Solar, Wind and Nuclear Energy Prices

Solar is getting competitive on power generation costs but on average is more costly than nuclear power.

NV Energy’s had solar power pricing of $24.99/MWh in Arizona. This was 2.5 cents per kwh. A dual nuclear power plant in the US has 2.4 cents per kwh in pricing. The average pricing for new solar in the US is about 3.6 cents per kwh.

The O&M costs of solar have been flat for four years at 0.8 cents per kwh. The land for solar is about 1.6 cents per kwh. Solar projects get a 30% tax write-off and has other subsidies.

Panels are less than half the price of the installation.

How low can solar prices go? GTM predicts under $15/MWh in 2022 (1.5 cents per KWh).

U.S. Solar Market added 10.6 GW of PV in 2018.

Xcel Energy received a median bid price of $21/MWh for wind-plus-storage projects and $36/MWh for solar-plus-storage projects.

The most recent comprehensive US solar costs I can find is a 2018 Berkeley Lab report of 2017 prices and projects.

For a typical solar installation, the general rule of thumb is that for every 1kW of solar panels needed, the area required is approximately 100 square feet. A 1mW solar PV power plant, the area required is about 2.5 acres or 100,000 square feet. This amount of land is just for the panels themselves and more space is needed for other solar equipment, which can bring the total closer to about 4 acres for a 1mW farm. Land for solar is leasing out at about $4000 per acre per year. Each Megawatt needs $16000 per year in land. This is $16 million per year for land leasing on a gigawatt. A gigawatt of solar generates about 1 TWh. This would be 1.6 cents per KWh for land charges.

Installed Prices: Median installed PV project prices in 2017 were $2.0/WAC (or $1.6/WDC) for projects completed in 2017.

The median price is $2000 per KW (AC) install/. $1600/KW (DC).

The lowest 20th percentile of projects within our 2017 sample (of 76 PV projects totaling 2,303 MWAC) were priced at or below $1.8/WAC, with the lowest-priced projects around $0.9/WAC.

Operation and Maintenance (“O&M”) Costs: What limited empirical O&M cost data are publicly available suggest that PV O&M costs were in the neighborhood of $16/kWAC-year, or $8/MWh, in 2017. These were costs incurred to directly operate and maintain the generating plant, and should not be confused with total operating expenses, which would also include property taxes, insurance, land royalties, performance bonds, various administrative and other fees, and overhead.

Capacity Factors: The cumulative net AC capacity factors of individual projects in a sample of 392 PV projects totaling 16,052 MWAC range widely, from 14.3% to 35.2%, with a sample median of 26.3% and a capacity-weighted average of 27.6%. This project-level variation is based on a number of factors, including the strength of the solar resource at the project site, whether the array is mounted at a fixed tilt or on a tracking mechanism, the inverter loading ratio, degradation, and curtailment.

Nuclear Costs in the USA

US Nuclear capacity factor has been about 90% for nearly 20 years. All nuclear reactors. The US nuclear industry has reduced shutdowns for refueling to 25 days on average. The US Nuclear capacity factor has been over 91% for 5 years running.

The US Nuclear industry reports generating costs without capital of 2.5 cents per kwh for dual plants or 3.4 cents per kwh for single unit plants. Vogtle and others are dual plants. 4.3cents per kwh looks like it includes capital costs for single plants.

60 thoughts on “Solar, Wind and Nuclear Energy Prices”

  1. The future will always be fusion. If you had a vat of hydrogen and took it up to the same temperature and pressure as the centre of the sun, the average half-life before a hydrogen nucleus fuses with another is about four and a half billion years – about the same age as the sun is. That’s a good thing, or the sun would have burnt out long before we evolved. But it means your super hot, super dense hydrogen vat would have about the same energy output per cubic centimetre as a well aerated compost heap. So it would have to be huge, hence unaffordable. To raise the reaction rate to plausible levels, you have to go to about five times hotter than the sun’s core, and to deuterium/tritium reactions. That can be done – by letting off a uranium fission bomb next to your hydrogens – but for some strange reason the fusion enthusiasts don’t talk about that. So expect the Iter programme, and the other hopeful startups, to keep producing more PhDs, and hot air, than power.

