How much solar would it take to power the U.S.?

Critics claim that there simply isn’t enough land in the U.S. for solar to power the country. While it’s not an immediately practical question, it’s still fun to ponder. So, ignoring practical constraints like storage and grid technology, let’s explore whether we can fit enough solar to electrify the U.S.

One approach would be to start with data points like the size of solar panels, the wattage they produce, and the number of sunny hours per year. Unfortunately, this approach produces wildly different answers depending on assumptions. A better approach is to look at real-world data captured from real-world solar arrays and just extrapolate from there.

Starting with some conservative assumptions from a 2013 National Renewable Energy Labs (NREL) report, we know that it takes, on average, 3.4 acres of solar panels to generate a gigawatt hour of electricity over a year. Given the U.S. consumes about 4 petawatts of electricity per year, we’d need about 13,600,000 acres or 21,250 square miles of solar panels to meet the total electricity requirements of the United States for a year.

The yellow square is approximately 21,000 square miles. Map courtesy of Google Maps.
Are you interested in the shift from fossil fuels to clean energy? Sign up for regular news analysis and cutting edge insights.
We respect your privacy. You can opt-out at any time. Your email will only be used for communications from Freeing Energy.

How much are 21,250 square miles?

This may seem like an impractically large amount of land but not when you put it in perspective. 21,250 square miles is a square about 145 miles on each side . The U.S. has 3,797,000 square miles of land. Only about half a percent of that would be needed to provide enough solar energy to power the country.

Here are some other examples of land use in the range of tens of thousands of square miles:

  • 40,223 square miles – this is the size of the land leased by the oil and gas industry (according to the US Bureau of Land Management).
  • 18,500 square miles – the amount of federal land offered for lease to the oil and gas industry in 2017 alone.
  • 33,750 square miles – this is the land set aside to grow the corn used to make ethanol, a gasoline substitute.
  • 62,500 square miles – the total amount of U.S. land used for lawns.
  • 22,000 square miles – the size of the Mojave desert, located in southeast California.
  • 3,590 square miles – a best guess at how much land is used for parking lots.
  • 16,000 square miles – this is an estimate of the total surface of U.S. roads, including highways (4.12 million miles of roads that are an average 20 feet wide).

Actually, we probably only need about 10,000 square miles

Fortunately, we wouldn’t need to use all 21,250 square miles. NREL has another report conservatively estimating that rooftop solar alone could generate 34% of all U.S. electricity requirements.

Additionally, the solar arrays in NREL’s 2013 survey had efficiency levels of 13-14%. Modern solar panels average 16-17% efficient with widely available models easily exceeding 20%. Revising the estimates using higher efficiency and including rooftop coverage, only 10,000 square miles is required.

The blue square is approximately 100×100 miles or 10,000 square miles. Maps courtesy of Google Maps.

Interestingly, Elon Musk shared a nearly identical metric during a speech to the National Governors Association.

“If you wanted to power the entire U.S. with solar panels, it would take a fairly small corner of Nevada or Texas or Utah; you only need about 100 miles by 100 miles of solar panels to power the entire United States. The batteries you need to store the energy, to make sure you have 24/7 power, is 1 mile by 1 mile. One square-mile. That’s it.” — Elon Musk

Elon Musk speaking at the National Governors Association in 2017

The Freeing Energy Perspective

Between land and rooftops, there is more than enough space to build all the solar panels necessary to power the United States. Realistically, though, the future of clean energy won’t be entirely solar. Hydro, geothermal, and, particularly, wind, will contribute their fair share as well. But, if we want to move quickly, solar is the fastest path to clean energy. It can be installed as small, easier-to-finance projects. It can be built faster than any other kind of energy. Solar’s small footprint projects are easier to permit and there are more places to put them. Unlike wind and hydro, solar can be built in virtually every state, town, and county in the country. And, as the cost of battery storage continues to decline, it will quickly become feasible to store daytime solar energy for use at night. 

Additional reading

  • LandArt analyzed the amount of solar required to power the world and produced a great infographic (read here).
  • The original NREL analysis (here).

6 thoughts on “How much solar would it take to power the U.S.?

  1. What about using solar to power all cars, trucks, etc., if we converted all gas vehicles to electric, and how do we get solar energy to remote locations for all vehicles?
    How much more solar than 101 sq miles?

