The Energy Collective Group

This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.

9,751 Subscribers


How much does nuclear power actually cost?

This the final article in the SA Mines & Energy Journal series on nuclear energy (issue 24, pg 34), about the economic bottom line for nuclear. Ben Heard, my co-author, has also blogged about this on DecarboniseSA. And if you want a second opinion on, read what Columbia University’s Jeff Sachs has to say (one of the smartest economists out there — I’d strongly recommend his 2011 title “The Price of Civilization“).


It does not take long in any discussion of nuclear power before people want to talk turkey. How much does nuclear power cost?

It’s odd that when it comes to nuclear power alone, some environmentalists morph into incredibly hard-nosed economic rationalists. If the solution can’t pay its own way from the get go, bad luck.

That suggests a misunderstanding of not so much nuclear economics, but of energy economics more generally. It also hints at an ideological position if the same criteria are not applied elsewhere.

In considering nuclear at all, we are looking to replace baseload fossil fuels at 100s or over 1,000 MW at a time. Take your pick of technology, including modern fossil fuels: that is never going to be a cheap task. There is no way around the “sticker shock” of a modern power facility.

If we want new, large-scale energy generation in Australia, there is a large price tag, comfortably in the billions of dollars range. If, as we would argue, response to climate change demands that any new baseload is zero-carbon generation, then the options are (currently) restricted to the more expensive end of the range for capital costs (fuel is cheap or free for these technologies).

So, what, in that context, can low-carbon options offer in terms of up-front cost?  Let’s take some real-world examples (for details of the following calculations, see TCASE 15: Comparison of four ‘clean energy’ projects).

If we take the oft-quoted Olkiluoto nuclear new build in Finland (oft-quoted because it is suffering major cost and time over-runs), we find that the new EPR design, with 1600 MWe of generation capacity, looks to be coming in at a cost of EU6.4 billion. That normalises to $6.0 bn per GWe when capacity factors are accounted for.

Dome 3 being lowered onto the Olkiluoto nuclear power plant in Finland. Cost is $6 billion per GWe, but with very high capacity factor.

A large (600 MWe peak) planned wind farm in South Australia, with a proposed 120 MWe biomass generation as back-up, will cost $1.2 billion, plus and extra $0.2 billion for the connecting infrastructure. That’s about $6.9 billion per GWe.

When we turn and face the sun,costs jump. Based on the proposed Moree Solar Farm, this large solar PV facility with no storage or back-up (i.e. not a true baseload solution) comes in at $19.6 billion per GWe. A concentrating solar thermal plant (based on the Spanish Gemasolar plant) with molten salt storage back-up can be had at a cost of $25.1 billion per GWe.

The lesson is clear.

Costs, like any other number, mean nothing sitting on their own. This is a question choosing the best option. Even using a notoriously expensive ‘first-of-a-kind’ nuclear example, new nuclear is still the best value for zero-carbon generation.

The proposed Moree solar PV farm comes in at $19.6 billion per GWe – with no electricity storage.

If we look beyond the infamous Finnish example to some of the other 60 new reactors under construction or more than 200 currently proposed,the picture becomes even clearer. South Korea is undertaking a substantial program of new nuclear build. Indeed, the South Koreans have sold their technology and expertise to the currently non-nuclear United Arab Emirates at a contracted price of $3.5 billion per GWe with 6 GWe to be delivered by 2018. Meanwhile the Chinese are delivering new nuclear based on the Westinghouse AP 1000 design for reported domestic cost of as low as $1.7 billion per GWe. So, if we want zero-carbon generation at scale, we would be foolish in the extreme to reject nuclear from consideration on capital cost grounds.

But  what we really want is the product of the power plant, not the plant itself: that is, dependable electricity. Here, nuclear excels, delivering electricity at an excellent price, with capacity factors now exceeding 90% in the U.S. and South Korea. Perhaps even more importantly, this price will be reliable. Thanks to negligible fuel costs and no carbon emissions in the generation, nuclear power is almost completely insulated from two of the biggest incoming pressures on power prices: carbon prices and fuel scarcity. When we are building expensive infrastructure with long life, such considerations matter a great deal.

So where does that leave us? Real-world experience tells us that nuclear can provide well-priced and reliable electricity. In capital terms, nuclear is the best-value form of zero-carbon generation, with miles of daylight to the competition. That may be a surprise, but this industry has learned. New designs are predominantly more standardised in design, and more reliant on passive, rather than engineered safety systems, and come in a range of sizes. All of this brings cost down.

