Is bigger always better? Nuclear generators are no longer quite so sure. In fact, in the past few months, three of the ten largest nuclear energy-producing nations have set their sights a little lower.    Most notably, the US Department of Energy inked a $1.35 billion public-private deal to fund a proof-of-concept project for small modular reactors (SMRs). Its partner, the Carbon Free Power Project, will itself be partnering with NuScale, Inc. to install twelve reactors at Idaho National Laboratory.  NuScale has been a big player in SMRs for around a decade, but this project will finally be the proof of concept the industry has long been eyeing. Elsewhere, the UK is currently considering a $2.6 billion investment in SMR technology , while Canada recently put up $15.1 million. 
Nuclear proponents have long touted nuclear energy for having many of the conveniences of fossil fuels with very little of the carbon. Indeed, the technologies are similar in the abstract. Both involve consuming a fuel in order to heat water and turn a turbine to generate electricity. But instead of combusting fossil fuels, which creates greenhouse gases, nuclear fission reactors generate heat by splitting atoms.
But what does it mean for a nuclear reactor to be small or modular? While there’s no accepted industry definition, the International Atomic Energy Agency defines a “small” reactor as one generating under 300 megawatts (MW).  In some ways, the small reactor defies the nuclear trends of the last half-century, in which reactor capacity was aggressively increased as much as possible. Since the 1950s, relatively modest reactors producing hundreds of MW increased to nearly 1000 MW. Today, the APR-1400, designed in South Korea and topping out at 1.4 GW, is considered on the cutting edge.  But with the advent of SMRs, there’s been a deliberate return to nuclear’s lower capacity roots. Each NuScale reactor generates 50-60 MW — the idea is that a dozen or so of these smaller reactors can be run in concert to reach the total generating capacity of a typical plant.
Modular is an even hazier term. A recent paper in Physics Open explores five different definitions of “modular,” from scalable power production to interchangeability in design.  But when most energy experts say “modular,” they are referring to a streamlined style of factory production. Nearly every conventional nuclear reactor in the US is assembled on site. Standardizing the process by shipping ready-to-install reactors fabricated at a single location would greatly shorten production time and reduce the possibility of cost overruns. 
Or that’s what nuclear producers hope, at least. It’s no secret that nuclear energy has been in decline. Today, 11% of global electricity comes from nuclear power, down from a peak of 17.6% in the mid-1990s. Nuclear’s fall is driven by the rise in other energy sources and the dwindling interest that erstwhile top nuclear countries have shown in building new plants. While the US is still the world’s largest nuclear producer, its production is in decline.  At present, 39 reactors of the US’s approximately 120 reactors are in some stage of shutdown while only two nuclear reactors are under construction. No new nuclear plants have broken ground in the US since 1980. Only five reactors have come online since 1990, all of which began construction back in the 1970s, including one at Watts Bar in Tennessee which only connected to the grid in 2016. 
It’s not just the cost overruns that nuclear generators are hoping SMRs can overcome. SMR manufacturers believe their technology can also confront other dark clouds over the industry: proliferation risks, the severity of meltdowns, and the spent fuel dilemma. Nuclear is the prototypical dual use technology — the same fuel used in reactors can, when highly-enriched, be used in nuclear weapons. And while every sensible reactor manufacturer has long moved away from the graphite-based technology so catastrophic in Chernobyl, it can’t be ignored that the safety technology involved in the 2011 Fukushima meltdown was state-of-the-art. Finally, the vast majority of spent nuclear fuel and other radioactive waste in the US still sits in temporary cooling tanks while the government agonizes over where to site long-term waste storage.
More thoughtful and flexible design, SMR proponents argue, can convincingly reduce these risks. Large, conventional reactors need “active” security features (e.g. the pumps that failed to keep water circulating in Fukushima). But smaller reactors can incorporate “passive security” into their design. That way if a natural disaster or terrorist attack took the control room out of commission, the reactor would remain stable possibly indefinitely.  NuScale is in the process of getting such a feature approved, although not without controversy.  Similarly, some SMRs are designed to arrive on site with all the nuclear fuel they require for their full lifetime, and be transported back to the factory when the fuel is spent. This would greatly reduce the exposure of dangerous radionucleotides to the environment and consolidate it for long-term storage. 
But design can only go so far: SMR proponents want out of certain safety regulations made obsolete by their new designs (or so they believe). Under current regulations, a single control room can monitor a maximum of two reactors, hence why conventional reactors tend to come on line in pairs. If the two-reactor cap stays in place, SMR plants (which use bundles of 10-12 small reactors to match typical plant capacities) will greatly multiply capital and labor expenditures. Beyond that, SMR proponents argue that the reduced risks of disaster mean that these reactors could safely be placed nearer to occupied areas without, for example, the onerous requirement of making a full-scale evacuation plan for the plant’s environs.  But besides a floating installation completed in Russia earlier this year, there’s very little hard evidence to determine the actual risk posed by new SMR plants.
This brings us to the Department of Energy’s recently approved SMR project. The plan is to begin construction on twelve SMRs at Idaho National Laboratory in 2025 and send power to the grids of five states by 2029.  The goal is to achieve a levelized cost of energy (LCOE, a holistic assessment of the cost of production) of $65/MWh with an eye towards bringing the price down to $55/MWh. This would make SMRs competitive with conventional nuclear reactors. But they would still be more expensive than wind and solar, which have LCOEs of $34.10 and $38.20, respectively. 
SMRs are coming to the fore at a time where the US is transitioning towards low-to-zero carbon energy sources but away from nuclear power. The SMR business and production model seems to cure many of the problems that have ailed the nuclear industry for decades. But its cost competitiveness may be predicated on rolling back safety protections. More than half of today’s reactors will exceed their initial 40-year licenses in the next decade, which means they’ll either have to apply for extensions from the Nuclear Regulatory Commission or shut down.  Something will need to fill the coming gap of low-carbon, around-the-clock power. Will SMRs revive the industry, or merely further entrench it in economic unviability? All eyes are on Idaho to find out.