Clean power from deserts

Overview of CSP
History of CSP
Activities and events
Press releases
Letters and comments
Mailing list
Site map

Desertec: COSTS

Back to Concentrating Solar Power

This page contains information about the cost of CSP (per kWh and per watt of capacity) and the cost of building the proposed EUMENA-wide supergrid.

Two things are relevant to an assessment of the cost of CSP in relation to other sources of power:

With regard to the second point, it is well established that the cost of new technologies falls as volumes increase—via mass-production of standardised parts, rationalisation of designs to speed up construction and to minimise the need for skilled labour, and other measures. This is discussed in the TRANS-CSP report from the DLR and in reports by Sargent & Lundy from 2003 and 2005. The following chart (from the MED-CSP report in 2005) shows how the cost of CSP electricity is likely to tumble compared with the cost of electricity from fossil fuels, assuming a 1% annual rise in the costs of fossil fuels (a predicted rate of increase that is smaller than the rate at which those costs have actually been rising).

xx Click the image to enlarge it

Dr Franz Trieb, Project Manager for the TRANS-CSP and MED-CSP reports, said (in August 2006) that "The cost of collecting solar thermal energy equivalent to one barrel of oil is about US$50 right now ... and is likely to come down to around US$20 in future."

Public support

In the chart shown above, from the 'MED-CSP' report (published in 2005), it is suggested that, in the period up to about 2017, public support of about US$ 75 billion will be required for CSP in the Middle East and North Africa. But a more recent study for the Center for Global Development (Desert power: the economics of solar thermal electricity for Europe, North Africa, and the Middle East, PDF, 710 KB, December 2008) says that "implementation will require international clean technology subsidies of about $20 billion over ten years. By the end of the program, the expected profitability of unsubsidized CSP projects is competitive with that of coal and gas power generation in Europe. Over the full life of the program, the present value cost of CO2 averted is as low as $14 per ton. This is a very modest carbon shadow price, even by conservative standards."

Cost per kilowatt hour

Currently, opinions seem to vary about the cost of electricity from CSP:

  • On the high side:
    • A report in the New Scientist (2007-05-05) says the current price is about 15 cents per kilowatt hour but suggests that it is likely to fall in the coming years.
    • A report from BBC News (2007-05-02) says "Is it true that this power is three times more expensive than power from conventional sources? Yes, but prices will fall, as they have with wind power, as the technologies develop." However, power from fossil-fuelled conventional sources is, at the present time, artificially cheap because it is using the atmosphere as a free dumping ground for CO2. And the real cost of nuclear power is disguised by overt and hidden subsidies (see Why we don't need nuclear power).
    • A report in Guardian Unlimited (2007-04-03) suggests that CSP electricity costs twice as much as electricity from fossil fuels. The same remarks apply as for the BBC report.
    • According to a report in (2008-09-02) "Jose Alfonso Nebrera [of ESTELA] estimates the current cost per kilowatt hour in Spain (assuming insolation of 2100 kWh/m2) at 27 eurocents/kWh—identical to Spain’s current feed-in tariff. In North Africa (assuming 2600 kWh/m2) he says this would be nearer 17.5 eurocents/kWh."
    • In Energy Technology Perspectives (2008), the International Energy Agency says that CSP plants under construction are expected to generate electricity at between 12.5 and 22.5 US cents/kWh.
  • On the low side:
    • In a press release dated 2007-11-27, Google Inc. announced a new strategic initiative to develop electricity from renewable energy sources that will be cheaper than electricity produced from coal. The initiative, known as RE<C, will focus initially on advanced solar thermal power (ie CSP), wind power technologies, enhanced geothermal systems and other potential breakthrough technologies.  RE<C will begin with a significant effort on solar thermal technology (CSP), and will also investigate enhanced geothermal systems and other areas. Larry Page, Google Co-founder and President of Products said: "Solar thermal technology ... provides a very plausible path to providing renewable energy cheaper than coal."
    • Bill Gross, CEO of eSolar, said (eSolar, NRG Energy to build solar power plants, Green Tech, 2009-02-23):

      eSolar's breakthrough modular power plants use more software and less steel to allow solar energy to be competitive with fossil fuels for the first time ever.