  2. The “Powerwall” battery you mention was $350/kWh. But the research-level quinone flow battery has a projected bulk component cost of $50/kWh, and ESS Inc claims the bulk cost of *their* electrolyte can go below $20/kWh. So the potential is there for costs to drop a lot. The nuisance at the moment is that lithium tech is so high-volume, it makes it hard for new grid-specific tech to get launched.

  3. Untill we have effective battery technology for mass storage of electricity for long periods of time nuclear will stay more competetive. Once we get batteries though-I would expect all kinds of renewables, especially solar, to become disruptive technology-especially if paired with blockchain and localized energy network so that micro transactions can happen on nationwide scale.

  4. I have heard of people with woodburning heaters deliberately subscribing to lots of junk mail. Don’t know if it worked out worth the hassle.

  5. 1) I never claimed ITER is anything but an amazing waste of money. I was talking about companies like General Fusion and LPP.
    2) it all depends on how much extra capacity is needed to make this work.
    3) When the sun isn’t shining the wind is most often blowing.
    4) Pumped hydro is great where the geography is right and there is water to spare. I’m thinking flow batteries or compressed air will probably be more applicable in general. Note, heated rocks is just an old idea I tossed out there to show there are things to do with extra, super cheap because it’s temporarily being over produced, electricity.
    5) A lot of the modern parking lots are heavily wired as is because they have overhead solar panels. By the way, there is plenty of wind power at night.
    6) Fission works but it’s as expensive as all get out to build new plants… just like all other large infrastructure projects in this country.

  6. I really like the arcs!
    I assume the H demand will grow. Seems to have had false starts but now have low platinum fuel cells and lots of stations in Calif, and esp trucks in Long Beach port. If even part of the energy goes thru H, there will be use for “excess” wind power, but the big problem is no wind, in which case the power would have to come from stored H, instead of batteries or whatnot. Need either very cheap elect or some sort of artificial photosynth to really make it work, as that first step, splitting the water, is hard and inefficient. Now made from methane, in fact!
    But, after the H is made, it is ideal.

  7. Author is making statements based upon nuclear industry lobbyists data! There isn’t a nuclear plant in the US that achieves anything near 90% capacity factor. Nuclear plants are dispatched as must run units and despite costing more that ISO wholesale cost they can’t be shut down. Nuclear plants capacity decreases as the fuel is consumed and typically drops down more than 20% from nameplate. The data showing unbelievable 90% capacity factors are a reflection of must run condition and reducing capacity. No other generation type fossil or renewables make the illogical assumptions to compute Capacity Factor with anything other than nameplate capacity!

  8. Sparce array idea is that you “should” be able to beam power from separate sats, much as you can see great detail from separate sensors, like the radio or optical ‘scopes familiarly do. Trouble is, the focus spot only carries a small amount of the total expended power, so they DO NOT work!
    Another thing to consider is that we are way behind making the signal as opposed to seeing interference. We can control the phase of radar frequencies, but not visible, (YET!), so we can make phased array radar, but not with visible. Still wondering about your “dish” comment.
    Flat phased array radar can send beams any forward direction, as many as needed, if you can do the math or have a pilot signal to follow back.
    As to the referenced info in your other comment, looks like quantum stuff, so who can really say!
    When they talk of climate, or really almost any Earth problem, they should be talking Space. Or quit acting like they care!

  9. My comment is highlighted under “Eugene Dsky” nickname. Have never heard of sparse arrays. Interesting. Yeah, I agree that future in energy is lunar (space) solar stations. It is an enormous source! Unfortunately, they won’t talk about it every week even on Fox News.

  10. I’ve thought that using excess wind power to create H2 on site to be used in turbines is an interesting idea. Nobody is doing this so I assume it is nowhere near cost effective.

    Nuscale wants to use electricity that isn’t needed for load balancing to desalinate water, for process heat, etc. I prefer giant electric arcs because that would look kinda cool.