That means the hurdle is up-front capital. Economist Tony Owen is clear about this situation, saying this:

If the CEO of, say, Origin Energy said to the board “I’ve got a great idea. Let’s spend $5bn of the company’s money, for which we will not start seeing a return for at least 5 years” he would be laughed at. In fact he would probably be sacked.

Tony’s point is a deadly serious one. He is not saying it’s difficult for fully private investments in nuclear, or other multi-billion dollar energy technologies for that matter. He’s saying it’s impossible.

If Australians want the best energy outcome as we undertake the challenging replacement of our aging fossil baseload, we will need to remember something: such projects are nation-building works.Whether we like it or not, some Government involvement will be required, to ensure a public good. This could be as simple as a loan guarantee (which protects the lender, not the vendor. The project must still stack up on financial grounds) such as the U.S. Government is providing. Or it could be something more complex, like an emissions-trading scheme and power-purchase agreements.

Our point is this: we can’t have something for nothing, least of all major infrastructure. The “barrier” of nuclear cost is one of our own creation, born of a lack of context and comparison, and our collective amnesia regarding nation building. We have a job to do. It is going to cost a lot of money, so we had better make sure we get the best result. If nuclear technology is a financial lemon, it won’t get up. There is no reason to exclude it from making its case, on a fair and level playing field.

Content Discussion

Tami Kennedy's picture
Tami Kennedy on August 16, 2012

As often in the argument for nuclear power the waste issue is ignored. Too often reactor sites turn out to be waste storage facilities. I'm not sure enough effort is put into engineering that as a permanent function.

You do hit the good point of government loan guarantees for initial investment. But the GOP suggests government is too big. The provision is eventually made with both sides taking responsibility for the federal support.

Solve the waste issue and be prepared to initiate federal funding for upgrades as we take the plants beyond the engineered life as normal procedure. NRC shouldn't be continually underfunded.

Mike Hanson's picture
Mike Hanson on August 16, 2012

Nuclear waste is a political, not a technical issue. 

Adam Clifford's picture
Adam Clifford on August 16, 2012

The in-your-face issues of storage,and duration of storage,of waste,management of waste effluent and decommisioning,both the cost and the disposal of the plant and it's radioactive parts,and the management of the waste-sites over long periods of time,do not seem to go away or become marginal concerns,especially with uranium nuclear reactors.

'The world's present measured resources of uranium, economically recoverable at a price of US$130/kg, are enough to last for some 80 years at current consumption.'[NEAIAEA (2006). Uranium 2005 – Resources, Production and DemandOECD PublishingISBN 978-92-64-02425-0.-Wikipedia]

is a popular statistic.The lifespan of a nuclear reactor ' Under the US licensing system, reactors are originally offered an operating licence for 40 years'.[with 20-year extensions available].

This article offers another aspect of the issue:

What little I know of thorium-fluoride molten-salt reactors suggest that they are more promising with much shorter-lived by-products,though having read this:

I remain as uncertain and unreassured.

Tami Kennedy's picture
Tami Kennedy on August 16, 2012

Totally agree. Selling the population on the technology is difficult. Not many schools offer even basic courses. There isn't much communication of ongoing testing and independent verification of results. I didn't see a government statement regarding the safety of our waste storage pools following Fukushima. Think of the jobs building and operating the high-level waste processing at each site.

Bill Hannahan's picture
Bill Hannahan on August 16, 2012


Excellent report Barry. We have yet to design the Model T of nuclear power plants. The potential cost reductions of a standardized design, factory mass produced are huge.

Imagine that the government forced you to plug up the sewer pipe from your house and store all your wastewater and solid waste on site for sixty years or more. That would be almost impossible, hugely expensive, and very risky. Imagine a similar requirement applied to a coal plant, natural gas plant, gas fracking operation or stockyard.

 The fact that a nuclear plant can store 60 years of spent fuel in a medium sized pool or a modest number of air cooled casks indicates how small and simple the problem really is. We have many solutions for nuclear waste, we have simply not selected one for implementation. The most practical approach is to bury the waste under the seabed.

We need to restart an all out R&D program to develop safe breeder reactors, especially molten salt reactor technology, that can split all the uranium and thorium atoms mined to fuel them. They will produce about 6 ounces of fission products per 80 year lifetime, requiring only 300 years of isolation.



douglas card's picture
douglas card on August 17, 2012

I dont know where you get the $20 per watt for the Solar PV, but in California, the cost is below $5 per watt.  You can get a 10 kW PV system on your roof for less than $100K without any subsidies.