    • The US Department of Energy (in Big solutions for big problems (2007?)) says:

      ... existing CSP plants produce power now for as low as 12¢ per kilowatt hour (12¢/kWh) (including both capital and operating costs), with costs dropping to as low as 5¢/kWh within 10 years as technology refinements and economies of scale are implemented. Independent assessments by the World Bank, ADLittle, the Electric Power Research Institute, and others have confirmed these cost projections.

    • The TRANS-CSP report from the German Aerospace Centre (DLR), published in 2006, gives the cost of electricity from CSP plants in Spain as 15 cents / kWh, falling to about 7 cents / kWh by 2020 (Figure 2-24, p. 63). At that time, it is anticipated that electricity imported from CSP plants in North Africa will be about 5.5 cents / kWh, mainly because North Africa has more sunshine.
    • A report from Newswire Today (2007-11-05) says:

      These facilities [hybridised CSP plants] are to be developed at a levelized cost of about 6 cents per kilowatt-hour (kWh), which is about the same as that of coal plants, when taken into consideration that coal plants have to pay for their heavy emissions, construed as their additional levelized costs.

    • A report from CNET (2007-09-09) says:

      Ausra executives said that the [Ausra Fresnel mirror] system can now deliver electricity at 10 cents per kilowatt hour, more than the 9 cents per kilowatt hour that natural gas power plants cost.

    • Once Ausra's manufacturing operations are working on a large scale, its production costs and cost of capital will go down below the price of coal-fired plants which are 6 cents per kilowatt hour, he said.

    • In a report from Business Week (2006-02-14), the CEO of Solel is quoted as saying "Our [CSP] technology is already competitive with electricity produced at natural-gas power plants in California". At that time, the price of producing a kilowatt hour of electricity in California using Solel's parabolic trough CSP technology was quoted as approximately 10 cents, close to the cost of electricity from a newly-constructed gas plant.
    • Speaking about CSP at Solar Power 2006, legendary venture capitalist Vinod Khosla said "... we are poised for breakaway growth—for explosive growth—not because we are cleaner [than coal-fired electricity] but because we are cheaper. We happen to be cleaner incidentally." From remarks he has made elsewhere it seems that the comparison he had in mind was with so-called 'clean coal' that includes carbon capture and storage. There is also an interesting interview with him on ABC TV (Australia). There are further details on another page.
    • In February 2006, Professor David Mills of Solar Heat and Power (he is now at Ausra) said: "Our costs look to be about 10 times cheaper than photovoltaics and we believe now we are rapidly approaching cost-competitiveness with fossil fuels—advanced fossil fuel systems such as natural gas combined cycle and advanced coal technology. And we hope to be in that place probably about 2009 or 2010." This quote is from a short WWF film called "Renewable energy in Spain". The technology he was referring to is a version of the Fresnel mirror system.
    • John S. O’Donnell, President of Tsugino Co., speaking before the Public Utilities Commission of the State of Colorado on the 29th of September, 2006, said "Concentrating Solar Power can be provided as firm dispatchable power using thermal storage at a cost per kWh under $0.10 for deployments over 50MW in plant size, and under $0.09 for deployments over 200MW in size. I have just returned from meetings in Australia with Solar Heat and Power which confirm these costs."
    • The Solar Power Group says that, with the relative simplicity and cheapness of their Fresnel mirror CSP system: "... for big power plants (e.g. 200 MW) placed in remote locations with high solar irradiation, the cost of electricity will be highly competitive with that of conventional fossil fuel power plants". In a paper entitled The Solarmundo line focussing Fresnel collector. Optical and thermal performance and cost calculations, the cost of electricity from the Fresnel mirror system is calculated as 0.0750 €/kWh, a little lower than the cost which they calculate for electricity from a parabolic trough system: 0.0845 €/kWh.
    • In a press release issued on 2007-06-01, SCHOTT AG say "Parabolic trough power plants have the lowest electricity production costs of all types of solar technologies. That combined with the extremely high efficiency gained through technological advancements from companies like SCHOTT, will soon give parabolic trough power plants the potential to generate electricity in regions inside the Earth’s Sunbelt at costs comparable to those of power plants that run on fossil energy sources."

