People of the world, look at France
I am a firm believer that the best power mix anywhere in the developed world is a combination of nuclear power with renewables like wind and solar. My claim has been for a long time that if a country covers some 50% of its power with nuclear to provide baseload capacity and the rest with wind and solar (plus battery storage), it will have not only decarbonised its power system, but is guaranteed to have enough power at all times without the need for batteries or other energy storage facilities—case in point: France.
Ok, French politics is not something that I recommend emulating anywhere in the world, but the French power system may well be something that other countries want to copy. At the end of 2024, France’s nuclear fleet had a capacity of 63GW while its wind and solar fleet had a combined capacity of c.50GW. Hence, France almost reached that 50/50 split between nuclear and renewables.
Now, in this note, I am not going into concerns about the safety of nuclear power plants, nor am I going to discuss whether wind and solar really are as cheap as life-cycle cost analyses indicate. For the record, I think nuclear power is much less harmful and safer than burning gas or coal, while wind and solar tend to be the cheapest sources of energy all over Europe and in much of the developed world.
Instead, I want to focus on the question of whether such a combination of nuclear plus renewables can deliver energy reliably (i.e. without the risk of power shortages or blackouts) and economically. The economics are particularly important because the problem we have in some markets in Europe (notably Germany and Spain) is that sometimes, wind and solar energy are so abundant that there is an oversupply of power and electricity prices become zero or even negative.
That, of course, is good news for industrial consumers, but bad news for the companies that run wind and solar farms, because they start losing money while operating them. If power prices are too low too often, wind and solar energy cannot be run profitably, and the development of new wind and solar capacity will cease.
Nicolas Astier and Frank Wolak analysed the French power grid from 2012 to 2024. One of the key findings of the study is that EDF, which owns and runs all French nuclear power plants, has significantly changed the way they are operated. In the past, nuclear power plants were run close to full capacity at all times. This is why some people still think that the drawback of nuclear power plants is that they cannot change their power output in response to changing supply and demand balances.
But every nuclear power plant has load following capabilities, i.e. the power output can be regulated up or down depending on the needs of the power grid. This variability of nuclear power plants is slow; the power output can typically change by 5% per minute and up to 140MW per minute. Plus, the nuclear power plant needs to always run at a minimum 30% capacity to keep the nuclear chain reaction in the reactor going. But even this relatively slow adjustment, where power output is changing in the period of an hour, is typically enough to ensure there isn’t an oversupply of electricity.
The chart below shows how the Belleville 1 reactor in France has become more flexible in its power output between 2014 (a year with little to no renewable power supply) and 2021 (a year with substantial renewable power supply).
Hourly output of the Belleville 1 reactor in two different weeks
Source: Astier and Wolak (2026)
This flexible approach to running nuclear power plants allows for adjusting energy supply to changing demand and changing supply from renewables to avoid situations of oversupply. In 2024, the nuclear power fleet of France reduced its output by 0.6MWh for every 1MWh increase in power supply from wind and solar.
As a result, power plants (both nuclear and renewables) can be run much more economically, and power prices are higher and more stable. France has many fewer days of low or negative power prices than neighbouring Germany or Spain. To be sure, there still are days when the day-ahead power price becomes negative, but in all of 2024, this was just one day, as the chart below shows.
Day-ahead power price in France as a function of the remaining flexible capacity of nuclear power supply
Source: Astier and Wolak (2026)




I'm thinking there should be some industrial arbitrage opportunity out there. Some kind of energy-intensive production that could fully utilize intermittent negative energy costs. Sadly, capital and personnel costs would negate most applications, it seems. Perhaps very cheap storage à la molten salt batteries is the only perspective. Or smaller-scale pumped storage.
These nuclear plant operating strategies are interesting and I'm all for innovation, but I work in the renewables sector, and my first thought was, "but they weren't designed for this kind of load-following. What is this doing to reliability?"
Sure enough, the problem is mentioned right there in the abstract: "However, we find that an additional load following event is associated with a slightly higher likelihood of a unit failure."
What's that cost? What kind of failures are we talking about? And what about delayed degradation effects? What are the risks?