⚛ How small nuclear reactors fit into my pro-progress Up Wing vision
Also: 5 Quick Questions for ... energy expert Adam Stein on nuclear energy regulation
Quote of the Issue
“Technology is a gift of God. After the gift of life, it is perhaps the greatest of God's gifts. It is the mother of civilizations, of arts and of sciences." - Freeman Dyson
The Essay
⚛ How small nuclear reactors fit into my pro-progress Up Wing vision
Item: Oklo, a nuclear-fission startup backed by Sam Altman, plans to go public through a merger with his special-purpose acquisition company, company officials said Tuesday. The deal would add to recent SPAC mergers involving nuclear companies and test investor appetite for clean-energy startups, which surged in 2020 and early 2021 before falling out of favor. Companies such as Oklo trying to build a new generation of smaller nuclear power projects must prove they can deliver on time and on budget, unlike the fleet of large nuclear plants that preceded them. Altman, the chief executive of OpenAI—the artificial-intelligence startup behind the viral chatbot ChatGPT—said the nuclear-energy industry can make electricity that is “a way better deal than anything else out there.” …
Dozens of developers globally are testing small modular reactor designs, though there are no SMRs making electricity in the U.S., and none under construction. Supporters say the smaller-scale reactors could prove cheaper and faster to build than their massive predecessors; skeptics say the effort is a gamble on a technology with unproven economics. - The Wall Street Journal, July 12, 2023
Why is it often the case that if you know someone’s views on, say, tax policy or Social Security reform, you can make a pretty accurate guess about their views on a host of other policies and issues? It’s a phenomenon, one more about intuition than deep policy analysis, that economist Thomas Sowell attempts to explain in his 1987 book, A Conflict of Visions. As Sowell writes, “Visions are like maps that guide us through a tangle of bewildering complexities. Like maps, visions have to leave out many concrete features in order to enable us to focus on a few key paths to our goals.”
Following on that definition, the eponymous “conflict” is one between the Constrained Vision of human nature and societal processes (we are flawed and self-interested creatures whose limitations must be controlled through a limited government of checks and balances whose institutions evolve over time) and the Unconstrained Vision (human nature is both fluid and improvable/perfectible through collective action via expansive government action and intentional policymaker design). The classic examples of these differing visions: The American Revolution represents the Constrained Vision, while the French Revolution represents the Unconstrained Vision.
This notion of visions as roadmaps came back to me when heard about the Oklo news. But first things first: This seems like another bit of encouraging news on the small modular reactor, or SMR, front. This financial fusion Oklo could use $500 million in AltC cash, according to the WSJ, to expand its business, including its supply chain and procurement operations, and build a pilot-scale production facility for its Aurora reactor. Oklo’s current plan is to build a plant at the Idaho National Laboratory by 2027 and deploy two commercial plants in Southern Ohio by 2030. In addition, Oklo must get regulatory approvals from the U.S. Nuclear Regulatory Commission. The company is also developing nuclear-fuel recycling technologies with the U.S. Energy Department and U.S. national labs so “used fuel” can power its advanced reactors.
As it happens, JPMorgan last week published some takeaways from a chat with Oklo CEO Jacob DeWitte that gets at the company’s business model and future plans:
Oklo looks to decrease nuclear development risk by maximizing hours in the factory versus on the construction site. Moreover, smaller builds leads to lower capital risk, allowing for different risk parameters, flexibly and potential less outsourcing. As opposed to selling reactors, Oklo seeks to retain reactor ownership to sell power and heat to utilities and large energy customers under an IPP-type business model, with PPAs unlocking project level financing similar to renewable builds.
The company initially targeted 1-2MW designs, and now focuses on 15MW based on customer interest, with ultimate capabilities to scale to 100MW. Extensive engagement for licensing continues with the NRC following Oklo’s initial denial in 2022, which carried meaningful pandemic impacts to the in-person review component of the NRC’s process.
