Nuclear Power is Better than Fossil Fuels, but Renewables are Better than Nuclear
Nuclear power is much better than fossil fuels, but renewables are better still. Why? Nuclear power is safe and clean, but not cheap and not fast. Wind and solar are safe, clean, cheap, and fast.
Nuclear power is a complex and controversial subject, and in this Deep Dive, we’re only giving a hyper-brief summary. The core thesis of this article, and the one key takeaway, is this: nuclear power is much cleaner and safer than fossil fuels, but it’s also not nearly as good as renewable energy, which is equally clean and safe but much faster and cheaper to build. Both parts of this statement have a wide array of supporting evidence, but both are surprisingly little-known. In this writer’s view, understanding why this statement is true is really important context for any discussion about the future of energy.
Why Nuclear Is Much Better Than Fossil Fuels
Pretty much the core argument for nuclear power is summed up in the chart above from Our World in Data. Contrary to popular belief, nuclear power is one of the safest and cleanest energy sources around, right up there with wind and solar.
Nuclear, wind, and solar (and almost certainly geothermal too, although when this chart was made it was rare enough it hadn’t been formally evaluated) are the only major energy sources with less than 10 tonnes of greenhouse gas emissions per gigawatt-hour and less than 0.1 human death per terawatt-hour of electricity production. Let’s expand on what this means in practice.
Worst-Case Scenario Nuclear Accidents Kill Fewer People Than Fossil Fuel Plants Working Normally
Nuclear power is immensely more safe than fossil fuels, but to most people it doesn’t seem that way. Why is that? Nuclear accidents are rare, big, and dramatic: they cause mass evacuations, make international news for months, and inspire documentaries for years to come. There have really been only two serious nuclear power accidents (Chernobyl and Fukushima, the only two Category 7 events on the International Nuclear Event Scale), but it was enough to brand nuclear power as “very very dangerous” in the eyes of the global public. However, fossil fuel plants functioning normally produce lots of air pollution every day, causing a constant toll of deaths from asthma, COPD, lung cancer, and a wide range of other ailments (see the chart above from the European Environment Agency). This isn’t nearly as cinematic or headline-grabbing, but is killing hundreds of times more people.
A 2021 study from Harvard calculated that fossil fuel-caused air pollution killed over 8 million people in 2018, making fossil fuels responsible for one-fifth of all human deaths worldwide. And that’s just direct fossil fuel deaths, from soot and particulate matter in the air getting in people’s lungs. Indirect fossil fuel deaths, from the wildfires, storms, heatwaves droughts, floods, and famines exacerbated by climate change, are harder to calculate but are also considerable. (For more, check out this amazing map from CarbonBrief of how exactly climate change has influenced specific major weather events in recent years).
By comparison, the nuclear power safety stats in the Our World in Data chart at the top of this article include all deaths from the Chernobyl and Fukushima disasters, which top out in the hundreds to low thousands depending on how you calculate increased cancer risk. (The Three Mile Island meltdown, probably the third most famous nuclear power plant disaster, didn’t even have any directly associated deaths, it was just scary).
It follows that replacing nuclear power with fossil fuels seems safer, but ends up killing people. A 2020 study found that Germany’s phase-out of nuclear power after the 2011 Fukushima disaster directly caused over 1,100 additional deaths from air pollution each year!
“The [German nuclear] phase-out resulted in more than 1,100 additional deaths per year from increased concentrations of SO2, NOx, and PM (particulate matter). The increase in production from hard coal plants is the key driver here, making up roughly 80% of the increase in mortality impacts.”
-Jarvis, Deschenes, and Jha, 2019.