    1. ITER/DEMO will never produce cost effective power.
    2. Overbuilding does not lead to effective power.
    3. Overbuilding an unreliable power source just means that sometimes you have far too much power but you still won’t have enough when the sun doesn’t shine.
    4. You can store energy quite efficiently with pumped hydro. 1/10th lithium ion costs, GWhr scalability, easy to add more storage. Hot rocks transfer heat poorly, look at underground compressed air storage if you don’t like pumped hydro.
    5. Solar doesn’t fit mass EV adoption well. People will charge at night and need 20-30kwhr to do so. Deploying chargers for daytime charging is costly. Wiring parking lots = $ adds up quickly.
    6. Fission is a working technology.
  11. People will stay warm by asking for paper copies of their utility bill and tax bills and they can burn them in their wood fired stoves.

  12. Video about interference(could not find your comment yet), but going the “other way” to send the beam without power loss is much more difficult than seeing the direction accurately, as in telescope pairs, for example. The whole limiting diffraction area has to be covered, not just a few points around the edge. See the “sparce array” dream for more.
    So, dish would not work as being too big, and not needed for phased array radar. Or do I not understand the dish question? Criswell uses Moon tidal lock to get screens to “look” like one big area by having all the edges backed up by other screens behind, on slopes near the limb.
    Again, Earth fusion will never be as cheap as existing solar fusion(free), for the source of energy, and then has distribution limits that Space power beaming can solve economically.
    And it does not get us into Space!
    BTW, H economy automatically solves many of the problems w/ wind, solar and Space Solar that nuclear points out.

  13. Irrational fear won’t go, I guess. Modern industry is too complex for a laymen.

    Yeah, phased arrays are a miracle! I wrote a comment on a video describing the phenomena on which they were based:

    Though it is interesting why Dave Criswell didn’t offer dish antennae instead.

    Regarding tokamaks, there is a British non-government company which claim that with more powerful magnets (obtained by increase in size and use of modern superconductors) they are _close_ to it being commercially available!

  14. Of course, with redirecting sats, solar and wind power can be redistributed with power beams to make such an absurdity work! An added benefit to LSP!

  15. Very interesting, but “death ray” is a main objection to Space Solar in general, of which Lunar Solar is a version. In fact, the Criswell beam design is 20% sunlight intensity, but can still work as low as 2% if issues (or irrational fears) arise. Whatever Solar Power Sats use, so will LSP, I suspect. Particularly the rectennae. Keep using that cell phone, you are proving the safety of Power Beaming! An old saw is that there is good news and bad news about power beaming: bad news is that long wavelength requires big, low intensity beams. Good news is that long wavelengths prevent weaponization.
    Also, there is a huge benefit to being able to use phased array radar tech to hit thousands of receivers at once(simply follow pilot signals from the receiver positions), eliminating long ground transmission needs. (I think I remember Mankins objecting to LSP based upon his mistaken assumption that there would only be one beam).
    I want the cost estimates for (non-solar!) Earth based fusion, 20-200 TWe, delivered to small loads worldwide. LSP’s 1 cent per KWh-e is a profit making charge, not the cost, by the way.
    A very important topic, esp for Bezos/O’Neill future!

  16. What’s the issue, you can’t access google? Try “REPDO”. nuclear weapons? lol. Sure thing, like U-235 in a power reactor is a suitable path to Pu-239. Besides, it is far cheaper to buy weapons from places like Pakistan and Russia than to try and build a breeder (and convince the U.S. that it’s for “research purposes”).

  17. The space lost is fairly trivial, while the power generated is double and fairly steady at least in the deserts. The component to move them is very mature, as the technology is borrowed from C-band satellite dishes that used to be popular 25-35 years ago. It just needs occasional grease.

    One thing we need is efficient TV sets to reduce electric demand from 7PM to midnight to get the demand curve to more closely match available sunlight.

    Laser projection can be fairly efficient. Maybe set a limit of 200W on new TVs and 150W on new home computers would make a difference.

