Blog Post

Nuclear Power By the Numbers

With all of the attention on nuclear power this week — due both to the $8.3 billion in loan guarantees awarded to build the first nuclear plants in the U.S. in almost three decades, as well as the attention Bill Gates is paying to nuclear project TerraPower — we’ve been mulling over the numbers for nuclear. There’s some very big numbers (it’s one of the most expensive clean power technologies out there) and some smaller figures (there’s been a recent industry focus on small reactors).

Here’s a look at nuclear power by the numbers — from big to small — with figures courtesy of the Energy Information Administration, the Wall Street Journal, Greentech Media and the Nuclear Energy Institute.

$54 billion: The proposed amount for federal loan guarantees for nuclear power reactors by the Obama administration.

$10 billion: How much a large nuclear reactor can cost to build.

$8.3 billion: The amount of the U.S. federal loan guarantee for the first two nuclear power plants built in the U.S. in 30 years.

$750 million: An average cost for a small nuclear reactor.

$25 million: The cost of Hyperion Power Generation’s nuclear battery reactor.

$4,000-$6,000: The capital costs (per kilowatts electrical (kWe)) to build a nuclear reactor.

1,000 MW: The amount of power that an average nuclear reactor delivers.

700-1,000: The amount of permanent jobs created by building a large nuclear reactor.

125-140 megawatts of power: The amount of power that can be produced by a small nuclear reactor.

104: The number of nuclear plants operating in the U.S.

100: The number of nuclear reactors planned and under construction in Asia-Pacific region.

80 percent: The amount of France’s power supply made up by nuclear.

15 feet by 60 feet: The size of nuclear startup NuScale’s reactor.

20.2 percent: The amount of the U.S. power supply made up by nuclear, according to the Energy Information Administration.

6 years: Minimum amount of time it will take before the next wave of nuclear reactors will come online in the U.S.

5 feet by 5 feet: About the size of Hyperion’s nuclear battery — the size of a hotub.

0: The amount of carbon emissions nuclear power releases.

For More on Greentech from GigaOM Pro (subscription required):

Facebook’s Coal-Powered Problem

New Opportunities in the Smart Grid

Image courtesy of exquisitur’s photostream Flickr Creative Commons.

16 Responses to “Nuclear Power By the Numbers”

  1. Nuclear power plant thermal efficiency lower than the average fossil fuel power plants, which emit more heat into the environment, so nuclear power plant thermal pollution more serious

  2. Official govt. statistics dispute the article’s statement that nuclear is “one of the most expensive clean power technologies out there”. According to the energy information administration (link below), all non-emitting generation technologies are as or more expensive than nuclear. Wind and (especially) solar are highest of all. And this is just the raw, per kW-hr cost, before costs related to transmission and intermittentcy are considered.

    Concerning transmission costs, this is something faced by all centralized generation options, and it must be noted that the great majority of new renewable generation, including wind and solar, is in the form of large, centralized, utility owned plants. In fact, the transmission costs are actually much greater for large wind and solar plants, due to their being spread out over a large land area, and being very remote from demand centers.

    In terms of nuclear’s life cycle emissions, I agree with everything Mr. Hoffman says except the part about decreasing ore grades. Uranium is a ubiquitous element in the earth’s crust that we’ve barely started to look for. There are centuries worth of high-grade ore deposits out there that we haven’t found yet. Note that in the 1920’s or so, the “official” endowment of oil was ~1% of what we now know is there (i.e., that was found later in the 20th century, after making the effort to look). This is where we’re at with respect to uranium.

    Nuclear emits ~10-20 grams per kW-hr. Of note is the fact that the link Bilsko provided, it states that the studies that yielded the high CO2 results were commissioned by “green” groups. Nuff said. Nobody has any respect for the Storm analysis.

    • On life cycle green house gas emissions analysis –
      There have been over 100 life cycle greenhouse gas emissions studies of nuclear power and other technologies. The results of the study Bilsko cited are well outside the typical range of results reported for nuclear (more often 10-50 g/kWh, in line with wind and run-of-the-river hydro).

