Venture Capital's Nuclear Error?
Venture Capital's Nuclear Error?
Correcting the investor’s blind spot to build a true atomic age.
This piece is a bit special, as I co-authored it my friend Carl Perez. I credit Carl for my interest in the nuclear opportunity, and for everything I know about it. From our days as roommates, we have been engaging in late night conversations about the molten salt reactor technology he was developing. Having launched his first nuclear startup at the age of 21 in 2015, he was an early mover in nuclear tech and dealt with a much less favorable environment. Today, you can find Carl building Exodys, a new startup promising to disrupt the nuclear waste recycling sector. In full disclosure, I have personally invested in Exodys and Carl, BWXT, and Sprott’s physical uranium trust.
It makes us especially happy these days to see nuclear energy poised for massive growth globally. What seemed a contrarian bet, when our conversations started in 2017, has now become an obvious and bipartisan reality for the future of energy.
In the US, nuclear energy has become the chosen enabler to fuel AI’s insatiable energy needs. The bi-partisan ADVANCE Act was just voted into law. Talen Energy sold a nuclear-powered data center to AWS earlier this year. Microsoft has secured 5GW of energy and signed a partnership with fusion developer Helion Energy. We’re at a multi-decade high point in public and government excitement for nuclear energy. 25 countries pledged to triple their production capacity during the most recent U.N. Conference of the Parties (COP) to fulfill Net-Zero objectives by 2050.
From the earliest stages to public markets, nuclear has completed a stellar return to fashion. Year to date, the S&P500s top performers after NVIDIA and SuperMicro are 2 nuclear utilities. YC’s request for startups includes advanced energy generation from SMRs and fusion.
The future is bright for this primary energy source in the United States, with 70+ advanced reactor (AR) concepts spanning several neutron spectrums, fuel and coolant types. However, this also creates challenges for the investor community in identifying which companies to fund and support.
Of the reported $4.4B in funding received by nuclear startups across 181 funding rounds according to Crunchbase, a vast majority has been focused on reactor construction startups.
These reactor vendors present risk profiles that make them a challenging venture investment. Indeed, the vendor approach does not enable enough value capture to justify the risks taken:
Longer timelines with the potential for cost overruns and project delays
Regulatory risk at the permitting level, which counters the “move fast and break things” tech mentality
Concentrated industry with established incumbents, large suppliers, risk-averse utilities and government friction throughout the supply chain – from procurement to certification
As a result, a large percentage of these companies present multi-decade timelines to first revenue, with limited visibility and control over said timelines. revenue will remain limited until a reactor is actively running. Quite the task, if you ask.
There are no shortcuts when it comes to reactor commercialization. Nuscale, which has raised $469M in total, took 16 years to go from pre-application to approval from the NRC. The two largest awardees of the Advanced Reactor Demonstration Project solicitation from the U.S. Dept of Energy – who are deemed closest to construction of their first reactors – are TerraPower and X-Energy, respectively founded in 2008 and 2009. In addition, these companies have been fully financed throughout the years by its serial and successful founders, Bill Gates and Kam Ghaffarian. Kairos Power has been particularly effective since it was created in 2016 and has secured several NRC permits to construct their first demonstrations near Oak Ridge National Laboratory. Nevertheless, Kairos Power is a spin-off from a DOE and university-led Fluoride-Salt-Cooled High-Temperature Reactors research program that was initiated in 2010.
Newly launched startups are using the same playbooks as the previous generation. But if this approach showed so much potential to disrupt the industry, shouldn’t we expect more investment activity from incumbents, as a way to preserve their previous market share? After all, some of the most active biotech funds are the venture capital arms of established pharmaceutical companies (Leaps by Bayer, Pfizer Ventures, Sanofi Ventures…). Biotech presents some interesting similarities given the reliance on core IP and delayed time to revenue. But this hasn’t happened in the case of nuclear investing.
The answer is, because they don’t need to:
While the startups take all the regulatory risk to get a reactor to market, their off-the-shelf approach requires them to work with trusted, established providers. The off-the-shelf playbook popularized by SpaceX creates stronger dependencies on a limited set of suppliers. If a reactor startup needs a power conversion system, a merchant provider will build it for them. This provider will get paid upfront, and they’ll capture most of the financial value while the startup absorbs all the regulatory and execution risk. No bueno.
It would be very surprising for a utility to partner with a new entrant with little to no construction track record. They are inherently risk averse. No one wants to be blamed for choosing the small, risky startup if something bad happens, or costs go through the roof. In the world of nuclear energy generation, distribution is a MOAT, and the industry is deeply protected by barriers to entry.
Therefore, the companies best positioned to capture the ramp up in nuclear reactor construction are likely to be the incumbents. Startups lose the advantage of speed and execution, as they depend on the very same competitors to their own reactors to get their essential components. It could greatly delay the technology’s deployment, and subsequent promise of an abundance of energy for all. Thankfully, a lot of the answers we are looking for in a better model can be found by looking into the past.
