On an August afternoon in Washington, D.C., typically miserable for its heat, humidity and stillness, reporters gathered at a downtown hotel not known for its air-conditioning. Stuffed inside a windowless conference room that was being heated still further by the television people’s lights, we waited for Michael J. Wallace, who had been trying, in fits and starts, to unveil nuclear power’s second act.

On arrival, Wallace, a meticulous manager not known for ad-libbing, looked out over the sweating reporters and smiled. “It’s days like today that highlight the real need for new, emissions-free, baseload power,” he said. Unless we get started soon, he added, rolling blackouts could become the norm.

Wearing a suit and tie and seeming to enjoy the heat, Wallace announced that his company, UniStar Nuclear Energy, a partnership between Constellation Energy and the European nuclear consortium Areva, was looking to build a new kind of nuclear power plant in the U.S. and elsewhere. “I’m pleased to say I played a role in the last round of nuclear power plant development, and I’m really pleased to be involved,” the chairman said, calling to mind a graying astronaut who walked on the moon years ago and now wanted to do it again.

That was in 2006. Since then, Wallace has intermittently made new announcements about incremental progress toward building a new reactor about 45 miles south of Washington, which could be the first U.S. nuclear plant put on order and built since 1973. Wallace’s original feat was leading the start-up of two of the nation’s last big nuclear plants, completed in 1987 in Illinois. Like another moon shot, the launch of new reactors after a 35-year hiatus in orders is certainly possible, though not a sure bet. It would be easier this time, the experts say, because of technological progress over the intervening decades. But as with a project as large as a moon landing, there is another question: Would it be worthwhile?

A variety of companies, including Wallace’s, say the answer may be yes. Manufacturers have submitted new designs to the Nuclear Regulatory Commission’s safety engineers, and that agency has already approved some as ready for construction, if they are built on a previously approved site. Utilities, reactor manufacturers and architecture/engineering firms have formed partnerships to build plants, pending final approvals. Swarms of students are enrolling in college-level nuclear engineering programs. And rosy ­projections from industry and government predict a surge in construction.

Modern competitive pressures complicate the matter, however. For one thing, in much of the country any new construction would be by “merchant generators”—independent companies rather than large, monolithic utilities. Nuclear power was simpler two decades ago, because utilities built their own plants and could usually pass costs through to captive consumers no matter how big the overruns. But in states such as Texas, Maryland and New York, where the public service commission has separated the generation of electricity from power transmission and distribution, there is no longer a cushion for a generation company that guesses wrong. Such plants must sell electricity at whatever price the market will bear.

That number is hard to predict, because although reactors would exploit current technologies and techniques, so will modern coal and natural gas plants. Gas, especially, has much lower up-front costs, a big consideration if credit remains tight. And gas plants can be built in small units in only three or four years, as compared with six or eight for mammoth reactors.

For nuclear power, the modernization is intended to produce dramatic differences: plants that will run more than 90 percent of the hours in a year and last for 60 years or longer. The ones in service today ran only about 60 percent of the time when they were new and were assumed to have only a 40-year life. But utilities are already signing long-term contracts for large solar generators, and wind turbines are being erected at an unprecedented rate. Those alternatives operate fewer hours of the year, but with no burden of fuel cost or fuel-disposal problems the price of power they produce could be low enough to squeeze nuclear power out of the mix.

[break]

Perhaps even more of a question is the shape of the market that reactors would serve. Some states have a goal of zero electric growth, achieved by replacing lamps, pumps, blowers and everything else that runs on electricity with updated equipment that does the same work with less energy. If growth stopped—an ambitious prospect—new plants would still be needed to replace old ones as they wore out, but far fewer orders would result.

By almost all accounts, cutting demand is a lot cheaper than building capacity. Dan W. Reicher, a former assistant secretary of energy for conservation and renewables, has complained repeatedly that companies will invest in solar plants that produce electricity at 20 or 30 cents per kilowatt-hour, while ignoring fixes that would save comparable energy at a cost of four cents per kilowatt-hour.