    This is a compilation of the results of about 10 different studies of the topic performed in the USA, UK, Switzerland and Japan by government agencies, national laboratories, utilities, and environmental groups. It’s a very good cross cutting and shows surprising consistency between the very different groups with very different motivations.

    The life cycle greenhouse gas emissions of nuclear power plants depend greatly on a few different assumptions:
    -How many years the plant operates
    -What fraction of the time the plant is running
    -What enrichment technology is used and whether fuel is recycled
    -What grade ores are used
    -What mining technology is used (pit vs. in-situ leeching)
    -What the energy source is for enrichment and mining

    When a paper reports life cycle emissions over 100 g/kWh, it is though making pessimistic assumptions with respect to these values; when it reports less than 10 g/kWh, overly optimistic assumptions are involved.

    In many cases the pessimistic assumptions are based on the application of old technology, be it diffusion enrichment technology, prototype plant life times, or a coal-heavy electrical grid. These assumptions may be reasonable if one wants to assess the historic emissions of nuclear, but they’re not appropriate as the basis for deciding what we will do in the future. In this case, more recent data is applicable, which generally trends toward the optimistic results.

    The exception to this is uranium ore quality, which will tend to degrade into the future. However, as ores become of increasingly poor quality, recycling used fuel becomes more attractive, as does thorium. I think it is extremely doubtful we’ll ever produce uranium from ultra-low grade ores, although Bilsko’s study assumes as much. If we put a price on carbon, it becomes still less likely.

    • On the cost of nuclear –
      Nuclear power plants are expensive in one respect: they have an enormous up-front sticker price. Something like $5-10 billion is about right for a 1,500 MW reactor, which is typical of the new large reactors on the drawing board in the US. (Current reactors average about 1,000 MW.) This is really expensive compared to coal and natural gas, but relatively cheap compared to the price per kW generated of solar and wind (removing subsidies). However, about 90% of the utilities in the US are so small that a single reactor is worth more than their entire asset base, which makes building one equivalent to betting the farm.

    Nonetheless, once they’re built they become the cash-cows of the utility fleet. For seven years running US nuclear plants have been the cheapest electricity source in the US at about 1.7 cents per kWh. This is why utilities are so eager to build them. Still, even if wind (or small nuclear) is a “poorer” investment in terms of the ultimate return, at least utilities can afford to make it. That’s why these loan guarantees are so significant. The Administration has said that it’s intent is to help the utilities build about 5-10 of these plants to give private investors the confidence to make loans available for them. If the first few plants go poorly and overbudget, this “nuclear renaissance” will be short-lived.

    • On considering added transmission costs –
      As another poster has pointed out, additional transmission will be required irrespective of the technology adding electricity to the grid. This problem is actually somewhat worse for wind and solar because they tend to be located in remote areas, well away from population centers.

    Transmission requirements for renewables are further exacerbated by the fact that renewables are intermittent. As a result, more capacity has to be built into the grid and backup generation (generally natural gas) has to be incorporated.

    On the other hand, improvements in energy efficiency actually reduce the grid requirements.

    • I’ll just repeat my comment on avoiding T&D costs above…if one adopts that tried and true utility mindset that the only way to provide power to consumers is by producing it at big central power plants then jamming as many electrons through the T&D system as possible, then sure, T&D upgrades will be required.
      If we can shed that outmoded way of thinking about electricity generation & delivery and recognize the benefits of distributed generation, then no, T&D won’t require such large capital outlays.

      Just as the IT paradigm shifted from the simple server-client setup of the 70s and 80s to the peer-to-peer dynamic of the 90s and this decade, electricity will see sideways flows on the grid.

