Dear Aventine Readers, 

We’ve written a lot about AI, the massive amount of data it requires and the mad scramble for new sources of energy it has caused, like a collective scrounging for quarters in the couch. One of the most novel sources of new energy might lie in restarting dormant nuclear power plants. On the face of it, it seems pretty straightforward — just turn everything back on for clean, always-on power. The reality is far more complicated, burdensome and costly. And we’ve never tried this  before in the U.S. Read on to learn about what’s required and what success would mean for our energy needs. 

Also in this issue: 

I hope you all had a lovely Thanksgiving and thanks for reading, 

Danielle Mattoon,
Editorial Director, Aventine

Subscribe

Subscribe to our newsletter and be kept up to date on upcoming Aventine projects

Almost unimaginable even five years ago, the prospect of restarting shuttered nuclear power plants — something that has never been tried in the U.S. — is now edging closer to reality. In late 2022, the energy company Holtec International and the governor of the state of Michigan announced that it planned to restart the recently closed Palisades nuclear power plant on Lake Michigan, a project that is now moving forward with the intention of getting it back online in October 2025. This September, Constellation Energy announced its intention to resuscitate the one usable reactor at the notorious Three Mile Island facility in Londonderry Township, Pennsylvania, which stopped producing power in 2019. And in October, John Ketchum, CEO of the energy company NextEra, said that his company was considering reopening its Duane Arnold nuclear power plant in Palo, Iowa, which stopped producing power in 2020.

The appeal is understandable. Nuclear energy is carbon-free, provides huge quantities of always-on power — or baseload power, in energy industry parlance — that complements renewables such as wind and solar, and restarting a plant should be far cheaper than building a new one. But these facilities are complex systems and must meet strenuously high safety standards to reopen.

“It's not a washing machine that you flip a switch on,” said Eugene Shwageraus, professor of nuclear energy systems engineering at the University of Cambridge, U.K. “Restarting, it’s not some trivial thing.”

Aventine spoke with experts, including Shwageraus, to understand the complexities of and potential barriers to restarting a nuclear power plant. They described daunting schedules of paperwork and maintenance, as well as high levels of uncertainty around costs and timelines, to bring these plants back online. They also explained that reinstating nuclear power plants is unlikely to become a trend — only the three plants named above are relatively straightforward to restart — but expressed enthusiasm about the additional power these facilities will provide to the grid as well as the potential for using infrastructure at existing sites to efficiently deploy small modular reactors or SMRs.

Not just any power plant will do

The desire to restart nuclear plants reflects a confluence of factors that has massively increased the demand for electricity in the U.S.: the electrification of vehicles, the rising use of air conditioning to battle hotter summers and — most consequentially — the proliferation of data centers needed to support the soaring demands created by AI. The International Energy Agency predicts that global annual electricity consumption from data centers, AI and cryptocurrencies could double by 2026 compared to 2022, from 460 terawatt-hours to as much as 1,000 TWh — about as much electricity as Japan consumes per year. 

Against this energy-hungry backdrop, nuclear power plants that had been closed over the past decade for economic reasons developed a fresh allure. Many had years of operational life left in them or as Shwageraus put it, “juice in the tank.” So if the economics of energy pricing could be made to work, restarting them would make financial sense and help ease energy demands. “Since it takes so much time and money to build a new nuclear power plant, the proposition to restart idle reactors is attractive,” explained Jacopo Buongiorno, professor in nuclear science and engineering at MIT. For context, the two new reactors built at Plant Vogtle in Waynesboro, Georgia, which went fully online this year and contribute up to 2,200 megawatts of power to the grid, cost $35 billion and took 15 years to build. Constellation, meanwhile, plans to spend about $1.6 billion on getting the 835-megawatt Three Mile Island reactor back up and running, and Holtec has a loan guarantee of up to $1.52 billion from the U.S. Department of Energy for the work it’s undertaking at the 805-megawatt Palisades plant.