Looking at the past to better understand the path forward
GE-Hitachi, Westinghouse, and BWXT are reactor vendors with decades of experience designing, testing, delivering, and maintaining nuclear power systems. GE-Hitachi and Westinghouse service the commercial sector, while BWXT primarily supplies reactors for the U.S. Navy’s submarines and aircraft carriers. With that said, all three companies share two commonalities: 1) they are the only companies currently completing reactor sales, 2) they are more than reactor vendors.
The last point regarding value proposition is perhaps the most important element. All three companies provide components, fuel services, and plant equipment. Therefore, whenever they sell a reactor, these firms generate multi-phase revenue:
Before reactor construction: revenue from procurement
During reactor construction: revenue from technology licensing/plant construction performance fees
After construction and during plant operation (40-80 years): revenue from fuel supply and maintenance services.
Nevertheless, how significant is the difference between non-reactor vs. reactor revenue?
The two Westinghouse AP-1000s in Georgia encountered significant cost overruns, however a recent MIT study indicated that “the next AP-1000” could cost $6.8 billion. Assuming a 15% performance fee (very optimistic), this would result in ~$1 billion in reactor sale revenue over the years of construction. If the NPP operates for 80 years, it will require 20 core loads of ~ 100 metric tons (MT) of fuel. With low uranium prices at $42.6/lb (although $84.25/lb as of 06/30/2024), fuel would cost $1,663,000/MT and result in $3.3 billion in fuel revenue. Without accounting for all other services (e.g. components, refueling outages, O&M support), it is clear that selling a reactor is the path, not a finality, to significantly more revenue from the same client.
For a scenario in which a startup sells a reactor, the performance fee linked to its sale and delivery is only a fraction of the revenue generated by its consortium partners. They provide the fuel, components, and plant services. However, all risks associated with the project are concentrated within the startup’s balance sheet. They develop the reactor, pay for the development of its components, absorb the reactor design’s regulatory risk, and obtain the clients – yet are the least financially rewarded.
A different approach to the opportunity:
To solve this risk/reward misallocation, there are two potential approaches:
1) Increase incumbent participation in R&D stages through in-kind contributions for initial phases and strategic investment closer to deployment.
2) Startup reactor vendors must limit their “off-the-shelf” approach to ensure the revenue they fought so hard to capture does not also end up on the shelves. This could also be in the form of tackling an overlooked portion of the value chain (simulation software, new delivery models, components with broad use cases). Exodys moved from developing a molten-salt reactor to using similar technology for its fuel recycling modules.
In regard to Approach #1, incumbents find themselves in a similar conundrum as investors. Which companies to expend internal resources and time on? Especially if some of these incumbents are developing their own reactors and vying for the same customers. BWXT is building its own micro-reactor project, directly working with the government on the PELE project. Prolonged engagement with a prospective partner and setting expectations from the start is crucial to ensure proper risk pooling. For example, tying in royalties for components developed by the incumbent in partnership with the startup ensures that the startup generates revenue, even when the same component is sold to a competing reactor vendor.
The predominant due diligence factor for investors assessing reactor startups is the perceived regulatory risk. This has incentivized companies to focus almost exclusively on “off-the-shelf” strategies that attempt to minimize this risk – to the detriment of the company’s long-term IP protectability and thus economic viability. In an industry focused on working with capable contractors with successful past performances, auxiliary services or products are market entry points and the foundation of an engineering firm’s brand. By generating earlier revenue, it also boosts company bankability as it attempts to secure project financing for the reactor construction.
The nuclear industry is due for a revamp in a variety of fields and many current needs remain unaddressed. The US government has been heavily investing in erecting a domestic supply chain, but there are significantly more opportunities to contribute to a successful nuclear sector. Holtec has introduced its SMR-160 reactor concept, but they didn’t do it right away. It comes after decades of thermal-hydraulics consulting, nuclear waste storage cask supply, and decommissioning services that have built its financial foundation.
There is no “one size fits all” strategy for commercializing reactors. However, diversified value propositions remain the only commonality across all currently successful nuclear companies also involved in reactor development. Reactor startups should stay focused on their product. But they should always keep in mind that along the development pathway, they are building tools that can serve as market entry points and should be marketed, not hidden. Otherwise, what’s the fall back?
— Carl & Julien
 | A guest post by
|
I come from the aerospace industry (former spx, built a company serving rocket companies afterward) and SpaceX did some interesting things in the early days to overcome these revenue issues. Initially, they had large amounts of cash down on a fixed price launch, which gave them money to run on. I repeated this strategy for my startup. Later, they just had so much momentum from delivering that there was investor money galore to eat the lumps between delivery cost and revenue. All of that to say, I think that there are many advantages the nuclear legacy companies have, but I think it's rare to build $6.8B plant. Someone in the SMR world will crack the code using upfront but fixed pricing and COTS components, is my bet.
No fall back. Shoot for the moon!