We observed optimism around more recent process trends, where the company continues to address final application gaps from its first attempt. Moving forward, Oklo targets a pre-application process commencement in early 2024 and full application filing later in 2024. Separately, the company anticipates its first commercial project will achieve initial power in 2026-2027. Despite the ability to efficiently utilize 90-95% of the energy content from uranium, Oklo noted that securing a supply of enriched uranium remains an across the board challenge. As a result, Oklo looks to establish a nuclear fuel recycling facility pilot by the end of the decade. Finally, Oklo continues to develop a growing pipeline, as evidenced by the recently announced partnership with the Southern Ohio Diversification Initiative.
All good stuff. But things are just as interesting once you get past the headline news, things that brought to mind the Sowell “conflict of visions” framing, as well as the potential of the technological advances we’re currently seeing across of broad range of sectors. Check out this quote from Altman, who, by the way, is also an investor in nuclear fusion startup Helion Energy (which already made a power supply deal with Microsoft):
Altman said the goals of making artificial intelligence and energy cheap and abundant are tied together. “The AI systems of the future will need tremendous amounts of energy and this fission and fusion can help deliver them,” Altman said, adding that he thinks that as AI advances it will contribute to nuclear-system designs.
The power of combination
Bingo, that’s exactly what I’ve been talking about. AI/machine learning/generative AI and nuclear fission/fusion represent two kinds of general-purpose technologies — indeed, smarts and energy may be the ultimate GPTs — that are combinatorial technologies with each other, augmenting each other. Progress multipliers, if you will. This symbiotic effect, especially with GPTs, is a key component of my three-dimensional Up Wing Vision (rising above mere Left and Right) of a more abundant and prosperous future for humanity.
The concern about how AI uses more energy than previous forms of computing — training a single model uses more electricity than a hundred American homes use in an entire year — isn’t much of a concern if that energy is from nuclear rather than fossil fuels. And AI deep-learning models have been taught to control the superhot plasma inside a tokamak nuclear fusion reactor. And in another WSJ piece, Helio CEO David Kirtley pointed out that Helion employees have been experimenting with ChatGPT to see how it might accelerate engineering work.
But just as the combinatorial effects of AI smarts and nuclear energy are embraced by my Up Wing Vision, one could see those effects rejected by those with a degrowth, anti-progress Down Wing Vision. From that perspective, any technology that promotes greater consumption of the finite resources of Spaceship Earth must be opposed. Back in the 1970s, some environmental opposition to nuclear energy wasn’t that it couldn’t be made to work safely and economically, but rather that it could!
An unhelpful alliance
If so, nuclear-powered techno-capitalism would only accelerate a global “limits to growth” resource crisis. It’s not hard to imagine the AI doomers joining forces with the anti-nuke environmentalists to form a broad new Down Wing front, all sharing a devotion to the safety-first, “mother may I?” precautionary principle. It’s also been my experience that at least some AI doomers and energy scarcity activists aren't fans of humanity having a future as a multiplanetary civilization — especially since it might be AI enabled and nuclear powered.
Hard pass. We’re going to need more clean energy to raise global living standards down here and also journey out there. We’re lucky to be living in a time where some critical technologies seem to be advancing a such a pace to turn those dreams into reality.
5QQ
💡 5 Quick Questions for ... energy expert Adam Stein on nuclear energy regulation
Adam Stein is the director of the Nuclear Energy Innovation program at the Breakthrough Institute. Recently, I sat down with Adam to ask him a few questions about nuclear energy regulation, small modular reactors, and what Congress should be doing to make a nuclear renaissance possible in the US. Check out his answers below.
1/ Vogtle Unit 3 is the first reactor that has started with the Nuclear Regulatory Commission and moved all the way to completion. How big a deal is that?
It's a very big deal. It's the first reactor that the NRC has ever licensed that has gone all the way to commercial operation. It's also the first reactor to come online in the US in decades. Each unit represents more than a gigawatt of new power production that's clean around-the-clock energy. And it shows that even though a first-of-a-kind of reactor can have some challenges, and there have been many with that project, we still can build new nuclear plants.
