The Future of Nuclear Fission: Picking Suitable Applications to Optimize Outcomes

Chris Robinson: Hello, everyone, and welcome to the webinar, “Future of Nuclear Fission: Picking Suitable Applications to Optimize Outcomes.” My name is Chris Robinson. I’m a senior director here at Lux, and I’ll be moderating today’s session. Presenting today is my colleague Karthik Subramanian, an analyst here at Lux Research. Throughout the webinar, you can type any questions you have into the question box on your screen. Time permitting, we’ll get to as many questions as we can. If your question isn’t answered, send an email to [email protected], and we’ll make sure to respond. Lastly, if at any point you have technical difficulties, such as a frozen screen, first try refreshing your browser and checking that you have a strong internet connection. This usually solves most problems. Now, before we jump into today’s presentation, a quick word about who we are. At Lux, we help organizations make more confident innovation decisions about what matters next. As an independent research and advisory firm, we work with many of the world’s leading companies to identify emerging opportunities, evaluate new technologies, and ultimately ensure that they make high-impact innovation decisions. People like Karthik and me have backgrounds in science and engineering. We’re analysts and industry experts who combine original research and methodologies with practical decision-making frameworks to help people like you separate signal from noise. Every day, we help clients answer three critical questions: Where should we focus innovation efforts? Which technologies deserve investment? And finally, which partners can accelerate innovation? Today’s webinar is an opportunity to share some of this thinking with you. With that, let’s jump in. Over to you, Karthik.

Karthik Subramanian: Thank you very much, Chris, and good morning or good afternoon to everyone tuning in today. Nuclear energy has reached its highest level of political momentum in years, but the sector has always faced implementation challenges. In this webinar, we will review what those challenges are, what they mean for the sector’s future, and how nuclear reactors can be deployed to maximize outcomes. Before we begin, let’s take a quick look at the landscape. As you can see, many countries, such as Germany, decided to move away from nuclear energy after the Fukushima incident in 2011. As a result, the number of nuclear reactors scheduled to come online fell drastically. Despite the rise in momentum over the past few years, the number of reactors expected to come online is not high enough to offset the closures anticipated over the next 25 years. Ultimately, this means that approximately 204 gigawatts of nuclear capacity will be retired during that period. To put this in perspective, as of 2026, we have around 375 gigawatts online. Thus, more than half of today’s nuclear capacity is expected to go offline over the next 25 years. One reason nuclear has been difficult to scale quickly—and will remain so in the current market—is its negative learning curve. Generally speaking, the more a technology is deployed or built, the more learning effects arise. Developers learn where to optimize costs and reduce time, ultimately helping them produce products more quickly and cheaply as they scale. Unfortunately, nuclear has shown the inverse effect: the more reactors are deployed, the more expensive and slower they become to build. One major reason for this negative learning effect, or lack of a learning effect, is that increased public scrutiny following nuclear incidents such as Chernobyl in 1986 and Fukushima in 2011 has led to stricter, evolving safety regulations. This increases regulatory costs and prompts design changes, which also results in a lack of design standardization. Consequently, scaling nuclear reactors requires extensive state support. The state must absorb much of the risk associated with building these reactors, while projects continue to face delays and cost overruns. However, as mentioned earlier, nuclear has returned to the conversation, primarily because of recent geopolitical events—for instance, the ongoing conflict in Ukraine or disruptions around the Strait of Hormuz in the Middle East. Natural gas was considered the backbone of the energy transition, complementing the rise of intermittent renewables. Unfortunately, recent geopolitical events and supply-chain volatility have brought nuclear energy back into the conversation. At COP28, 31 countries—highlighted in green on this chart—announced their willingness to work together to triple existing nuclear capacity, reaching one terawatt of operational capacity by 2050. In fact, this number has risen to 38 over the past couple of years. We have seen that scaling nuclear capacity is difficult in the current market. However, several developments aim to help achieve this goal. The first approach is to go modular—small and modular, specifically. Many new developers, especially startups, are targeting the production of nuclear reactors and equipment on factory lines. Series production of these components off-site allows them to