Nuclear Power is a Clean Energy Source, Replacing It With Fossil Fuels Hurts the Climate
It’s fairly common knowledge, but it’s worth clarifying as well that nuclear power is also a clean energy source from a climate perspective, as it doesn’t emit major climate change-causing greenhouse gases like carbon dioxide or methane. See the picture above, of Leibstadt Nuclear Power Plant in Switzerland? That white stuff coming out of the cooling tower is water vapor, and does not contain carbon dioxide. The water vapor is there because like fossil fuels, nuclear power normally generates energy by creating heat that turns water into steam that turns a turbine1. Water vapor is technically a greenhouse gas as well, but we don’t worry about it much because it falls out of the atmosphere as precipitation in a matter of days or so2 (instead of lingering for decades, like methane, or centuries, like carbon dioxide) so we don’t get a big climate change-causing backlog.
It follows that replacing nuclear power with fossil fuels means making climate change worse. Germany’s unnecessary closing of its last nuclear power plants in 2023 has been widely criticized for causing a short-term increase in coal burning. This has slowed decarbonization, with a “treadmill” effect where the new wind and solar power coming online is replacing nuclear power going offline, swapping one clean energy source for another without decreasing the use of fossil fuels. To be fair, Germany has also committed to ending coal burning by 2030 and replacing it with even more renewables, so this isn’t a permanent issue-but it’s already caused years of unnecessarily high emissions during a climate crisis! Not okay!
“On April 16, the day after the final nuclear plants shut down in Germany, the country recorded a carbon intensity of 476 grams of CO2 equivalent for every kilowatt-hour of electricity produced. About half the nation’s electricity came from renewable sources, but coal made up about 30% of the supply.
Meanwhile, in France, only 30% of electricity came from renewables. Add in nuclear, though, and low-carbon power sources made up 93% of the electricity supply. So France’s emissions for every unit of electricity were lower than Germany’s by a factor of nearly 10, at 51 grams CO2-eq/kWh, largely because of its heavy reliance on nuclear power.”
-MIT Technology Review
Why Nuclear Isn’t As Good As Renewables
So nuclear sounds great, right? Why are we even talking about wind and solar power when we’ve had this relatively clean and safe power source since the 1950s3?
Because nuclear power is expensive and hard to build, and renewables are cheap and easy to build.
Renewables Are Fast and Cheap. Nuclear Is Neither.
The above chart displays one of the most important trends in the world4, arguably the major source of progress in the fight against climate change. Both solar and wind power got incredibly cheap, incredibly fast in the 2010s due to rapid advances in technology and a surge in investment: from 2020 onwards, solar has been the cheapest source of electricity in history.
During this same period, nuclear power got more expensive, with the price of building a nuclear power plant increased by 26% from 2009 to 2019, while the equivalent prices for wind and solar declined by 70% and 89% respectively. There are many reasons why this is the case. In large part, it’s due to solar and wind having a higher “learning rate”: companies can build lots and lots of little solar panels and wind turbines and learn how to make them better with each new generation, but can generally only build one or two new nuclear power plants per decade, so it takes a long time to learn from previous mistakes and make technological improvements5. Renewables are also just inherently a cheaper technology in some ways: for example, they don’t need expensive uranium as a fuel input, or water as a coolant. You can build them turbine-by-turbine or panel-by-panel as funds become available, instead of sinking a huge amount of capital into building a nuclear power plant that only generates energy once it’s fully complete. (Also, nuclear power plants are generally extremely heavily regulated, which some have argued is disproportionate and stems from public fear rather than realistic cost/benefit analyses, but is still an on-the-ground reality slowing nuclear power construction that the global energy market has to deal with). As seen in the chart above, all this has made wind and solar substantially cheaper to build overall than nuclear, as well as most other sources of energy.
Unsurprisingly, this has led to the building of lots more wind and solar and very little new nuclear. There is currently only one nuclear energy project under construction in the United States: adding two additional reactors, each with a capacity of “about 1,117 MW each” (or 1.117 GW, note the decimal point instead of the comma, for a total of 2.234 GW), to the Vogtle nuclear power plant in Georgia. It’s massively late and over-budget due to bankruptcies and contractor disputes: construction began in 2009 (!), and it’s cost over $28 billion so far.
