Physical Insights

An independent scientist’s observations on society, technology, energy, science and the environment. “Modern science has been a voyage into the unknown, with a lesson in humility waiting at every stop. Many passengers would rather have stayed home.” – Carl Sagan

The real costs of wind (and solar) power (and nuclear, too).

with 18 comments

I was involved in a blog-based discussion recently where somebody told me this:

But it will take at least 10 years to build a 1GW [nuclear] plant and cost 3-4b dollars.

In this same time frame, you could be installing 1GW of wind power a year, 10GW in the same time frame, for a cost of around 2-3M /MW 66% of the cost, ten times the amount in the same time frame.

Personally, I think that’s rubbish.

The largest wind farm in Australia is the Wattle Point wind farm in South Australia. It took about one year to build, between June 2004 and June 2005.

It cost 180 million dollars, and has a total nameplate capacity of 91 MW. But we’ve got to remember the capacity factor – I don’t know exactly what it is for this wind farm, but I assume it to be somewhere around 20%.

So, the “real” average power output capacity of the wind farm is 18.2 MW – a capital cost of $9.9 Million per MW.

So, to build 1 GW of real energy output capacity might be expected to cost 9.9 billion dollars – maybe a bit less if we can see some economy of scale – and take 55 years – maybe less than that, since you can build more of them at once if you need to.

Of course, a single typical modern nuclear power reactor can easily deliver 1 GW of “real” average power output – say, from an AP1000 with 1117 MWe nameplate capacity and 90% capacity factor, that’s over 1 GWe of average capacity – and capacity factors well over 90% are certainly achievable.

One wind farm just doesn’t compare to one nuclear plant – you’ve got to appreciate the actual amount of energy output being generated.

The actual construction time for a modern nuclear power reactor – say, a Westinghouse AP1000 – is three years. Three years, that’s all. Plus some number of years of approvals, planning and red tape.

Now, I’m not saying we shouldn’t have regulation and oversight of nuclear power – of course we should – but with these figures of 10 years that get waved around, how many of those years are going to be essentially political bickering, greenpeace rubbish and other nonsense that could easily be done away with?

In the Australia-specific context, how many years did it take ARPANSA to approve and licence the OPAL (research) reactor at Lucas Heights? I don’t recall exactly, but it wasn’t 10 years.

Yes, it might cost, say, $2 Billion – quite a bit of money. But the electricity has got to come from somewhere – and where ever it comes from, it costs money to install that capacity. (That’s a high-ish estimate, but let’s make a conservatively high estimate for Australia’s first nuclear power reactor, in Australian dollars) That’s certainly a lot less than the $10 billion (perhaps somewhat less) real-world capital price tag for 1 GW of wind capacity!

There was significant interest and excitement in Australia earlier this year after plans were announced to build what will be the largest solar photovoltaic power station in the world, using advanced concentrating heliostatic photovoltaic solar collectors.

The proposed solar generation facility would cost $420 million, and have a nameplate capacity of 154 MW, with a 20% capacity factor – generating 270 GWh annually. It will take about four years to construct – from 2009 to 2013.

I discussed this solar energy project and how much it will cost for the amount of energy generated in an earlier blog post – but I thought I’d briefly repeat those figures here for the sake of comparison.

So, if we scale that cost up, to the equivalent of a gigawatt of actual average power output capacity, that’s a cost of 13.6 billion dollars to generate the same amount of energy, (assuming no economy of scale comes into it, I admit – we’re talking about quick, imperfect analyses here)

Additionally, assuming you can’t scale up the rate of construction, it could be expected to take 130 years to construct that much solar capacity!

So, to summarise these estimates:

Solar: Probably around 10 billion dollars or more, and perhaps 100 years or more. (For 1 GWe of real, reliable output.)

Nuclear: Perhaps around 2 billion dollars, and perhaps around 4 years.

Wind: 10 billion dollars, possibly a bit less, and 40-50 years.

I deliberately base these estimates on technologies and designs that are around and are ready to be rolled out today – not on technologies or designs that are especially immature.