be prefabricated and then installed on-site, reducing the bespoke or one-of-a-kind nature of today’s nuclear projects. The second option many developers are considering is applying nuclear energy to industrial heat and, more broadly, cogeneration applications. Nuclear energy is generally thought of as a power source, but the additional revenue generated by selling heat alongside electricity changes project economics. In addition, several power-hungry sectors have emerged in recent years, particularly data centers. These data centers are scaling rapidly and require large amounts of power, making them willing to pay a premium to secure electricity for their operations. This premium also improves the economics of nuclear power generation. Finally, both market and regulatory factors are changing—not only through new targets, but also through licensing reforms intended to alter the current state of nuclear deployment. A good example is the United States, where new licensing pathways have been introduced to speed up deployment. Together, these factors offer a good indication of what the industry lacks and how it plans to address those gaps in order to scale successfully. The question remains, however, whether these factors will have a positive impact on the nuclear industry. That is what we plan to address today. Can these developments change the story of nuclear fission and help it scale? And how can an entity interested in deploying nuclear reactors maximize project outcomes? We have already reviewed the landscape; next, we will examine each factor in detail: what has changed or is changing today, how those changes affect nuclear economics, and what they mean for the sector going forward. The first factor is the small and modular approach, which aims to introduce economies of numbers rather than economies of scale. Nuclear reactors—and sectors more generally—have traditionally targeted larger operating capacity to create learning effects. Recent developments, by contrast, have adopted a smaller approach by building reactors that can be prefabricated on a factory line. This enables series production because all the components are modular and identical. It also allows prefabrication and plug-and-play construction: you can imagine a nuclear project assembled like Lego blocks, with multiple pieces connected to create a system ready for deployment. Together, these factors reduce deployment time. Nuclear reactors take about 10 years to build on average today, while these developers target two to three years. Finally, because the components and reactor are identical across multiple iterations, type approvals become possible; each reactor produced does not need to undergo a separate design assessment to obtain construction or operating permits. On paper, these changes drastically alter how a nuclear reactor can be built and therefore directly affect its economics. Looking at the small modular approach, many startups are pursuing this avenue. Most small modular reactor, or SMR, developers fall into five categories. The dark-gray categories shown here are water-cooled designs: pressurized water reactors (PWRs) and boiling water reactors (BWRs). The categories in green represent fourth-generation reactors, which introduce inherent safety features and are considered safer to operate than water-cooled designs because they operate near ambient pressure and have negative coefficients of reactivity that reduce the likelihood of thermal runaway. These technologies include liquid metal reactors (LMRs), molten salt reactors (MSRs), and high-temperature gas-cooled reactors (HTGRs). What is interesting is that most startups are targeting fourth-generation designs, while established companies are sticking with water-cooled designs. Startups have identified an opportunity to disrupt the consolidated incumbent market by producing reactors capable of supplying higher-temperature heat. Fourth-generation reactors can produce primary-coolant temperatures ranging from 550 to 750 degrees Celsius, while water-cooled reactor designs heat their coolant to only about 350 degrees Celsius. In addition to their safety advantages, this temperature advantage is expected to help generate additional revenue by enabling cogeneration. The opportunity is not only power sales, but also heat sales. This unlocks new offtakers, such as industrial companies that can use both power and heat, improving reactor economics. One example of this dual-revenue model is a Dow Chemical project in Texas. Dow is partnering with a high-temperature gas reactor company called X-energy to build nuclear reactors co-located with an industrial facility, supplying both power and heat. As Dow’s CEO noted, a key reason for pursuing this approach is that nuclear energy has a compact footprint and can produce power and steam reliably compared with some other sources the company has considered. This represents a significant shift from the traditional view of nuclear reactors as solely power producers. Moving on from small modular reactor design and the ability to cogenerate heat, one of the biggest developments in the United States since the new administration took office has been changes