For the contrast with renewables, check out this data from the US Energy Information Administration. In the year 2021 alone, 15.5 GW of solar energy (15,500 MW) and 17.1 GW of wind energy (17,100 MW) came online! Combined, that means that in 2021 we got over 14 times more new wind and solar energy than all the new nuclear energy being built during that year6. And that nuclear project had taken eleven years already and wasn’t even finished in 2021! As of July 2023, a headline read “Georgia regulators to hear Plant Vogtle progress report after nuclear expansion stalled again.” The one US nuclear project is still delayed, while renewables are getting built like crazy all over the country.
To be fair, there is some movement on building new nuclear power capacity overseas. As of April 2023, China’s state-run news claims that the country currently has 24 nuclear power units (presumably reactors) under construction, totaling 26.81 gigawatts of capacity, as part of its broad-based effort to move away from coal. And coal-heavy Poland recently signed a $20 billion deal with US company Westinghouse to build its first nuclear reactor by 2033, with a potential six reactors by the 2040s.
But these are tiny sideshows compared to the vast global renewables boom. For context, China is on track to have over 1,200 gigawatts of capacity from renewables by 2025. Also, notice the timescales here: when fishing for nuclear progress news, we end up talking about maybe some nuclear reactors built by the 2040s, while gigantic wind and solar projects are going up every month.
As the invaluable Our World in Data summarizes, renewable energy is rapidly growing worldwide, providing more and more of human civilization’s primary energy every year. Nuclear, on the other hand, is stagnating, verging on declining. It would have been great if humanity hadn’t been so freaked out by Three Mile Island and Chernobyl, and had built a lot more nuclear power plants and a lot less coal plants in the late 20th century. But that’s not what happened. And we now have even better clean energy sources to power us into the future.
(Also note that primary energy is different from electricity, where the picture is even better for renewables, as in the chart above. Primary energy includes all forms of energy use, counting non-electricity things like gasoline burned in cars and coal burned to melt steel in industrial processes).
Prioritizing renewables over nuclear isn’t an ideological decision, it’s an economic one. We should bet on the winning clean energy technologies, the ones that can actually be built at scale fast enough to make a really big difference in the climate fight. And that’s not nuclear.
To sum up: nuclear power is safe and clean, but not cheap and not fast, and it’s not growing as a share of humanity’s energy. Wind and solar are safe, clean, cheap, and fast, and they’re growing exponentially.
Also, “Next-Generation Nuclear” is a Slow, Expensive Distraction (So Far)
2024 Update: I was too pessimistic in this section, as of fall 2024 the TerraPower reactor is getting built in Wyoming and a domestic HALEU supply is on the way from Oak Ridge in Tennessee. Next-generation nuclear looks considerably more viable than I thought when writing this article in 2023: the more clean electrons, the merrier!
A quick interlude: when faced with the glacial pace of traditional reactor construction, nuclear advocates often point to “next-generation nuclear7,” a loose constellation of startups and new “small modular reactor” designs that have promised to upend the energy market with fast, cheap, scalable nuclear energy. However, next-generation nuclear has so far proven to be an expensive distraction that still hasn’t produced so much as a kilowatt of usable commercial power. Maybe someday it’ll work, but this writer wouldn’t bet on the current “next-generation nuclear” technologies ever being seriously cost-competitive with renewables.
The farthest along and generally most acclaimed among the next-generation nuclear hopefuls is TerraPower, a next-generation nuclear startup founded by Bill Gates and widely lauded for its innovative small-scale, liquid sodium-cooled reactor design. However, TerraPower recently had the viability of its entire premise crushed: its bespoke reactor design can only use high-assay low-enriched uranium (HALEU), and the world’s only current source of HALEU is in Russia. Russia! Any technology that depends on sourcing a key component from Russia right now is not going to be fast, cheap, or scalable, to say the least. To their credit, TerraPower announced in December 2022 that would have to delay the opening of their much-hyped demonstration reactor in Wyoming due to this glaring problem.
TerraPower, right now, represents years of work and hundreds of millions in funding that has produced no useful clean energy, and has no prospect of doing so anytime soon. Their goal was admirable, but we can’t afford outcomes like these when we need to be decarbonizing as fast as possible.
