This was supposed to be America's nuclear century.
The Three Mile Island meltdown was two generations ago. Since then, engineers have developed innovative designs to avoid the kinds of failures that devastated Fukushima in Japan. The U.S. government was earmarking billions of dollars for a new atomic age, in part to help tame a warming global climate.
But a remarkable confluence of events is bringing that to an end, capped in recent days by Toshiba's decision to take a $6 billion loss and pull Westinghouse, its American nuclear power subsidiary, out of the construction business.
The reasons are wide-ranging. Against expectations, demand for electricity has slowed. Natural-gas prices have tumbled, eroding nuclear power's economic rationale. Alternative-energy sources like wind and solar power have come into their own.
And, perhaps most significantly, attempts to square two often-conflicting forces — the desire for greater safety, and the need to contain costs — while bringing to life complex new designs have blocked or delayed nearly all of the projects planned in the United States.
"You can make it go fast, and you can make it be cheap — but not if you adhere to the standard of care that we do," said Mark Cooper of the Institute for Energy and the Environment at Vermont Law School, referring to the U.S. regulatory body, which is considered one of the most meticulous in the world. "Nuclear safety always undermines nuclear economics. Inherently, it's a technology whose time never comes."
In the process, the U.S. could lose considerable influence over standards governing safety and waste management, nuclear experts say. And the world may show less willingness to move toward potentially safer designs.
More from The New York Times:
"I'm concerned that if the U.S. is not seen as a big player, and doesn't have that kind of market presence, that we won't be in a competitive position to bring those standards back up," said Richard Nephew, a senior research scholar at the Center on Global Energy Policy at Columbia. "If you've got more lax safety standards worldwide, I think that's a problem from an industry perspective as well as just a human standard."
This may be an advantage for state-owned nuclear industries worldwide. Often they benefit from long-term national policies in places like Eastern Europe, Asia and the Middle East.
By contrast, the Toshiba-Westinghouse withdrawal from nuclear construction shows how daunting it can be for the private sector to build these plants, even with generous government subsidies like loan guarantees and tax credits. Projects take decades to complete. Safety concerns change along the way, leading to new regulations, thousands of design alterations, delays and spiraling costs for every element.
In one case, even the dirt used to backfill excavated holes at the Westinghouse project in Georgia became a point of contention when it did not measure up to Nuclear Regulatory Commission standards, leading to increased costs and a lawsuit.
Thus far in the U.S., only the Tennessee Valley Authority, itself a government corporation, has been able to bring a new nuclear reactor into operation in the last 20 years.
Of the dozens of new reactors once up for licensing with the Nuclear Regulatory Commission, only four are actively under construction. Two are at the Alvin W. Vogtle generating station in Georgia, and two at the Virgil C. Summer plant in South Carolina. Both projects, which plan to use a novel reactor from Westinghouse, have been plagued by delays and cost overruns, some stemming, paradoxically, from an untested regulatory system intended to simplify and accelerate their development.
The projects, more than three years late and billions over budget, are what pushed Westinghouse — one of the last private companies building nuclear reactors — and its parent, Toshiba, to the brink of financial ruin, resulting in Toshiba's chairman stepping down.
The company has said that Westinghouse will complete the reactors for the projects it already has underway, including two in China. But the fate of other projects in the United States and abroad that plan to use the Westinghouse reactor, known as the AP1000, are in doubt, along with the role of the United States in the future of nuclear energy. It is also unclear how President Trump will approach nuclear energy development, which has broad and overlapping implications for tax and trade policies, economic development and national security.
The AP1000 is considered one of the world's most advanced reactors, with simplified structures and safety equipment which were intended to make it easier and less expensive to install, operate and maintain. It has been designed with an improved ability to withstand earthquakes and plane crashes and is less vulnerable to a cutoff of electricity, which is what set off the triple meltdown at Fukushima.
The industry has lurched through boom and bust cycles before.
Nuclear construction had all but disappeared in the United States, particularly after the partial meltdown at Three Mile Island in Pennsylvania in 1979. Concerns over climate change led to renewed interest in building new plants under the administration of George W. Bush, however. The Bush-era energy policy acts authorized $18.5 billion in loan guarantees, plus tax credits like those available for wind and solar.
Determined to avoid the delays and ballooning costs that were common as plants were built in the 1970s and '80s, federal regulators had devised a new licensing process.
