The Exponential Gains in Solar Power per Dollar

My post on the Moore’s Law-like exponential gains in solar power per dollar went up at Scientific American yesterday.  Reprinting here with permission.

The sun strikes every square meter of our planet with more than 1,360 watts of power.  Half of that energy is absorbed by the atmosphere or reflected back into space.  700 watts of power, on average, reaches Earth’s surface.  Summed across the half of the Earth that the sun is shining on, that is 89 petawatts of power.  By comparison, all of human civilization uses around 15 terrawatts of power, or one six-thousandth as much.  In 14 and a half seconds, the sun provides as much energy to Earth as humanity uses in a day.

The numbers are staggering and surprising.  In 88 minutes, the sun provides 470 exajoules of energy, as much energy as humanity consumes in a year.  In 112 hours – less than five days – it provides 36 zettajoules of energy – as much energy as is contained in all proven reserves of oil, coal, and natural gas on this planet.

If humanity could capture one tenth of one percent of the solar energy striking the earth –  one part in one thousand –  we would have access to six times as much energy as we consume in all forms today, with almost no greenhouse gas emissions.  At the current rate of energy consumption increase – about 1 percent per year – we will not be using that much energy for another 180 years.

It’s small wonder, then, that scientists and entrepreneurs alike are investing in solar energy technologies to capture some of the abundant power around us.  Yet solar power is still a miniscule fraction of all power generation capacity on the planet.  There is at most 30 gigawatts of solar generating capacity deployed today, or about 0.2 percent of all energy production.  Up until now, while solar energy has been abundant, the systems to capture it have been expensive and inefficient.

That is changing.  Over the last 30 years, researchers have watched as the price of capturing solar energy has dropped exponentially.  There’s now frequent talk of a “Moore’s law” in solar energy.  In computing, Moore’s law dictates that the number of components that can be placed on a chip doubles every 18 months.  More practically speaking, the amount of computing power you can buy for a dollar has roughly doubled every 18 months, for decades.  That’s the reason that the phone in your pocket has thousands of times as much memory and ten times as much processing power as a famed Cray 1 supercomputer, while weighing ounces compared to the Cray’s 10,000 lb bulk, fitting in your pocket rather than a large room, and costing tens or hundreds of dollars rather than tens of millions.

If similar dynamics worked in solar power technology, then we would eventually have the solar equivalent of an iPhone – incredibly cheap, mass distributed energy technology that was many times more effective than the giant and centralized technologies it was born from.

So is there such a phenomenon?  The National Renewable Energy Laboratory of the U.S. Department of Energy has watched solar photovoltaic price trends since 1980.  They’ve seen the price per Watt of solar modules (not counting installation) drop from $22 dollars in 1980 down to under $3 today.


Is this really an exponential curve?  And is it continuing to drop at the same rate, or is it leveling off in recent years?  To know if a process is exponential, we plot it on a log scale.


And indeed, it follows a nearly straight line on a log scale.  Some years the price changes more than others.  Averaged over 30 years, the trend is for an annual 7 percent reduction in the dollars per watt of solar photovoltaic cells.  While in the earlier part of this decade prices flattened for a few years, the sharp decline in 2009 made up for that and put the price reduction back on track.  Data from 2010 (not included above) shows at least a 30 percent further price reduction, putting solar prices ahead of this trend.

If we look at this another way, in terms of the amount of power we can get for $100, we see a continual rise on a log scale.


What’s driving these changes?  There are two factors.  First, solar cell manufacturers are learning – much as computer chip manufacturers keep learning – how to reduce the cost to fabricate solar.

Second, the efficiency of solar cells – the fraction of the sun’s energy that strikes them that they capture – is continually improving.  In the lab, researchers have achieved solar efficiencies of as high as 41 percent, an unheard of efficiency 30 years ago.  Inexpensive thin-film methods have achieved laboratory efficiencies as high as 20 percent, still twice as high as most of the solar systems in deployment today.


What do these trends mean for the future?  If the 7 percent decline in costs continues (and 2010 and 2011 both look likely to beat that number), then in 20 years the cost per watt of PV cells will be just over 50 cents.


Indications are that the projections above are actually too conservative.  First Solar corporation has announced internal production costs (though not consumer prices) of 75 cents per watt, and expects to hit 50 cents per watt in production cost in 2016.  If they hit their estimates, they’ll be beating the trend above by a considerable margin.

What does the continual reduction in solar price per watt mean for electricity prices and carbon emissions?  Historically, the cost of PV modules (what we’ve been using above) is about half the total installed cost of systems. The rest of the cost is installation.  Fortunately, installation costs have also dropped at a similar pace to module costs.  If we look at the price of electricity from solar systems in the U.S. and scale it for reductions in module cost, we get this:


The cost of solar, in the average location in the U.S., will cross the current average retail electricity price of 12 cents per kilowatt hour in around 2020, or 9 years from now.  In fact, given that retail electricity prices are currently rising by a few percent per year, prices will probably cross earlier, around 2018 for the country as a whole, and as early as 2015 for the sunniest parts of America.

