The Underlayer
Monday 30 September 2024
A blockbuster deal between Microsoft and Constellation Energy to restart Unit 1 at Three Mile Island looks like a real signal to the future of nuclear energy. The deal will put back in operation a single 819 MW capacity reactor that had been operating as recently as 2019, and Microsoft will purchase 100% of its generation. Significantly, that represents 7.3 TWh of annual clean electricity output that currently does not exist on the US powergrid. Compare that to the annual output of the nation’s largest utility-scale solar plant, Solar Star in California, that generates about 1.6 TWh per year from 747 MW of capacity.
Notice the unsurprising difference in annual generation between Solar Star and TMI, despite having closely matched nameplate capacity. That characteristic was likely highly motivating for Microsoft, which no doubt understands that it would take 5 Solar Stars to create the same output as a single unit at TMI. While it’s not clear how much of TMI’s generation will actually be used locally by Microsoft, the company can simply broker the unused generation to the US grid. And in doing so, Microsoft can improve its own clean-energy balance sheet.
The US tech giants have of course been sourcing new, clean energy for years—either by contracting for new solar capacity, or siting data centers in clean energy corridors like the eastern portion of the Columbia River, where Oregon and Washington have built significant wind capacity. But with the advent of AI, the pace of power demand growth is further boosted—on top of electric vehicle growth, and electrification of buildings. The pace therefore may be too much to match with renewables only—and this mirrors the same problem on the global level: total demand is starting to grow too quickly to catch with wind and solar alone.
Cold Eye Earth has warned for two years now that natural gas, both in the US and globally, will continue to win market share as global power generation growth outdistances clean energy growth. This warning has also applied to coal, which also keeps hitting new all time high in the global powermix. The solution offered here has centered on a new role for nuclear—not as the leader of grid decarbonization—but as a supporter. Which is why the TMI deal is fascinating: it’s precisely the solution we need.
The signal coming from the Microsoft-Constellation deal is that the world is about to turn towards the unused capacity represented by economically recoverable nuclear. These sites also contain an increasingly precious resource: an on ramp to the grid, at a time when large bottlenecks are preventing large volumes of new power from being deployed. We will still need to build many Solar Stars, and we are! Solar and offshore wind are going gangbusters from China to Texas. But the opportunity to quickly add a very large 5-10 TWh of clean power from resurrected nuclear is too juicy to let slip by. The revived TMI nuclear station, which will be renamed the Crane Clean Energy Center, is expected to be operational in relatively short time, by the year 2028. This counters one of the more tedious objections to nuclear that flogs the admittedly long development timelines in the OECD which do, unfortunately, sometimes run towards 20 years. This popular retort however tends to leave out the much shorter newbuild timelines in Asia, at roughly half that length.
There are many well known voices in the world of energy analysis that have been adamant for years that we can decarbonize global power through wind, solar, and storage alone. You wonder, can they hear the signal now starting to arise from nuclear?
The restart of Three Mile Island would be helped, in part, by tax incentives already embedded in the Inflation Reduction Act. This may be why owners of other, shuttered US nuclear plants may be looking at restarts. This summer, for example, it was reported that Next Era Energy, a developer of large renewable projects, was taking a second look at its 600 MW Duane Arnold Energy Center located north of Cedar Rapids, Iowa. In Michigan, meanwhile, the US government is providing a loan to Holtec International to restart the 800 MW Palisades Nuclear Generating Station.
Let’s put these potential re-openings in context. For example, combined wind +solar power in the US is on course to grow from 663 TWh to 755 TWh in 2024, an advance of 92 TWh. Restarting all three of Palisades, Duane Arnold, and Three Mile Island would roughly add 20 TWh to the US grid mix. The proportions here are illustrative. First, the world is unlikely to decarbonize electricity if we were to restrict ourselves to using nuclear, only. This is why wind and solar are the leaders, and will not be surpassed. They are cheaper, but mainly, they are faster—which only adds to their ROI and affordability. But at the same time, unless we build or resurrect some nuclear, then the gaps will be filled by natural gas and coal. Wouldn’t you rather the gap be filled by nuclear? If the answer is no, you must still be telling yourself something false: that the world will decarbonize by adding renewables alone, without additional help on the demand side, and additional help on the supply side—in the form of nuclear.
Why is it so hard for the world’s energy system to attack the fossil fuel underlayer, which presses onward despite larger and larger additions of clean energy? An initial, general answer is that the fossil fuel foundation is composed of multiple cohorts or generations, which do experience turnover, but not at a particularly fast rate. There are coal plants that were built just after WW2, there are natural gas plants that were built in the 1970’s, and then subsequent generations of both as we move through time up to the present. In the US, when that oldest generation of coal plants started to retire about 15 years ago, it happened just as utility scale wind and solar were becoming economically viable. Unfortunately, neither was able to scale up quickly enough to take 100% of the market share left behind by coal, so natural gas filled in the remainder.