    • And as for the life-cycle emissions analysis, I’m sure the piece that the authors over at the LightBucket blog put together is a well-reasoned summary of the 10 or so studies, but I’m going to stick with the academic work that actually made it into the Energy Policy journal.
      As for the timing of the data, it looks to me like the studies cited in the Beerten/Laes/Meskens/D’haeseleer study are considerably more recent than those on the LightBucket blog

  3. My Father founded American Nuclear Corp. (then Gas Hills Uranium Corp.) in the mid 1950’s. I remember him talking about the French plan to produce all of their electricity from Nuclear (80% now). I also remember Three-Mile Island, which of course never came close to being the environmental disaster claimed, but ended similar ambitions and bankrupt our domestic nuclear energy industry. Had we the backbone to have done as the French (who’d have thunk the French had more courage than us) – not only would we be more competitive globally, but would have delayed the tipping point of climate change.

  4. Nuclear Power Should Be a Part of U.S. Power Generating Portfolio

    The answer to our country’s (and our world’s) energy needs is not nuclear power. It’s not solar power. It’s not wind energy. Or natural gas. Or clean coal. It’s all of the above. The most efficient, most effective way to meet our energy needs, now and in the future, is to do it with a portfolio of generation methods.

  5. If you include their massive manufacturing, maintenance, and land use requirements, Wind and Solar energy are huge carbon emitters too.

    If you look at all costs, Wind and Solar are MORE expensive than nuclear by far, take MASSIVE amounts of land, have MASSIVE short-term reliability problems, and are painfully slow to build up.

    If you include the problem of storing or re-routing energy when the wind fails (constantly) or the sun going down (every night) then the carbon emissions and pollution problems of Wind and Solar are even WORSE.

    The only reason Wind and Solar operate successfully in Europe is that they have the huge French nuclear power system providing “back up” energy to all the other nations.

    The question is, when are the anti-nuke, green-energy advocates going to get real and start talking about these facts?

    Apparently never, based on past history.

  6. So if not nuclear, then what? Wind and solar? The new plants in GA will be about 1200 Megawatts each. Except when the units are in refueling outages every 2 years for a couple months, that’s is reliable power on the grid 24×7, day and night. Solar is only on the grid when the sun is shining if someone keeps them clean. Wind is only on the grid when the wind is blowing.

    How many wind turbines will have to be built to in place of the 2 units in GA? Or, how many “square miles” will have to be covered with solar panels to produce the same capacity.

    As far as a new distribution system, that is a general problem with the entire infa-structure in the US. If the study says it will peg transmissions costs are 2200/kw, it isn’t any cheaper if the power source is fossil fuel, wind or solar. They don’t built a separate grid for nuclear electrons, electrons they all come from a generator.

    • T&D costs are a general problem if utilities insist on putting more and more centralized power on the grid… large scale deployment of distributed generation is one way to avoid such massive investments in T&D upgrades.

  7. Katie,
    Nothing like a list of numbers to put things in perspective! Well done.
    Two observations:

    The zero emissions figure, while technically correct if only considering direct emissions, is a bit misleading. Obtaining fuel for nuclear plants is not an emissions-free endeavor so its important to remember that there are emissions associated with nuclear power. Life-cycle emissions analysis of nuclear power vary, but this recent paper ( ) from researchers in Belgium weighs in with a fairly balanced assessment – CO2 emissions associated with nuclear power are probably around 110g/kWh…or abour 0.12tons/MWh. While this is considerably less than natural gas combustions (around 0.5-0.6 tons/MWh), its not negligible either.

    The second thing worth noting is that you’re missing a huge cost component of new nuclear generation: transmission and distribution upgrades. This report from the Brattle Group ( ) from last year pegs transmission costs at $2200/kW and distribution upgrades at $4400/kW. At $6600/kW thats more than the /kW cost for the generation construction!! And while I understand that you’re focusing on the nuclear plant costs in particular, you simply can’t disassociate the two costs…you can’t build the plant and not connect it to end users!

  8. Nuclear power in its present form is not clean. It’s dangerous, water intensive and expensive. Perhaps you can argue it has a reasonably light carbon footprint but that shouldn’t be the only consideration.