to permitting procedures. We are highlighting the United States because it is one of the few countries where such changes have been made to existing licensing pathways. The U.S. Nuclear Regulatory Commission, or NRC, introduced the 10 CFR Part 53 licensing pathway, which allows technology-inclusive permitting. Previously, nuclear reactors were benchmarked against water-cooled designs. That requirement has now been removed, allowing each technology to be considered on its own merits in terms of safety and value. By value, we mean its target application. This reduces redundancies in the licensing process and much of the back-and-forth that occurs when new nuclear technologies are benchmarked against water-reactor designs, whose safety and operating parameters are very different. In addition, the Trump administration is considering changes to certain on-site worker safety requirements in the United States. For example, it has proposed allowing nuclear maintenance workers to spend more time in a nuclear environment than current limits permit. These subtle changes could also affect licensing timelines, licensing costs, and operating costs over the project life cycle. The final factor is the rise of premium offtakers, particularly data centers. News reports over the past year and a half show hyperscalers such as Meta, Microsoft, and Amazon proactively signing agreements with nuclear power plants to support their data-center operations. In the United States alone, Google, Meta, Amazon, and Microsoft have contracted more than 12 gigawatts of capacity. This includes both existing reactors and new builds. For example, Microsoft plans to work with the operators of Three Mile Island to restart an existing reactor to support its data centers. Amazon, meanwhile, plans to work with X-energy to deploy 500 megawatts of X-energy reactors in the 2030s. The trend is not limited to private investor confidence through power purchase agreements; the amount these companies are willing to pay is also noteworthy. Many hyperscalers are willing to pay premiums of up to 30%, with the average closer to 20%, above traditional PPA prices. This premium not only demonstrates their willingness to pay more to secure power quickly, but also indicates that they are willing to bet on nuclear technologies to support future hyperscale projects. Although project timelines remain uncertain, this private investment ultimately lowers risk for developers seeking project financing. Now that we have reviewed the four factors, will they have a tangible impact? Before examining each factor independently, it is important to note that the small and modular approach has not yet delivered major gains in construction time or cost. This graph shows cost-escalation percentages relative to a 100% baseline for three projects: two that are operating as of 2026 and one that is still under construction. These escalations were estimated well before the projects came online. The Chinese Shidao Bay-1 high-temperature gas reactor project, which came online in December 2023, was estimated to be three times more expensive than originally projected when construction began in 2012. The same was true of Russia’s floating small modular reactor project, which was four times more expensive by the time it came online. Argentina’s CAREM-25 small modular reactor is still under construction and was already estimated five years ago to be seven times more expensive. This does not bode well for a sector that claims smaller, modular designs will reduce construction costs. The same applies to construction time. On average, operating small modular reactor projects took about 10 years to build. Even newer technologies faced challenges. China’s fourth-generation Shidao Bay-1 reactor had a cumulative load, or capacity, factor of only 20.7% in 2024. This was primarily because the operator was using it as a test bed to study fuel handling, helium coolant performance, and turbine performance. These examples suggest that today’s small modular reactors are inherently simply smaller reactor projects; modularity and prefabrication have not yet reduced timelines. Fuel supply for advanced reactors is also highly centralized. Most reactors expected to be built in the future will rely on uranium, specifically high-assay low-enriched uranium (HALEU), in which uranium is enriched to contain up to 20% uranium-235. Advanced reactors require HALEU because they are designed to be compact, refueled less frequently, and achieve higher fuel burnup rates. The objective is to build a compact reactor that can run long enough on a single fuel load and produce energy as efficiently as possible. Existing low-enriched uranium, or LEU, which has lower enrichment levels than HALEU, cannot meet these targets. For most advanced reactor designs, HALEU therefore becomes a supply-chain bottleneck. As of 2026, only China and Russia have large-scale capacity to produce and sell HALEU into the nuclear market. In the Western Hemisphere, the United States is still building first-of-a-kind, small-scale projects. In Europe, the United Kingdom is also building an enrichment facility, but these projects are further down the