Steven Scannell's picture
Steven Scannell on August 17, 2012

I have a radical approach to nuclear energy.  I think the US Navy should be in business to operate nuclear plants.  Each plant would be a Navy base.  The standardization of plants could save money, funding would not be so much of an issue, and the plants power could be sold onto the market. This is certainly a radical idea, but I think it may be practical.  For other countries our credibilty, faith, and credit would be welcomed. It would be a good industry for us.  A special training base could be used to train people in foreign countries, and they could serve under the US Navy, as special employees, in foreign plants.    I would recommend three sizes of reactors, which can be paired in multiples.  In conjunction with these core nuclear plants wind to CAES could augment the base loads, to offer a peak not attainable under nuclear alone.   Any steam plant can be supercharged with CAES from the wind, wave, geothermal, etc.  Tripe System, or Track-Pipe, five or six feet in diameter thick walled multiple (13 one large, twelve small imbedded) conduit composite (  provide the CAES,  which carry monorail systems and or futuristic wide gauge train systems.   Compressed air at sea is like pumped hydro on land, in many ways.  But unlike the pumped hydro, 8,000psi CAES using track pipes, both as conduits and storage, offer realistic green energy on tap, where we need it. CAES is the common denominator, easily comodity economic applications, and will work well as an interface with the old and new.  Of course the uses of CAES are basically uncharted terrirory, but they include refrigeration and AC, car and truck hybrids, various water and sanitation pump applications, major public works national water systems, and more.  Using CAES/nuclear we can build smaller plants.  Co-Gen systems are a bonus, using very high pressure steam, able to ship in moderate distance loops. 

So I'm advocating this raw idea, from a systems perspective, as credible to model up. In this way nuclear power could be very viable. 



Robert Bernal's picture
Robert Bernal on August 17, 2012

$25 a watt for solar? Really? Plain old solar panels are way cheaper even considering their low capacity factors (about $2 per watt times 4 for capacity).

Molten salts for solar and wind may be the best choice despite inefficiency because it is sooo much cheaper than batteries. However, batteries are not yet mass produced in robotic factories for pennies on the dollar, as they should be mandated to be.

Resulting jobs would be far in excess of any jobs lost to automation due to increased installations.

However, (LFTR or LMFR) nuclear is much better in terms of MWh's per acre and parts per MW. Their only problem is of political and monetary backing, as they are old technology.

Conventional light water reactors are just plain stupid in light of these other, proven, vastly more efficient and inherently safe nuclear energy choices (except that I like that Navy idea posted earlier!). LWR's will cause major problems if the grid gets "knocked out" by a solar storm (too). Not so with LFTR or LMFR because they don't rely on explosive water for cooling and will passively cool in an emergency.

Thus, a major collaborative needs to be formed to substantially reduce manufacturing costs of solar, wind and batteries by use of advanced machine automation, and to kickstart re-development of LFTR's and LMFR's which, bty, can solve the (un)spent nuclear waste issue.

Paul Ebert's picture
Paul Ebert on August 21, 2012

Adam, you might be interested in this rebuttal to the Jonathon Porritt paper:

wind smith's picture
wind smith on August 21, 2012

Barry, maybe you or others can answer this nuclear power plant question.

One of the main reasons ,besides it's nearly carbon free power, to use fission is it's 24/7 base load capability. Just today our Monticello and 1/2 of the Prarie Island nuclear plant was restarted after a 7 day shut down. To replace the 1150 MW, Excell purchased or increased output at other plants. So is it true that the main cost difference between renewables and nuclear is the capacity factor which gives one the CO2 reductions over time because back up is needed in either case? Is it a matter of how much back up is running that makes the difference?

Nathan Wilson's picture
Nathan Wilson on August 21, 2012

I belive the author is refering to cost per average watt delivered, in other words, cost/capacity_factor.

So for your $5/Watt example, assuming 20% capacity factor, that's $5/.2 = $25 /Watt_avg.

$5/W is good for residential, $4/W more typical for utility scale.  

20% capacity factor would be realistic for fixed PV in California with professional mirror cleaning; maybe 10% in Germany with homeowners doing the cleaning (or neglecting).  

For solar thermal plants with energy storage, the capacity factors vary.  With 5 hours of storage and enlarged collection area, expect a capacity factor of 40%.  With 15 hours of storage and even more area, the plant runs 24/7 for the summer months, but still get a daily shortfall in the winter, plus get zero output on dozens of cloudy days per year, yeilding about 70% capacity factor.

Nathan Wilson's picture
Nathan Wilson on August 21, 2012

Batteries are indeed made by the millions in robotic factories (just think of all of the cellphones, laptop computers, and cars that are built every year - they all have batteries).

