Nuclear vs. Solar

One has promise for the future. The other is changing our world right now.

Many of my friends in the techno-optimist community spend a lot of time thinking about nuclear power. When the question of how regulation holds back technology and growth comes up, the first example that tends to get named is how safety regulations made nuclear uncompetitive in the 70s and 80s. Jason Crawford cited this in our interview the other day, and Jim Pethokoukis talks about it too. And when it comes to the future of energy, it’s often nuclear that techno-optimists talk about — small modular reactors, the thorium fuel cycle, and so on. Recently, there’s been a huge amount of renewed excitement around fusion.

I have no problem with this — I would love to see technological breakthroughs that solve fission power’s basic problems, and fusion, if it works would be even more exciting. Furthermore, France’s experience powering itself largely with nuclear reactors (thanks to massive government financing and coordination) shows that we could have done so safely and efficiently — and reduced our carbon emissions enormously in the process — had we chosen to do so. Furthermore, scrapping existing nuclear power plants, like the one at Diablo Canyon, is crazy, and will increase carbon emissions for no good reason. As the excellent Jesse Jenkins and his co-authors remind us, “clean firm” power is key to rapidly decarbonizing the grid, and nuclear power is one of the ways we get that.

In sum, I am pro-nuclear.

But I also feel that many of my techno-optimist friends, in their focus on the promise of nuclear, are giving short shrift to the even greater technological revolution happening right under their noses — the unbelievable progress in both solar power and battery storage. If you are a true techno-optimist, you should definitely be excited by this! It’s not the kind of thing that happens once in a generation — it’s the kind of thing that has never happened before.

If you’re not excited about solar and batteries…

From 1976 to 2019, solar modules dropped in cost by a factor of 250. Total costs (including design, marketing, installation, permitting, and so on) fell at a less eye-popping but still very rapid clip. For many years, solar was so expensive that even these massive, sustained cost drops weren’t enough to make it competitive with other forms of electricity. But that is no longer the case. Once again, it is time to post my favorite graph:

Note that nuclear got more expensive over this time, which represents the direct or indirect impact of increased regulation.

This is why everyone is building a lot of solar now. Here’s a graph from a recent report by the Electricity Markets and Power group at Berkeley Lab showing the electricity projects that are currently proposed:

Solar now dominates, while gas and wind are slowly falling off and nuclear barely registers.

It utterly astonishes me that anyone who loves technology and is excited about the prospect of cheap energy wouldn’t be jumping up and down in excitement over this trend. Not only will solar power let us save the world from climate change, but it will actually give us energy cheaper than we have ever had before. The story here is not that solar has inched past fossil fuels in terms of cost — it’s that solar still has plenty of room to get cheaper. A new paper by some economists at the Institute for New Economic Thinking shows that forecasts for solar cost drops have consistently been too conservative; using updated methods, the economists forecast that renewable costs will continue without a plateau for at least the next 10 to 15 years. Other analysts are now issuing similar predictions. (Solar costs have risen a bit this year due to the supply chain disruptions and shortages that are hitting much of the global economy, but other energy sources have suffered even more, and these disruptions should be temporary.)

This will be world-changing. Electricity “too cheap to meter” will enable cheap things like large-scale desalinization, cheap public transit, cheap aluminum smelting, and a whole host of other things that were science fiction in a world of fossil fuels and nuclear. But what will really change our physical world will be the other energy tech revolution: Batteries.

Solar power has two basic limitations — intermittency and transportability. Batteries can potentially solve both of those problems. Lithium-ion batteries, of course, are solving the transportation problem, revolutionizing the auto industry, making drones possible, and generally making energy much easier to cart around. The cost drops in lithium-ion batteries are a more recent phenomenon than in solar, but are every bit as amazing:

And the most amazing thing about batteries is that we keep inventing new kinds. For example, lithium-ion batteries aren’t that good for overnight storage or during a storm, but engineers have now invented iron-flow batteries that may be perfect for those applications. As Bloomberg’s Akshat Rathi reports:

SB Energy Corp., a U.S. renewable-energy firm that’s an arm of Japan’s SoftBank Group Corp., is making a record purchase of the batteries manufactured by ESS Inc. The Oregon company says it has new technology that can store renewable energy for longer and help overcome some of the reliability problems that have caused blackouts in California and record-high energy prices in Europe…The units, which rely on something called “iron-flow chemistry,” will be used in utility-scale solar projects dotted across the U.S., allowing those power plants to provide electricity for hours after the sun sets.

(This doesn’t solve solar’s seasonal problem, of course. But assuming the INET forecasters and others are right about continued cost declines, that can be partially solved by overbuilding; If solar PV systems get cheaper at a rate of 7% a year, then in 10 years you can build twice as much solar for the same cost. That will dramatically reduce, though not eliminate, the long-term need for non-solar “clean firm” power sources.)