Comparing apples to oranges is fine – as long as you’re comparing a gigawatt-year of apples to a gigawatt-year of oranges. This is what I think many people out there overlook when getting really optimistic about the costs of these “clean, renewable” energy projects.


18 Responses

Subscribe to comments with RSS.

  1. As a professional nuclear engineer with thiry some years experience in the field, I would add a couple of comments: Your estimates on solar and wind seem realistic (we in the nuclear field have been saying it for years); However, I think you’re a little optimistic about the nuke plant. Given that Australia has never built a large nuke and the predicted bottlenecks in the supply chain, ten years and four billion Aussie dollars are more realistic. Still nuclear is the best available alternative.


    April 17, 2008 at 5:06 pm

  2. Do you know where the ubiquitous “wind is now cheaper than coal” claim originally came from? I’d like the opportunity to scrutinize the math behind it.


    April 18, 2008 at 3:40 am

  3. I hope to persuade some of the submarine nuclear propulsion groups, such as Westinghouse/Toshiba with their S6W, Areva or General Electric with their S8G (an eigth generation reactor)_, to bring their nuclear ship propulsion reactors into the free market for small (30MWe to 100 MWe) power stations. Russia is developing a very interesting ‘floating power station’for an estimated US$350m selling price.
    In ships they can run for 10 years between refuels, but I would assume a 3-year fuel load is mre appropriate. However, licensing and construction of less than 4 years per plant should be a target.
    50 MW electrical is a nice size for a town of approximately 4000 homes and some small industries. It is also a nice figure for a remote mine. That should leave another 50 MW process heat for other pruposes. The market is there, the technology is there, the fuel is available, it is small AND IT IS CLEAN!!!
    Do you think the idea will fly?


    April 18, 2008 at 10:57 am

  4. Those are certainly practical ideas to think about. 10-100 MWe power reactors are certainly do-able, providing nuclear energy on a smaller, more decentralized scale than the “business as usual” of big centralized multiple-GW units.

    Those reactors designed for naval propulsion are certainly quite different from the “usual” power reactors we’re familiar with at nuclear power plants – and they have a lot of potential advantages.

    Even if you could get it licensed, you can’t just buy a naval reactor from GE or Westinghouse though! Given the security and secrecy which surrounds information concerning the engineering, design and operation of US Naval nuclear propulsion technology, clearly licensing those reactors for commercial use presents problems.

    Luke Weston

    April 18, 2008 at 11:42 am

  5. You know, I can be a tree hugging, granola munching hippie from time to time. I am always a realist thought. I wish that ALL subsidies were removed from the whole energy generation sector and the ideas therein could compete in free market without any sort of distortion. Coal plants produce radioactive waste. Unlike nuclear plant waste the EPA allows coal waste to be disposed of as industrial rather than hazardous waste. So nuke plant customers have to pay disposal costs that nuke plant customers don’t. But solar is the worst. Don’t get me wrong, I love solar power. It’s a beautiful idea and it should be pursed, BUT… solar is enormously subsidized in its waste disposal. Though making semiconductors is energy intensive, (it takes modern solar cells 10 to 20 years to produce as much energy as they consumed to be produced.) that fact pales in comparison to this: making semiconductors is a highly polluting process. Silicon Valley has more EPA super fund sights (sights who’s clean up is more than a billion dollars) than any other area on earth. It also has more billionaires than any other area on earth. Solar cell makers can pawn off the their hazardous waste disposal off on taxpayers, and other energy sector business cannot.
    I’m a big believer in fair play. If we got the subsidy out of it all and let the technologies compete, I suspect nuclear would win.


    April 18, 2008 at 11:52 am

  6. Compare the costs of renaturation for atomic power plants with solar, wind and water – wait, there is no need for them the be renaturated.

    Instead compare the range of wind, sun and water with Uranium.

    Compare the reprocessing of used Uranium with… with what?

    Compare the level of atomic power reactor efficiency with coal plants.

    Compare the risc which rises with every new reactor.