The world right now is doing a better and better job each year covering marginal growth with either clean, or cleaner, energy sources. But we don’t seem to be able to work away at all against the bedrock layer of coal, natural gas, and oil that still governs the world’s economy. This means that, overall, renewables have built up their own layer that roughly sits on top of legacy energy. This layer grows larger, it grows quickly, it’s without question its own form of accomplishment, but global emissions tell the tale of the bottom line: we have slowed, but still not halted emissions growth.
A second answer is that for the world to attack the underlayer, it would have to retire economically viable natural gas and coal generation. And in the transportation sector, the world would have to retire viable buses, trucks, and cars. Now you see the problem. Sure, US power generation from coal has been cut in half the past 15-20 years. But the other half remains, and is composed of much younger generations of coal infrastructure. This is even more true—more painfully true—with natural gas, which has seen rapid capacity growth the past decade.
Another way to see the problem is that once you have retired a ton of very old power capacity, the average age of your coal plants and natural gas plants starts to decline, sometimes alot. And that leads to a slowdown in retirements, which the EIA has said has started this year, 2024. If we imagine a world of stasis, in which 100% of all new power demand is matched by wind and solar, but also in which legacy fossil fuel power generation never retires because it’s too young, then there would be one and one way only to make further progress: shuttering fossil fuel generation well before it reaches senior citizen status.
In this context, can you see how pointless it would be to prospect for peak emissions, especially if one assumes those emissions would eventually turn downward? They wouldn’t. And this is a problem that Cold Eye Earth identified years ago: even in regions that see consumption peaks, effectuating actual decline is hard to come by.
For example, there are several institutes in the world that have been calling for peak emissions in China for years, but this project has no method, and is nothing more than serial guessing. Worse, peak emissions prospectors always imply that declines will follow peak when we know that this is not the case at all. Plateaus and barely declining consumption are the norm after peak. The Paulson Institute recently weighed in on this question and correctly identified both of these problems.
Beijing officialdom so far doesn’t seem to share in the enthusiasm of early peaking, or they’re certainly not showing it. In August, the National Energy Administration (NEA), the closest thing China has to an energy regulator, released a white paper that offered little clarity on the road ahead for China’s energy transition. When asked about an early emissions peak, an NEA official made it clear that China is holding to the 2030 peaking target.
While China will certainly earn plaudits for achieving peak carbon ahead of schedule, what matters much more is how fast emissions fall after the peak. If the country’s emissions level stagnates after the peak, then hitting that point will quickly lose its punch.
A great example of the powerful fossil fuel underlayer is the use of coal in global power generation. Initially coal use in global power peaked in 2014. (This is also when wider coal use in all applications also peaked—at least initially). And while after 2014 there is no question that coal growth in global power slowed substantially, we are a full 5% above those decade-old levels. Importantly, the ten year period from 2014 to 2024 has seen a starburst of new wind and solar globally, especially in China which is driving the bulk of the coal growth.
If your forecast is that emissions from global power generation are about to enter decline, or are about to peak and decline, you’ve got to come up with a more plausible explanation than “the world is building an insane amount of new wind and solar.”
Here is another example of an underlayer that’s been in decline for some time, yet whose progress is slow considering the total set of policies adopted in this region. Despite terrific uptake of electric vehicles, buses, trucks, and high petrol taxes and cities well served by public transit, the EU’s oil consumption 21 EJ in 2023 is not that much lower than 2014’s, at 22 EJ. Again, there is no argument that the EU has tackled oil consumption growth. But it looks like a weak victory.
Finally let’s take a look at yet another underlayer story, in US power generation. If you toggle over the histograms (bars) you can see that roughly for the past 5-6 years the US is holding sources ex-wind+solar to a flatline. Again, that is a form of accomplishment, but given the extraordinary retirements of coal capacity the past decade, we must conclude the US has already let one opportunity pass by when it built a ton of new, young, natural-gas fired power over the past decade.
There are several routes to attacking the fossil fuel underlayer. First, we could build so much wind, solar, and storage (on a stepped up basis that mimics a war-time effort) that it would render younger coal and natural gas plants uneconomic, by forcing their capacity factors lower. Once you sentence coal and natural gas plants to a life in which they only run some of the time, you will have ruined their economics, triggering a new wave of retirements. Second, we could tax natural gas and coal usage at a punishing rate, which would accomplish relatively the same thing—though, a tax on fossil fuel combustion doesn’t necessarily translate into a bigger buildout of clean generation.
Perhaps another route would be, for a time, to step up nuclear deployment. How so? By resurrecting shuttered coal plant real-estate to host new nuclear instead. Remember, old coal sites possess the coveted on-ramp to the powergrid already in place. The distribution of these plants geographically is also widespread. A recent US DOE report actually looked at this potential, and found that 145 coal sites could host from 128 GW to 174 GW of nuclear. If just half of that potential was realized, along with the continued buildout of wind and solar, we would definitely see a steady erosion of the dreaded fossil fuel layer.
Further reading: Evaluation of Nuclear Power Plant and Coal Power Plant Sites for New Nuclear Capacity (opens to PDF). This follows an informational posting from the DOE earlier this year, 8 Things to Know About Converting Coal Plants to Nuclear Power.
—Gregor Macdonald