An article from RenewEconomy gives some costs for nuclear power in Australian Dollars, but gives no sources for these figures at all. I’m checking into its rough accuracy. We need to note that this investigation is into failed projects, or projects with cost overruns, this may not be normal – but it usually seems to be.

The cost of the two reactors under construction in the US state of Georgia has doubled and now stands at A$20.4‒22.6 billion per reactor. In 2006, Westinghouse said it could build a reactor for as little as A$2.1 billion ‒ 10 times lower than the current estimate….

RenewEconomy (all references to the article linked to above)

These must be referring to Plant Vogtle units 3 and 4. According to the new builders they use:

the Westinghouse AP1000 advanced pressurized water reactor technology. This advanced technology allows nuclear cores to be cooled even in the absence of operator interventions or mechanical assistance. The AP1000 is the safest and most economical nuclear power plant available in the worldwide commercial marketplace, and is the only Generation III+ reactor to receive Design Certification from the U.S. Nuclear Regulatory Commission (NRC).

Georgia Power

Note this is not a fabled Gen 4 reactor.

Westinghouse was originally doing the build but went bankrupt, and other companies took over. There are a number of different prices for the project circulating, although the price has clearly increased and the build has been delayed, and it is not yet (Mar 2020) finished.

The two new reactors were expected to cost a total of about [US]$14 billion when the expansion was approved by the PSC in 2009, but the latest estimates from analysts put the current cost at [US]$27.5 billion. Units 3 and 4 originally were expected to come online in 2016. The current timetable calls for one reactor to enter commercial operation in November 2021, with the other following in November 2022.

Powermage 19 Feb 2019

Costs have ballooned from an initial budget of about [US]$14.1 billion…. [it] has doubled in price and is running more than five years behind schedule News 6 Mar 2019

The estimated total price for the project is expected in the [US]$18.7 billion range. 

Construction Equipment Guide 26 Dec 2019

Obviously there is some conflict over the cost. The project has been heavily backed by taxpayers in the form of loan guarantees. Reactors don’t seem to be buildable without heavy public financial risk – and remember one major US company, Westinghouse, has already gone bust over nuclear projects. And this taxpayer cost, or heavy loans, may increase the general price of electricity.

U.S. Energy Secretary Rick Perry… announced… another $3.7 billion in financing guarantees… on top of $8.3 billion in earlier loan guarantees made during the Obama administration, for a total of [US]$12 billion in guarantees backed by taxpayers. 

In addition to safety concerns about Plant Vogtle, critics have assailed the project’s future hit to customers’ wallets in Georgia. Already, typical residential customers of Georgia Power have paid hundreds of dollars over recent years to cover financing costs and company profits on the project….

Unit 3, the first of two new nuclear reactors, was supposed to be finished and operating almost exactly three years ago. Instead, it is now projected to be in service in November 2021. The second reactor, Unit 4, is slated to go in service a year later, in November 2022.

A project that had been underway in South Carolina [also being built by Westinghouse] using the same new reactor design was canceled in recent years as costs soared.  

Atlanta Journal Constitution 22 March 2019

The organisation ‘Taxpayers for Common Sense’ remarks:

If the Vogtle co-owners default on their DOE-guaranteed loans, the loss to taxpayers would be 24 times greater than the $500 million DOE lost on the now-defunct solar power company, Solyndra.

[The] latest estimate means the project is [US]$14 billion over budget and more than 5 years behind schedule….

Taxpayers for Common Sense 21 Mar 2019

There is a cheerful promotion piece from the US Department of Energy, which does not discuss the costs.

The successful completion of Vogtle Units 3 and 4 will set the tone for what could be a nuclear resurgence in the United States….

The new units are the first new reactors to begin construction in America in more than 3 decades. No date

This last fact alone, clearly indicates nuclear has not been popular with power companies, and the company building the reactors obviously has second thoughts about building more of them..

It probably will be in the 2030s or 2040s before Atlanta-based Southern Company attempts another nuclear construction project, Southern CEO Tom Fanning told analysts Wednesday.

Atlanta Journal Constitution 1May 2109

Further, there is a class action against Georgia Power, accusing them:

of overcharging customers millions of dollars in “cost recoveries” associated with the nuclear expansion project. The lawsuit charges Georgia Power has artificially raised municipal franchisee fees that appear on customers’ monthly bills based on the costs of the nuclear expansion.

