Solar electricity: Why it's expensive now & where it's heading


Solar photovoltaic is currently the most expensive form of electricity available ... which is weird, right? I mean, the fuel is free and the intermediate technology is made out of, essentially, sand. So what's the deal?

In this article for National Geographic News, I probe the details to try and understand what makes solar power so spendy most places—and why it's actually already a reasonably-priced option in others.

So why is solar so expensive?

Converting light into electricity with no moving parts is a profoundly different enterprise than turning a turbine to make power--the technology that is at work in coal, natural gas, nuclear, hydropower plants and, most visibly to the public, at wind farms.

"Wind power is the same technology as it's been for 1,000 years," said Tom Meyer, a professor of chemistry at the University of North Carolina at Chapel Hill. "There's nothing to invent. It just needs to be improved." The makers of wind turbines have made huge cost reductions in recent decades with relatively small tweaks to an otherwise familiar system.

That's not yet true for solar, experts say. Most solar cells are made from silicon--the same semiconductor material that is at the heart of computers. The cells are expensive to produce because it takes a great deal of energy to purify the silicon. And, while the computer industry has made enormous strides in making cheaper silicon devices, those advancements don't translate to the solar industry.

National Geographic News: Shining a Light on the Cost of Solar Energy

Image: Some rights reserved by Eleaf


  1. Not true. A silver-zinc watch battery is the most expensive form of electricity. Not sure what comes after that, but silver is way out ahead of the others.

  2. Your article missed the point. The point is that solar has gotten cheaper at approximately the predicted rate for 20 yr. Solar grid parity is coming in 2- 8 yr depending on where you live. Once it meets grid parity (roughly the cost of our coal-gas-nuke system) it will be rapidly deployed. For the US, widespread grid parity will correspond with an INCREASE in foreign energy reliance. Mainly, our peaking domestic natural gas will be replaced with Chinese PV.

    Thanks to our climate denying / fossil fuel subsidizing/ fat and simpleminded America, this rise of solar will really and clearly signal the death of American economic leadership. Why? Well approximately ~25% of the global economy is for fossil fuel. Presently, this GDP is dominated by the US and other western countries. China will have the infrastructure to replace it and they will gladly sell it to everyone. China will develop the rest of the world. (FYI, in 1 – 2 yr they will be producing PV at a faster rate than we can build base load plants of any kind, 2 – 3 yr, the same for peaking gas plants) The US (already chatter in congress) will probably combat this with tarrifs (can’t have chinese PV cheaper than US natural gas), thus exacerbating the problem by yielding the economic advantage to everyone abroad while depleting our dwindling resources and destroying our own environment.

  3. Why is nobody making phones with solar panel backs? So when it starts running out of juice, you flip your phone over and leave it on your desk in the sun.

    1. Because in the current state of the art, the phone would consume battery at a rate faster than the solar panel could recharge it.

    2. PV panels on the back of the phone is still a good idea – even if leaving it in the sun for 30 mins lets me make a 5 min phone call or send 20 texts, that’s a win compared to “crap, my charger’s back at home, guess I’ll have to wait till the end of my work day then”.

  4. What Anon said. Same is happening in some European countries and it is horrible. Energy independence and reduction in consumption is not something that world leaders are hot for. Establishment

  5. I’m not a huge fan of photovoltaics (PV). Primarily because all of the solutions seem to require rare-earth minerals, which are scarce, and subject to political pressure (see China’s export embargoes) and human-rights issues (see conflict mining in Africa).

    I’m a bigger fan of solar-thermal technologies. These do not require exotic materials, mainly mirrors and structural materials (metal, plastic, wood).

    See this story, which describes an interesting government/private research and development program that looks to use America’s idle automotive manufacturing capacity to produce solar-thermal generators.

    Note, there is a scale and decentralization issue as between PV solar and solar-thermal. PV is probably more appropriate for decentralized, or small remote deployments. Solar-thermal is more suited for medium/large scale generation, and then transmission.

    1. YUP YUP

      There is definitely thin film technology that isn’t based on rare-earths, and silicon (obviously) isn’t either.

      That said, you’re on the right track with the rest of this. There wasn’t enough room in this one story to get into solar thermal and decentralization, but there’s a lot of cool stuff there.

    2. Solar cells only use very small amounts of rare metals.

      27.7% of the Earth’s crust is made out of silicon, the base material for solar cells. Silicon is quite easy to purify, the 7 largest factories produced about 114.500.000 kg during 2009, and even larger factories are under construction (in China). Modern solar cells are also made form very thin slices of silicon, so less base material is needed.

