The Breakthrough Institute’s Nuclear Notes, in a piece by Matthew L. Wald and Spencer Toohill, walks through the nuclear side of the new trade arrangement with South Korea. According to the White House, Korea will invest up to $200 billion in U.S. energy projects. Up to $120 billion of that is earmarked for eight large reactors: six Westinghouse AP1000s and two Korean APR-1400s. The APR-1400 has never been built in this country.
Those two are the ones to watch.
The APR-1400 is sold by the Korea Electric Power Corporation (KEPCO) and its subsidiary, Korea Hydro & Nuclear Power (KHNP). It is a pressurized water reactor descended from Combustion Engineering’s System 80+, the same design family as Arizona Public Service’s Palo Verde plant. Westinghouse bought Combustion Engineering’s nuclear business in 2000. Nuclear Notes reports that eight APR-1400s run around the world, and that the NRC approved the design in 2019.
The Koreans’ calling card is Barakah in the United Arab Emirates. Construction began in 2012, and the fourth and final unit entered commercial operation in September 2024. That record is a big part of why KHNP is sought after for project management.
The problem was who owns the underlying technology.
There is a catch, though: the Koreans did not have permission to use Westinghouse’s intellectual property associated with the design of the 80+ reactor, and U.S. law banned re-export of U.S. technology without permission. That meant a one-time deal to get the reactors built in the United Arab Emirates.
https://www.nuclearnotes.org/p/outlines-of-a-korean-nuclear-deal
On the settlement that followed:
In January 2025, Westinghouse announced a “global settlement” with the Korean companies, but kept the details confidential. According to published reports, the settlement requires Korea to pay Westinghouse royalties for each APR-1400 that Korea exports, and bars KEPCO and KHNP from North America and parts of Europe. The blog Neutron Bytes recently cataloged the public reports.
https://www.nuclearnotes.org/p/outlines-of-a-korean-nuclear-deal
Westinghouse announced the settlement on January 17, 2025. Korean reporting puts the royalty at around 200 billion won, about $143 million, per export project. Other reporting says the exclusion covers new large reactor bids across North America, the EU outside Czechia, the UK and Japan unless KHNP partners with Westinghouse.
And what the new deal changes:
But the new deal appears to make space for the first two Korean reactors stateside. If the APR-1400 goes head-to-head with the AP1000, the Korean companies will have to pay royalties to Westinghouse. However, the APR-1400 would have to be significantly less expensive to build for the Koreans to turn a profit. Other issues include Korea’s ability to bring large numbers of experts here to manage the work. In September 2025, armed American immigration agents raided a Hyundai battery plant in Georgia and arrested 300 technicians.
https://www.nuclearnotes.org/p/outlines-of-a-korean-nuclear-deal
The raid was at the Hyundai and LG Energy Solution battery site in Ellabell, Georgia, in September 2025. Agents arrested 475 people, some 300 of them South Korean nationals.
Korea may end up in the AP1000 business too. In June, the Energy Department’s Office of Energy Dominance Financing offered $17.5 billion in conditional loans to buy long lead-time equipment for up to ten AP1000s, and several utilities have applied. Nuclear Notes on where that equipment would come from:
But that hardware, for example reactor pressure vessels and steam generators, would most likely be built abroad, quite possibly in Korea. HD Hyundai Heavy Industries and Doosan Enerbility are likely candidates.
https://www.nuclearnotes.org/p/outlines-of-a-korean-nuclear-deal
Then there is ownership. Westinghouse is headed for an initial public offering, and Korea is weighing a stake, reported by Nuclear Notes at 5 to 20 percent. The U.S. government has floated a “golden share” with veto power over major decisions. Brookfield Asset Management, the Canadian firm that brought Westinghouse out of bankruptcy, holds the majority, and Cameco owns the rest.
H/T Dave Burton
This deal may depend more on how much Donald Trump dislikes Mark Carney than anything else.
Even tho Korea may have a big cost advantage in 3gen reactor technology, the Voglte 3-4 cost disaster makes me dubious about any further 3gen US nuclear generation investment. Ditto the Finnish disaster with the lastest French 3gen EdF design.
The US has abundant fracked shale natgas. CCGT uses much less capital, is much faster to build, and is easier to operate. The CCGT factory warrantied useful life is 40 years.
IMO, the wiser (US) nuclear move is to thoroughly investigate and then prove out with at least pilot scale builds over the next decades one or more of the several 4gen nuclear fission concepts currently in discussion (including SMR and MSR), then go nuclear with 4gen plus Japanese proven MOX fuel recycling. Let CCGT replace old coal during the next few decades while 4gen is sorted out.