Under the old system, companies received construction permits based on incomplete plans and then applied for an operating license, often leading to rebuilding and lengthy delays. The idea for the new system was that companies would submit much more complete design plans for approval, and then receive their operating licenses as construction started. That way, as long as they built exactly what they said they would, the process could move more quickly.
In the meantime, companies like Westinghouse and General Electric were developing a new generation of reactors intended to operate more safely. With the AP1000, for instance, emergency cooling for the reactor mainly relies on natural forces, like gravity, to propel the coolant, rather than relying on mechanical pumps powered by electricity. The problem is that electricity can fail, as it did at Fukushima, which can lead to disastrous overheating in a damaged reactor of an older design.
Read more from The New York Times:
In addition, Westinghouse was engineering its equipment so that large components of the plants could be made in sections at factories, then welded together and lifted into place with cranes at the construction site. In theory, this approach would save money and time, requiring far less skilled labor than the old, bespoke approach, in which workers assembled more parts onsite.
By 2008, Westinghouse had deals to expand two existing plants with the electric utilities Georgia Power and South Carolina Electric & Gas. Little went as hoped.
Because nuclear construction had been dormant for so long, American companies lacked the equipment and expertise needed to make some of the biggest components, like the 300-ton reactor vessels. Instead, they were manufactured overseas, adding to expense and delays.
One reactor vessel, headed for Georgia Power's Vogtle plant from the Port of Savannah, almost slipped off a specialized rail car. That led to a weeks-long delay before a second attempt was made to deliver it.
And, in a separate snafu, while working on the plant's basement contractors installed 1,200 tons of steel reinforcing bar in a way that differed from the approved design. That triggered a seven-and-a-half month delay to get a license amendment.
To some extent, the unexpected delays were to be, well, expected, given the novelty of the design and the fact that builders were decades out of practice. Any large undertaking involving so many first-of-a-kind complexities would be likely get tripped up somewhere, said Daniel S. Lipman, vice president of supplier and international programs at the Nuclear Energy Institute, which represents the industry.
"Whether you're building a nuclear power plant or providing a new locomotive or a new fighter jet complex for the Defense Department, the first of a kind almost always takes longer to be deployed," he said.
And then there was Fukushima, when an earthquake and tsunami knocked out both grid and backup emergency power at the plant, disabling its cooling systems and leading to the meltdown of three reactors. The plant remains shut down, and the decommissioning and cleanup process is projected to take as long as 40 years.
The Japan disaster prompted regulators to revisit safety standards, slowing approval of the Westinghouse designs and resulting in new requirements even after the Nuclear Regulatory Commission gave the go-ahead for the Georgia and South Carolina projects. That led to more costly delays as manufacturing orders had to be changed.
As all of that unfolded, Westinghouse was having troubles with the contractor it chose to complete the projects, a company that struggled to meet the strict demands of nuclear construction and was undergoing its own internal difficulties after a merger. As part of an effort to get the delays and escalating costs under control, Westinghouse acquired part of the construction company, which set off a series of still-unresolved disputes over who should absorb the cost overruns and how Westinghouse accounted for and reported values in the transaction.
Toshiba, which would like to sell all or part of its controlling interest in Westinghouse, has said it will continue to look into Westinghouse's handling of the purchase.
"Certainly they underestimated the amount of liability or cost overruns that these projects were in," Robert Norfleet, a managing director at Alembic Global Advisors who has followed the machinations, said of Westinghouse. "I don't really know how they can't take the blame for that. That's something within their own due diligence that they needed to do."
In the meantime, the main stage for nuclear development will move overseas to places like China, Russia, India, Korea and a handful of countries in the Middle East, where Westinghouse will have to find partners to build its designs.
In China, plants using an earlier model of the AP1000 are moving toward completion. If they are successful, that may stir up more interest in the technology, and future installations may go more smoothly. But Toshiba's ambitions of installing 45 new reactors worldwide by 2030 no longer look feasible.
Indeed, despite the much-ballyhooed ingenuity of a new generation of reactors designed by the likes of Westinghouse and G.E., countries may stick with older technologies that they can produce and install more quickly and cheaply. "Until several of these new designs — including the AP1000 from Westinghouse — come online on time and on budget," said Brent Wanner, an analyst at the International Energy Agency, "it will be an uphill battle."