      Only very small amounts of (rare) elements are needed to turn the silicon into a semiconductor, in the order of 0.001%.
      More metals are used for the contacts on the solar cells, but these are getting thinner and narrower, and there are many alternative materials for the contacts.

      I think a shortage of rare elements will have no effects on the production limits of solar panels, and even the price of rare elements will have a neglect-able effect on the overall price of solar pannels.

      The biggest usage of rare elements are batteries and magnets, not the semiconductor industry, and even these elements are not rare because we used them all, they are rare because most mines have closed down, because they could not compete with the cheap Chinese mines, and now it takes a while to start these mines back up.

      Because of high consumer electricity taxes in the Netherlands, solar energy is already cheaper than grid electricity if you install the panels yourself, even if they are not subsidized (like in Germany).
      So, Solar grid parity is coming in 0 – 8 yr depending on where you live…

  6. “The most expensive form of electricity” — them’s fightin’ words.

    For starters, you could burn mummies. You could pay homage to original “Mark Twain said that Egyptians trains are powered by burning mummies” story, with the train fixed in place, driving a generator. Now that would be some expensive electricity.

    We could make it more expensive by stipulating inefficient generators, but my in my fantasies I demand the best of everything — the best generators, the best resinous mummies, everything.

  7. Good article. My only complaint would echo Anon, above, that the question of external environmental and human costs of conventional power generation is just going to get bigger and more pressing, and could have had a more prominent place in the piece. It’s not just the global warming, although that would be bad enough — you’ve got everything from childhood asthma to mountaintop removal to account for if you want to be clear about the costs of energy generation in our economy. Plus all those dead folx over in Iraq, obviously.

  8. Facts like this are the reason why libertarianism and private property/the free market are not necessarily anti-environmental, and also why I am not so worried about things such as peak oil.

    To put it simply, the federal government is a tremendous inhibitor of environemntal progress. There are many reasons for this – subsidizing corporate polluters, allowing reckless logging and mining on federal lands, enforcing IP laws that allow new technologies to be buried, polluting itself (#1 global polluter is US military). But this oil thing..if we did not militarily subsidize the price of oil, it would be so much more expensive. Alternative energies would then be much more economically competitive, and we likely would have seen their widespread adoption years ago.

    Smothering something by altering the market against it and promoting it through government research at the same time is just stupid.

    Likewise with “peak oil” – there are thousands of finance professionals around the world, working in oil commodity futures, working every day to predict future supply and demand for oil. When oil actually does start to run low, then, there will be a several year anticipation through a spike in prices. Then, alternate technologies will become necessary, and through simply economics and entrepreneurship will receive vast R&D money. It will certainly create a shock, and cause some pain, but it won’t destroy civilization.

    1. i see a lot of anti-government from libertarians. But what about anti-corporations? Imo, they are just as bad. This because once a corporation gets a foothold, the way it operates basically demands that everything goes static and monopolistic. Or is it that, in the eyes of the libertarian, corporation and government are two sides of the same coin?

  9. Thom Hartmann is fond of mentioning that Germany runs a deal like Gainesville does, so that the home owner ends up owning the PV, while utilities buy your surplus electricity and your costs are underwritten as well. The figure he quotes is that these home units generate 20% of the energy needs of the nation.

  10. Congrats on the article, Maggie; extra bank for the heating bill is always nice. Too bad the pub date worked against you here; if this new silicon vapor-deposition method can be scaled up, PV prices are gonna drop like Cooks Source ad rates ;-) Good breakthroughs on CIGS thin-film this summer, too, but these “silicon wires” might just be the deal-sealer for solar….

  11. S2
    There’s a lot of “if” in that. Did you read the part about the long lag time between lab bench and market, and how not everything CAN be scaled up? I’d love to see a breakthrough. But it will take more than someone just saying they’ve found a one-off, works-in-the-lab breakthrough.

    That’s a pretty incorrect statement, then, unless I’m misunderstanding Hartmann’s argument. Germany’s own figures (quoted in the intro of my piece) show all solar as generating only 1.1% of their total electricity needs. Renewables, in general, were pushing 17%, if I recall correctly.

    1. Absolutely, iffage runs high. I’m just happy to see competition increasing within the field — between CIGS v. CIGD thin-films, Konarka and their magic inkjets, Atwater’s work, and more. Throw in advances in large scale solar thermal plants, and coal starts to lose some of its economic edge. I suspect the climate change debate will become moot for more than a few deep-pocketed deniers once it’s possible to make serious money from sunbeams. (Yeah, my glasses are rose-tinted; I’m a midwesterner!)