To be fair, the Inflation Reduction Act does include $700 million to try and build up a domestic American supply chain for HALEU (the Inflation Reduction Act is seriously just the best). And there’s a company in Ohio trying to make HALEU in America, but they’re currently making tiny non-commercial-scale amounts. Domestic HALEU will take years to make happen-and there’s honestly not that much of a reason to try to make it happen, when we have domestic wind, solar, and battery manufacturing advancing by leaps and bounds.
So What Should We Do?
We should prioritize new nuclear over old or new fossil fuels, but less than new renewables. If a country, utility, or city is looking for a new energy source, renewables are the obvious first choice. However, if a country is looking to replace coal with nuclear (like Poland) and for local political or socioeconomic reasons that looks more doable, that’s an OK second choice, even if it’s probably not going to be as fast, cheap, or efficient as going with renewables.
We shouldn’t shut working nuclear plants down, because they’re a reliable, safe source of low-carbon energy and it’s crazy to take that off the grid in the age of climate change. Activists working to close down active nuclear plants are working to make climate change worse and make everyone less safe, as far from their intentions as that might be. For example, when California’s legislature passed a five year extension to prevent an early closure of Diablo Canyon Nuclear Power Plant, that was a good thing, and the pending federal 20 year license extension for Diablo Canyon should go forward as well. An S&P analysis found that keeping Diablo Canyon online avoids the need for new gas generation and likely lowers California’s emissions 10% by 2030. This paradigm broadly holds for all of America’s other nuclear plants-it’s slow and costly to build new ones, but we should keep all the old ones as long as we can. Relatedly, if a country has shut down perfectly good nuclear power plants for “safety” reasons, it should reactivate them to help achieve cleaner air and lower emissions. Japan is doing this, and hopefully Germany will soon follow.
Countries that have managed to make nuclear power work at scale, like France, should keep doing what they’re doing, and go ahead and build more nuclear reactors if that turns out to be a cheap and feasible way to get more clean electricity. New renewables will probably still be cheaper, though! (France’s history with nuclear power really deserves an article all its own. Basically, they’re the one country to really fully realize the promise of nuclear energy8, due to having a standardized reactor design and many other factors).
Overall, in most circumstances in most countries, we should on balance focus new energy investment in building more renewables, not new nuclear plants. Nuclear is clean and safe, but it’s also slow, expensive, and with extra vulnerabilities that renewables don’t have to deal with9. Nuclear fission for electricity generation is fundamentally a dead end, a “road not taken” that was an underrated and unfairly maligned good choice once, but now has been overtaken by superior technology.
Life would have been better if we’d built lots of nuclear power plants instead of coal plants back in the 1970s and 80s, but now we’d be better off to build lots of solar and wind farms, backed up by battery storage, pumped hydro, and soon enhanced geothermal. Nuclear power plants will be helpful in many places on the margins, and if there are opportunities to cheaply and effectively build more we should take them, but they’ll ultimately be a sideshow in the epic planetwide quest to decarbonize human civilization this century.
This brings up a whole new problem: in the age of climate change, water for those steam turbines is getting harder to come by, with once-reliable sources often unseasonably dry. This is a growing problem for coal, natural gas, and nuclear power plants around the world. Notably, solar farms, wind farms, and batteries don’t need regular water inputs!
Fun fact: the average residence time of a water molecule in Earth’s atmosphere is nine days. For methane, it’s 12.4 years. For carbon dioxide, it can’t be summed up in one value that easily because some CO2 molecules are absorbed by the ocean very quickly while some stay in the atmosphere for thousands of years, but it averages out, broadly speaking, into impacts lasting for hundreds of years.
This writer has seen versions of this argument online a lot. There’s a certain “cool” factor to saying “The political right wants fossil fuels, the political left wants renewables, but I, ultra-logical and above the fray, support nuclear power!” However, this position fundamentally relies on a lack of understanding of the relative costs since renewables’ massive price decline in the 2010s.