Wild cards add even more uncertainty to how much power will be needed. Proponents talk about tens of millions of plug-in hybrid or even electric cars, each of which might use 10 kilowatt-hours a day from the grid to cover 30 or 40 miles of travel. That would substantially bump up demand, but the success of such vehicles is difficult to predict. If millions of the cars did sell, they would mostly recharge at night, which would change the shape of the “load curve”—instead of households consuming peak amounts of power during the day and far less at night, consumption would be at a more constant level across a 24-hour period, which favors technologies such as nuclear that are capital-intensive but operate around the clock with low marginal costs. Consequently, any power company planning a big generating station that takes six or eight years to build does so without a clear prediction of what demand will be by the time the plant is finished.

Potential carbon regulation adds even more guesswork. Governments are seriously considering a flat tax on emissions or a cap-and-trade system that would create a de facto surcharge for emissions. Either way, predicting the price is hard: the initial experience in European trading was a wildly unstable market. Still, economists predict such a system would result in a price that averages in the tens of dollars per ton of emissions. A $10 charge per ton would raise the consumer price of electricity by about a penny a kilowatt-hour. A new coal plant typically produces that much electricity for six or seven cents, so an addition of $20 or $30 a ton would create a huge advantage for carbon-free technologies such as nuclear power.

In fits and starts, a nuclear renaissance might actually be under way. Wallace’s vision is for standardized plants, identical right down to “the carpeting and wallpaper,” that could therefore be manufactured and approved for less than reactors of the past, almost all of which were custom-built. Teams of engineers and craft workers would construct the same plant again and again in different locations; just like assembling furniture from kits, practice would make perfect. The idea that mass production—or, at least, serial production—is cheaper than one-of-a-kind products is nearly universally held in the industry. John Krenicki, president and CEO of General Electric’s Energy Infrastructure division, says site-by-site construction will never create a cost-effective solution.

Wallace’s idea seems to be catching on. The first standardized plant is planned as a third unit beside Constellation Energy’s two existing Calvert Cliffs plants, about 45 miles south of Washington. In July, AmerenUE, a big Midwestern utility, also filed a license application for a cookie-cutter unit. More applications are waiting in the wings: one in Pennsylvania, one in upstate New York, one in Idaho and a twin-unit plant in Amarillo, Tex. All would be built by the UniStar joint venture, in partnership with a local utility or generating company. UniStar has not listed precise costs, but in recent briefings Wallace has pointed to other studies of standardized plants that quote “overnight costs” (not counting interest for construction) of $4,000 to $6,000 per kilowatt of capacity. His plants would be in the upper end of that range, he says. An up-to-date coal plant costs about $3,000 a kilowatt, but charges levied on carbon dioxide emissions, or extra equipment to capture the gas instead, could add substantially to that.

[break]

The possibility of a series of new reactors is a stunning turnaround for the industry, which bankrupted some of its customers in the 1980s because of huge cost overruns and which looked so bad in the early 1990s that some completed plants were shut down after only a few years of operation. Proponents say that today energy utilities find greater benefit in a technology that puts the financial risk up front, in the construction cost, and has little vulnerability to later swings in the price of fuel, as natural gas does, or to changes in emissions regulations, as coal faces. Consequently, companies around the country are spending tens of millions of dollars to explore their nuclear options, conducting engineering studies and preparing license applications, even if no one has ponied up the billions of dollars that an actual reactor would require. “There’s a huge sense of déjà vu for me personally,” Wallace says.

More Viable Than Clean Coal
To no one’s surprise, cost will loom large in any decision to plan on a reactor. The first installation of UniStar’s standardized model, known as a European Pressurized Reactor, or EPR, is under way in Olkiluoto, Finland. The project is now behind schedule and over budget, after quality-control problems early in the construction period.