  12. It is hard to see PVs becoming a major player anytime soon. Not mentioned here is the other problem of energy storage. The lack of cost effective way to store power will hold PVs back whether you consider either the central or the distributed generation model.

    I know it is considered boring, without the appeal of new technology or any more than a hint of foreign intrigue, but the obvious first and easiest step is one the United States has consistently refused to take, energy conservation.

    How can anyone suggest we could commit to an expensive program of energy research, one that cannot possibly supply the instant gratification the American people seem to currently require when we haven’t been able for more than thirty years to come up with the will to enact the simple steps needed to reduce our energy consumption? Steps requiring no huge technological gains. Steps which actually save money in the short as well as the long term. Steps which are enacted with little or no opposition in virtually every other country in the world.

  13. You know maybe one of those over unity things will work out…. :-)

    But on a serious note, PV could have a nice place in the power grid. If it’s cheap enough and in my opinion recyclable enough.

    Granted current PV design really just extends from most silicon design work, so of course it’s going to be expensive. And most PV setups have an estimated 15-20 year life span. That’s not really cutting it in my book, but perhaps by then the efficiency will have improved enough to make replacement worthwile.

    Now if we end up with something really simple for PV, something relatively “green”, then recycling would be less of an issue. But if it ends up still needing rare metals and other limited supplies we really do need to find a why to make recycling cheap and convinent.

    And in terms of storing the power generated, hopefully carbon nanotube capacitors and or low temp super conductors make an appearance around the time all this cheap PV floods the market from China…

    The real key to getting any tech mass market uptake is to make it cheaper and better in the long run. Look at CFL’s, yeah they are more expensive up front…but (at least in theory) they will last 4x as long, plus reduce your power bill by enough to pay for themselves in that time period. Bonus on both sides.

    Frankly if the government cared we could have a green energy policy in place in less than 5 years. Hell we built “the bomb” in less time than that. In my opinion the government should offer an X prize type of contest for making mass produceable low cost PV. (Or insert your green energy source of choice.)

  14. Eleven years ago, I put solar panels on my roof, here in the Bay Area – received Oakland’s second building permit for photovoltaics. Did all the work myself. Over the past decade, I’ve added and rearranged the panels until my home is a net power generator. Very much a fun hobby.

    Rate of return? Oh, the 36 panels cost about $15,000 (I watched for the cheapest ones); the system should pay for itself in another 7 to 10 years.

    The expensive part isn’t the silicon in the solar panels. Naw … it’s the climbing around the roof, mounting waterproof risers through the shingles, hauling supplies up a slope. Worrying about getting killed when walking over a ridge – a loose shoelace could result in a hospital bill that dwarfs the cost of all the panels.

    And it’s not zero maintenance. Since 1999, I’m now on my fourth inverter (Xantrex) … three of these have burnt out. Changing from inverter #2 to inverter #3 meant rewiring the panels from 48 volt strings to 350 volts. Along the way, I converted from flex conduit to MC-3 to MC4 connectors. Added grounds that weren’t required ten years ago.

    And don’t get me started on raccoon poop, bird droppings, and general dust. Monthly cleaning of the panels results in an 8% increase in electric output (which is to say, not cleaning them monthly results in a gradual reduction of power output. And no, rain isn’t as effective as a hose and squeegie. Ever walk around your roof with a squeegie on an 8 foot pole?). Meanwhile the plum tree growing next to my house casts a shadow on the panels in the fall. Which do you want: Solar power or plum jam?

    I distrust the estimates of 25 year lifetimes for solar panels. I began with Photowatt PW1000 panels. After ten years, three of them have significantly dropped in power output – due to cracks in the silicon. What about Photowatt’s 25 year guarantee? Oh, the company’s gone from America. *sigh*

    Throughout this, I’ve tried to keep the solar panels unobtrusive – visible to the sun, but not to pedestrians on the street. This results in less-than optimum electricity generation.

    Yes, the technology has matured. Inverters are more reliable, solar panels are more robust, and plenty of companies are now experienced in roof installations. Many more residential PV installations are happening, and I cheer them on. For me, it’s been a fun hobby.

  15. When Barry Commoner ran for president in 1980 he had a fairly comprehensive plan for converting the country to solar energy. I don’t recall the exact specifics of the plan, but it did show a way to reduce the cost within a reasonable period of time. But we got Reagan instead, and it was decided that it was better to regulate pollution instead of eliminating it.

