ($32.05/MWh), and once more in 2023 ($31.76/MWh): https://www.nei.org/CorporateSite/media/filefolder/resources/reports-and-briefs/2024-Costs-in-Context_final.pdf#:~:text=Operating%20costs%20increased%20from%20$22.87%20per%20MWh,reduction%20compared%20to%20the%20peak%20in%202012.

  • The visible (and rather precipitous) drop in r values between 2022-2023 and 2023-2024 looks to be intertwined less with operational efficiencies, and more with federal tax incentives:

    • The Zero-Emission Nuclear Power Production Credit (45U) of the IRA took effect at the start of 2024.  45U proffers up to $15/MWh for energy supplied by qualified nuclear facilities placed in service before August 2022.

    • Vogtle Units #3 and #4 were only eligible for the 45J PTCs (providing $18/MWh) under the 2005 Energy Policy Act; the only two domestics reactors to qualify.

      • Vogtle #3 begin operations at the end of July 2023

      • Vogtle #4 commenced near the end of April 2024

  • Nuclear doesn’t really follow the Henry Hub Index price signals that natural gas, coal, and VRE adhered to in earlier examinations

    • No significant drop in nuclear power-to-electricity cost correlations were observed from 2020-2021 to 2021-2022 – unlike within the coal-fired and combined wind-solar power correlation graphs – when natural gas pricing surpassed $7-, $8-, and $9/MMBtu

  • The signals here insinuate that nuclear power functions as more of stand-alone power source for the US grid; almost as if nuclear operates in a vacuum

    • This is likely due to nuclear’s incredibly high capacity factor (low-to-mid 90%); signifying that it performs irrespective of the other platforms

It stands to reason that the impact of subsidies for a specific technology (in this case – nuclear power) would be more profound if said technology is divorced from outside influences.  This could help explain why a relationship between nuclear output and decreasing energy costs manifested so suddenly and distinctly as 2023 transitioned to 2024; especially when compared to the credit boosters that both wind and solar energy platforms received as a result the IRA.  As previously demonstrated – the correlation between VRE and retail costs for the grid are partially impacted by the index price of natural gas.  Consequently – the presumed benefits of augmented tax credits for both wind and solar assemblies on electricity costs are more difficult to discern than those of nuclear.

Just to summarize a couple of key findings here:

  • The nuclear industry really is the sole determinant of its economics: the more it can efficiently operate; the more attractive its electricity appears to the rate payer

    • The industry looks to still be streamlining its performance; even after decades of operation