Manufactured Pre-fabricated Nuclear

With the renewed attention on nuclear energy stemming from the debate over its treatment in the EU taxonomy, I thought it an appropriate time to propose a business strategy for nuclear that might actually transform it into the most impactful decarbonization solution available. Further, nearly every zero carbon technology conceivable has been receiving massive venture investment regardless of how far it is from commercial reality (e.g. direct air capture, superconducting transmission, electro fuels, nuclear fusion), but nuclear fission has found little support despite providing 10% of global electricity production today.

Problem 

While many individuals and political groups will lambast nuclear fission for its risks, waste products, or proliferation concerns, none of those topics are actually reasons why nuclear plants do not get built. The primary reason, arguably the sole reason, nuclear plants are not built in greater number around the world is because of cost. The last several nuclear projects in the U.S. and Europe have come with a sticker price in excess of $8,000/kW. Whereas, onshore wind can be built for <$2,000/kW and natural gas can be built for <$1,000/kW. The price for nuclear is so high and so uncertain (climbing 50-200% or more during construction) that no rational private market investor can justify financing one. Only when governments step in with massive loan guarantees, tax credits, or guaranteed cost recovery through rate basing (in the case of Southern Company), do projects ever move forward. Closely connected to their cost is their long and fragile construction schedules. There are so many types of contractors (>4,000) and trades involved at a nuclear construction site that any disruption or delay from weather, regulatory approvals, supply chain bottlenecks, mistakes, or other extraneous factors rapidly extends the critical path. Recent nuclear plants have taken approximately 10 years or longer from date of first construction to generating the first MWh. Large, expensive, specialized, and bespoke construction projects with highly uncertain timelines and cost overruns incur high interest rates on loans, which ruin the return targets of an investor.

The industry’s reposes to these fundamental problems has been small modular reactors (SMRs), essentially a smaller sized reactor that sacrifices economics of scale for economies of mass production. However, the realty is that SMRs have themselves taken 20 years to commercialize and have not been able to tell a convincing story on economies of mass production. The leading SMR design from NuScale has been fully licensed by the NRC but cannot find a buyer because it is still quoting a price of $5,000-$6,000/kW (the same price Westinghouse quoted on the AP1000 before it skyrocketed to >$8,000/kW at Vogtle) and 60% of that cost is still bespoke civil engineering work (i.e. not terribly modular nor reminiscent of mass production). The only advantage from being small is that more investor can hypothetically write a $5 billion check than a $10 billion check. There are very real economies of scale in nuclear energy (security, interconnection, balance of plant, mobilizing workers, earth moving, etc.) that SMRs sacrifice.

The nuclear industry needs a real manufacturing business model for deployment. The reason wind, solar, and natural gas are built in large number around the world is because they are largely pre-fabricated blocks that require minimal on-site work. They are manufactured off-site and bolted to the ground. From a purely raw materials and engineering standpoint, there is no reason nuclear energy cannot achieve costs below $1,000/kW, and if it did, it would change the trajectory of global decarbonization. 

Solution

As much as I find generation IV reactors, micro-reactors, and fusion technically exciting, they are all years away from commercial reality and none of them (except maybe microreactors) necessarily directly address the cost problems outlines above. The expensive part of a nuclear power plant, is not actually the nuclear reactor, it is everything around the reactor. Therefore, why not just use the PWR or BWR technology available today and innovate on the manufacturing methodology. Rather than moving the 4,000 specialized construction workers to a new site for several years, manufacture the entire nuclear plant in a centralized location where the same workers can repeat the same task over and over again (i.e. mass production). How do you manufacture an entire nuclear plant?

Answer: shipyards. 

Hyundai Heavy Industries constructing multiple container and tanker ships in parallel

I propose taking an “off-the-shelf” nuclear reactor technology (e.g. NuScale’s 77 MWe SMR), installing a pack of them in a ship hull in a dry dock, floating that hull to a coast in need of power anywhere in the world, and sinking the hull down so it rests on the coastal bed (i.e. it becomes just another coastal power plant). One might even be able to fit several reactors into a seawaymax hull to address the market around the Greta Lakes.

Mockup of a ship hull reactor, breakwater, and switchyard.

Wireframe of two SMRs in a seawaymax ship hull

With a pipeline of orders from multiple countries, existing modern shipyards could easily build multiple such nuclear plants in parallel, maximizing worker utilization, equipment utilization, learning, and raw material procurement efficiency. Large-scale components could even be farmed-out to different yards with lower labor costs and merged dockside. Nuclear plant construction phasing would no longer be constrained to bottom-to-top processes, they could be manufactured in the same vein as automobiles. 