Other reasons to be skeptical of nuclear’s price persist as well. Estimates submitted by utilities to regulators in Florida predicted $8,000 per kilowatt of capacity when transmission and loan interest costs are included. The cost of steel, concrete and labor have all risen, and the recent financial crisis may mean higher interest rates for construction loans, although that would affect the building of any kind of power plant.

Whether a reactor would be cost-effective depends on how it compares with other environmentally sound generation options. Coal plants that capture their own carbon emissions are one choice, but the leading demonstration plant that was being built in the U.S., known as FutureGen, has been scrapped.

FutureGen, originally planned for Mattoon, Ill., was overseen by a public-private consortium. The coal would have been cooked in a low-oxygen environment, creating a fuel gas made of hydrogen and carbon monoxide. The hydrogen would be burned for electricity and the carbon converted to carbon dioxide and pumped underground. The only emission would have been water vapor. But as the price of materials rose internationally, the plant’s cost went up even more. In early 2008 the U.S. Department of Energy pulled out. That move left China with the leading project, equally uncertain, which it calls GreenGen.

The Energy Department continues low-level work on so-called Gen IV nuclear reactors, fourth-generation technologies that use altered fuels or produce a more manageable waste stream. Other low-carbon coal technologies are being attempted, too. In Pleasant Prairie, Wis., the Electric Power Research Institute and Wisconsin Electric are testing a process that uses an ammonia-based chemical to bind carbon dioxide in a smokestack so it can be sequestered. But the test deals with only a little more than 1 percent of the plant’s emissions.

“We’re maybe 15 or 20 years behind where we should be for burning coal in an environmentally sound manner,” says Marsha H. Smith, president of the National Association of Regulatory Utility Commissioners. The bottom line is that although nuclear energy has obvious drawbacks such as cost and poisonously radioactive waste, it is far better demonstrated than coal with carbon capture.

More Dependable Than Wind or Solar
At the moment, the fastest-growing source of clean energy is wind. The American Wind Energy Association said in September that installations had reached 20,000 megawatts, double the capacity of 2006, with growth driven by generous tax incentives and state renewable energy quotas. But wind plants run far fewer hours of the year than nuclear plants do; 10,000-megawatt wind machines produce the energy equivalent of only two or three big 1,000-megawatt reactors. Because wind is not “dispatchable”—meaning the generators run only when nature allows, not when operators might order them to—the extent to which it can replace around-the-clock technologies such as nuclear is unclear.

[break]

Solar is more predictable, and with certain forms of energy storage may even be dispatchable, providing power during cloudy periods or during high-demand hours after sunset. Current solar facilities reflect the sun’s rays off of curved mirrors to heat water or mineral oil, but experimental systems use materials such as molten salt, which could run far hotter and be stored in insulated tanks for hours or days. Other companies are building massive arrays of photovoltaic cells that convert sunlight directly into electricity.

Generally, however, large solar and wind projects—the kind most likely to be cost-effective—are built in deserts or on remote mountaintops or plains, far from population centers that need the power. So transmission lines must be built to connect supply with demand. “You’re talking about immense amounts of transmission,” says John Rowe, chair of Exelon, one of the nation’s largest utilities. “It requires a really huge grid. I don’t see us going that way anytime soon.” Indeed, a recent Energy Department study concluded that wind could meet 20 percent of American needs by 2030 but would require a new transmission system costing $60 billion or more. Nuclear reactors can be located far closer to consumers and would require more modest additions to the existing grid.

Efficiency Could Forestall Reactors
One of the strongest competitors nuclear power faces is energy efficiency. Improvements in efficiency, driven by the need to reduce greenhouse gas emissions, could for many years offset increases in demand from a growing population with higher living standards, forestalling the need for reactors.

In December 2007 consulting firm McKinsey & Company determined that the U.S. could cut its output of global warming gases by more than 11 percent using conservation steps that were better than free: they would pay for themselves and earn a profit. These “negative cost opportunities” would require little or no technology innovation, the report said. And emissions could be cut by another 17 percent with efficiency improvements that had only a moderate cost.

