NuScale, the first nuclear energy company to get design approval from the Nuclear Regulatory Commission (NRC) for the new breed of reactors called Small Modular Reactors (SMRs), has just canceled its first planned buildout in Utah. In what some see as a setback for SMRs, NuScale lost some key utility customers fearing cost overruns such as when inflation increased the company’s construction costs almost 50 percent above the plan. That drove NuScale’s expected electricity pricing to 8.9 cents per kilowatt-hour, considerably above the wholesale price of electricity from fossil fuel plants.

NuScale’s SMR design is one of more than two dozen new designs for reactors that are much smaller, cheaper, and easier to build than large conventional power reactors such as those just completed at the Vogtle plant in Georgia. Most of the new designs use fuels that are what is called “walk-away-safe”; that is, if they go over temperature, the physics of the fuel automatically shuts down the reaction without human or mechanical intervention.

In NuScale’s case, short-term market considerations undoubtedly dictated the utilities’ decision. Unfortunately, the long-term effect will be more carbon in the atmosphere and eventually much higher electricity rates for customers. That’s because those cheap fossil fuel plants will have to be replaced by something clean. And if intermittent renewables are to be the replacement, the cost of their complete systems will be many times greater than the cost of that NuScale nuclear plant and its system.

Why is this so?

It’s the difference between looking at just the cost of the generator versus looking at the cost of the entire energy generation and delivery system. It is the system cost that is the problem. Utilities must deliver what is often called baseload power, reliable 24/7 power, regardless of the weather. So while it is true that wind and solar are, by far, the cheapest ways to generate electrons, they do so only when the wind is blowing or the sun shining, almost always less than half the time, often much less. The problem utilities have is incorporating that kind of unreliable intermittency into a system that must be reliable. That is the problem that makes a system that relies chiefly on wind and solar so extremely expensive.

But isn’t nuclear energy more expensive? No! But because of the way we compare plant costs, nuclear seems more expensive. Consider this real-life example comparing nuclear to solar. Plant build costs are calculated based on how much power they can produce—their capacity, measured in kilowatts (kWs). Constructing a solar array costs about $1,320 per kW of capacity. By comparison, the Southern Company’s new Vogtle nuclear plant made repeated headlines when the cost of the plant’s two reactors skyrocketed to over $30 billion, equal to over $13,000 per kW. It sounds from those two figures like the nuclear plant is ten times more expensive than solar. And based only on generator cost, it is.

But what matters is the full cost of the system that can deliver reliable power. The Vogtle plant cost includes everything needed to deliver 24/7 power to the grid. That solar array cost above includes only the cost of the solar generator. What is not included are the huge costs necessary to produce long-term 24/7 power for the grid. Vogtle will produce electricity 24 hours a day, 365 days a year. On the other hand, the best solar plants produce power less than one-quarter of the time (because of nights, clouds, winter). Therefore, to generate the same amount of power as one kW of nuclear capacity, you must build over four kWs of solar arrays.

The next problem is lifetime. Vogtle will produce power for at least 80 years. A solar farm’s average lifespan is 20 to 25 years. That means it is necessary to build four or five solar farms to deliver the same 80 years of power as the Vogtle plant. A little math shows that to equal the output of the $13,000 per kW Vogtle plant, it will be necessary to spend between $16,000 and $20,000 per kW on solar plants. But even that expense doesn’t cover the whole solar bill. Each of Vogtle’s reactors will generate 1.11 gigawatts of power. To get just one gigawatt worth of power from solar farms will require between 45 and 75 square miles of land, which must be bought or rented and cleared. Collecting that widely dispersed power will require massive new transmission lines, another added cost.

Then there is the cost of solar’s intermittency problem. In order to deliver 24/7 power on demand from intermittent generators, utilities build and maintain alternative means of generating power. Dense clouds can reduce a solar farm’s output to almost zero for a week or more. Since batteries at that scale are impossible, utilities are using coal or gas plants to fill the gap. This means the utility is maintaining two separate power systems—intermittent solar and baseload coal or gas. Germany and California are real-life examples of this. They have both built considerable new intermittent energy sources, and they have both also built new fossil plants. The cost of operating two parallel systems has driven their electricity rates well above surrounding areas.

When you add it all up, that NuScale SMR, even with its 8.9 cents projected cost, will over the SMR’s lifetime be several times cheaper than clean alternatives.

So why do utilities make decisions like the one they made in Utah that nixed the construction of NuScale’s SMR?

Part of the problem is the way expected electricity costs are traditionally computed. To compare the expected electricity costs of one plant versus another, they compute something called the Levelized Cost of Electricity (LCOE). The LCOE formula originated when most plants were fossil-fueled and there was no need to back up intermittency or to include extraneous or externalized costs such as powerlines, waste disposal, plant lifetime, or decommissioning costs. LCOE figures for solar and wind rarely include these costs. For nuclear plants, however, all these costs are automatically included as a matter of law. In other words, when comparing costs, traditional LCOE does not begin to tell the whole story and does not compare apples to apples.

The unfortunate result is that traditional LCOE figures along with high initial capital costs create the public perception that nuclear plants are much more expensive than other technologies. And that misperception drives public opinion, which, too often, drives decisions by boards of directors and utility shareholders.

Another important issue with intermittent generators is the timing problem. Solar and wind can produce too much energy some of the time and little or none most of the time. In the middle of a bright, sunny summer day, California solar farms produce so much more power than needed that the price of electricity actually goes negative. California pays other utilities to take it! As a result, and despite heavy subsidies, investors are losing interest in financing new sources of intermittent energy.

So while it is true that solar and wind are the cheapest way to generate electricity, their variability, their intermittency, makes them very expensive to incorporate into a grid system that must deliver 24/7 power on demand.

Two recent studies dive into all this in more detail. In May, the Bank of America released a detailed study of clean energy systems that says this: “Solar and wind look more expensive than almost any alternative…. Nuclear appears to be the cheapest scalable, clean energy source by far.” And just this month, KutakRock, an energy consulting firm, released an in-depth report that says this: “We found that with tax incentives and without tax incentives, SMRs are the most cost effective option based on LCOE.”

The renewable problems are serious. Combined with inflation, the real cost of intermittent renewables is changing some of our ideas for a clean energy transition. Just this month we’ve seen notices that giant planned wind farms are being canceled around the world. Indeed, recent news suggests that the clean energy transition is slowing down and has little chance of meeting the UN’s COP goals.

But don’t get us wrong. Solar and wind can be the fastest way to reduce greenhouse gas (GHG) emissions. When they are a relatively minor part of a large and resilient grid, they can reduce fossil burning and produce an immediate GHG reduction. But they cannot economically serve as large-scale replacements for baseload fossil-fueled plants. As the Bank of America and KutakRock reports confirm, nuclear is the least expensive, and least environmentally damaging, way to replace those fossil plants.

Another factor that will come into play is experience. Vogtle had what is called a First-of-a-Kind experience (a FOAK, in the argot). Other than the 2015 completion of a reactor in Tennessee, the United States has built no new reactors in the last 30 years. Vogtle started with no experienced personnel and no existing supply chain. Everything was from scratch. In 1977, when solar technology was starting out, solar cost a whopping $76,000 a kW! Today its cost has dropped to a little over $1,000 per kW. We can expect large reactors to see significant cost declines as we begin building a lot of them. SMRs, because they can be built in factories, should see even greater declines. So the good news is that nuclear, already the cheapest form of clean energy, should become even cheaper.

All of this means that if we really want to decarbonize, we must look past our love affair with ideologically favored technologies. We must evaluate the various clean-energy technologies from a system perspective, and not just focus on the generator. If we do that, we will see nuclear energy in an entirely different light.