China has adopted a similar strategy, except with USC coal instead of CCGT since they don’t have abundant natgas. They have already built and are operating a pilot scale MSR (15MWe) using the—if I recall correctly—uranium (not thorium) MSR fuel cycle. A big advantage of the uranium MSR fuel cycle (per MIT) is that it can use up spent 3gen fuel without recycling.
The MIT cost analysis for new-build nuclear in the US is that capital costs for American nuclear projects are driven mostly by the lack of a robust American nuclear industrial base in comparison with South Korea, China, and Russia.
Excessive regulatory impacts in the areas of LNT and ALARA certainly do have an adverse effect on capital costs. The Trump administration is now in the process of reducing and eliminating excessive LNT and ALARA regulation.
However, capital costs related to the NRC’s quality assurance requirements cannot and will not be reduced except through the reestablishment of a robust American industrial base which is capable of delivering QA-compliant nuclear systems and components on budget and on schedule.
The MIT analysis states that there is no inherent cost advantage per megawatt for a 4th generation reactor technology versus a 3rd generation technology. Both technologies currently run at roughly $18,000 per kilowatt capital cost. A reduction to a floor of perhaps $9,000 per kilowatt can be achieved for both 3rd Gen and 4th Gen technologies if enough firm orders can be placed to cover the costs of re-establishing the American nuclear industrial base.
We buy nuclear for its energy security and reliability benefits, for which we are obliged to pay a premium over what gas-fired generation costs. At this point in time, the only way for enough firm orders to be placed in order to reduce nuclear’s capital cost to a floor of $9,000 per kilowatt is for the federal and state governments to fund ten or so new reactors with taxpayer money.
What about the Vogtle 3 &4 cost and schedule disaster?
When we were doing cost and schedule estimates for new-build nuclear in the 2005-2009 timeframe, an assumption was being made that the nuclear project learning curve for professional staff, for component suppliers, and for in-field craft skills would have to be passed through for a second time.
Cost and schedule was added to the base estimates to cover the nuclear project learning curve, at which point it was assumed that new-build projects would be avoiding the mistakes made in the 1970’s and 1980’s and would be implementing all of the lessons learned from that period of nuclear construction history.
That’s not what was done at Vogtle 3 & 4. None of the many lessons learned from the 1970’s and 1980’s was addressed. Every classic nuclear project mistake that occurred in the 1970’s and 1980’s was repeated in the Vogtle 3 & 4 effort. They didn’t miss a one of the classic mistakes.
Very early on, it was apparent to those of us who have been around the block in nuclear project estimating and scheduling that Vogtle 3 & 4 was in deep trouble. Southern Nuclear, its prime EPC contractor, and its project management team were not prepared for the task that faced them.
I’ve said any number of times that going with nuclear power is strictly a public policy decision.
Georgia’s politicians, acting through Georgia’s PSC, decided that Vogtle 3 & 4 would continue forward regardless of how much the project would eventually cost, or how much time it would eventually take. And so the project did go forward to eventual completion. But only after the original EPC had been replaced and after a competent project management team had been installed.
I have read elsewhere that the hardware cost of Vogtle 4 did not change much as a consequence of the delay. Nearly all of the cost increase was due to poor QC and rework, over and over. Project management was very poor.
The driver for SMR now is two-fold, 1) provide a low carbon alternative to wasteful solar and wind proliferation, and 2) cash in on the $multi-trillion market for those places without access to inexpensive horizontally drilled and hydraulically fractured natural gas.
But can SMRs be built at a cost that makes their electricity profitable? So far, the answer seems to be no. Even NuScale, with their 70 MW electric PWR machine (NRC approved), using conventional proven technology that would likely work well, cannot be built at a cost competitive level and is reportedly near to bankruptcy. Several other SMR offerers have already gone bankrupt. Would it be cheaper to run one 1,000 MW PWR and one steam system with one big generator than to build 14 or 15 NuScale reactors or some other company’s reactors each with their own manning staff, reactor system, steam system, generator, cooling system and so forth? To believe so seems akin to believing that New York City could sell all of their near 6,000 city busses and replace them with 60,000 or more Ford Fiestas and expect the system to run on time at less cost. Economy of scale is real.
We have tried small reactors before. A portable US Army reactor, PM-3A at 1.8 MW electric, operated in Antarctica for about 10 years with constant maintenance and operating difficulty. It was finally shut down because of these difficulties. It was built to save the cost of acquiring and shipping diesel fuel and fresh water to Antarctica. It failed. Could a small reactor do better now? Likely yes. Would it be cheaper than diesel? Hmm.
Small reactors may have special applications where nothing else can do the job. The Navy’s Submarine NR-1 powered by a very small reactor at about 150KW electric ran successfully for 39 years. There may be other applications for small reactors for difficult jobs or in dangerous locations. But for supplying the grid? Don’t count on them.
It will take decades to accomplish, but future generations will be scratching their heads wondering why nuclear energy wasn’t pursued with vigor 1990 – 2030, instead of all the farting around with wind & solar.
