The agency needs a bigger boat!
Several of the 11 firms selected by DOE for its microtreactor program can be characterized, more or less, as what Admiral Rickover famously called "paper reactors." What he said in a nutshell is that it is easy to design a reactor on paper. It is hard to build one.
Also, some of these firms are going to have burn run flameouts. This is a case where the firm's cash runs out and investors balk at providing more because the work involved doesn't lead to the kind of progress they expect.
This is what happened to Ultra Safe Nuclear, a Seattle, WA, nuclear micro reactor startup, which last February declared bankruptcy. Investors walked away, and the firm saw its assets and promising projects in a microreactor and advanced fuels sold off to competitors.
DOE's microreactor program, with its July 2026 deadline for first criticality, diverts all 11 firms from the main goal of booking customers into building the equivalent of "research reactors." This is a regulatory fig leaf relative to the NRC because the reactors will not produce electricity on the grid.
The problem is that this work is not the same as building a first of a kind (FOAK) plant for a customer because DOE is not putting any money into any of the projects now or later. In other words, it doubles the amount of cash, at least, that the firms will need to cross the finish line to have their FOAK plant in revenue service - once for DOE, and then another time for their first commercial plant.
Factory Production of Micro and SMR Reactors Requires a Fat Order Book
I am also critical of DOE's program due to the agency's short term focus on micro reactors instead of energizing efforts to build large PWRs. All of the micros pledge to produce their creations via factory production methods. Cost reduction via manufacturing in factory setttings is the holy grail of micro reactors and SMRs. It is a worthy goal, but getting there requires real market demand.
The hurdle rate for booked orders to justify to investors to provide the funds needed to build the factory, and to spool up the supply chain, which is also a source of economies of scale, is in "fleet mode."
Rolls-Royce in the UK has a target of 16 of its 470 MW PWRs as a basis for factory production in the UK with additional strong prospects for multiple units in Poland and the Netherlands.
This example sends a signal to developers of microreactors and SMRs alike that having ink on the order book in at least double digits is a minimum for building and operating their factory scale production.
Decarbonization of the Grid Requires Big Iron to Meet Baseload Demand
The main point is that if decarbonization of the grid is to make progress, it has to occur with large baseload plants. These micro units won't make a dent in the problem. They are specialized plants for specific types of customers, e.g., factories, data centers, mines, electrified rail systems, etc. These customers make sense for developers of microreactors to pursue.
BTW: Hyperscale data centers will likely need much bigger sources of power than microreactors. Witness Microsoft's PPA with Constellation for 800+ MW of power from a PWR being restarted by the utility specifically for this purpose. The cost to complete the restart is $1.6 billion which is well out of reach of microreactor developers.
Containing Costs Are the Challenge
What DOE ought to be doing if the agency is really serious about deploying 25GW of nuclear power by 2050 is to figure out how to get the cost of construction of GW class reactors, e.g., 1,000MW or more, under control.
No utility worth its stock price is going to commit to new GW class projects without project controls that will be effective to prevent a repeat of the cost overuns and schedule delays that were recorded for building the twin AP1000s at Vogtle.
Supply Chains Are the Crucial Links to Success
Additionally, DOE’s priority endeavor also needs to include spooling up a supply chain that can cost effectively deliver the components that are used to build GW class reactors including LWR RPVs, steam systems, turbines, generators, and switchyard transformers. The current wait time for these components is stretching out in ways that make future large reactor project look to be unsustainable for publicly owned utilities.
On the point of supply chains, with the great variety of microreactor designs, every one of them will confront the challenge to their burn rates of the high costs of one-off fabrications of their first of a kind units with no guarantee for them or their suppliers of the economies of scale that would occur using much promoted concept of factory production of these designs.
In point of fact, the great variety of design specifications for 11 microreactors is a nightmare for suppliers. It will likely drive some suppliers to make choices about which customers they will work for based on which ones are willing to pay the highest premium for one-off fabrication purchase orders.
DOE's "No Sale" Bake Off
DOE’s concept of rolling out 11 of these projects looks like a old style DOD “bake off” for new weapons systems like fighter jets and munitions. The difference is DOD paid its contractors to play the government’s game.
DOE’s pursuit of three microreactors to achieve first criticality by next year (July 4, 2026) is a horse race without a purse and a distraction from the long term goal of decarbonizing America’s grid.
Earlier this year I wrote a blog post critical of DOE's microreactor efforts tagging the program as "wishful thinking." I have not changed my mind.
# # #