Expanding two power plants in Georgia and South Carolina with big, new reactors was supposed to spark a “nuclear renaissance” in the U.S. after a generation-long absence of new construction. 

Instead, Westinghouse Electric Co.’s state-of-the-art AP1000 design resulted in long delays and steep cost overruns, culminating in its bankruptcy in 2017.  The fall of Westinghouse was a major blow for an industry that the company had helped usher in at the dawn of the nuclear age. It was Westinghouse that designed the first reactor to enter commercial service in the U.S., at Shippingport, Pennsylvania in 1957. 

Two new AP1000 reactors at Plant Vogtle near Augusta, Georgia started operating in 2023 and 2024, turning the plant into the largest energy generation site of any kind in the nation and marking the first new operational nuclear reactor design in 30 years. But the reactors came online seven years behind schedule and $18 billion over budget.

In the wake of Westinghouse’s bankruptcy, utilities in South Carolina stopped construction in 2017 on two reactors at the V.C. Summer plant near Columbia after sinking $9 billion into the project. 

But today, interest in new nuclear power is reviving as the tech sector seeks reliable, carbon-free electricity to power its artificial intelligence ambitions, especially against China. Westinghouse emerged from bankruptcy in 2018 and was acquired by Canadian uranium miner Cameco and Brookfield Asset Management in November 2023

The changed environment means South Carolina sees an opportunity to finish the two reactors left partially built at V.C. Summer eight years ago. The state’s Santee Cooper public utility in January began seeking a buyer for the site to finish reactor construction, citing data center demand as one of the reasons to move ahead.

“We are extraordinarily bullish on the case for V.C. Summer,” Dan Lipman, president of energy systems at Westinghouse, told CNBC in an interview. “We think completing that asset is vital, doable, economic, and we will do everything we can to assist Santee Cooper and the state of South Carolina with implementing a decision that results in the completion of the site.”

Tech as a nuclear catalyst

The United States has tried to revive nuclear power for a quarter century, but the two reactors in Georgia mark the only entirely new construction across that period despite bipartisan support under every president from George W. Bush to Donald Trump.

A fresh start was supposed to have begun more than a decade ago, but was choked off by a wave of closures of older reactors as nuclear struggled to compete against a boom of cheap natural gas created by the shale revolution.

“We went from an environment in the aughts of rising gas imports and rising gas prices to fracking technology unlocking quite a bit of affordable natural gas here in the U.S., and companies didn’t really value the firm clean attribute of nuclear back then,” said John Kotek of the Nuclear Energy Institute, an industry lobby group, and former assistant secretary at the Office of Nuclear Energy under President Barack Obama.

What’s different in 2025 is the tech sector’s voracious appetite for power translating into a willingness to pay a premium for nuclear. But recent investments in nuclear have focused on restarting abandoned reactors and attempting to bring online smaller, next-generation modular reactors that many believe are the future, if they can be designed and built more cheaply.

The troubled nuclear plant at Three Mile Island near Harrisburg, Pennsylvania that almost melted down in 1979 is expected to resume operations in 2028 after owner Constellation Energy struck a power purchase agreement with Microsoft last September. Constellation wants to restart Unit 1, which shut for economic reasons in 2019, not the Unit 2 reactor that was the site of the accident.

Alphabet and Amazon invested in small nuclear reactors a month later. Meta Platforms, owner of Facebook and Instagram, asked developers in December to submit proposals for up to 4 gigawatts of new nuclear power to meet the energy needs of its data centers.

But while the recent focus in the U.S. has been on restarts and commercializing small reactors, Lipman said the extent of potential demand that has emerged from data centers over the past year has led to renewed interest in Westinghouse’s large AP1000 reactor design.

In any event, there are no operational small reactors in the U.S. today, though startups and industry stalwarts, including Westinghouse, are racing to commercialize the technology. And there only so many shuttered plants in the U.S. in good enough shape to potentially be restarted.

Gargantuan undertaking

Meanwhile, meeting the demand for power is a gargantuan undertaking. Meta’s need for new nuclear power, for example, is nearly equivalent to the entire 4.8 gigawatts of generating capacity at the Vogtle plant, enough to power more than 2 million homes and businesses. Large nuclear plants with a gigawatt or more of capacity — the size of the AP1000 — will be essential to power large industrial sites like data centers because of their economies of scale and low production costs once they’re up and running, according to a recent Department of Energy report.

Georgia Gov. Brian Kemp called for another reactor at Vogtle the same day he dedicated the plant expansion in May 2024. Southern Company CEO Chris Womack believes at least 10 gigawatts of large nuclear are needed. Southern is the parent company of Georgia Power which operates Vogtle.

