A picture is worth a thousand words. The graph above compares the price history of solar energy to conventional energy sources. This is what a disruptive technology looks like. While conventional energy prices remained pretty flat in inflation adjusted terms, the cost of solar is dropping,fast, and is likely to continue doing so as technology and manufacturing processes improve.
First, about the graph. I recently published an article, Bitcoin, Energy and the Future of Money, which explores the idea of using energy as the basis for money. One of the key concepts in this is to standardize the way energy commodities are measured, to measure them in terms of energy content rather than parochial units of measure (e.g. therms or cubic feet of natural gas, gallons of diesel, kilowatt hours of electricity, and so on). See also www.joulestandard.com for more information about this idea.
The graph above charts the inflation adjusted price of different types of energy, not in terms of gallons, but in terms of gigajoules of energy (a gigajoule is one billion Joules, the standard metric unit for energy). Pricing energy commodities in terms of their energy content makes it easier to compare the relative cost of different sources of energy.
Using data from the Energy Information Agency, I pulled together a history of retail prices for natural gas, crude oil, gasoline and residential electricity, all adjusted for inflation. For each energy source, I converted the prices to $/gigajoule, using conversion factors from engineering tables. (For example, a million cubic feet of natural gas contains 1.083 gigajoules of energy content).
Next, using data from the National Renewable Energy Laboratory and other sources, I looked up the price history for solar power, in terms of dollar per Watt of system capacity (a standard unit of measure for solar). With this data, I built a cost model to translate the cost of a solar cell into $/gigajoule. The basic idea is to amortize the system cost over its useful life, and divide this by the average amount of power it generates per month. This allows the cost of solar to be compared directly to other sources.
The comparison shows quite clearly that the cost effectiveness of solar power is increasing exponentially. In 1977, solar cells cost upwards of $70 per Watt of capacity. In 2013, that cost has dropped to $0.74 per Watt, a 100:1 improvement (source: The Economist). On average, solar power improves 14% per year in terms of energy production per dollar invested. Technological improvements are the primary driver of this trend, as manufacturers learn to produce panels using less energy and raw materials (the basic physics of solar panels are a century old), and to make systems easier to install.
While solar currently accounts for less than 1% of the energy supply, it is an exponentially improving technology, both in terms of price (14%/year) and pace of construction (60%/year). Already it is approaching parity with other energy sources in the Western US. Assuming this trend continues for another 10 to 20 years, and there’s no reason not to, solar power will become 5 to 10 times more cost effective than it is today. This raises an interesting question. What happens if solar becomes an order of magnitude cheaper than other sources of power?
This is the nature of disruptive technology. It represents such an improvement that it renders existing industries obsolete. We saw waves of disruption take place as the Internet upended entire industries. Expect to see a lot of this in the coming years.
UPDATE: a follower pointed out that not all Joules are equal, for example electrical energy versus thermal energy. Since the majority of electricity in the US is generated from carbon based fuels, we should expect it to cost roughly three to four times their amount. Why is this? The process of converting heat to electricity (by driving a steam turbine) is relatively inefficient. Most of the energy is lost as waste heat (this can be reclaimed in cogeneration, for example to use steam to heat buildings but this is not reflected in electricity prices).
UPDATE: several readers ask how I calculated solar prices. I did this by calculating how much power a 1 watt cell would generate per year, using average insolation of 4.5 hours/day (southern US), divided this by 12 for a monthly average. I calculated the system cost as three times the solar panel cost (installation and other components account for 2/3s of typical system cost). I then calculated the amortized monthly cost, assuming a 360 month term with a 5% cost of capital. Then divided this by monthly power output in gigajoules to get $/GJ. If the system is located in a less sunny area, or the cost of capital is higher, that will increase the cost per gigajoule. That said, the point of the article is the exponential decrease in unit costs, which trump everything over the next 10-20 years.
UPDATE: I am looking for inflation adjusted historical data for oil, natural gas and coal going back to the 1800s, so I’ll be updating the price history as I collect new data.
The World In Your LanguageUpdated April 24, 2013