Industry Overview:
Nuclear Energy is back in the spotlight after a lethargic 60 years. Big tech’s interest in SMRs, Small Modular Reactors, to fuel their power-hungry data centers has driven a resurgence in interest in Nuclear Energy. Microsoft piloted the case to reopen the infamous Three-mile Island plant and purchased the rights to 100% of its power for the next 20 years. (Source) Amazon led a $500M round of venture capital into SMR company X-energy and partnered up with Dominion Energy to explore using SMRs to power its data centers. Google is also in on the game and teamed up with Kairos Energy to hopefully get SMRs up and running by the decade's end. (Source)
The key driver behind this wave of next-gen nuclear energy investment stems from the proliferation of AI. Hyperscalers like Microsoft are expecting unrelenting demand for AI services in the upcoming years, which will rely on millions of power-consuming GPUs to crunch quintillions of matrix operations every second. Current estimates put a single ChatGPT query at using 2.9 Wh of electricity. (Source) Considering how many prompts are thrown around, and the tendency for these models to get larger, more complex, and subsequently more power-hungry to run as we innovate, it’s clear that the local supply of electricity will quickly become a bottleneck.
This is where nuclear saves the day as a green, consistent, and compact source of energy. ESG-aware companies like Google do their best at being sustainable where it's convenient for them, but relying on inconsistent renewable energy sources like wind and solar just doesn’t make sense for their data centers that run 24/7. They’d either need to dip back into a fossil fuel-powered grid or install unheard-of amounts of battery capacity on-site to keep those data centers saturated with power. Not to mention the sheer amount of land the company would need to install solar panels or wind turbines. Nuclear is a solution to all of these problems. These SMRs are small and also don’t care if it’s day, night, sunny, cloudy, or windy out. It’ll be able to generate power consistently.
Even though SMRs are the new kids on the block, we mustn’t forget about the old guard. The US has been using nuclear power plants to generate electricity since the opening of Pennsylvania’s Shippingport Reactor in 1958. (Source) Since then, companies like Constellation Energy, Entergy Corporation, Duke Energy, and TVA have been big names within the space, each operating multiple reactors. Barring Constellation's consolidated holdings of 21 of the nation’s 93 reactors, the industry appears quite fragmented.
(Data from Nuclear Energy Institute)
This picture painted by data from the Nuclear Energy Institute isn’t as clear-cut as it seems. They count operators like Virginia Electric & Power as separate from the entity in which it is wholly owned by Dominion Energy. (Source) Another operator, Florida Power & Light, is actually owned by NextEra Energy. (Source) This begs the question, are the little guys actually little? And are there any opportunities for smaller firms to cash in on legacy nuclear?
The answer to the first question is straightforward. 15 operators have 3 or fewer reactors in their portfolio.
However, after digging into whether any of these “Operators” are independent or small enough for mid-market firms to take a sizable stake in, the answer is unsurprisingly no. Even operators at the bottom of the list like Union Electric and Wolf Creek Nuclear are subsidiaries of larger utilities worth tens of billions. If you don’t have billions of AUM to spare, getting in bed with any of the legacy nuclear operators is nearly impossible. On a forward-looking basis—excluding the advent of SMRs—new legacy-scale power plants like Plant Vogtle are even putting corporations into bankruptcy. Toshiba paid $3.7B to walk away from the project after the plant's overrun construction budget bankrupted its subsidiary Westinghouse Electric. (Source)
This is not to say that it’s not possible to get involved in legacy nuclear as a sponsor. Carlyle had its foray into the industry via credit investments in nuclear plants, Brookfield bought a stake in Westinghouse in its emergence from bankruptcy, and Apollo is directly involved in financing the construction of Hinkley Point C, a new reactor for the Hinkley Point Nuclear Plant in the UK. (Source) Even still, it looks like sponsor involvement in legacy nuclear still seems concentrated at the mega-fund scale.
Mid-Market Investment Fronts for Exposure to Non-SMR Nuclear
Picks and Shovels of Nuclear: Uranium
Even though the natural monopoly-like characteristics of the nuclear power industry restrict active sponsor involvement to the mega-fund scale, industries tangential to nuclear power plants exist. Assuming an aversion to the venture-type risk that comes with investing in SMR companies, the construction of new SMRs and legacy-scale plants will likely be a tailwind for Uranium, which is in an industry with lower barriers to entry—on a relative scale of course. There’s even a public company that is carrying out this strategy right now, Uranium Royalty Corporation. The strategy Uranium Royalty employs is in the name, royalties. Their mission is to gain exposure to Uranium prices without imposing on themselves the operational and execution risk that a company would normally have to take on. They accomplish this feat by buying royalty interests in uranium mines and also trading physical uranium from time to time. The company, worth about $300M, buys royalties in mines owned by other companies, and the upfront price for these royalties can be as low as $300,000. Those prices are real, the company just signed a $300,000 deal with a collective of mines in Spain called The Salamanca Project in return for a 0.375% net smelter return—meaning they get 0.375% of the project’s net revenue less transportation and refining costs. Currently, they own royalty interests in 19 different mines, owned by various mining companies, across the US, Canada, Spain, and Namibia. The royalties can take form in a myriad of ways: net smelter returns, net profit, gross revenue, gross value, etc. Uranium Royalty Corp has shown that there are no restrictions on where, who, or how you structure a uranium royalty. This information is all made available through their 40-F, where management discusses why they see upside in risk-adjusted exposure to Uranium, in addition to some other strategies they are looking to employ in the future.
Nuclear Energy and the AI Play: Flow Control
While the uranium industry is quite literally the picks and shovels of nuclear power, another picks and shovels play is the industrial equipment required for data centers. If you’re underwriting a second golden age for nuclear via the effects of AI workloads on the power grid, you must also underwrite the need for industrial equipment at the data center level. In this case, perhaps a flow control company like Pfeiffer Vacuum might be of interest. They specialize in flow control generally but specifically offer solutions for the semiconductor, refrigeration, and energy industries. (Source) These 3 industries are crucial to the data center play. Of course, more AI demand means more data centers, but that comes hand in hand with more chip demand and additional semiconductor fabs to service it. Fabs need flow control equipment for a myriad of mission-critical processes—which you can find on the Pfeiffer Vacuum website. Refrigeration is another tailwind for flow control at the data center, 100% of the electricity pumped into any chip is eventually expressed locally as heat. Cooling is the utmost priority for anybody who wants their new Blackwell chip to function at anything higher than 1% utilization. The advent of new data centers will no doubt require flow control for both air and liquid refrigeration techniques. On the nuclear energy front, flow control equipment is necessary in every core function. The centerpiece of how a nuclear reactor generates power comes down to directing steam through a turbine, none of which is possible without specialized flow control equipment. Seeing as though a flow control company like Pfeiffer Vacuum is only partially exposed to the brunt of a possible downturn in data center demand, it opens the door for mid-market players to gain exposure to the AI-Nuclear energy growth trend without taking on VC levels of risk with SMRs. A strategic way to get involved in a company like Pfeiffer Vacuum that is worth $1.5B would either be to acquire a significant minority stake, or more realistically gain exposure via debt instruments like convertible bonds or even mezzanine financings.
Sources