“The people that are going to own and operate AP1000s traditionally are investor-owned electric utilities,” Lipman said. “When they look at the marketplace for a large reactor, AP1000 is where they turn because it’s got a license, it’s operational.”

Still, nobody in the U.S. is on the verge of signing an order for a new AP1000, he said. Westinghouse is focused on deploying reactors in Eastern and Central Europe, where nuclear projects are seen as a national security necessity to counter dependency on Russian natural gas after the invasion of Ukraine.

In addition to the two units in Georgia, Westinghouse also has four operational reactors in China.

But South Carolina’s search for someone to complete the partially built reactors at V.C. Summer will likely draw investment from Big Tech “hyperscalers” building data centers, and large manufacturers like the auto industry, Lipman said.

“That kind of asset attracts industry that relies on 24/7, 365 energy and that’s what you get with an AP1000,” Lipman said. There are ongoing discussions within the industry about whether the tech sector might act as a developer that invests capital in the upfront costs of building new plants, he said.

What went wrong in the South

Any attempt to build new AP1000s in the U.S. again will almost certainly meet with skepticism after the experiences in South Carolina and Georgia.

Lipman said the challenges that the AP1000 construction faced in the South have been resolved. Back then, Westinghouse agreed to the projects before the reactor design was complete, and supply chains weren’t fully formed due to a long period in which U.S. construction was dormant, he said.

“One big lesson learned, maybe the big lesson learned, is designs need to be complete before they hit the field, meaning they have to be shovel ready,” Lipman said. The design for the AP1000 is complete and Westinghouse has its supply chain in place, he said.

“We have winnowed over our list of suppliers,” Lipman said. “They are supporting us globally, and so it’s really easy then to have them make more equipment for deployment.”

“You’re getting economies of scale,” he said.

Ironically, given the overruns in Georgia, the original aim of AP1000 was reduce costs by creating a standardized design that requires less construction materials compared to older reactor types, Lipman said. Components of the plant are prefabricated before being assembled on site, he said.

“You basically assemble, kit-like, major portions of the plant in a modular fashion, a bit like aircraft and submarines are done,” Lipman said. “That was not fully shaken out completely at the Vogtle site.”

The Department of Energy under the Biden administration argued in a September report that future AP1000 builds should be less expensive because they won’t incur costs associated with the first-of-a-kind project in Georgia. Support from the department’s loan office, tax credits under the Inflation Reduction Act, and shorter construction timelines would substantially reduce costs, according to the report.

Trump plans for nuclear

While President Donald Trump is supportive of nuclear, it’s unclear whether the industry will receive support through DOE loans and the investment tax credit under the Inflation Reduction Act (IRA). Those tools were pillars of the Biden administration’s plan to help reduce the cost of new AP1000s.

Trump issued an executive order on his first day in office that directed federal agencies to remove obstacles to development of nuclear energy resources. The same order, however, paused all spending under the IRA. Two weeks later, Secretary of Energy Chris Wright made commercializing “affordable and abundant nuclear energy” a priority in a Feb. 5 order.

“The long talked about nuclear renaissance is finally going to happen, that is a priority for me personally and for President Trump and this administration,” Wright told CNBC in a Feb. 7 interview. Wright was previously a board member of Oklo, a nuclear startup that aims to disrupt the status quo of the industry by deploying micro reactors later this decade.

Wright emphasized commercializing small reactors and said private capital would drive the construction of new plants. Before the November election, Trump was skeptical of building large reactors, citing the cancelled project in South Carolina.

“They get too big and too complex and too expensive,” he told Joe Rogan in an October interview.

Lipman said the first Trump administration was pro-nuclear, and he expects the president will support the industry in his second term.

“If there’s going to be gigawatt scale deployment in the U.S., decision making needs to accelerate,” Lipman said. “The business model, the investment climate, any legislative changes that might be in the offing at the state level or the federal, now is the time to address those pertinent issues.”

— CNBC’s Gabriel Cortes contributed to this report.

Jack Dorsey’s Block got started as Square, offering small businesses a simple way to accept payments via smartphone. Affirm began as an online lender, giving consumers more affordable credit options for retail purchases. PayPal upended finance more than 25 years ago by letting businesses accept online payments.

The three fintechs, which were each launched by tech luminaries in different eras of Silicon Valley history, are increasingly converging as they seek to become virtual all-in-one banks. In their latest earnings reports this month, their lofty ambitions became more clear than ever.