Amory Lovins, a well-known efficiency expert, has long referred to such opportunities as being better than a free lunch, “lunch that someone pays you to eat.” But the steps are often not taken. One reason is that efficiency is usually number 11 on people’s top 10 to-do lists. For example, a high-efficiency air conditioner costs more than a standard model but will earn back the difference, in electricity savings, in a season or two. Yet many purchasers do not care, especially if they are landlords or builders who will never pay the electric bill.

Other steps might minimize convenience, even those that border on slothfulness. Lots of home appliances, for example, continue to draw power when the switch is “off” so that they are always warmed up and can come back to life instantly. Experts sometimes call this constant draw a “vampire load.” Around the house, all those vampires add up, but hardly anybody knows or cares. As Richard D. Duke, an energy expert at the Natural Resources Defense Council, quips, “What consumer, when buying a TiVo, is going to demand that the manufacturer make the standby power consumption a criteria? Nobody.”

Build before Memory Runs Out
Although individual consumer actions can help, major changes in carbon output will likely require better electricity-generation technologies, retiring much of the coal-fired capacity and replacing it with the most cost-effective combination of modern reactors, renewables and even clean coal. Around the country, players in the electricity business—regulated utilities, independent merchant generators, and municipal suppliers—are placing bets on which options will be the winners.

The competition is a bit like a high school track meet, however, in which competitors’ starting lines are staggered around the track. Nuclear has the longest path, because it takes more time to obtain site and building permits and to clear safety reviews. Yet anybody even thinking of a new reactor must pony up the entry fee—the cost of submitting an application and conducting preliminary studies. Given the uncertainties in future demand, carbon regulation and the price of fossil fuels, exploring the nuclear option makes business sense. Whether to actually build is another question.

[break]

One key factor is the price of loans. If a plant runs $5 billion in “overnight” costs and the money is spent over five years, interest on capital during the period of construction—the utility’s version of a home builder’s construction loan—could add hundreds of millions or even billions of dollars. To help, the federal government offered the nuclear industry loan guarantees worth $18.5 billion. It quickly received applications for more than $100 billion in funding.

Another factor is just how long that construction period will be. American builders could base their estimates on reactors built recently in Asia, but no one really knows how a project in Texas or Florida might compare with one in Japan or South Korea. If two or three reactors, such as Wallace’s, could get built in the U.S., the issues would become much clearer. Legislation that provides a predictable price for carbon emissions for the next few decades would also bring clarity.

The country needs a better way to manage nuclear waste as well. The federal government signed contracts with the electric utilities in the early 1980s that promised to take spent nuclear fuel off their hands beginning in 1998. But today, 10 years beyond that deadline, the Energy Department has only applied for a license to build one controversial waste repository, at Yucca Mountain in Nevada. Estimates of the opening date range from 2017 to never. An interim plan, such as long-term storage in aboveground casks in a few areas that are dry and sparsely populated, might be within reach. Many plants around the country have maxed out temporary storage in their spent-fuel pools, forcing them to put waste into huge, dry casks. Filled with inert gas to prevent rust, the casks are moved out to concrete pads surrounded by barbed wire, which look a little like basketball courts at maximum security prisons.

Still, advocates say the reactors are inevitable. At Areva, the company that Wallace’s firm has partnered with, the chief executive, Anne Lauvergeon, scoffs at the idea that there is any other choice. Could coal plants sequester their carbon? “It’s not ready at all,” she says. “You don’t know anything about the cost, and the technology doesn’t exist.” In the meantime, world demand is galloping ahead. Her company will build in China, she points out, and would like to build the first reactors in the Persian Gulf.

In the U.S., many power industry experts doubt that more than a few reactors will be built, at least until company executives see how the first ones go. But potential reactor builders sense that the world has changed enough to consider going back into business, with designs that are optimized and standardized versions of what they built more than 20 years ago. And people like Michael Wallace want to get going while those companies’ engineers still remember how.