It is the underlying dynamics of energy pricing that determine whether restarting these plants makes financial sense, and technological shifts over the past decade — even in the last two years — have turned those dynamics on their head. “The big tech companies, be it Microsoft or Amazon or Google, are willing to pay a premium above market price for electricity that is carbon-free and 24/7 reliable,” said Buongiorno. That much is evident in the deal that Microsoft has struck with Constellation: It has committed to exclusive purchasing rights of all of the electric generating capacity from the single Three Mile Island reactor over the next 20 years, according to the companies. While the details of the agreement are undisclosed, Morgan Stanley analysts estimate that Microsoft will purchase power at a cost of $98 per megawatt-hour, compared to market rates that are around $50 per megawatt-hour. Palisades also has power purchase agreements in place with regional electric cooperatives. “It provides a known stream of revenue,” said Patrick O’Brien, director of government affairs and communications at Holtec. Without a power purchase agreement in place, he added, “you're basically bidding into the energy market,” which is much harder to predict and makes financial forecasting much more difficult. Constellation did not respond to a request for comment on this article. 

The number of suitable nuclear plants for restarts is, however, limited. For a restart to be viable, said Scott Burnell, public affairs officer at the Nuclear Regulatory Commission, a plant needs to sit in a particular sweet spot: It would typically have time left on its original licence granted by the NRC; it should not be too far into the shutting down, or decommissioning process, as every step makes it more difficult to restart it; and it must be in a good state of repair to avoid renovation costs. Palisades, Three Mile Island and Duane Arnold are “the easiest paths to restart,” said O’Brien. “There are others that might be feasible, but you know, it's probably going to be more time and money.” 

File your papers, fix your faults

For Holtec and Constellation — and potentially NextEra in the future — the hard work begins now. Not only must these companies ensure that their respective plants are fully restaffed and able to produce power, they must also satisfy the Nuclear Regulatory Commission that the facilities are safe. When a nuclear power plant stops operating, its owner submits paperwork that changes the licence it holds with the NRC, meaning the plant can no longer officially run; both Holtec and Constellation are seeking to reverse that licence change, a request that has never before been granted. This is an onerous twin-track process, according to Burnell, that requires what can only be described as a mountain of paperwork as well as the physical work required to ensure that the plant is up to scratch, both happening at the same time. 

The NRC will require the operators to submit painstaking documentation on how the plant operates, how it would deal with unusual events that would require a plant to shut down quickly, and detailed emergency preparedness plans with the surrounding communities, among other contingencies. In a public meeting between the NRC and Constellation held on October 25 that Aventine attended virtually, NRC employees repeatedly urged the energy company to engage with the requisite paperwork early to avoid delays, underscoring the amount of work ahead.

In terms of ensuring that the plant is in perfect working order, the many hundreds (or thousands) of complex systems inside a nuclear power plant — from the reactor core to steam generators, pumps and motors to concrete containment vessels — need to be checked by the operator and the NRC and brought up to standard before the agency will provide an operating licence. The amount of work needed to achieve that will vary. There are “millions” of maintenance tasks required to keep a plant running, said Shwageraus, and as plants come to the end of their operational lives, many will have been “either delayed or cancelled altogether” to optimize the economic performance of the facility. That, combined with the length of time a plant has been sitting idle, could mean that some of the facilities need extensive refurbishment. 

While both Holtec and Constellation will have invested significant time and energy in plant inspection ahead of making public commitments to reopening, there is still much uncertainty over what exactly must be done so that each plant can operate safely again. A good example is the steam generators at Palisades — effectively huge heat exchangers that transfer heat from the main reactor to the turbine system and must do so without allowing any radioactive water to escape. Holtec and the NRC announced earlier this year that after close inspection this equipment, which contains hundreds of small pipes, would require extensive renovation in order to satisfy the NRC that the hardware is safe. 

Such issues are likely to be more common and more complex the longer a plant has been idle, the experts said. For context, Palisades stopped producing power only in May 2022, and since then Holtec has been “doing what's needed to keep the plant in a kind of condition that you could reverse the shutdown decision,” O’Brien said. Three Mile Island, meanwhile, has been closed since 2019.

Ultimately, costs could spiral dramatically — and there is a chance that the NRC won’t be satisfied with the condition of the plant and will deny a licence to operate. “At the end of the day, there's always that risk,” said O’Brien. “[But] that's why you build in contingency, that's why you build in time, and that's why we're going very systematically through the building.” He also argued that the restart represents an opportunity for deeper maintenance than the plant may have received if it had simply continued running: Typically, a nuclear plant might run for 18 to 24 months with a short break of a month or two during refueling when maintenance can occur. Now, he said, operators might have a year or more to address long standing maintenance issues, if there are any. 