  16. Interestingly enough, The Atlantic has a story on the inventor of the Super Soaker, who has been tinkering with solar, and who now possibly holds the key to affordable Solar Power:

    Regarding its potential to revolutionize energy production on a global scale, “It has a darn good chance of being the best thing on Earth.” says Paul Werbos, an energy expert and program director of the National Science Foundation.

  17. If you look around you can sometimes find new solar laminates for under a dollar a watt, atm I can find them about 1.25 a watt. The most expensive parts of a solar electric system now is the infrastructure needed to make use of the silicon, not the silicon itself. Panel frames, mounts, wiring (often massive cables!), charge controllers, inverters and batteries- it all adds up to several times the cost of the laminate… often several times the cost of the framed panels.

    The cost of the silicon cells is only a small part of the cost of an average residential solar electrical system. Reducing that cost will no longer significantly reduce the cost of the system as a whole.

  18. What interests me about this thread is the general assumption among many commenters that silicon and rare-earths are the only game in town.

    I know they hold most of the market right now, but research groups around the world are quite close to producing usable organic solar cells made of thin layers of flexible organic plastic. The cells can be printed on standard presses, and are being developed to also install as window film and as paint.

    I understand there are others, but I’m mainly aware of this project at the Canadian National Research Council:
    (See also this update: .

  19. I work in the energy industry, and I have personally calculated lifetime values for acquisition analyses on dozens of power plants.

    I would argue that merchant generators, who are not on regulated tariffs and are therefore not essentially guaranteed cost recovery plus a reasonable return on investment, look at power plants a bit differently from those EIA statistics.

    For instance, expected capacity factors for a power generating facility in a truly competitively dispatched market are way off from their study, and depend very much on correlation on an hourly level, or even a fifteen-minute level, between fuel prices and power prices. Natural gas combined cycle plants (CCGTs) don’t even approach an 87% capacity factor if they are optimally dispatched, it’s generally closer to 45%-65%. Of course it depends on the market, but 87% means they are running flat out except for scheduled and occasional unscheduled maintenance outages. This would imply a very high-priced market that lacks adequate generation resources and such a market would usually attract new generation until those power prices fall more in line with a developed/mature wholesale power market.

    Merchant generators also tend to value assets on an all-in capital cost approach, which can probably be unpacked from those EIA statistics (if there were a little bit more information provided on their assumptions) as capital cost plus transmission costs, but as total dollars per kilowatt of generating capacity, not dollars per total kWh generated. CCGTs traded at a peak of $800-$1,000+ in $/kW a few years ago, but whole plants have sold for under $200/kW in good working order in the last 5 years. Solar PV may still be in the $2,000-$4,000/kW capacity it was trading at a few years ago, but it has certainly been trending down over a longer period of time, largely due to continued research and investment. That price isn’t as uncompetitive as it would seem by comparison, as there is no fuel cost for PV, which (and this is where the EIA table is useful) comprises 65% of the lifetime costs (probably a bit less due to lower Capacity Factor) of a CCGT.

    PV may only have an expected 22% capacity factor, but when the sun is shining the brightest, prices tend to be peaking in western developed nations due to air conditioning load, at least in summer/hot conditions. So the average power price received by PV solar (read: return on investment) is better than, say wind, which is more levelized and often delivers its best output at night when power is cheapest.

    I could go on for several pages, but I don’t think anyone would read it all. Ultimately, PV solar is approaching affordability (i.e., sufficient ROI to spur merchant investment) without subsidies, but it isn’t there yet.

  20. I read Buckminster Fuler’s book “Critical Path” and his idea of linking a worldwide network of ~1000 solar panel power plants to give 90% of the world’s population free power on a Dymaxian map grid sounds pretty good. I’m not entirely convinced moving the AC juice from the day side to the night side of the globe via the “intergenerator connection grid” would be that efficient but a simulatede run-through the data couldn’t hurt.

    But to further remark upon this, I was thinking to make it more realistic each power plant should have a family-like group to live there, study there, and create an extra resource like grow food, hemp, algae for diesel, reprocess plastics, manufacture parts, etc.

    And I like the idea of using solar thermal power as a more industrial and efficient source than photovoltaic farms.

    But what I’m curious about is that should each of these plants have a giant bank of overflow-catching rechargable battery banks to charge, hold, and share when not “leeching” from the other grid sources on the day side of the world at night.

    Real questions, people.