Powermage 19 Feb 2019

There is a simple history of the project at Powermag 24 Sep 2018

Back to the RenewEconomy article

a twin-reactor project in South Carolina, was abandoned in 2017 after the expenditure of at least A$13.4 billion. Westinghouse filed for bankruptcy soon after, almost bankrupting its parent company Toshiba in the process.


This is presumably referring to the South Carolina reactor mentioned above, which helped Westinghouse go bankrupt. There are also allegations of corrupt processes.

Westinghouse bought thousands of hand-machined nuts that cost $114 each, rather than sturdier, off-the-shelf nuts that retailed for $2.20, according to The Post and Courier. There was a reason for that: Westinghouse got to charge 15 percent overhead on everything it spent. Every thousand nuts meant $17,100 in revenue for the company, rather than the $330 it would have collected if it used the cheaper version….

An audit by Bechtel Corp. two years ago found that the construction plans and design were faulty, and that the project was poorly managed. As one legislator put it, the entire project was “built to fail.” Jan 2018

It is obviously unlikely it was ‘built to fail’. This is more likely to have to do with the way public projects are carried out nowadays, with the aim of making as much profit, or cutting as many costs, as possible, rather than building the best that is possible, or being prepared to adequately deal with the complexity required of nuclear builds. Anyway, as stated previously Westinghouse went bust during this project, and we cannot assume that was deliberate.

South Carolina, in a bid to expand its generation of nuclear power in recent years, dropped [US]$9 billion on a single project — and has nothing to show for it…..

It started in 2008. SCE&G and Santee Cooper announced plans to add two nuclear reactors to the V.C. Summer Nuclear Station in Jenkinsville, South Carolina, and contracted Westinghouse Electric Company, owned by Toshiba, to handle construction. The state’s Public Service Commission (PSC) approved the plan in early 2009, with construction slated to begin in 2012, and the first reactor set to begin operating in 2016…..

Documents released as the project unraveled show that both SCE&G and Santee Cooper were well aware of shortcomings, mismanagement, and lack of oversight that eventually made the reactors impossible to complete, years before Westinghouse declared bankruptcy and both companies pulled out….

Thanks to a state law passed in 2007, residents in South Carolina are footing the bill for a massive failed nuclear reactor program that cost a total of $9 billion. Analysts say that corporate mismanagement and poor oversight means residents and their families will be paying for that failed energy program —  which never produced a watt of energy — for the next 20 years or more.

The Intercept 6 Feb 2019

Customers have already been billed some $2 billion for the reactors. Under current regulations, the utilities continue to collect $37 million per month. That means the average ratepayer is paying an additional $250 per year, or 18 percent of the bill. This could go on for 60 years. Jan 2018

So failed projects can increase electricity bills, as can successful ones…

former Nuclear Regulatory Commissioner Gregory Jaczko [spoke to] The Intercept… New plants, Jaczko said, take too long to build for the urgency of the climate crisis and simply aren’t cost effective, given advances in renewable energy. “I don’t see nuclear as a solution to climate change,” Jaczko said. “It’s too expensive, and would take too long if it could even be deployed. There are cheaper, better alternatives. And even better alternatives that are getting cheaper, faster.”

“They were allowed to charge the customers for all the money that they spent, plus a return,” Jaczko explained. “Even though they failed to deliver the project.”

The Intercept 6 Feb 2019

As with the Georgia reactor, there are a number of court actions trying to get the money Ratepayers have lost back. Plus court actions from shareholders and workers alleging gross mismanagement.

As an addenda, when looking for info on South Korea, I came across this causal remark about an earlier ‘successful’ venture:

The US’s $6.8 billion Watts Bar Unit 1 reactor in Tennessee had taken 23 years to complete, and cost more than 18 times its original $370 million price tag. 

MIT Technology Review 22 April 2019

Back to RenewEconomy Again.

The cost of the only reactor under construction in France has nearly quadrupled and now stands at A$20.0 billion. The cost of the only reactor under construction in Finland has nearly quadrupled and now stands at A$17.7 billion. The projects in France and Finland are both 10 years behind schedule, and still incomplete.


One independent Academic news report states, that, for the French case, Flamanville-3:

Construction started in 2007, with the final cost estimated at 3.3 billion euros. On October 9 the plant’s operator, EDF, annonced new delays, with costs now estimated at 12.4 billion euros and the opening pushed back to 2022 

The Conversation 28 Oct 2019: Stéphanie Tillement & Nicolas Thiolliere

The French Reactor Builders have recently made what seems like an unlikely claim to justify cost increases. If it is to be believed, quality control was almost non-existent:

Electricite de France SA said repairs of faulty welds at a nuclear plant under construction in western France will boost the project’s cost by 14% to 12.4 billion euros ($13.6 billion), adding further financial strain to the cash-strapped atomic power giant…..

EDF has increased its estimated bill for the project by 1.5 billion euros in the latest assessment, it said Wednesday in a statement. The almost-completed plant, which is already seven years behind schedule, won’t be able to load nuclear fuel before the end of 2022 as EDF needs to repair 66 welds, it said….

The budget for Flamanville-3 has more than tripled since construction started in 2007. The repeated setbacks, which have forced EDF to sell assets to curb debt in recent years, contrast with tumbling costs for solar and wind projects…..

The French government has asked EDF to prove by the middle of 2021 that it can build competitively priced nuclear plants to replace some of its 58 aging reactors. Competition from other clean-energy sources is stiff. France, like Britain, is working to step up the pace of building offshore wind farms.

Bloomberg 9 October 2019

Even before this, state environment agency ADEME (Agence de l’Environnement et de la Maîtrise de l’Énergie) said

Building new nuclear reactors in France would not be economical…, contradicting the government’s long-term energy strategy as well as state-owned utility EDF’s investment plans…..

“The development of an EPR-based nuclear industry would not be competitive,” ADEME said, adding that new nuclear plants would be structurally loss-making.

Building a single EPR in 2030 would require 4 to 6 billion euros of subsidies, while building a fleet of 15 with a total capacity of 24 gigawatt-hour by 2060 would cost the state 39 billion euros, despite economies of scale that could bring down the EPR costs to 70 euros per megawatt-hour (MWh), ADEME said…..

The gradual increase of renewables capacity could reduce the pre-tax electricity cost for consumers – including generation, grids and storage – to about 90 euros per MWh, compared to nearly 100 euros today….

EDF – which generates about 75 percent of French electricity with 58 nuclear reactors – declined to comment.

Reuters 11 Dec 2018

The prospects for Nuclear in France does not look good with the recent announcement that:

PARIS (Reuters) – France’s CEA nuclear agency has dropped plans to build a prototype sodium-cooled nuclear reactor, it said on Friday, after decades of research and hundreds of millions of euros in development costs.

Confirming a report in daily newspaper Le Monde, the state agency said it would finalize research in so-called “fourth generation” reactors in the ASTRID (Advanced Sodium Technological Reactor for Industrial Demonstration) project this year and is no longer planning to build a prototype in the short or medium term.

“In the current energy market situation, the perspective of industrial development of fourth-generation reactors is not planned before the second half of this century,” the CEA said.

Reuters 30 August 2019

The Finish Reactor called Olkiluoto 3

The EPR reactor in western Finland is already more than a decade behind schedule and had been due to start producing electricity in January 2020.


in October 2003, TVO announced that Framatome ANP’s 1600 MWe European Pressurized Water Reactor (EPR) was the preferred reactor on the basis of operating cost. Siemens was contracted to provide the turbines and generators. TVO signed a €3.2 billion turnkey contract with Areva NP and Siemens for an EPR unit in December 2003, with commercial operation expected in mid-2009…..

The [new 2018] agreement states that the supplier companies are entitled to an “incentive payment” of €150 million “upon timely completion” of the project. At that time the schedule was for grid connection in December 2018, and commercial operation in May 2019 – some 10 years behind schedule…. In April 2019, the target date was pushed back again to March 2020, and in July 2019, it was moved to July 2020….

The Areva-Siemens consortium was claiming €3.52 billion against TVO in relation to the delay and cost overruns of the project. The claim included payments delayed by TVO under the construction contract, and penalty interest totalling about €1.45 billion and €135 million in alleged loss of profit. TVO counterclaimed costs and losses of €2.6 billion to the end of 2018, having revised its loss figure from €1.8 billion to the end of 2014. 

World Nuclear Feb 2020

This again implies that costing is complicated, and companies cannot agree on what each other should pay towards it.

There were initially two power plants proposed for Olkiluoto, but:

TVO decided not to proceed, since “the delay of the start-up of Olkiluoto 3 plant unit … [makes it] impossible to make significant Olkiluoto 4 related decisions necessary for the construction licence application within the current period of validity of the decision-in-principle.” “Olkiluoto 4 is important for us and therefore we will be prepared to apply for a new decision-in-principle.” In June 2015 TVO shareholders resolved not to proceed with plans for unit 4.

World Nuclear Feb 2020

There is also a Russian design reactor to be built in Finland. The processes were agreed in 2010 but beginning production has been delayed again to 2028.

Back to RenewEconomy:

The cost of the four reactors under construction in the United Arab Emirates has increased from A$7.5 billion per reactor to A$10‒12 billion per reactor.


There are a lot of articles pointing to the dangers of nuclear energy in a politically unstable region and people have already fired missiles at them. I cannot find anything about cost increases. Most reports say that they are predicted to cost, in total, US$25b or currently about A$37.6b, not quite A$10-12 billion per reactor. It is possible the original cheapness arises from corruption on the side of the Korean builders.

South Korea ‒ which is supplying the UAE reactors ‒ is held out to be a model for the global nuclear industry. But South Korea is slowly phasing out its nuclear reactors, its nuclear industry is riddled with corruption (the courts have dispensed a cumulative 253 years of jail time to 68 offenders), and its business model clearly sacrifices safety in order to improve economics…


wikipedia reports on faked documents and certificates for components. Yes you really want fake parts for a reactor.

Korea is dismantling its nuclear industry, shutting down older reactors and scrapping plans for new ones. State energy companies are being shifted toward renewables.

South Korea’s reactors… are mostly packed into a narrow strip along the densely populated southeastern coast. The density was a way of cutting costs on administration and land acquisition. But putting reactors close to one another—and to large cities—was risky…. there are four million people living within a 30-kilometer radius of the Kori plant alone.

On September 21, 2012, officials at KHNP had received an outside tip about illegal activity among the company’s parts suppliers. By the time President Park had taken office, an internal probe had become a full-blown criminal investigation. Prosecutors discovered that thousands of counterfeit parts had made their way into nuclear reactors across the country, backed up with forged safety documents.

KHNP insisted the reactors were still safe, but the question remained: was corner-cutting the real reason they were so cheap?

MIT Technology Review 22 April 2019

Safety additions which came in after Chernobyl were abandoned. This helped the speed of the build, and partially explains why the Korean estimates for the UAE reactors were half the price of the competitors.

One informant told Technology Review:

“You’d have a group of white-haired executives from competing firms sitting across from each other, playing rock-paper-scissors to decide who would take certain contracts.” Dummy bids would then be supported by fake documents, doctored to ensure that the designated loser would fail.

“I personally knew of around 300 cases where those [load] transformers caught on fire. They’re incredibly unstable.”

MIT Technology Review 22 April 2019

Earthquake surveys had not been properly carried out, and there were earthquakes in areas of dense reactor concentration.

In May 2012, five engineers were charged with covering up a potentially dangerous power failure at South Korea’s Kori-1 reactor which led to a rapid rise in the reactor core temperature. The accident occurred because of a failure to follow safety procedures…..

[In General] The Korea Institute of Nuclear Safety reported:

# A total of 2,114 test reports were falsified: 247 test reports in relation to replaced parts for 23 reactors, an additional 944 falsifications in relation to ‘items’ for three recently commissioned reactors, and 923 falsifications in relation to ‘items’ for five reactors under construction.

# Results were ‘unidentified’ for an additional 3,408 test reports ‒ presumably it was impossible to assess whether or not the reports were falsified.

# Twenty-nine of the forgeries concerned ‘seismic qualification’, and the legitimacy of a further 43 seismic reports was ‘unclear’.

# Over 7,500 reactor parts were replaced in the aftermath of the scandal…..

The situation in South Korea mirrors that in Japan prior to the Fukushima disaster ‒ i.e. systemic corruption

Wiseinternational 23 Sep 2019

In total, 68 people were sentenced and the courts dispensed a cumulative 253 years of jail time. Guilty parties included KHNP president Kim Jong-shin, a Kepco lifer, and President Lee Myung-bak’s close aide Park Young-joon, whom Kim had bribed in exchange for “favorable treatment” from the government.

MIT Technology Review 22 April 2019

Casual attitudes to safety seem to persist:

an incident at the Hanbit 1 reactor on 10 May 2019 [when t]he reactor’s thermal output exceeded safety limits but was kept running for nearly 12 hours when it should have been shut down manually at once.17 The thermal output rose from 0% to 18% in one minute, far exceeding the 5% threshold that should have triggered a manual shutdown.

Wiseinternational 23 Sep 2019

The Korean government now seems in favour of a slow faze out of nuclear energy, and not to build any more reactors.

The Kori nuclear plant, north of Busan, has already had one of its four reactors decommissioned in June 2017, with the other three facing closure between the years 2023-2025.

Other nuclear plants facing decommissioning are Yonggwang 1, formerly known as Hanbit in 2026 and the Wolsong nuclear plant, which will be retired by 2022. Nuclear plants are also being put on hold or cancelled altogether: the Cheonji nuclear plant, formed of four reactors, was cancelled by the current government in June 2018, while the expansion of Shin Hanul from two reactors to four was delayed from May 2017 to February 2019 due to government policy.

Power Technology 1 Aug 2019

To the UK.

In the UK, the estimated cost of the only two reactors under construction is A$25.9 billion per reactor. In the mid-2000s, the estimated cost was almost seven times lower. The UK National Audit Office estimates that taxpayer subsidies for the project will amount to A$58 billion, despite earlier government promises that no taxpayer subsidies would be made available…..


I’ve written about the UK reactor before, but to reiterate:

It is currently estimated to cost around 22.5 billion pounds and is still at least five years away from producing electricity.

A document which seems to be from 2013-14 by EDF estimates the cost of the reactor at 16 billion pounds. {It talks about something being created in 2013 as if that is the past, and its earliest appearance in is 2014. The document is frequently said to be from 2012, but I cannot see any evidence for this.}

The price tag is expected to exceed £20bn, almost double that suggested in 2008 by EDF Energy, which is spearheading the project alongside a Chinese project partner.

At the time, EDF Energy’s chief executive, Vincent de Rivaz, said the mega-project would power millions of homes by late 2017. He pegged the cost at £45 for every megawatt-hour.

[Now EDF] will earn at least £92.50 for every megawatt-hour produced at Hinkley Point for 35 years by charging households an extra levy on top of the market price for power…. The average electricity price on the UK’s wholesale electricity market was between £55 and £65 per megawatt-hour last year.

The Guardian 14 Aug 2019

This type of reactor has been built elsewhere, and:

“It’s three times over cost and three times over time where it’s been built in Finland and France,” says Paul Dorfman, from the UCL Energy Institute. “This is a failed and failing reactor.”….

British electricity consumers will pay billions over a 35-year period. According to Gérard Magnin, a former EDF director, the French company sees Hinkley as “a way to make the British fund the renaissance of nuclear in France”. He added: “We cannot be sure that in 2060 or 2065, British pensioners, who are currently at school, will not still be paying for the advancement of the nuclear industry in France.”….

[The UK Government] offered to guarantee EDF a fixed price for each unit of energy produced at Hinkley for its first 35 years of operation. In 2012, the guaranteed price – known as the “strike price” – was set at £92.50 per megawatt hour (MWh), which would then rise with inflation. (One MWh is roughly equivalent to the electricity used by around 330 homes in one hour.)… The current wholesale price is around £40 per MWh…..

 if EDF abruptly sold a lot of electricity on to the market at a pre-planned time, the wholesale price could drop substantially. The lower the wholesale price, the bigger the difference from the fixed strike price, and therefore the higher the “EDF tax” paid by consumers.

The Guardian 21 Dec 2017

More recently:

The cost of building the UK’s first new nuclear power plant in a generation has risen by up to £2.9bn and the total bill could be more than £22bn….

The Guardian understands that the latest cost increase brings EDF Energy’s internal rate of return down to between 7.6% and 7.8%. The project originally offered a 9% return on investment, which slipped to 8.5% after its 2017 cost review….

The cost of supporting new offshore windfarms from the mid-2020s fell to record lows of about £40 per megawatt hour of electricity last week in an auction for government contracts, less than half the cost of Hinkley Point C.