That reminds me of an incidental conversation I was in with a fellow that was teaching classes on ISO 9000 (Quality System Management). It was around 1986, and he was an unemployed nuc engineer, making ends meet being a quality consultant. He made a comment that he thought the nuc industry was about to make a comeback. The fellows listening, comprised of engineers of various sorts, were pretty skeptical. The vibe about nuclear energy then was so negative, whether you supported or opposed, you knew building new capacity was dead.
The luddites, followed by the press and the activists, killed the commercial nuclear program. Undeterred the US Navy continued to invest in and upgrade their nuclear plants on both the subs and the carriers, whose reactors are essentially commercial-sized plants although most of their output is used to drive the ship and launch aircraft. In a prior life I was a reactor operator on our nuclear subs and was involved in the startup of an experimental reactor outside Saratoga Springs (designated as MARF, Modifications and Additions to a Reactor Facility, aka S7G) and was the reactor operator who took it to criticality for the first time. If there was a problem with the commercial reactor plants was that, unlike the Navy which had standardized plants, each plant in a different location was very different. For me, when I qualified on the Fast Attack Submarine USS Tunny (SSN682) for all watch stations so after transferring to the FBM submarine USS Robert E Lee (SSBN601) I was able to step right in fully qualified.
You posted: “Undeterred the US Navy continued to invest in and upgrade their nuclear plants on both the subs and the carriers, whose reactors are essentially commercial-sized plants . . .”
Not true. The largest naval reactors are rated at less than one-fifth of a large U.S. commercial reactor.
U.S. Navy nuclear reactors typically generate hundreds of megawatts of thermal power compared to large civilian (i.e., commercial) single reactors that typically generate 3,000-4,500 MWt.
The largest thermal power rating (MWt) for a U.S. Navy aircraft carrier nuclear reactor is approximately 700 MWt per reactor, found in the each of the two A1B reactors used on the Gerald R. Ford-class supercarriers. So, yeah, the largest US Navy nuclear reactor in use is only about one-fifth the output of the typical largest commercial reactor in use.
You give an example of how nuke should be done, standardized and built in factory settings. Sub reactors are more like SMRs, which then also shows the logic of the current buzz around SMRs, which also have the advantage of being smaller, easier to manage projects, with quicker payouts. E.g.: imagine 3 SMRs built one after the other with start-up and grid connection after each completion, versus waiting 10 years until a huge 1G reactor is finished to get to money rolling in.
1986. Just in time for Chernobyl. (Not a negative, just a timing fact.)
No need to scratch heads for future generations –
The war on cheap energy started when the Club of Rome was created – as deindustrialisation won’t happen with abudance of cheap energy.
It took about a generation+ tschernobyl to shift to reverse gear on an international scale and by 1992 it was set in motion with the Rio UN earth summit.(and such an ambitious project needs a global crisis, let’s say AGW.
And crisis is just another word for centralization of power.)
Vogtle 3&4, it surprises me that so many people make nice detailed smart sounding comments about the cost yet they always fail to state the correct reason why.
4 AP1000’s were approved under a Republican administration, construction was started.
After the election the Democrats won and set out to destroy the project.
Amendment 19 and the Aircraft Impact Assessment rule, retroactively applied.
This is what cost all the time and money.
People have no idea how this rule impacted the costs. The rule change caused construction to stop, components to be thrown away and or taken apart and rebuilt, re-inspected. The rule change also caused billions in new paperwork. Operating procedures, rules and regulation compliance, inspection procedures had to be created re-wrote.
Voglte 3&4 were a success, for Democrats destroying the public’s perception that nuclear power is competitive.
South Korea certainly has an impressive record in building nuclear – 13 reactors built since 1996 at an average time to construct of 4.5 years, the more recent ones faster than that.
Let’s work out the math for the above article:
Up to $120 billion of that is earmarked for eight large reactors = $15 billion per reactor.
That appears to be about right considering an estimate of about $35 million for Plant Vogtle Units 3 and 4 commercial nuclear power reactors.
However, it also says the US and Korea have learned little beyond the rate of inflation for increasing cost efficiency in nuclear power plant design, siting, and manufacturing in the ~30 years since Vogtle reactors were first designed/engineered (in late 1990s and early 2000s).
Of course, beyond this, we all know how reliable bureaucrat “earmarks” are for predictions. /sarc
Oh, well.
Ooops . . . correction to typo in my third paragraph above:
“. . . considering an estimate of about $35
millionbillion for Plant Vogtle Units 3 and 4 commercial nuclear power reactors.”South Korea built 4 reactors for the UAE in 12 years. The total cost has not been fully disclosed but it is of the order of only $30B. Sounds like you Americans are about to be ripped off big time.