Don't expect lead-acid batteries (i.e car batteries) to come down at all, they are mature.  High tech lithium ion batteries may continue to fall in price to maybe reach parity with lead acid, but that still leaves them several times more expensive than pumped hydro or molten salt energy storage.

Nathan Wilson's picture
Nathan Wilson on August 21, 2012

Here is the US government's prediction for electricity cost by source:

It shows nuclear for $.11/kWh, solar PV for $.15/kWh, solar thermal for $.24/kWh.

The price of renewables always assumes that there is plenty of dispatchable fossil fuel back to quickly ramp up when the wind dies or the sun goes down.  If you try to design a purely solar energy system, a big chuck of the total energy has to come from short-term storage (15 hours), which is really expensive.  Worse, another big chunk has to come from long term storage (many weeks), which can only be done with hydrogen or other synthetic fuel (assuming there are no more available sites for big hydro, as in the US).  Burning synthetic fuel to make electricity is only about 40% efficient, and therefore horribly expensive.

With nuclear, very little of the total energy comes from fossil backup or storage.  This is because most of a reactors downtime is the refuelling outage, which is done every 18 months or less, and is planned for a time of year when demand is low (spring or fall).

douglas card's picture
douglas card on August 22, 2012

I am unfamiliar with the term cost/capacity.  I was obvioulsy (to anyone who knows what they are talking about) referring to the actual cost of installation.  There is nothing else that matters except the watts produced, which is lowered only by transmission thru cables, inverter, and the amount of insolation.  A 5 kW system will generate an average of 30kWh on a summer day and 20kWh on a winter day with no major clouding at the SoCal latitude.

We were only talking about the cost to install if I remember corectly.


The only ongoing cost is minimal.  Cleaning once or twice a year and a new inverter at 12 to 15 years.

Robert Bernal's picture
Robert Bernal on August 22, 2012

Perhaps automation and further economies of scale will bring down the costs of whatever best lithium based car battery (such as the lifepo4 which have less thermal issues and longer cycles).

I mean, we need companies that own the mines, own the mining equipment, own robotic production equipment, own the trucks and ultimately, even the station where batteries are swapped in a matter of seconds... just like the oil companies.

Nichol Brummer's picture
Nichol Brummer on August 28, 2012

Let me start by stating that I'm not against keeping nuclear energy as a viable option for the future, and certainly continuing development of those options that promise to be safer, may be able to 'burn' the waste from old reactors, and may also be safer when it comes to proliferration of nuclear bombs.

As things stand, however, I don't see nuclear energy quickly scaling to world scale. Which countries can we trust enough to sell nuclear plants to? If we are serious, that is a precious small part of the world. If we can sell a nuclear plant to Iran, then I'll be convinced it is safe and sound. We have seen even a very advanced and reliable country like Japan make mistakes, and get into big trouble during a natural disaster. For the time being, I'd feel a lot more comfortable if nuclear energy remains a relatively risky niche solution that isn't considered safe and ready for the world stage.

One huge advantage of solar PV on a roof is that it is usually produced close to where it will be used, and therefor does not need much additional infrastructure to transport. Even better: adding distributed solar PV will free up some of grid capacity while we're still thinking about upgrading the grid to what we need today.

About Nuclear energy: the story above shows the other disadvantage of nuclear in that it only comes in very large increments, needing massive investment. Wind and solar need many more parts, but these are totally independent, and can be built separately. They also start working immediately after having been installed and connected to the grid. That is a huge advantage.

It may be that new nuclear energy must be optimized for being flexible, in combination with energy storage, to be complementary to wind and solar energy?

Adam Clifford's picture
Adam Clifford on August 31, 2012

Thanks for that,Paul.i staggered through that and got the drift.Basically the Porritt article was slanted and information omitted and not developed to cast thorium-fluoride molten-salt reactors in a bad light.I think they could be a game-changer-I have always been against nuclear fission because of the cost,the waste,and it's duration,it's dispersal of effluents,the accidents,big and small,and their decommissioning.Reducing the impact/density/radioactivity[danger and it's duration] of the nuclear fuels and waste brings the nuclear reality in to greater focus.

I'm increasingly finding the mixture of science,politics,ideology and commerce absolutely,well,dangerous,and I'm sure contributes to the uncertainty about things scientific in the population as epitomised by the 'creationists presence and capability of totally ignoring science.

It's a huge issue.

Bob Meinetz's picture
Bob Meinetz on November 6, 2015

TImothy, somehow your misguided comment was “featured” on TEC’s homepage, so I suppose it deserves a response. 

