Westinghouse, headquartered in Pennsylvania and jointly owned by private equity group Brookfield and uranium miner Cameco, could not have hoped for anything better than an executive order from the US president to build 10 nuclear reactors in the US with the stated aim of reviving nuclear energy as a source of clean energy (i.e., without CO2 emissions). The order is motivated by an expected construction boom in the United States, which is estimated to increase electricity demand by 25% on its own, but above all by the need to power more and more data centers dedicated to the development of artificial intelligence (AI), a sector in which the United States risks losing the challenge with China. In addition to the DeepSeek case, which is only a wake-up call, there are much more serious signs that justify well-founded concern about China’s strategic positioning on this front: the construction of ten new large nuclear reactors last year alone by the yellow giant. In fact, the President’s executive order refers to 10 new reactors at an estimated cost of $75 billion, which Westinghouse plans to produce in their entirety using AP1000 models, for which it claims to have solid experience in design, construction, delivery, and on-site installation. The reactor has the capacity to produce 1,000 MW per year, enough to power approximately 500,000 homes.
This is a unique opportunity because the company has virtually no rivals. China General Nuclear Power Group and Rosatom, a leading Russian state-owned company in the field of nuclear technology, are automatically excluded, mainly for geopolitical reasons. GE Vernova (in a joint venture with Hitachi) has not produced large reactors for decades; similarly, France’s EDF withdrew from the US market for large reactors a decade ago, and South Korea’s Kepco has only one project approved by the US authorities but has never actually built one.
However, there are also some shadows behind this opportunity, consisting of construction delays compared to the planned schedule and, above all, significant cost overruns. The AP1000 at the Vogtle nuclear power plant in Georgia took seven years longer than expected to complete, with costs exceeding the budget by $17 billion. This has dampened enthusiasm for large reactors, especially among major investors such as Microsoft, Google, and Amazon, which are interested in powering their AI research, and above all, it risks losing this challenge with China as well.
But how is the rest of the world moving? Just last week, we witnessed a historic turning point at the World Bank, which lifted its ban on financing nuclear energy projects: the last (co-)financing provided by the institution for the nuclear power plant on the Garigliano River in Campania (Italy) dates back to 1959 and amounted to $40 million. Since then, fears of nuclear proliferation for military purposes (the war had just ended) and various disasters, starting with the Three Mile Island meltdown in the US in 1979, followed by Chernobyl in 1986 and most recently Fukushima (2011), have statutorily prohibited this important financier from supporting any nuclear project in the world. So what has changed? First of all, pressure from its major shareholders, primarily the United States, but no less importantly Japan, joined by Germany, which has long been opposed to this type of initiative, with the recent change of government. And then there is the forecast doubling of electricity demand in developing countries expected by 2035: a very short time frame indeed. The World Bank has entered into a cooperation agreement with the International Atomic Energy Agency (IAEA), which means that nuclear power is also being targeted for ‘rapid and deep’ decarbonization. This means that nuclear energy (precisely because of its absence of CO2 emissions) is perceived as clean energy. We therefore expect that green certifications such as guarantees of origin (or their international counterparts, I-RECs) will soon be issued for each MWh of electricity produced by nuclear power. Of course, nuclear power also has another fundamental characteristic: it is a continuous source of production, which helps to create a constant base supply that contributes to stabilizing distribution networks and can be integrated into its peaks by alternative production or modulated according to the erratic and unpredictable nature of these sources.
But beware of premature enthusiasm. The World Bank’s openness is actually very cautious. While the agreement with the IAEA demonstrates a particular focus on environmental issues, it also demonstrates a form of protection that the institution intends to adopt against a possible escalation that could favor the use of these funds for military purposes. To corroborate this caution, we believe it is important to mention that the World Bank will not finance new plants, but only the reactivation of shut-down plants or the modernization of functioning but outdated ones, in addition to explicit support for small or micro reactors, which we will discuss in a moment.
This decision is also intended as an incentive for other institutions, such as the African Development Bank and the Asian Development Bank (ADB), to do the same. In fact, the latter has also taken action, spurred on by its two largest shareholders: the US and Japan. This institution also had a ban on nuclear financing, which its leaders are now considering lifting. Japan’s moral suasion seems particularly effective. After Fukushima Daiichi, Japan shut down all 54 active reactors that supplied 30% of the country’s energy needs. To date, about a quarter of the reactors have been reactivated, and the aim is to make extensive use of nuclear energy between now and 2040. The main motivation lies in the growing demand from the semiconductor industry and the development of artificial intelligence.