  21. Cadmium is a heavy metal and toxic, so I believe we must act to stop cadmium telluride based tech from entering our environment. We don’t want a bunch of cadmiun based panels over our heads (on houses and buildings) slowly degrading and leeching posion into our living space.A recyclable tech, possibly like Konarka’s power plastic seems to be a better alternative.

  22. That was one of the most disturbingly misinformative articles I have read on BoingBoing. Did you not talk to one person who actually understands PV pricing before putting together that collection of myths? Here are a few:

    1. As noted in the comments on the article itself, the LCOE table is completely out of date. Utility-scale PV system LCOEs are already below $0.17/kWh and are dropping quickly.

    2. PV costs are now below solar thermal costs on the large scale. It also uses much less water, which is a critical issue in desert states. Finally, it is easy to build PV plants in the 20MW range, which states like AZ and NV are realizing makes more sense as far as grid stability.

    3. As noted, we are only financing silicon PV systems on a 20 year timeframe; that is exactly what makes the cost look high. It is well accepted that the systems will last 30-40 years, with a degradation of 0.5% per year. That means a 40 year-old system will still be putting out ~80% of its first-year output. The galvanized steel an aluminum structures will last well over 40 years in desert environments. That means that after the 20 year finance period, these systems will be putting out power that is essentially free. Every system we build is like money in the bank. If you tried to finance nuclear plants on a 20 year horizon, it would cost more than any other form of electricity.

    5. You wildly exaggerate the uncertainty with regard to long-term production. PV system production is far more reliable than sources like wind. The LCOE is, in fact, much more stable than fossil fuel LCOEs, relying, as they do, on fuels that fluctuate greatly in price. Even with global warming, irradiance and temperature variation at a given location is far less extreme than fossil fuel pricing – and global warming is actually expected to improve the value proposition of Southwestern systems.

    I am really disturbed by the fact that you pretend to have expertise on this subject, when you so clearly don’t, and when articles like this one can do so much harm. And why do you never even reveal the cost drivers in anything more than a general sense? A budget of costs could do much more toward revealing the costs of solar systems far more than a bunch of fluffy generalizations – but it looks like your research didn’t take you that far.

    I hate to be so negative, but your article reminds me of the articles I was reading three years ago about how solar panels didn’t pay back the energy required to make them – a “fact” which hadn’t been true for over 30 years. It’s regurgitation, not research, and it’s does not represent good scientific thinking.

  23. As a mini-datapoint, when I was travelling in Tajikistan this year I saw a lot of yurts with Chinese-made solar panels on the side. This gives the nomadic owners enough electricity to run e.g. a DVD player (leading to surprising conversation-openers like “Have you seen Avatar yet?” from these remote tribespeople).

  24. A few thoughts from someone in the PV business…

    Solar panels on phones or other electronic devices are generally a bad idea, if your main concerns are environmental in nature. It takes solar panels a fair amount of time to generate more energy than it took to make them (~1yr or so, in many situations), and given the rate that people discard or lose phones, and the relatively small amount of time that they would actually be sitting around charging, the panels would probably never become a net-positive to the environment. PV needs to be in a place where it is always working. Of course, PV on phones would be nice for other reasons – remote use, fewer cables, etc., just not for environmental benefit.

    As to the cost of PV – it is true that it is hard to quantify at the moment, because there are so many variables. When you consider the cost of utility scale PV, the main variables you miss when comparing directly to traditional energy sources is the distributed cost to society of repairing damage caused by traditional energy sources – which at the moment everyone pays for, just not in the form of an electric bill. The cost to society of air pollution, water pollution, water table depletion, etc., are all hard to pin down, but are very real.

    When you consider PV on homes and businesses, you add another set of cost variable that isn’t accounted for – decreased cost of operation to the utility through distributed generation (as simple as transformers wearing out slower because they get used less on solar-heavy neighborhood), decreased transmission losses, and increased grid reliability relating to shaved peak loads. Right now most electric billing plans don’t differentiate much (or at all) between the extremely expensive energy produced during peak hours, compared with the very cheap energy produced during non-peak hours. Shaving peak loads, thus preventing the need to activate expensive “dispatchable” power sources, has a huge cost savings – one that is currently taken by the utility, but not passed along to energy consumers.

    The issues with battery storage are real, but as far as their effect on distributed generation (home and business, rather than utility scale), studies are currently suggesting that you wouldn’t have any problems with grid stability until you reach 10-15% penetration – which we are nowhere near.

    It is clear that PV has a lot of improvements to make, but it is closer than most people realize.

Comments are closed.