    Just subsidise the new energy like the reactors were and are subsidised.


    July 8, 2008 at 12:51 am

  7. In 2006, Business Week magazine stated, “…,the [US] industry is aiming to build new plants for $1,500 to $2,000 per kilowatt of capacity,…”. However, they also added, “Trouble is, the cheapest plants built recently, all outside the U.S., have cost more than $2,000 per kilowatt.” 2007 estimates have considerable uncertainty in overnight cost, (over night cost being the current price for construction , does not allow for inflation, extended construction time , wage increases etc) conservative estimate of between $5,000 and $6,000/kW. In June 2008, Moody’s estimated the cost of installing new nuclear capacity in the U.S. as potentially exceed $7,000/kWe.

    Electricity generated from wind power within Australia is currently more expensive than that from coal plants. The Challicum Hills wind farm, one of the first in Victoria, cost $76 million for a 52 MW peak output. After operating for several years, the average capacity of the facility has been measured at 30%. The Capital Cost of the facility is therefore $USD 3660.00 (AUD$ 4880.0) per KW.

    S.T.S. student

    September 4, 2008 at 11:18 am

  8. Sorry these have been referenced from

    Luke these credible resources sort of contradict you estimates.
    I have never seen an estimate for nuclear power plant less than 10 billion.( but have only been studying this topic for the last couple of months).
    How did you arrive at a 2 billion dollar estimate?

    S.T.S. student

    September 4, 2008 at 11:26 am

  9. S.T.S. Student: Good comments. Thanks for the comments.

    To quote the following from the web page in question:
    “The cost of Nuclear Fission Power is dominated by the capital cost of construction of the plant. Proponents of next generation, (3rd generation), Nuclear Power plants project construction costs of around $1000 – $1200 per KW (Westinghouse AP1000, Advanced Boiling Water Reactor, Advanced CANDU Reactor, South Africa HTGR), significant increases in Uranium efficiency and substantial increases in operating life of the plant (60 years).”

    So, they said $1000-1200 per kW of nameplate capacity, and I said $2000 per kW of nameplate capacity – the estimate I used, $2 billion for 1000 MW of nameplate capacity, is actually a little higher than what they’ve used, and more conservative – but still, I fail to see much of a discrepancy.

    I think it would probably be fair for you to say that the figures for the projected cost of advanced Gen. III+ LWR builds are perhaps a little out of date – the general consensus I’ve seen over the last year or so is certainly more than $1000/kW, you’re right about that.

    That said, however, the site is a good site, and I do recommend it for your studies of this topic.

    Thanks for the additional wind farm example. Data with real capacity factor measurements rather than guesstimates is always good.

    As you’ve shown, if you look at the pessimistic high end of the costings for new nuclear build, you’re looking at costs where wind power would become competitive. In such a scenario, I’ve got nothing against wind power, but it remains to be seen if it would or would be practically scalable up to this scale of electricity generation, where it is actually replacing coal or nuclear generation.

    If you wish to compare wind with solar, for example, it certainly tends to be clear that wind energy is a more economically competitive option, between the two.

    Luke Weston

    September 4, 2008 at 6:58 pm

  10. Where does the information from Luke Wilson come from about solar panels “it takes modern solar cells 10 to 20 years to produce as much energy as they consumed to be produced.” I buy it but would like to see the study if there is one, thanks, steve

    steve goodall

    October 1, 2008 at 8:45 am

  11. “One wind farm just doesn’t compare to one nuclear plant – you’ve got to appreciate the actual amount of energy output being generated.”
    Of course not. One wind farm and one nuclear power plant don’t compare in cost, either. Per MW the cost difference is not that great. Most recent estimates blow nuclear out to around US$7,000 per kW, or A$10b for a 1GW reactor.

    As others have pointed out, there’s also the cost of inputs/waste processing to consider. For nuclear/coal/gas, not only do you have to pay for your input materials, but as we’ve seen in 2008 this exposes you to great financial risk as you are exposed to price movements of your inputs. The world economy was brought to its knees by $120/bbl oil, but nobody seems to see energy input costs as a risk that it’s better (and perfectly possible) to avoid.