2/ Why are there no small modular reactors running in the US today? Is it about the technology or the regulation?
Depending on how you define "small modular reactor," there are hundreds running in the US right now: on military sites. They power all of our subs and our aircraft carriers. However, regulations through the NRC for licensing new designs are challenging. The regulations were intended for very large light-water plants like Vogtle and not for very different technologies, which is what most SMRs are. Licensing new SMRs is challenging both for the developer and for the NRC itself internally, because they have to justify why the new applications deserve exemptions from the existing regulations. In short: It's for many reasons, but regulation is definitely one big component of it.
3/ When I read about the regulation of SMR reactors, it always seems like there's more to the story. It's never quite done. The NRC has approved the NuScale SMR design. Is that sort of the end of the regulatory story? Does that mean that we'll see one of these reactors plugged into the grid sooner rather than later?
No. NuScale has decided that they will never build the design that they actually have an approval for, which is a 50-megawatt design. Instead, they have decided to up-rate that design to 77 megawatts and build it in different configuration of the plant. They are currently working through the process of submitting a new application for that design of plant instead. But instead of doing what's known as the design certification, which is what they did initially, they're doing a standard design approval, which is a different pathway through the licensing frameworks.
Each developer is forging their own pathway through the existing regulations for how they think would be best to license their particular technology. Because they are each blazing their own trail, there isn't a significant amount of learning between them on what's the most efficient way through the regulations, nor does it a lot of opportunity for the NRC to revise those regulations down to be logical for everybody. Because they're not experiencing a unified approach to licensing that they can then simplify for everybody. It's a little bit of everybody is doing their own thing.
To have a major nuclear scale up, the regulations just don't allow for that efficiently, whether you like the existing frameworks or not.
4/ If we're hopefully on the verge of an expansion of nuclear power in this country, do we have the regulatory framework in place that will allow that? Or does it need to be substantially different?
It needs to be substantially different no matter how you look at it. There are developers that are currently going through the licensing process that say, "No, we can get through the existing frameworks as they are." But that would be for their initial design. It would be for a one-off plant. It has no efficiency if you consider building a dozen or a hundred of these plants. You would license each one in isolation. To have a major scale up, the regulations just don't allow for that efficiently, whether you like the existing frameworks or not. And in part because of the challenges of adapting the existing frameworks to new technologies easily, in 2019 Congress passed the Nuclear Energy Innovation and Modernization Act telling the NRC to design a new framework from scratch that is fundamentally different in its approach to regulation, that would be flexible and allow innovation for new designs. That's still working its way through the regulatory process right now and so isn't available to current applicants.
5/ Is the NRC fundamentally about regulating nuclear power with the goal of having lots of safe nuclear power? Or is it fundamentally, at least the facto, about there just not being any more nuclear power?
The NRC's mission, as they interpret it, is to provide public safety and eliminate environmental impacts from nuclear power. The Atomic Energy Act specifically says "reasonable assurance of adequate protection" to the public and environment. They can be overly conservative and risk averse with "reasonable assurance." It is not specifically defined anywhere, so they interpret it in a very risk-averse way. And they don't consider the potential benefits of nuclear energy. They consider that to be the Department of Energy's job: to promote nuclear energy.
I think the NRC is currently trying to figure out how to license new reactors. Some of the staff at the NRC want to see a nuclear renaissance, but the regulations aren't designed to allow that easily. On the other hand, they don't consider the benefits to the general welfare of the public, which is also in the Atomic Energy Act. They don't actually make that calculation in their decision-making. I've been pre vocal to that that they should, and if they did their risk aversion would be skewed towards building more nuclear, because that's what actually protects the public more.
Bonus: What should Congress be doing right now to promote the construction of new nuclear?