line. Consequently, geopolitical relationships will influence whether HALEU can be obtained quickly enough to bring reactors online on schedule. Fuel recycling has also been considered as a way to reduce dependence on this centralized supply chain. Unfortunately, over the years, recycling has proved much more expensive than enriching virgin fuel and procuring it from external suppliers. This has reduced investment in both enrichment and recycling capacity. Only a few countries—specifically India, Russia, France, and the United Kingdom—have recycling capabilities, while Japan is still constructing its first-of-a-kind recycling facility as of 2026. This creates a chicken-and-egg problem for the fuel supply chain, not only for recycling but also for enrichment. New nuclear reactors require either enriched or recycled fuel, depending on their design. Investors, however, are reluctant to commit significant capital to fuel production because they are uncertain how quickly demand will grow and when nuclear reactors will come online. A lack of investment in fuel production will directly slow reactor startups, which in turn will further discourage investment in the supply chain. Unfortunately, this chicken-and-egg problem will persist and lead to lower nuclear deployment rates. Having reviewed the positive factors and some barriers, what do they mean for advanced nuclear economics? Looking specifically at small modular reactors through a power-generation lens, it is clear that first-of-a-kind reactors are far more expensive because their capital costs resemble those of large-scale reactors and they face the same timeline challenges. The second graph shows that even as projects move from first-of-a-kind (FOAK) to nth-of-a-kind (NOAK) iterations, SMRs make only limited gains. In some regions, they remain more expensive than large-scale reactors because of local financing conditions. Lux’s analysis defines an nth-of-a-kind design as one achieved after 10 iterations of an existing first-of-a-kind reactor. Even after 10 deployments, SMRs still fail to demonstrate the economic benefits they claim under current conditions. Although we have discussed power generation, we must also examine heat because newer nuclear designs are also considered heat sources. If we analyze newer nuclear designs, especially SMRs, purely as industrial heat sources, their levelized cost of heat can be much lower than that of a fossil-fuel boiler. In fact, Lux’s analysis accounts for higher natural-gas and fossil-fuel costs resulting from recent conflict in the Middle East. Although this points to a potential economic benefit from switching to nuclear, the primary challenge is not levelized cost, but system integration. Nuclear steam generators produce very high-energy steam—that is, steam at temperatures ranging from 350 to 550 degrees Celsius and pressures ranging from 70 to about 140 bar. Most industrial applications, such as steam cracking, require heat or steam at much lower pressures. Depressurizing the steam wastes much of its useful energy. Lux has also studied integration in which waste heat is recovered after power production, at the turbine outlet. In that case, the pressure and temperature are very low, and the waste heat would need to be upgraded to meet the requirements of industrial facilities, particularly chemical plants. These findings indicate that system-integration challenges outweigh the economic benefits of nuclear reactors. Smaller reactors—below 100 megawatts electric—could be built to supply these loads, but their cost becomes so high that the economic tipping point is no longer reached. This makes nuclear energy difficult to justify for heat generation alone. Given these power-generation and system-integration challenges, what is the best way to maximize nuclear reactor performance? One of nuclear energy’s primary benefits is its energy density. Nuclear reactors are compact and can avoid some of the land-acquisition and resource-availability challenges associated with renewable energy. A fossil-fuel power plant has an energy density of about 500,000 megawatts per square kilometer. Offshore wind can provide only a fraction of that, requiring a much larger or oversized system, potentially deployed in a different geography with adequate wind resources. Nuclear reactors are relatively geography-agnostic and compact, and can therefore support energy-intensive industries. Another option is to scale in the conventional way. China offers a useful example, as it is building more than half of the nuclear fleet currently under construction. China carefully identifies sites where nuclear plants work well, particularly along the coastline, where water resources are nearby and load centers can be supported. The state absorbs substantial risk by providing low-interest loans. Borrowing costs for nuclear projects are much lower in China than in Europe or the Americas. This type of support is important for nuclear energy’s future success. The most successful projects also deploy multiple units to climb the learning curve quickly. As shown in this figure, the levelized cost of electricity falls as a project moves from one unit to two and then four units. Building