Block was the last of the three to report, and the high-level numbers were troubling. Earnings and revenue missed estimates, sending the stock down 18%, its steepest drop in five years. But to hear Dorsey discuss the results, Block is successfully implementing a strategy of offering consumers the ability to pay businesses by smartphone, send money to friends through Cash App, and access credit and debit services while also getting more ways to invest in bitcoin.

“In 2024, we expanded Square from a payments tool into a full commerce platform, enhanced Cash App’s financial services offerings, and restructured our organization,” Dorsey said on Block’s earnings call on Thursday after the bell.

Block and an expanding roster of fintech rivals have all come to see that their moats aren’t strong enough in their core markets to keep the competition away, and that the path to growth is through a diverse set of financial services traditionally offered by banks. They’re playing to an audience of digital-first consumers who either didn’t grow up using a brick-and-mortar bank or realized at an early age that they had no need to ever set foot in a physical branch, or to meet with a loan officer or customer service rep.

“Longer term, we see a significant opportunity to grow actives, particularly among that digital-native audience like Millennial and Gen Z,” Block CFO Amrita Ahuja said on the earnings call.

As part of its expansion, Block has encroached on Affirm’s turf, with an increasing focus on buy now, pay later (BNPL) offerings that it picked up in its $29 billion purchase of Afterpay, which closed in early 2022. Block’s market share in BNPL increased by one point to 19%, while Affirm held its position at 17%, according to a recent report from Mizuho. Both companies are outperforming Klarna in BNPL, the report said.

Block’s BNPL play is now tied into Cash App, with an integration activated this week that gives users another way to make purchases through a single app. With Cash App monthly active users stagnating at 57 million for the last few quarters, the company is focused on engagement rather than rapid user acquisition.

“We think that there is significant opportunity for growth longer term, but there are some deliberate decisions we’ve made as part of our banker-based strategy in the near term” that have kept user numbers from increasing, Ahuja said. “This is a part of our continuous enhancements to drive healthy customer engagement as we bank our base.”

Compared to Block, Wall Street had a very different reaction to Affirm’s earnings earlier this month, pushing the stock up 22% after the company’s results sailed past estimates.

Affirm founder and CEO Max Levchin, who was previously a co-founder of PayPal, built his company with the promise of giving consumers lower-cost and easy-to-tap intstallment loans for purchases like electronics, jewelry and travel.

Affirm is also partnering with FIS to bring its debit card functionality to traditional banks.

Levchin left PayPal in 2002, after the company was acquired by eBay. It was a decade before he’d start working to help popularize the modern day BNPL market.

Now his former employer, which spun back out from eBay in 2015, is in on the BNPL game.

Under the leadership of CEO Alex Chriss, who took over the company in September 2023, PayPal is in the midst of a turnaround that involves working to better monetize products like Braintree and Venmo and joining the world of physical commerce with a debit card inside its mobile app.

Investors responded positively in 2024, pushing the stock up almost 40% after a brutal few years. But the stock dropped 13% after its earnings report, even as profit and revenue were better than expected. PayPal’s total payment volume for the quarter hit $437.8 billion, slightly below projections, while transaction margins rose to 47% from 45.8% — a sign of improving profitability.

One of Chriss’ big pushes is to get more out of Venmo, which has long been a popular way for friends to pay each other but hasn’t been a big hit with businesses. Venmo’s total payment volume in the quarter rose 10% year-over-year, with increased adoption at DoorDash, Starbucks, and Ticketmaster.

PayPal is also promoting Venmo’s debit card and “Pay With Venmo,” which saw 30% and 20% monthly active growth in 2024, respectively. The company is introducing new services to improve merchant retention, including its Fastlane one-click checkout feature, designed to compete with Apple Pay and Shopify’s Shop Pay.

Last year, the company launched PayPal Everywhere, a cashback-driven initiative designed to boost engagement within its mobile app. Chriss said on the earnings call that it’s “driving significant increases in debit card adoption and opening new categories of spend.”

As with virtually all financial services products, the new offerings from Block, Affirm and PayPal are designed to produce growth but not at the expense of profit. Banks operate at low margins, in large part because there’s so much competition for lower-priced loans and better cash-back options. There’s also all the costs associated with underwriting and compliance.

That’s the environment in which fintechs have to operate, though without the costs of running a network of physical branches.

Levchin talks about helping customers spend less, not more. And Block acknowledges the need for hefty investments to reach the company’s desired outcome.

“This is a part of our continuous enhancements to drive healthy customer engagement as we bank our base,” Ahuja said. “We’ve made investments in critical areas like compliance, support and risk. And as we’ve done that, we’ve progressed more of our actives through our identity verification process, which in turn, unlocks greater access to those actives to our full suite of financial tools.”

WATCH: CNBC’s full interview with PayPal CEO Alex Chriss