The Guardian 26 Sep 2019

Here is some new stuff:

The government has confirmed plans for consumers to begin paying for new nuclear reactors before they are built, and for taxpayers to pay a share of any cost overruns or construction delays….

The new funding structure could be used to prop up EDF Energy’s £16bn plans for a new nuclear reactor at Sizewell B in Suffolk, which was left in doubt after fierce criticism of the costs surrounding the Hinkley Point C project in Somerset.Advertisement

It could also resurrect the dormant plans for a £16bn new nuclear reactor at the Wylfa project in North Wales, which fell apart last year due to the high costs of nuclear construction…..

Dr Doug Parr, the chief scientist at Greenpeace, said: “The nuclear industry has gone in just 10 years from saying they need no subsidies to asking bill payers to fork out for expensive power plants that don’t even exist yet and may never.”

“This ‘nuclear tax’ won’t lower energy bills – it will simply shift the liability for something going wrong from nuclear firms to consumers,” he added.

The Guardian 24 Jul 2019

Backers of mini nuclear power stations [SMRs] have asked for billions of pounds of taxpayers’ money to build their first UK projects, according to an official document…..

the nuclear industry’s claims that the mini plants would be a cheap option for producing low-carbon power appear to be undermined by the significant sums it has been asking of ministers.

Some firms have been calling for as much as £3.6bn to fund construction costs, according to a government-commissioned repor

The Guardian 1 Oct 2018

SMRs apparently have a context which also needs to be built.

No company, utility, consortium or national government is seriously considering building the massive supply chain that is at the very essence of the concept of SMRs ‒ mass, modular construction. Yet without that supply chain, SMRs will be expensive curiosities.

All or almost all SMR projects are either dependent on government handouts or they are run by state-owned agencies. The private sector won’t bet shareholders’ money on SMRs to any significant degree but governments have “a once in a lifetime opportunity” to bet taxpayers’ money on private-sector SMR frolics and to offer SMR developers “full and ongoing Government support”…..

In the US, government SMR funding of several hundred million dollars is an order of magnitude lower than subsidies for large reactors (several billion dollars for the AP1000 projects)…

Of course, it could be argued that government funding for SMR programs is excessive given the strong likelihood of failure. A case in point is the mPower project in the US, which was abandoned despite receiving government funding of US$111 million.

World Information Service on Energy

Back to RenewEconomy

A December 2019 report by CSIRO and the Australian Energy Market Operator finds that construction costs for nuclear reactors are 2‒8 times higher than costs for wind or solar. Costs per unit of energy produced are 2‒3 times greater for nuclear compared to wind or solar including either two hours of battery storage or six hours of pumped hydro energy storage.


The Draft report states:

there is no hard data to be found on nuclear SMR. While there are plants under construction or nearing completion, public cost data has not emerged from these early stage developments.

draft Report p3

[and] nuclear SMR costs are very uncertain due to the lack of public cost data on completed international projects,


If this capital cost reduction pathway is achieved then nuclear SMR is competitive with CCS….. in other scenarios, nuclear SMR capital costs remain high.


The cost tables are complicated, and not reproducible here, but it looks to me as if the costs of energy of SRMs is higher than solar and wind, but I stand to be corrected on this


It is not sensible to accept the construction times or costs issued by contracting companies. It will likely cost much more and take much longer to build – especially if they are built properly and safely. The construction costs can add massively to power bills. At the moment, nuclear power seems to be more expensive than renewables with storage.

We are also not factoring in the CO2 costs. According to one source: Nuclear construction produces up to 37 times the CO2 emissions of renewable energy sources, some of this because of the mining and refining of uranium. As we know in Australia, Uranium mining is not always safe for the surrounds, being sensitive to floods and droughts. Then we have the costs of waste storage for long periods of time, and finally the massive cost of decommissioning old nuclear reactors. It was this cost which led the Conservative Government of Mrs Thatcher to take nuclear power back into public responsibility. No company would pay the massive costs of decommissioning, and so they figured the nuclear industry would collapse without government and taxpayers’ support.

Can nuclear power get better? Of course, everything can get better, but will it? At the moment this seems unlikely to be the case in time to help reduce climate emissions in a way that counts. And no one seems particularly interested in lowering decommissioning costs.