If the reactor is located sufficiently low in the hull, one could also use the surrounding body of water as an unlimited passive heat sink for safety-grade cooling (the fundamental technical failure at Fukushima, and one of the primary safety concerns in any fission reactor design). The entire plant could be pre-fabricated and self-contained in a single ship hull. Multiple hulls could be chained together for larger power needs. The balance of plant could also be outfitted to provide steam in addition to electricity for industrial customers or district heating (substantially improving plant economics where available). 

Minimal work would be needed at the final generation site. Other than a substation/ switchyard for transmission connection (require at any generator) and some minor civil work to prepare the site, one should only need to conduct some basic seismic and environmental studies to permit the site. The host country would only need to concern itself with regulating the operations and security of the plant as opposed to its full life cycle of construction, fuel procurement/enrichment, and decommissioning. The manufacturing country would handle regulations and quality control of manufacturing, fuel, and possibly even decommissioning (by floating the nuclear plant to a centralized decommissioning and recycling center. Spent fuel could also be moved as one lump sum with the plant during decommissioning through international waters (much easier than getting legal approvals through multiple border crossings and process that is already carried-out today). As long as host countries bought into the regulations and quality control of the manufacturing nation, regulatory overhead and costs could be substantially streamlined.

This proposal is supposed to sound simple for a reason: I am only introducing innovation (i.e. risk) on the manufacturing and delivery process of nuclear energy, everything else (technology, materials, regulations, manufacturing facilities, etc.) is off-the-shelf. This strategy offers the fastest path towards deploying nuclear energy at scale and low cost. I have spent several years studying additional innovations that could be made in offshore nuclear, and I am convinced the best path forward is the simplest one.

Business Model

Now let’s talk about the business model implications of this form factor and delivery paradigm. I have already drawn the analogy to automobile manufacturing but the comparison do not stop there. At scale, a manufacturing company with a pipeline of orders could consider financing new plants for countries itself. The manufacturing company would have greater confidence in its ability to deliver than external banks and could mitigate the risk of default from the purchasing country by selling the plant to another customer in the pipeline or even repossessing the plant. While repossession would presumably be unlikely, the mere fact that it is technically possible substantially reduces lending risk. The manufacturing company can also bundle-in services throughout the plant’s lifetime (like serving your car every 10,000 miles) such as outage planning/refuelings, security services, decommissioning, and even operations. Managing a fleet of SMRs distributed globally would have massive economies of scale in procuring uranium, SWUs, replacement parts, and managing security and operations. On-site operating teams could be minimized to keep fixed costs low and plants could be managed by centralized engineering teams. A standardized fleet design would also allow parts sharing/pooling. Centralizing fuel operations and security would also mitigate concerns over proliferation from the IAEA. One could even take this automobile analogy to the extreme and lease the power from the nuclear plant to the host country rather than transferring ownership of the plant, further reducing regulatory/proliferation concerns for emerging markets as well as default risk. This could further enable verticalization of the industry, fully merging operations with manufacturing, and realizing greater cost synergies. Alternatively, the nuclear OEM could setup a network of operations companies (i.e. dealerships) to manage the leases, possibly realizing a lower cost of capital, higher leverage, or higher P/E ratio if publicly traded by the same logic as YieldCos.

Even if there was no reduction in overnight capital cost, shrinking the delivery time from 10 years to 36 months (or less) and the financing interest to 2% would make nuclear far more accessible and affordable for global markets. In addition, transitioning nuclear power to a real manufacturing paradigm has the potential to reduce overnight capital cost 75% or more. 

Nuclear startups have struggled in the past because they have devoted all of their resources towards passing NRC regulation, which has historically taken 10 years and $500 M - $1B. A nuclear manufacturing company needs to focus the resources from its first capital raise on establishing a pipeline of orders (I.e. build conviction on the market size) and a manufacturing cost model through vendors (i.e. build conviction on cost reduction potential). The first several projects will likely need highly structured project financing with government backing, but as the cost reduction milestones are proven, more equity investors can be brought on for scale-up. One could even consider borrowing some tactics out of Tesla’s playbook by taking early deposits on the first several orders from interested countries. For countries with low cost debt markets, they could issue green-labeled sovereign debt to fund specific projects.

With the service bundling, each plant sold becomes an annuity to build reliable future cash flow for decades, enabling the manufacturing company to accept lower margins on the up-front manufacturing cost. Countries are highly dependent on their power providers so this would be a highly reliable earnings stream. Achieving this business model would likely require some cooperation with the IAEA and the U.S. government regarding 123 agreements, but it would give Western countries confidence over safety and proliferation.