A significant amount of analysis has been devoted to determining the lifecycle carbon emissions of various generation technologies, and the Intergovernmental Panel on Climate Change (IPCC), after consulting much of it, has estimated the lifecycle emissions of nuclear to be four times less than rooftop solar – and sixty-eight times less than coal.

Please stop spreading this irresponsible nonsense – nuclear energy is the most promising avenue we have for addressing climate change. Though I have no doubt of your good intentions, your actions are working at cross-purposes with them.

Nathan Wilson's picture
Nathan Wilson on November 7, 2015

Maintenance outages at nuclear plants are scheduled for a time of low demand (daily demand peaks are typically 1.5 times higher during peak seasons compared to Spring & Fall), so the fleet average output actually follows the seasonal demand pattern.  

For nuclear and fossil fuel power plants, normally the amount of backup capacity (“reserves”) required is about 15% of the total.  This will accomodate the failure of one or two major generators in a typical “balancing area”.  For solar or wind, grid operators have to be prepared for a weather system that zeros out everything at once, so the total reserve requirement is much higher (although the “spinning” reserves are typically only somewhat higher).

Nathan Wilson's picture
Nathan Wilson on November 8, 2015


“30% of the emission of coal” is lower than what you get using solar and wind with fossil gas backup.

Willem Post's picture
Willem Post on November 8, 2015


New nuclear plants, CF about 0.90, would be designed fo 60 y, whereas wind tubines, CF about 0.30, need major refurbishing well before their 20th y.

Calculate the production over 60 y for both, and you will see you need 3 times the MW of wind turbines.

But that is not all. The wind turbines require support of the other generators for filling and balancing 24/7/365, and in case of higher wind energy penetration, energy storage systems, plus extensive grid build-outs.

Germany’s wind energy sometimes is only a few percent of capacity any time of the year.

That means all other generators and energy storage are needed to meet demand Any time of the year.

To compare the cost of nuclear with wind and solar is pure nonsense, apples to oranges.

Bob Meinetz's picture
Bob Meinetz on November 9, 2015

Nathan, you can likely expect criticism for quoting an industry source with your graphic. IPCC’s more recent (and thorough) analysis, however, shows nuclear with a median value one-half that of NEI’s chart, which references a 2002 study.

Because when picking nits, the importance of calculating median nit-mass can’t be overemphasized.

Keith Pickering's picture
Keith Pickering on November 10, 2015

Timothy, you’ve been misinformed. Full lifecycle meta-analysis by the National Renewable Energy Laboratory has found the following total emissions (in gCO2e per kWh) for various technologies:

Hydro 7

Ocean 8

Wind 11

Nuclear 12

Photovoltaic (CdTe) 14

Photovoltaic (a-Si) 20

Geothermal 25

Solar thermal (CSP) 25

Photovoltaic (CIGS) 26

Biomass 30

Enhanced geothermal 57

Natural gas CC 450

Natural gas CT 670

Coal 980

Similar full lifecycle meta-analysis by the IPCC has found similar numbers.
So no, that’s not 30% of a coal plant, and not even close. It’s about the same as wind, and significantly better than solar in all its forms.
From the recent analysis by Weissbach et al (2013), wind uses about 10 times more steel per kWh generated, and about 3 times more concrete per kWh generated, than does nuclear.
wind smith's picture
wind smith on November 10, 2015

Keith, do you know if the the steel and concrete used for wind versus nuclear is based on a 20 year life for wind and a 60 year life for nuclear? If this is the case, do they consider the footing and tower under the 20 year life span? (I assume all the concrete is in the footing and the greater share of steel used in the tower and footing rebar) If so, is that do to the fatigue loading of either one, or is it something else that limits the life of the footing and tower? If not, is there empirical data or modeling for the life of the footing and tower that would allow continued multi decade use of footing and tower, rebuilding with new blades and generator?

At the end of the 40 to 60 year life of the nuclear plant is most of the steel recyclable or will most of it need to be considered low level nuclear waste?

wind smith's picture
wind smith on November 12, 2015

Keith, do you know if the the steel and concrete used for wind versus nuclear is based on a 20 year life for wind and a 60 year life for nuclear? If this is the case, do they consider the footing and tower under the 20 year life span? (I assume all the concrete is in the footing and the greater share of steel used in the tower and footing rebar) If so, is that do to the cyclic fatigue loading of either one, or is it something else that limits the life of the footing and tower? If not, is there empirical data or modeling for the life of the footing and tower that would allow continued multi decade use of footing and tower, rebuilding with new blades and generator?

At the end of the 40 to 60 year life of the nuclear plant is most of the steel recyclable or will most of it need to be considered low level nuclear waste?