The number one thing that Congress can do is to clarify the NRC's mission does include considering the general welfare of the public and protecting the environment, which as I previously stated, is already in the Atomic Energy Act. But the NRC doesn't think it's their job, so Congress should clarify that it is in fact their job. And if the NRC did consider what's best for the public in decision-making, the process would be substantially different.
There are other things that Congress can do as well to enable new nuclear, such as help to enable the construction of a domestic fuel supply to get away from challenges that we have with oversea fuel enrichment, to revise other areas of the NRC's operations such as the advisory committee and reactor safeguards. Another area that I've been working on that I think would be extremely helpful with building out new nuclear in a timely manner is using early site permits to preauthorize construction of nuclear plans on retired fossil fuel facility sites. So those sites are already ready to go, the community already supports the project. All you need is a utility that this says they want to build.
Micro Reads
▶ Bill Gates isn’t too scared about AI - Will Douglas Heaven | Gates doesn’t dismiss existential risk entirely. He wonders what may happen “when”—not if —“we develop an AI that can learn any subject or task,” often referred to as artificial general intelligence, or AGI. He writes: “Whether we reach that point in a decade or a century, society will need to reckon with profound questions. What if a super AI establishes its own goals? What if they conflict with humanity’s? Should we even make a super AI at all? But thinking about these longer-term risks should not come at the expense of the more immediate ones.”
▶ Nick Clegg: Openness on AI is the way forward for tech - Nick Clegg, FT Opinion |
▶ Elon Musk Announces New Company xAI as He Seeks to Build ChatGPT Alternative - Tom Giles, Bloomberg |
▶ Is the AI-fueled stock market rally already over? - Capital Economics | There are several reasons why AI-related stocks may have come off the boil. First, some investors could fear that valuations are getting stretched. The price over estimated earnings ratio of the S&P 500 IT sector is nearing its pandemic peak, which was its highest level since the Global Financial Crisis. Second, there seems to be growing willingness from public authorities to regulate AI. In June, MEPs adopted the EU Parliament’s negotiating position on the AI Act, opening the way for talks between EU countries on the final form of the law. And the UK government announced that Britain will host this autumn the first global summit on AI, with the objective of agreeing on “safety measures to evaluate and monitor the most significant risks from AI”. Third, excitement over artificial intelligence “chatbots”, such as ChatGPT, seems to have waned a bit. According to internet data firm Similarweb, worldwide web traffic to the ChatGPT website dropped by nearly 10% in June. Finally, rising real rates may have prompted investors to revise down the premium they are willing to pay on AI stocks. It was, after all, quite unusual that the IT sector kept outperforming the overall S&P 500 index in May while real rates were rising in the US.
▶ Will superintelligent AI end the world? - Eliezer Yudkowsky, TED2023 |
▶ AI Safety and the Age of Dislightenment - Jeremy Howard, Fast.ai |
▶ Google and IBM push for increased govt resources to support AI innovation and transparency - Nihal Krishnan, Fedscoop |
▶ Bottlenecks: Sectoral Imbalances and the US Productivity Slowdown - Daron Acemoglu, David H. Autor, and Christina Patterson, NBER |
▶ Google’s AI for Medicine Shows Clinical Answers More Than 90% Accurate - Davey Alba and Julia Love, Bloomberg |
▶ Deep-sea mining is a watery wild west - Anjana Ahuja, FT Opinion | One draw is the Clarion-Clipperton Zone, an area spanning at least 4.5mn square kilometres in the equatorial Pacific. The abyssal plain, more than 4km down, is studded with trillions of potato-sized “polymetallic nodules” containing manganese, nickel, copper and cobalt, which are used in rechargeable batteries for electric vehicles. The nodules build up around small objects such as shells or teeth over millions of years. The seabed elsewhere also features seamounts, or underwater mountains, draped in metal-heavy crusts; and sulphide ores laid down around hot, deep-sea vents. The crusts are rich in precious metals such as platinum and molybdenum; the ores contain copper, gold and silver. All are sought-after commodities in electronics, construction and transportation.