  18. That means you need more area than for fixed panels, or they’ll shade each other, installation and maintenance costs will go up, and the ramp rates for whatever has to take over in the evening will be much steeper. It also lowers the value of power from Robbie Rooftop, who’s stuck a few panels on his house -why buy off him when you can get a couple of hours more per day from the utility ? Cloudy weather will knock them both out about equally, though.

  19. “I would not go as far as to say it subsidizes nuclear (or other full time power generation), it just suplements it”

    I think you misunderstood What I saying. I saying that nuclear in fact subsidizes renewables.

    If there was no nuclear or hydro the solar power would cost 5-6 times more.

  20. If you look around you will find that there ARE people calling for 100% wind and solar.

    To be achieved by forking out heaps of taxpayer dollars.

    This includes some politicians who release ill thought out position papers (I am thinking of some local examples but I think there are some in the USA as well.)

    This is what people are reacting against. The prospect of freezing in the dark during a cloudy winter huddled over their drastically increased tax bill.

    So, if someone is all emotional and doesn’t bother reading what they are responding to, they will go off about that prospect, and not notice that you are presenting a much different picture.

  21. I was thinking that just about every rural landowner I know would give their left leg to be able to lease out their property for $4000 per acre per year

    Maybe that is the price that someone is asking if they happen to own land right next to an existing solar plant that wants to expand, and there is nothing else available in the area.

  22. It’s not like you can’t use the extra electrify for anything. That’s what storage is all about also. In addition to the standard ways of storing electricity there is a type of heating that involves storing heat in an insulated chamber full of rocks. Once are hot they stay that way until you need to draw on them for heating. Say, if the over capacity needed to make solar works is say 50% (actually I believe is a lot smaller) and solar is 1/2 the cost of other methods of power generation then solar would still be 25% cheaper.

    Note, I like the idea of fusion too, but at the moment solar and wind are working technologies. We are probably at minimum 5 to 7 years from any of the alternative fusion technologies being at the point where they can demonstrate they are truly viable (assuming they can do so,) At which point they’ll build a full scale demonstration plant which will probably take another decade. All this means that we are probably 25 years from commercial generation on any scale.

  23. The project was started in 2009, and they the example of Units 1 and 2 which were built for $9 billion. Solar in 2009 cost three times the then cost of nuclear, so they made a rational choice at the time. I’m sure they didn’t expect solar to drop by a factor of 9 and nuclear to go up by 50% in the decade since, nor Westinghouse, which had experience building nuclear plants, to be so messed up they went bankrupt.

    The two reactors are due to be finished in two and three years respectively. They are so far along at this point, and they need their output so badly, that they are finishing despite the cost. But this project also is nuclear’s tombstone in the US, until someone can *prove* they can build it for radically lower cost. No one in their right mind will accept a cost projection any more.

  24. If they had known Vogtle 3&4 was going to cost 27.3 billion before starting no way would they have built nuclear. You could have set up the whole state with air source heat pumps for that much, with a massively better result.

  25. Once again, pay attention to basic economics folks. How smart is it to just overbuild? This is why Gary isn’t running a company. Solar and wind will have a place, just not the base. And, it ain’t happening for a long time. Also, its not the future if its happening. Its just now. The future is fusion!

  26. Biofuels are massive producers of GHG’s over the next few critical decades – not a solution.

    Since it is a no brainer to build a Allam cycle methane plant why haven’t we seen any? No shortage of methane.

    As a noted illiterate the last 7 Candu’s were built for $2.7B/GW $2019 similar to Chinese and Korean efforts – as I’ve pointed out you many times. I’m sure the US could learn how if it applied itself.

    Since we don’t know what competitive factors went into PPA’s we can along go by actual costs. Compared to the 60 year life of nukes at $4B/GW for Nuscale we have current builds over the same life of wind/solar at $8B/GW.

    Xcel storage numbers still come to 2 bucks a kwh to cover the 16 weeks of storage real science tells up is necessary but only covers 20 years of battery life. Real numbers comes to $5 a kWh added to your power bill to compare to a nuke .