    Secondly, your calculation of decades to get wind or solar capacity built rests on the assumption (and at least you’ve acknowledged it) that you can’t build two or more plants at once. In other words, it’s utter nonsense, as no such build-one-at-a-time-only limitation exists in the real world.


    December 19, 2008 at 3:10 am

  12. But you’re only taking into account one-off capital expenditure!

    How about the cost of fuel? I’d say it’s zero for wind and solar. How much is it for coal or nuclear, anyone?

    How about maintenance costs? How do you reckon nuclear power plants compare to wind farms? Not very favourably, I’d suppose.

    How about the security costs? Shall we put a figure on those as well?

    I’m fine with the idea of comparing apples with apples, but you could do that yourself too.

    Jimbo Jazz

    January 15, 2009 at 2:42 pm

    • The brief answer is that the capital cost is by far the majority of the cost of nuclear power.

      Mantainence costs for wind farms as compared to nuclear power? I haven’t seen detailed analysis, but remember that you’re comparing the cost of maintainence for a single nuclear power reactor plant to the total maintainence costs for approximately 1200 wind turbines.

      The cost of fuel, for nuclear energy, is relatively minimal, since the nuclear fuel is such a concentrated source of energy. Similarly, the sensitivity of the cost of nuclear energy to the uranium market is minimal – however, on the other hand, energy systems based on coal, and especially natural gas, are quite sensitive to the cost of the fuels.

      Luke Weston

      January 20, 2009 at 12:29 am

  13. Wow. 20%? try changing your figure there to about 60% and you’ll still be shooting way low for many areas. Some of our wind parks here in Cali produce around 95% and are told by the local electric company to stop producing because they don’t have enough transmission lines to transfer all that power. Also, what kind of turbines are we talking about here? How old are they? There is a huge difference in cost and potential production between a Nordtank 75 and a V90, try over 2900 kw of power difference. So obviously if you build a 1 GW park of tiny 75 kw turbines, you’ll pay an enormous amount in manpower and parts constructing, maintaining, and repairing all those turbines. However, if you construct a 1 GW park of GE 1.5’s or V90’s, your costs will fall dramatically while your ideal production will skyrocket. So instead of making assumptions lets post some facts.

    Rachel L.

    July 9, 2009 at 2:43 pm

    • 20% might be a little low, but between 20% and 30% is about right. For example, the Spanish electricity grid operator reported an output from wind farms that was 24% of peak capacity for 2008. And Spain has a lot of wind turbines.

      95% might be a peak output figure that some group of windfarms will reach occasionally, and as you illustrate, that will cause problems for the stability of the grid, because it is a huge power swing to go from say 20MW to 400MW of power input in a short time.


      August 4, 2009 at 11:00 pm

  14. Rachel, please show me one country that is able to demonstrate a capacity factor of over 25% for its wind turbines.

    Denmark cannot. Germany cannot. Spain cannot.

    These countries are the leaders in Europe for this technology. If they cannot make it work better than that, then why should you expect anyone else to?

    The burden of proof is on you.

    Brian Mays

    July 13, 2009 at 5:53 am

  15. How much will be paid by kwh to an insurance company to ensure the consequences of a nuclear disaster? How much will be paid by 7.000 years of custody of nuclear waste from a nuclear plant? How much will be paid by the dismantling of the nuclear plant at the end its life?


    July 19, 2009 at 10:13 pm

  16. I’d love to see your comments on hot rock technology since although it would not generate as much electricity from one place as nuclear, it could be highly dispersed, therefore less transmission lines.

    Utilizing existing oil and gas wells – I’m from Alberta, Canada, and applying a recycling amount of water first injected into earth crust to be turned into steam that is captured from a parallel well. That generates electricity, heat, supports itself. What’s wrong with that?

    B A
    PS – I’d love a direct reply to my email.

    B A Phillips

    July 23, 2009 at 5:31 pm

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: