How the Lights Stay On · The Atom Comes Back

The Atom Comes Back, part 2

These three pressures, together, are what have made 2024 and 2025 the most consequential years for American nuclear power since the 1970s.

Before going further, let’s be straight about what nuclear is and is not.

Nuclear is one of the cleanest sources of electricity humans have invented, measured by carbon emissions over the entire life of the plant. Studies put nuclear’s lifecycle emissions at around 12 grams of CO₂ per kilowatt-hour, roughly comparable to onshore wind, and far below natural gas (around 490) or coal (around 820).

Nuclear runs almost all the time. The average US nuclear plant operates at over 90 percent of its maximum capacity. It takes a few weeks off every year and a half for refueling, and that is about it. No other source of electricity comes close to this kind of reliability.

Nuclear is dense. A single one-gigawatt nuclear reactor occupies maybe a few hundred acres of land. To produce the same amount of electricity over a year with solar would require roughly 30,000 acres. To produce it with wind would require roughly 80,000 acres.

Modern reactor designs are dramatically safer than the ones built in the 1960s and 1970s. Passive safety systems use gravity and natural circulation to cool the core even if all the pumps and electricity fail. Newer fuel forms — particularly TRISO fuel, in which tiny pellets of uranium are coated in ceramic layers — can withstand much higher temperatures without releasing radiation.

Nuclear waste is a real problem, but a smaller one than it sounds. All of the spent fuel ever produced by the entire US nuclear fleet, over more than sixty years, would fill a single football field about ten yards deep. Compared to the waste streams from coal (millions of tons of ash, mercury, and CO₂ per year) or even from solar (lots of glass and silicon to recycle), nuclear waste is volumetrically tiny. It is, however, dangerously radioactive for tens of thousands of years, and the United States has not yet built a permanent repository to store it. The waste sits, at every operating nuclear plant, in concrete-and-steel casks on site.

Now the other side.

Nuclear plants are expensive to build, and the United States is bad at building them. The clearest case study is in Georgia.

The Vogtle Electric Generating Plant, near the town of Waynesboro, has had two operating nuclear reactors since the late 1980s. In 2009, its owners broke ground on two more: Units 3 and 4. They were the first new commercial nuclear reactors started in the United States since 1979.

The project was originally projected to cost $14 billion and to be completed in 2016 and 2017. It cost more than $30 billion. It was completed seven years late. Unit 3 came online in July 2023. Unit 4 came online in April 2024.

The reactor designer, Westinghouse, went bankrupt during construction. The cost overruns will be paid for by Georgia Power customers, in the form of roughly $14 to $20 per month on their electric bills, for decades.

Defenders of nuclear say that Vogtle was a first-of-a-kind project, that the next plant of the same design would be cheaper, that we lost the muscle memory of building reactors during the four-decade pause. They are not entirely wrong. China and South Korea build similar reactors faster and at a fraction of the cost, suggesting it can be done. But the Vogtle experience is what it is. It is what every new American nuclear project has to argue against.

Nuclear waste storage has been politically stuck for forty years. The Yucca Mountain repository in Nevada was designated as the permanent national waste site in 1987. After billions of dollars of study and partial construction, it was suspended in 2010 over political opposition. As of 2026, the United States still has no permanent repository for high-level nuclear waste. The waste continues to accumulate at every operating plant.

Accidents have been rare but psychologically enormous. Three Mile Island in 1979 killed nobody but ended a generation of American nuclear construction. Chernobyl in 1986, in the Soviet Union, killed dozens directly and likely thousands more from cancer, contaminated thousands of square miles, and is still being managed today. Fukushima in 2011, after the Tōhoku earthquake and tsunami, displaced hundreds of thousands of people and led Germany and Japan to make sweeping decisions about their own nuclear programs.

Modern reactors are far safer than the ones involved in these accidents. The newer designs being built now have passive safety systems that would have prevented all three. But the memory remains. Public support for nuclear in America has, however, shifted dramatically: a 2024 Pew Research poll found that 56 percent of Americans favored expanding nuclear power, up from 43 percent in 2020 — a 13-point shift in four years. Most Americans still favor solar and wind more strongly, but the gap is closing.

So that is the honest balance. Nuclear is clean, reliable, dense, and safer than it used to be. It is also expensive, politically fraught, and saddled with a waste problem we have not solved.

The most interesting bet in nuclear right now is that the future does not look like Vogtle.

For most of the history of nuclear power, the trend has been bigger. The reactors of the 1960s were a few hundred megawatts. The ones built in the 1970s and 1980s were a thousand megawatts or more. The two new ones at Vogtle are about 1,100 megawatts each. The logic was simple: economies of scale. Bigger plants are more efficient. Bigger plants spread the fixed costs over more output.

All 18 chapters of How the Lights Stay On · Library · All courses