Last but not least, Europe. The old continent has an obvious geopolitical and strategic problem: its energy dependence on Russia. While it has found a solution to its fossil fuel supply (primarily gas), which should be eliminated by 2027, the uranium supply chain is much more complex, and dependence on Russian nuclear technology is proving problematic. The EU has a fleet of 101 nuclear reactors, 19 of which are Soviet-made VVERs concentrated mainly in Hungary, Slovakia, the Czech Republic, Bulgaria, and Finland. Repairs, upgrades, and replacements of modules for these reactors are Russian industrial secrets, creating total dependence. In addition, Rosatom is extremely competitive in both uranium extraction and conversion and enrichment, and VVERs only work with enriched uranium.
In reality, when it comes to the construction of large power plants, the EU can count on solid French (and German) know-how, while—in terms of the uranium production cycle—despite lagging behind its Chinese and, above all, Russian competitors, it is rapidly catching up, mainly with companies such as Orano and Urenco. Although it does not have uranium mines, the Union is diversifying its supply by entering into agreements with Kazakhstan, Canada, and Niger (the world’s seventh largest producer). The problem is therefore contingent and consists of the opposition, albeit justified, of the above-mentioned countries that depend on Russian technology. To convince them, the Union has two weapons at its disposal: sanctions or trade incentives. The first measure has two flaws: it requires unanimity and, like other sanctions, it is not very effective (import bans could be circumvented with China’s support). What remains is a policy of duties and trade tariffs capable of generating cash flows to be reinjected into the system to finance the breakaway from the Russian supplier (and this does not require unanimity to be approved).
So far, we have talked about large nuclear reactors. Yet, until a few days ago, there was much talk about small reactors, the new generation ones. What happened to them? They have simply been ‘overshadowed’ by recent news (some of which is noteworthy, as we have seen), but they are still there, they have not disappeared, and indeed the debate, especially among major financiers, sees them as protagonists on a par with the many companies we mentioned earlier, starting with Westinghouse and its eVinci project.
Imagine a battery the size of a shipping container that lasts up to 20 years, can be buried underground or installed on the seabed, and generates 20 MW of power per year (enough to supply electricity to around 20,000 homes). This is a micro-reactor, i.e., a nuclear power plant that does not require dedicated personnel, is not installed near waterways or reservoirs for cooling purposes, does not occupy large areas of land, and does not need to be defended from military attacks. A micro-reactor does not have all these ‘flaws’; it only produces less energy, but there is nothing to prevent multiple micro-reactors from being used together to produce the same power as a large nuclear power plant. Furthermore, micro-reactors are modular, which means that parts can be replaced either on site or in the factory (but the reactor can be transported), and they are not cooled with water but with liquid sodium, so they have zero water footprint. Finally, it works with fuels such as Triso (in the case of eVinci, which has a very high melting point and produces little radioactive waste) or other fuels that can be regenerated (using self-fertilizing reactors). All these characteristics also apply to Small Modular Reactors (SMRs), which are also new-generation nuclear reactors but develop around 300 MW of power per year, or about one-third of a large conventional nuclear plant. Finally, they reduce the two classic vulnerabilities of large reactors: delivery delays and budget overruns. Cybersecurity is certainly one of the critical factors, because the plant is monitored by sensors that send information via the internet to monitoring centers and therefore requires a secure communication channel. Physical security is less of a problem (for example, in the event of bombing during a war—something we certainly cannot overlook these days), as these are small plants and therefore easier to manage from this point of view.
According to experts, these reactors seem to be particularly suitable for powering or co-powering data centers, as well as offshore oil (or gas) extraction platforms (which cannot always be connected to the grid) or mines and all those sites such as ports, terminals, petrochemical plants, islands, and remote locations where gas or diesel fuel supplies are prohibitively expensive. We believe that the current 80 projects underway for the construction of SMRs, in light of the situation described above, have much broader areas of application and therefore sources of demand. This is also demonstrated by the recent examples of TerraPower and Oklo, two companies dedicated to the development of SMRs to support the development of AI, funded by Bill Gates and Sam Altman, respectively. This is an investment theme that we find interesting, but certainly not yet mature. Development times of 3-5 years from now are expected. In the meantime, we intend to explore in subsequent posts some technical characteristics that we consider particularly relevant for making prudent investment choices.
Disclaimer
This post expresses the personal opinion of the Custodia Wealth Management staff who wrote it. It is not investment advice or recommendations, nor is it personalized advice, and should not be considered an invitation to carry out transactions on financial instruments.