Together, solar and batteries will change our world, and not just by allowing us to stop climate change. They’ll make energy cheaper to generate and more easy to store and carry around than ever before. And that is likely to unlock a wave of physical innovation — “atoms” tech, as they say — of the kind we’ve been lacking since oil got expensive and nuclear was regulated to death. Finally, humanity might be able to start using more energy per capita again, instead of having to always do more with less as we have since the early 70s:

And we did this all with the pure force of human ingenuity. If it’s impressive to gain access to near-limitless energy by splitting the atom, how impressive is it to gain access to even cheaper near-limitless energy simply by laying some slabs of melted sand out in the sun? Fundamentally, technology isn’t about finding harder ways to do great things; it’s about finding easier ways.

Nuclear vs. solar

Which brings us to the question of nuclear vs. solar. This isn’t entirely a race, of course; even a mostly-solar grid will still need a modest amount of non-solar “clean firm” generation as backup, and nuclear will be a part of that. But in terms of new generation capacity getting built, it’s not even a contest — as the Berkeley Lab graph above shows.

Nor is this a matter of regulation. Yes, with more sensible regulation, nuclear would be cheaper than it is. But even if deregulation cut U.S. nuclear costs in half — which would make us cheaper than any other country on the planet — it still wouldn’t even come close to beating solar on the margin at this point. When solar rises to be a huge percentage of the grid, its cost will rise due to seasonal intermittency, and at that point nuclear may become competitive; until then, it’s simply no contest.

The INET paper (technical details here) shows this very clearly, using similar forecasting methods to predict the cost of nuclear power that they used for other energy sources, and gaming out the scenario of a nuclear-driven transition:

We constructed an additional scenario in which nuclear plays a dominant role in replacing fossil fuels, but this is substantially more expensive than the baseline. For example, using a 1.4% discount rate the mean cost is about 15 trillion dollars more than No Transition and 27 trillion more than the Fast Transition [scenario].

Nuclear supporters on Twitter — who angrily insist that nuclear, and only nuclear, can save the world from climate change — are thus completely wrong. Relying chiefly on nuclear to eliminate carbon emissions would not only be much more expensive than relying chiefly on solar, it would actually cost money relative to doing nothing at all. Whereas relying on renewables, thanks to the amazing progress in solar and batteries, would actually make us richer, even without taking climate change into account.

But it’s not just cost that makes solar so much more attractive than nuclear right now. Finance is a factor too; nuclear plants are still very big and take a very long time to pay off, meaning that it’s very hard to raise the money to build one without government assistance (small modular reactors would change this, but…they’re not here yet, and won’t be here until the 2030s).

And then there’s the time factor; nuclear plants take incredibly long to build. Worldwide, the time to build a nuclear plant tends to range between 5 and 10 years; this compares to less than one year for a solar plant. In addition to the fact that this makes nuclear way too slow to be our main method of decarbonization, it also exposes nuclear builders to a huge amount of risk. For example, if the trend of the 2010s continues, solar system costs will continue to fall by about 7% a year; this means that if you started building a nuclear plant today, by the time you finished, solar would be 30-50% cheaper than when you started. But the trend could accelerate, making solar an even better bet. And advances in long-duration batteries or other storage technologies during the 5-10 years you were building your nuclear plant could make it obsolete the moment it comes online.

And yes, that long build time is partially due to regulation; but try changing the laws to deregulate nuclear power, and tell me how long that takes.

As for fusion, there’s huge and exciting progress in the field, but still enormous technical problems to be solved. One day fusion may give us something even better than solar (which is actually just spaced-based fusion power beamed to Earth for free!), but if so, it’ll be after the current energy transition has completed.

Anyway, none of this is to say nuclear sucks. It does not! Like I said, we need to keep all our nuclear plants online, to provide clean firm backup for the emerging solar-and-battery-based grid. And we should even build more, though it will be expensive and an uphill battle to do so. And progress in small modular reactors and alternative fuel cycles can’t come fast enough.

But what I am saying is that the great energy transition of the age — the early 21st century triumph of humankind over energy scarcity and energy stagnation — will mostly not be about nuclear. It will mostly be about solar and batteries. In the 1970s and 1980s and 1990s, nuclear lost out to fossil fuels because of regulation; in the 2020s, and probably the 2030s, nuclear will lose out to renewables because that’s simply how the technology happened to work out.

We don’t know which things will work out before we try them. It wasn’t obvious a priori that propeller-driven ships would be better than paddle steamers, or that airplanes would be better than airships, or that internal combustion cars would beat out electrics in the 20th century. We innovated in a bunch of directions, and we found the path that worked, and we took that path.

Because that’s what humans do; we find a way. Technological innovation isn’t about our ability to decide on a particular solution to a problem ahead of time. It’s about our ability to find some solution, by hook or by crook. To explore, to probe nature and scarcity for every weak point, to seize on the smallest crack in its armor and pull ourselves through. Nuclear power is amazing and cool, but solar power and batteries will be the combo that drives this round of energy improvement. And that’s OK. More than OK, in fact — it means we won.

---