Old sites with new tricks

If all goes well and the Palisades, Three Mile Island and Duane Arnold plants can all be brought back online, the U.S. energy grid will have gained more than 2.2. gigawatts of carbon-free, baseload electricity — about enough to power up to 2 million homes — provided at a fraction of the cost of building equivalent new nuclear facilities. Even so, far more capacity will be required to keep up with demand, which is projected to increase by 27 percent in 2050 compared to 2022.

Given its bipartisan support, nuclear energy in general looks set to become an increasingly important part of keeping up with that demand. Recent legislation known as the ADVANCE Act directed the NRC to take steps to support the industry, including increasing staffing to accelerate the review process and hastening reviews for new kinds of small or advanced nuclear reactors. It is currently unclear what decisions President Trump will make around nuclear policy, though more affordable options such as restarts are expected to be unaffected. “I think the reactor restarts are safe,” Buongiorno wrote in an email to Aventine. “[But] more government spending on new nuclear is at risk.”

Still, there are other routes to building out more nuclear capacity across the U.S. At the Palisades facility, Holtec plans to exploit the efficiencies of existing infrastructure and augment the 800-megawatt reactor with at least two small modular reactors, each capable of producing 300 megawatts of electrical power. These SMRs may produce only about a third of what a large nuclear reactor produces, but they are somewhere between a tenth and a quarter of the physical size and are designed to be factory-assembled and delivered to site, theoretically making them easier and cheaper to manufacture and install. 

It’s also anticipated that many power plant sites not suitable for restart — where reactors are being actively decommissioned, say, or licences have expired — as well as old fossil fuel power plants, will house new nuclear facilities, many of which are likely to be SMRs. “You can use cooling towers, you can use transmission lines, you can use access roads, you can use administrative buildings. And, more importantly, you could rehire part of the workforce,” said Buongiorno. “If you can't bring back a shut-down or mothballed plant, the next best thing you can do is to build a new one at an existing site.”

 Image by Unsplash

Want bigger, sweeter tomatoes? CRISPR may have the answer. When it comes to tomatoes, there’s often an inverse relationship between taste and size, with larger tomatoes, which are easier to cook with, package and store, lacking the flavor of smaller ones. But research published in the journal Nature has shown that this problem can be overcome by using CRISPR to delete two genes that are associated with the regulation of enzymes that produce sugars. When those genes were disabled in a variety of tomato called Moneymaker, sugar levels in the fruit were increased by up to 30 percent with no decrease in yield compared to the unedited fruit. The researchers are now working with commercial partners to bring the CRISPR-edited tomatoes to market, according to New Scientist. But there is a potentially larger benefit: the edited fruit could significantly reduce the quantity of crop required to make tomato-based products like ketchup and sauces, allowing more efficient use of land, water and other resources. This CRISPR breakthrough follows others that enable crops like rice to grow with less water, and corn to better survive storms. The benefits of these modifications may prove relatively modest in isolation, but together they could dramatically increase the efficiency in our food systems.

Bionic limbs that work like the real thing are edging closer. The act of picking up a warm cup of coffee is a surprisingly complex sensory experience: A hand comes in contact with smooth hard porcelain, detects heat, and — thanks to your nerves — communicates this information to your brain, letting you know not to drink it too fast. For individuals whose lost limb has been replaced with a prosthesis, those sensations are just a memory. But this story from Nature describes several new avenues of research that are bringing the reality of more natural and responsive bionic limbs ever closer. Brain-computer interfaces, a technology that has made significant strides in the last two years, are increasingly able to communicate sensations of touch to a user’s brain, providing them with more feedback and greater control. Separately, skin grafts onto a limb stump — say, ankle skin on a knee stump — show signs of being able to rekindle some sensory experience from the lost limb, giving people a sense that a prosthetic is more like part of their body. Finally, various forms of “e-skin,” which attempts to cover prosthetics with so-called smart materials that can provide rich and complex sensation, have taken huge leaps forward, and may be able to achieve levels of sensation that go beyond what humans are currently capable of — which would, for example, allow braille to be read more efficiently. So far these technologies remain confined to labs and small-scale trials, but they offer hope for the near-term future.