multiple units significantly improves project economics and should be considered at the outset. What does this mean for the nuclear sector going forward? Lux expects most large nuclear power plants to operate through restarts or life extensions over the next zero to five years, with new builds coming online only 10 to 15 years from now or later. Given their economic and system-integration challenges, small modular reactors are expected to remain one-of-a-kind projects, with many developers either going bankrupt or pursuing mergers and acquisitions to consolidate expertise and deploy solutions as larger entities. Offtakers interested in nuclear projects should consider large reactors primarily for grid power because other applications have loads that are too small. SMRs, on the other hand, are often too large for a single offtaker’s power or heat needs. Offtakers and project developers should therefore consider SMRs as cogeneration solutions for industrial clusters, maximizing the number of premium offtakers. For example, an SMR could serve a chemical facility, a data center, and a desalination plant. The goal is to maximize the number of customers willing to pay a premium for heat and electricity while ensuring technical compatibility. Thank you for tuning in to this webinar. I would like to close with several key takeaways. First, despite political momentum and policy tailwinds, nuclear energy will not scale at the anticipated rate, and extensive state support is required to bring projects to fruition. The main challenges lie in the supply chain, including fuel, prefabrication, approvals, and licensing. Lux’s analysis indicates that nuclear economics are largely form-factor- and technology-agnostic. Regardless of technology type or reactor size, most projects are expected to remain expensive. These reactors should be chosen for their energy density and their ability to alleviate land-acquisition challenges. Finally, maximizing the outcomes and revenue of a nuclear project requires multiple offtakers. Developers can also consider coupling nuclear energy with storage, including batteries or thermal energy storage, to support other applications such as grid services and frequency regulation. Revenue from balancing markets or grid services can then flow directly into the nuclear project, which is important given its capital intensity. Once again, thank you very much for tuning in. We are happy to take questions.

Chris Robinson: Thank you, Karthik. I’ll take some questions about the presentation. As a reminder, there’s a question box, so feel free to submit them there. If we don’t get to your question on this call, we’ll follow up after the webinar. A couple of questions have come in about microreactors. Can you explain why they were not considered in this webinar and share your thoughts on where they fit into the landscape?

Karthik Subramanian: Yes. We have seen many developments in the microreactor space, particularly the large amounts of funding that startups have raised. We did not specifically consider microreactors in our analysis because our internal analysis suggests that they are far too expensive to deploy for industrial applications and loads. For example, if you have a 250-megawatt industrial load and want to power it with 10 or 15 microreactors, the project economics become unmanageable. We expect microreactors to support primarily defense applications or niche microgrids where the load is small and alternatives are limited. That is why we did not include them in this analysis.

Chris Robinson: Understood. I think we have time for one more question. The webinar touched on system-integration challenges. How do you view progress at the Dow facility, and could that project change the industry’s perspective on nuclear integration?

Karthik Subramanian: The Dow Chemical project is, of course, the first of its kind in seeking to co-locate a nuclear power plant with a chemical facility. We expect that, even if the project obtains construction permits, adding more nuclear reactors on-site will be difficult to justify because system-integration challenges make compatibility across industries difficult. In fact, Lux has spoken with multiple industry stakeholders who suggest that X-energy’s value proposition to Dow is that the nuclear reactor will serve primarily as a source of baseload power and heat. It will not directly support all of Dow’s chemical processes, and the heat will also require some form of upgrading. Nuclear is therefore only one piece of the puzzle. Other alternatives may better match the temperature or demand requirements of different chemical facilities. We still expect this project to remain one of a kind.

Chris Robinson: Thank you. That concludes our webinar for today. The slide presentation and recording will be emailed to you later today. After leaving the webinar, you’ll be prompted to complete a survey about today’s presentation. We appreciate your feedback; it helps us improve and inform future webinars. Please also take a moment to check out our upcoming webinars and website. Thank you very much for joining us, and have a great day.

Karthik Subramanian: Thank you, everyone.