    Enough green storage to levelize over a year, would add over $800B/GW with the cheapest storage envisioned to the $8/GW for wind/solar – clearly not possible.

  27. I would not go as far as to say it subsidizes nuclear (or other full time power generation), it just suplements it. It is a poor choice to fully power the grid. Anyone who has done the math can see that. But that is not to say it has no role. It is great for those hot summer days when we need more AC power. Unless, we can undercut it with nuclear anyway. We probably can if we design well, regulations do a rational rethink, and funds are made more available. Perhaps if you invest in nuclear, those investments can be made tax free? That might make some fat cats rethink their portfolio.

  28. The “capacity factor” of solar in the US averages 25%. This is due to a combination of weather, night, and atmospheric absorption. When the Sun is lower in the sky, much of the energy is absorbed or refracted away before it reaches ground level.

    Capacity factor is (actual average annual output)/(rated capacity), expressed as a percentage. Nuclear is more like 92%, because they tend to run at full power, except when they shut down for maintenance and refueling.

    In the sunny southwest, solar is working at exactly the time it is needed most – on hot sunny days. So its OK it doesn’t run all the time, as long as it runs when it is needed most.

  29. That’s exactly what Georgia Power is doing. They already had nuclear from the Vogtle Units 1 & 2 reactors, but the Atlanta metro is rapidly growing. So much so, they need 150 MW of additional power every year. Thus the Vogtle units 3 & 4 are the only new reactors being built in the US, and on top of that they are building a lot of solar.

    Wind isn’t very effective in the southeast. All the trees and hills cause friction and slow down the wind speeds. That’s why the flat midwest and offshore are better places in terms of wind resource. We have hydro on the Tennessee River (TVA) and a few of the smaller rivers, and biomass from all the lumber production waste.

  30. Brian,

    Solar O&M costs are $8,000 per megawatt-year. $8 per kilowatt estimated for 2020. 800 cents. A 25% capacity utility solar plant will generate 2191 kWh per year. 800 divided by 2129 is 0.38 cents per kWh for solar O&M, less than half your number of 0.8.
    And of course your land cost of 1.6 cents per kWh was off by a factor of 53.3 and should be 0.03 cents per kWh (a rounding error). You need to update your numbers and hopefully also your extreme cognitive biases.

  31. Iowa has the most expensive farm acreage in the USA is why I mentioned it re: Brian’s gross misinformation on land costs (off by a factor of 50).

    Saying that Iowans would freeze in winter is a 100% wind and solar absurd strawman that you introduced, not anybody else.

  32. It’s cheaper to burn ethanol (or even better, butanol or gasified biomass) in a gas turbine than build new nuclear in the USA.
    And you don’t even need to go that extreme, because an Allam cycle methane plant can carbon capture for less than half the price per kWh of new nuclear. And has a cheap capital cost so can be run at 40% capacity factor.You can even run on gasified biomass or landfill gas if you want to go carbon negative.
    On top of that 4 hour storage was 0.29 cents per kWh in that Xcel competitive bids for wind for 2019. You could bid in 80 hour storage and still crush nuclear. The bids for 2023 start date for storage were less by 32.5%.

  33. You are confusing local power with the rest of the grid. Your neighbor still needs power, so does the car factory, and your local 7/11. Even mighty sunny Saudi can’t make the equation work and their solar PPA’s are issued at 1.2 CENTS. So meager are the revenues that only heavily subsidized state-owned utilities build it, and now they’ve completely stopped. Why? Cause they realize that it gets dark at night, and storage costs are exorbitant. Wind is nutty, as well.

  34. My understanding that: if Sun stays low on the horizon ( morning, evening ) solar panels not working at their full capacity, also there are clouds, rain etc. Perhaps 6 hours a day is too conservative estimations. Let say 8 hours at full power.

    However the moral is: Base load power like nuclear actually subsidizes renewables. If there is no base load power, cost for solar would go up at least 5 – 6 times. We would need 3 times more at least solar panels per kilowatt, plus expensive $10500 battery to support 1kW around the clock power.