AI may beat quantum computing to the punch. The shorthand sales pitch for quantum computers typically suggests that these devices, once built, will dramatically speed up certain kinds of calculations, enabling advances in fields such as chemistry and materials science. What the pitch generally fails to address is how long it will take to build useful quantum computers, exactly what problems they will really be applied to, and whether other emerging technologies could help find solutions in the interim. This story from MIT Technology Review digs into these issues and explains how advances in artificial intelligence could ultimately solve many of the problems assumed until recently to be solvable only by quantum computing. Three factors stand out as particularly important. First, quantum computing, while making impressive progress, is still far from being a commercial technology, and it could be a decade or more before it becomes particularly useful. Second, in recent years it has become apparent that the set of problems for which quantum computers will offer meaningful advantages in speed isn’t as large as people previously thought because quantum computing has inefficiencies that undermine some elements of its advantage. And finally, the rate of recent progress in AI has been ferocious, and machine learning approaches are already tackling some of the simpler quantum problems, achieving results that seemed impossible several years ago. None of this is to say that quantum computing is irrelevant — it could still help humanity solve some otherwise intractable problems, especially in complex scientific areas such as high-energy physics — but it might not be necessary in such a hurry. 

The Making of a New American Epidemic, from Noema
4,500 words, or about 18 minutes

Blistering heat. Frequent drought. Mile after mile of dry red earth. And seemingly nonstop construction. The combination has created a new health hazard in parts of the American Southwest, where dust can carry spores of a potentially deadly fungus, known as Coccidioides. When inhaled, the dust containing the spores can cause Valley Fever, a fungal infection of the lungs often mistaken for pneumonia and occasionally even for cancer. The disease can ravage the lungs and, if untreated, spread to other parts of the body, sometimes requiring a lifetime course of antifungals. And the threat of Valley Fever is intensifying. Not only are areas of the West getting drier, with more dust kicked up by increasing development, but rising temperatures and shifts in rainfall patterns mean that the areas of the U.S. in which Coccidioides could reside might more than double by the end of the century. This story attempts to understand exactly how Valley Fever spreads — and what we can do about it.

Why we now think the myopia epidemic can be slowed — or even reversed, from New Scientist
2,500 words, or about 10 minutes

Rates of shortsightedness are rising, with some estimates suggesting forty percent of the world’s population could be myopic by 2050. Many factors, this story explains, have been proposed as explanations, but one of the most plausible is quite simple: People now spend less time outdoors. While it isn’t clear why time outdoors reduces the incidence of myopia, it seems to be associated with the intensity of light, which is typically ten times greater outdoors than indoors. Still, can people be expected to spend two hours of their day — the amount of time the studies recommend as beneficial — outdoors? Probably not, which means that we need to do a better job of replicating outdoor light — which is more intense and has a different spectrum than typical indoor light — in our homes, schools and offices, as well as employing a variety of eye drops and peripheral-vision-altering eyewear to delay the onset of myopia in some people. Until those interventions gain traction, therapies — both established, such as laser eye surgery, and cutting-edge such as repeated low-level red light therapy — may be required more often.

The counterattack, from Science
3,600 words, or about 13 minutes

Sufferers of autoimmune disease face a hard truth: Their immune systems, which are supposed to fight off threats to their organs and tissues, instead attack their healthy cells. Given this dynamic, throwing yet more immune cells into the mix might seem like a strange idea, yet that is exactly what an emerging field of research is attempting. This Science story reports on the progress in using CAR-T therapy — a treatment already approved for use against some blood cancers — in autoimmune disorders. The idea is fairly simple: Patients with autoimmune disorders have a version of their T immune cells genetically engineered so that they are able to fight off rogue B immune cells that cause the autoimmune response. In early trials, success rates have been high. But, as the story points out, significant questions linger, from how long the effects last to whether CAR-T therapies, until recently some of the most expensive medical treatments on the planet, can ever be made cost-effective enough for wider use.