10 years later, in 2030, solar electricity is likely to cost half what coal electricity does today.  Solar capacity is being built out at an exponential pace already.  When the prices become so much more favorable than those of alternate energy sources, that pace will only accelerate.

We should always be careful of extrapolating trends out, of course.  Natural processes have limits.  Phenomena that look exponential eventually level off or become linear at a certain point.  Yet physicists and engineers in the solar world are optimistic about their roadmaps for the coming decade.  The cheapest solar modules, not yet on the market, have manufacturing costs under $1 per watt, making them contenders – when they reach the market – for breaking the 12 cents per Kwh mark.

The exponential trend in solar watts per dollar has been going on for at least 31 years now.  If it continues for another 8-10, which looks extremely likely, we’ll have a power source which is as cheap as coal for electricity, with virtually no carbon emissions.  If it continues for 20 years, which is also well within the realm of scientific and technical possibility, then we’ll have a green power source which is half the price of coal for electricity.

That’s good news for the world.

For an update on these trends, see here.

You can also read about how battery prices are dropping exponentially too.

I write much more about solar, wind, energy storage, and why the pace of innovation in them is critical – and hopeful – for both fighting climate change and for long term economic growth in the book I originally did this research for, The Infinite Resouce: The Power of Ideas on a Finite Planet 

Sources and Further Reading:

Key World Energy Statistics 2010, International Energy Agency

Tracking the Sun III: The Installed Cost of Photovoltaics in the U.S. from 1998-2009, Barbose, G., N. Darghouth, R. Wiser.,  LBNL-4121E, December 2010

2008 Solar Technologies Market Report: January 2010, (2010). 131 pp. NREL Report TP-6A2-46025; DOE/GO-102010-2867

16 thoughts on “The Exponential Gains in Solar Power per Dollar”

  1. Pretty charts and a nice speech, but does either – especially the charts – address either maintenance costs (significant in larger installations) and power storage costs? If it addresses the latter at all, does it also include the secondary costs of battery disposal?

    Finally, what are the odds that solar power will just replace oil with rare earths as the key resource and turn China into the next Saudi Arabia?

    1. Jonolan: The price per kwh does include maintenance. It does not include storage. Without storage, solar can replace perhaps half of current electrical production, as peak demands are consistently on summer afternoons. Low cost storage systems will be needed before solar can do the rest. Watch for a post on that subject specifically in the coming weeks.

      In terms of rare earths, they’re neither rare nor limited to China. China has dominated exports of them because they’ve mined for them, not because they are any more available there. They may also be replaceable in many applications if their supply does become limited (which is in and of itself unlikely to happen for any long term period).

      For one look at the topic of rare earths, see:

      1. OK then. I’m just naturally distrusting of any good looking figures for ANY power generation mechanism; too many people have too much money at stake for me trust easily.

  2. Ramez – thanks for this article. I’ve worked my entire career in the oil and gas industry and it’s certainly welcome to see hope for energy optimism considering the myriad difficulties my industry is facing in growing supply.

    I’m curious to know how the cost per kwh of solar is computed. It seems to me there must be several assumptions baked in to compare solar electricity costs to the residential grid. Solar seems like it’s all or almost all fixed and little to no marginal cost which is certainly not the case for electricity generated by fossil fuel. I understand that the main thrust of your article is that costs of solar are declining, but I think you still ought to have included a brief discussion about this point with some detail about the assumptions and sensitivity of the cost comparison to those assumptions. It’s not like you’re targeting USA Today readership with this kind of article, we can take it.

  3. Thanks – a great synopsis and analysis. One question – what assumption is made for the lifetime of a PV module? TO get a cost per kwh there is an assumption on the number hours – how long before you have to replace the module. (As well as the hours of insolation per day)
    Thanks again.

    1. Quentin: To get the cost per kwh, rather than make assumptions about lifetime, I took current average price per kwh from new Solar PV installations and scaled it to future Solar PV costs. I used prices in Chicago, which has fairly average yearly insolation for the US (perhaps a bit on the low side).

      You get similar costs if you assume a 20 year lifetime of the modules and amortize their cost over that time.

      In the highest insolation areas of the US, the price would be about 1/3 less. It’s possible that with High Voltage DC lines (which have very low loss over long distances), the most sensible thing to do is deploy large Solar PV installations in the sunniest areas (generally but not always the South) and distribute the power to less sunny regions.

  4. I love your page.

    I was wondering what these charts looked like for wind production? Are their similar improvements being made in that technology and its cost-efficiency?

    Just curious.

  5. The life of PV crystalline panels as far as I know is for as long as they are not damaged. The PV Panels NASA made back in the 50’s still work to this day. The new thin film massed produced plastic PV’s have an average life of about 8-10 years from what I’ve read about them. They can be mass produced like newspapers. That should bring the price of them down immensely.

Comments are closed.