    I do not want even go to the maintenance and power distribution…

  35. Large nuclear is much cheaper:
    (operating prices are given in the link 1.33 cents in 2002 but sold for 2.5 cents. Selling for 6.33 cents in 2007. No comment on cost…which is probably lower as they have probably finished paying into the fund for decommissioning). And it could be made much cheaper than that. There is no good reason to require 2000+ full time workers. All political. You don’t get these things built without promising lots of good paying jobs.
    I would build at least 3 times this size, and automate just about all of it. Maybe 100-200 workers max. Cost should be below 1 cent/ KWh.

  36. You would not build solar plants where you only had 6 hours of sunlight…unless you had no choice. It is more like 10+

  37. “For a typical solar installation, the general rule of thumb is that for every 1kW of solar panels needed, the area required is approximately 100 square feet. A 1mW solar PV power plant, the area required is about 2.5 acres or 100,000 square feet.”…

    Well if there is no nuclear, hydro or coal we need some corrections. Assuming that reliable harvesting of electricity last let say 6 hours a day. We need at least 5-6 kW ( 500- 600 square feet per KW ) )  of installed solar panels to provide base load 1KW power 24 hours a day. We need harvest and store energy for the times that Sun does not shine – 18 of 24 hours a day  .

    We need safety margin in energy storage if there is heavy clouding , rain etc. So we need about 30 kW/hr storage capacity for 1 kw   uninterrupted base load. ( we can not completely run battery dry   ) Tesla for example advertised their 10 kW/hr battery for 3500. So it would be 10500 only for storage. However it last only 500 cycles so it would be about 1 year 4 months of everyday use.

    And it just small part of costs for using solar as base load…

  38. One easy solution is to overbuild. With decreasing costs it is going to be viable to do so. Note, you’d still a grid update, maybe a modest amount of storage and gas turbine peak generators for the worst case scenarios. By the way, we’ve got those peak plants now and we desperately a grid update regardless what the future holds. The solar, wind and storage will keep happening on their own and there is little anyone can do about it… This is not a what if, it’s a what will. Welcome to the future.

  39. There are two things you seem to be ignoring. The capital costs for solar and wind power are decreasing in cost each year and for nuclear power the costs are rising. Also it costs billions to build a new nuclear plant but you can hire an installer to climb up on your own roof and add solar power capacity for a trivial cost.

  40. Not dispatchable, not going to charge the EVs at night, not going to heat homes in wintertime.

    Good for summer Daytime AC usage which is handy.

    A part of the grid that cannot become the whole of the grid.

  41. Iowa’s insolation is only slightly better than… Minnesota. Sure land is cheap but you won’t generate power near wintertime, aka freeze to death time.

    Maybe you’ll propose adding lots of wind turbines for winter time. Congrats you almost doubled your capital costs to deliver the needed GWhrs.

    Even still neither are dispatchable. You’ll need something to follow whatever load is available. This is where UAMPS is at and why they want the FOAK Nuscale nuclear plant. To smooth their wind and solar power output.

  42. Nuclear in USA is 18 cents per kWh.
    Austin Energy was looking to sign a power purchase agreement. Article is from March 2014.
    ”The 5-cent price falls below Austin Energy’s estimates for natural gas at 7 cents, coal at 10 cents and nuclear at 13 cents.”
    That is before the 50% cost overruns on the AP1000 reactors subsequently started.So 3 cents operating 10 cents capital cost. increase capital cost by 50%…18 cents per kWh for a new pair of AP1000s.

  43. Read the actual link page 4 that describes what they mean by capital costs

    How many bullshit articles is this now? Six?
    EDIT: and what’s with the insane made up land leasing costs? Land literally costs $12,000 per acre in the USA to buy and yet would cost $4000 per year to lease?! EDIT2: $75 per acre per year. Prime farmland in Iowa is $300 per acre per year.

  44. It doesn’t matter what it costs if you cannot turn it on when it is needed or use it for baseload. If it is not dispatchable or baseload it is garbage energy and should be trashed.


Leave a Comment