The TRANS-CSP report calculates that CSP electricity is likely to become one of the cheapest sources of electricity in Europe, including the cost of transmission.

Cost per watt of installed capacity

An alternative measure of costs is the cost per watt of installed capacity. That kind of measure can be derived from figures in news reports referenced in the table below. It seems likely that some of the figures given in these reports are not very accurate. Also, in these early stages, costs are probably higher than they will be in a few years time when economies of scale and refinements in the technologies have made things cheaper.

Another point to consider is that, a proper assessment of costs must include both the capital cost of the plant and the cost of the fuel over the lifetime of the plant. Plants that run largely on fossil fuels are cheap to build but expensive to run, and those running costs are likely to increase in the future. Plants that are powered mainly by the sun may, currently, be a bit more expensive to build but the fuel is free and it will last for ever!

US venture capitalist Vinod Khosla has said: "I think CSP, leveraging the large investment in traditional, steam-based power generation, and using passive mirrors to concentrate heat, can get to 35 percent efficiency today at $500 per kilowatt." (MIT Technology Review, March/April 2007).

Date link to news reports Name or location of project Cost (millions of US $) Capacity (MW) US$ per watt
2008-12-05 Tracy, California   1.5 1
2007-10-17 Kern County, California 100 109 0.9
2007-08-01 Nevada Solar One 266 64 4.2
2007-07-22 Seville (complete project, inc. PS10) 1618 (€1,200m) 300 5.4
2007-07-11 Andasol-2, Granada 351 (€260m) 650 0.54 (?)
2007-06-21 Beni Mathar, Morocco 632 (€469m) 470 1.3
2007-06-07 Andasol, Granada 332 50 6.6
2007-05-17 Extremadura province, Spain 809.3 100 8.1
2007-04-14 Victorville 2 hybrid power plant, California 450 563 0.8
2007-03-30 PS10, Seville 47 (€35m) 11 4.3
2007-03-28 Archimedes, Sicily 54 (€40m) 5 10.8
2007-03-27 Sacyr-Vallehermoso project, Spain 890 150 5.9
2007-03-15 Abu Dhabi, UAE 350 100 3.5
2007-03-14 Moree, NSW, Australia 161 (AU$200m) 73 2.2
2007-02-19 Kuraymat, nr. Cairo, Egypt 200 150 1.3

Reports on the cost of CSP

Reports about the cost of CSP are listed here with summaries and download links:

  • Economic, energy, and environmental benefits of concentrating solar power in California (PDF, 1.5 MB, prepared by Black & Veatch for the US National Renewable Energy Laboratory, April 2006). The report says that "The economic and employment benefits, together with delivered energy price stability and environmental advantages, suggest that the CSP solar alternative would be a beneficial addition to California’s energy supply. While early CSP plants are more costly than their traditional gas counterparts, subsequent plants are estimated to become nearly cost competitive on a levelized cost of energy basis."
  • Assessment of concentrating solar power technology cost and performance forecasts (PDF, 701 KB, prepared by Sargent & Lundy LLC Consulting Group for the US National Renewable Energy Laboratory, 2005). This report, which is an update of the one below, estimates that the cost of electricity from parabolic troughs is likely to fall from about 10.9 ¢/kWh in 2006 to about 6.5 ¢/kWh in 2020, while the corresponding costs for electricity from power towers is likely to be 15.0 ¢/kWh and 5.7 ¢/kWh, respectively.
  • Executive summary: assessment of parabolic trough and power tower solar technology cost and performance forecasts (PDF, 589 KB, prepared by Sargent & Lundy LLC Consulting Group for the US National Renewable Energy Laboratory, 2003. See also: the full report (PDF, 2.5 MB)). The report estimates that the cost of electricity from these two technologies in 2004 as being in the range of 10 to 14 ¢/kWh and then likely to fall to something between 3.5 and 6.2 ¢/kWh by 2020.