I know I’ve shown it in this Substack several times already, but it’s really cool and important! I love this chart!
Check out the work of energy analyst Ramez Naam, inter alia, for more on learning rates.
(15500 MW+17100 MW)/2234 MW=14.5927. As far as this writer can tell, this is an original The Weekly Anthropocene statistic, calculated in-house from EPA and Georgia Power data.
This refers to next-generation nuclear fission: nuclear fusion is a whole different kettle of fish outside the scope of this article. There has been some enticing progress for nuclear fusion at the lab scale recently, and at least one speculative commercial agreement (see The Weekly Anthropocene’s rundown) but it’s still a long way away from being a player in the global energy market or moving the needle on greenhouse gas emissions.
Yes, France’s nuclear fleet has been having some issues lately, with a bunch of them closing for repair and maintenance at the same time. I’m excited about the ongoing French progress towards President Macron’s goal of building lots of offshore wind farms, which will help diversify away from nuclear-related risks while continuing decarbonization. Still, France’s firmly nuclear-focused grid has been a much better deal for the country and the planet than Germany’s counterproductive attempts to shut down nuclear while trying to decarbonize.
For example, the aforementioned supply chain issues for uranium, strong public opposition due to perceived unsafe-ness (even when not logical, this counts as a real problem we have to consider and plan for), and vulnerability to water supply issues.
This is a nice summary article of the issues.
I wrote about this topic back in 2019:
https://www.researchgate.net/publication/342678933_Bad_Faith_Arguments_for_More_Nuclear_Power
One thing you could have talked about more is the structure of electricity markets. Nuclear plants are unlikely to be the best cost solution for much of our electricity generation, not just because it nuclear plants are highly regulated, but also because they're less likely to help with congestion and peak demand issues that are thornier than bulk supply.
As far as regulation goes, PV and wind construction in the US are far more regulated than might be publicly understood. I work in the PV industry and it can easily take 2 years (or more!) to develop, build and commission a 5 MW PV plant (or a <1 MW PV plant, for that matter). To get to 1 GW, which is your standard large nuclear reactor size now, that means it would take 400 years to build 200 PV plants one by one. It's not just learning-by-doing that matters, but also taking advantage of better opportunity costs by varying the scale of plants.
There was a great article in The New Atlantis a few years ago that might interest you:
https://www.thenewatlantis.com/publications/democracy-and-the-nuclear-stalemate
It goes into some detail about how technology lock-in has been a problem for nuclear, among other things.
When Right politics chose Doubt, Deny, Delay as their response to global warming and handed the climate issue and podium and media microphone to Left Environmentalists who were well known to oppose nuclear instead of showing leadership themselves - "you care so much, you fix it" and framing the issue as for, by and about extremists to discredit it - nuclear was the loser. It was the sacrifice people who supposedly like nuclear - just not more than fossil fuels - were willing to make to protect fossil fuels from climate accountability.
"THEY should support nuclear" has become the catch cry of opposition to zero emissions commitments - and what is there for a climate science denier not to like? Unpopular and widely distrusted, is expensive and slow to build and scale up, needs long term serious commitment to emissions reductions and strong government interventions to make happen. And with strong emphasis on how serious and urgent the climate problem is, rather than persistently denying or downplaying.
Sam, at least you appear to be sincere about wanting nuclear to reduce emissions - most of the prominent voices around here (Australia) who say "They should" don't have it as their own policy. They don't really mean it and it is purely rhetorical, a way to square the circle so that even their own lack of such policy is blamed on people who want strong action on climate.
It isn't just that the combination of absolute support for fossil fuels and lack of credible emissions reductions commitments comes across as insincere, it is like the insincerity is feature, not flaw - like they want climate science deniers and opponents of renewable energy they spent so much effort encouraging to know it is to stop renewables, not stop fossil fuels.
Nuclear can overcome anti-nuclear activism, when the largest bloc of support comes out from behind the Wall of Denial.