Other NREL reports about the costs and benefits of CSP may be found at See also the reports page.

The cost of building a large-scale HVDC transmission grid

Many people assume that it would be quite unacceptably expensive to build the kind of large-scale HVDC transmission grid that would be needed to transport CSP electricity from North Africa and the Middle East to Europe (see the page about Getting the energy to where it is needed). Here are the main reasons to think otherwise:

  • Quite apart from the import of CSP electricity, there are other powerful reasons for building a large-scale HVDC transmission grid throughout Europe and beyond.
  • Building a large-scale HVDC transmission grid need not be as complicated or difficult as it may at first sight appear. The supergrid proposed by Airtricity (see electricity transmission grids) would be created entirely from HVDC cables laid under the sea. This not only simplifies construction but it avoids objections on the grounds of visual pollution.
  • The HVDC transmission grid would work in conjunction with the existing HVAC transmission lines—it would not be a replacement for them. In other words, most of the transmission grid is already in place and the proposed supergrid would simply be an addition to that infrastructure.

Estimated costs

Here are some figures for the cost of building the kind of HVDC transmission grid proposed in the TRANS-CSP report:

  • In the report, it is envisaged that a total of 20 transmission lines would be needed between MENA and Europe, each one of 5 GW capacity—100 GW in all . The estimated total cost of this grid is €45bn.
  • Two 5 GW transmission lines (a 10 GW link) between North Africa and the UK would cost about €5bn.

As described below, these cost would be shared amongst several different countries and would be spread over several years.

Value for money

Studies show that, when the value of these several qualitative benefits of a supergrid are expressed in terms of money, they normally outweigh the cost of building the supergrid, and the difference can be substantial.

A useful source of information in this area is the article Interstate transmission superhighways: paving the way to a low-carbon future published by on the 30th of July 2008. Here are some quotes from the article:

  • Referring to a proposed 19,000 miles of new 765-kilovolt (kV) transmission lines at an estimated cost of US $60 billion, the article says "While the size and cost of the transmission superhighway may sound large at first glance, it is important to keep these numbers in perspective. Given that electricity transmission infrastructure typically remains in service for 50 years or more, the cost of the investment for the average household would be equivalent to about US $0.35 per month, less than the cost of a postage stamp."
  • "… a 2006 study by the Electric Reliability Council of Texas (ERCOT) found that over time an investment in new transmission infrastructure would produce benefits many times larger than the cost of the investment."
  • A more recent ERCOT study "found that the smallest transmission investment plan would bring enough new wind energy online to save US $1.2 billion per year in fuel costs—enough savings to cover the US $3.8 billion cost of the transmission infrastructure in a little over three years."
  • A study by the Midwest Independent System Operator (MISO) found that "Although the overall generation and transmission costs reached an estimated investment of US $13 billion, the project produced annual net savings of US $600 million over its costs. These savings are in the form of lower wholesale power costs and prices in the eastern U.S. resulting from greater access to lower cost generation in the western states such as Iowa and the Dakotas."

And so on. On the strength of this kind of evidence, it appears that large-scale transmission grids are good value for money.

Spreading the cost

Since the several benefits of large-scale transmission grids are spread very widely, it is not reasonable to put the burden of paying for transmission links on to the shoulders of individual project developers. Probably, the simplest and most effective way to spread the burden of paying for a EUMENA-wide supergrid is for the costs to be shared amongst national governments throughout the region, perhaps with contributions from the EU. No doubt, richer countries would pay more.

If the costs were spread amongst 30 or more countries in this way, and bearing in mind that the costs would also be spread over about 10 years, then the average annual expenditure by any one government would be €150 million or less.


Bearing in mind that large-scale transmission grids are good value for money, their cost is quite modest compared with other things that governments spend money on:


Back to Concentrating Solar Power

Last updated: 2009-09-22 (ISO 8601)