    1. Great question. I’ll do some quick math here. According to the US Bureau of Transportation statistics, vehicles drove 3.2 trillion miles in 2017. Using a Nissan Leaf as an average, an electric vehicle gets about 30 miles per kilowatt hour. This means it would take 106 billion kilowatt-hours or 106 terawatt-hours or drive those 3.2 trillion miles.

      Since it takes 21,250 square miles of solar for 4 petawatt hours (or 4,000 terawatt hours), each terawatt hour takes about 5.3 square miles per terawatt hour. So, very roughly, if all US cars and trucks converted to electric vehicles, it would take 5.3 square miles x 106 terawatt-hours, or 562 square miles of solar panels to power them all.

  2. I think that the title may be a bit misleading… there are lots of energy uses which are not included in that 4 petawatt-hours/year figure (which mearly looks at “electricity usage”.

    Total U.S. energy consumption (all end-uses, so this includes transportation that Steve was asking about as well) is about 97.3 quadrillion BTU or 28.519 petta-watt hours.

    That works out to needing about 151,500 square miles (not including battery area) which is nearly the area of California.

    I think this points to the need to pursue a balanced clean-energy portfoilio including solar, wind, batteries, hydro, and nuclear. (it would take around 3,255 square miles of nuclear plants, for example)

    1. Thanks for sharing these insights. I had hoped by using the term “power” rather than “energy” that it would be implied I meant electricity and not all energy sources. Either way, you raise a very intriguing point about the fundamental challenges of accelerating electrification in transportation and industrial heating. Can you point me to some sources that go through more of this math in depth? For example, how does energy to power transportation change (higher or lower) when it becomes electrified? Also, are the EIA data you cite the energy delivered or energy used? For example, for transportation is that data how much energy is used by the vehicles or is the potential energy available in the gasoline before it is combusted? I ask because converting to the equivalent electricity may or may not be 1:1 (because electric motors are more energy efficient than combustion engines)? Thanks for raising this and I look forward to exploring it more.

    2. (I don’t see a reply button on your comment – so I’m commenting above)

      That’s a good point – they seem to be measuring fuel consumption, not energy produced! For transportation (and for heating for that matter) you aren’t getting 100% of the possible heat out of the fuel. Typical modern ICE cars get around 25% thermal efficiency ( ) by contrast a modern natural gas power plant gets around 62% efficiency. That means that electrifying transportation is a massive savings in carbon even if those cars are charged with natural gas power plants. (around 2.5 times more energy per unit fuel)

      After generator, transmission, distribution, and charging losses – electric vehicles get around 85% of the power that is generated into their batteries (around 15-20% of the power is lost

      So if transportation makes up 27828 trillion BTU, at ~25% thermal efficiency, only around 2.038 pWh were generated (instead of 8.155). With charging/transmission losses, we’d need to generate about 2.4 pWh to actually charge 2 pWh worth of EVs.

      In other words, you save about 5.75 pWh through increased efficiency by moving away from ICE cars. Reducing your total generation need from 28.5 pWh to 22.75 pWh.

      Natural gas heating is pretty damn efficient – so you don’t get much savings there. Maybe 0.5-1 pWh.

      There is already around 300 tWh of hydro installed in the U.S, and around 800 tWh of nuclear. So that helps a bit.

      So the problem is a little less bad then I let on before… 109,393 sq-miles instead of 151,000. I still think this would be an envioronmental disaster to use that much land for solar/wind, not to mention the unbelievable amount of rare earth metals it would require ( ). I don’t think there is very much hydro resource to develop in the U.S. – so that leaves Nuclear (and realistically a little natural gas for flexability) to fill in the base load generation in my opinion.

    3. Great points all around. A few more thoughts…

      Heating, particularly industrial heating, strikes me as one of the largest challenges in transitioning away from fossil fuels.

      I think we’ll see electric vehicle charging become more and more local with community scale solar reducing the need for long haul transmission (and the associated energy losses from natural gas, distribution, etc). The fact that cars are idle most of the time, especially during the strong mid-day sun, means solar charging can be very flexible, somewhat mitigating the intermittency challenges. This isn’t a panacea but I think it’ll help a lot.

      Your point about hydro is spot on but one expert I interviewed reminded me that there are 73,000-ish dams in the US, the vast majority of which don’t have hydro. Collectively, these dams could add microhydro facilities and hydro could once again see its contribution rising.

      I too struggle to with how much land might be needed but given how much is already used for oil and gas leases as well as growing corn for ethanol (an economically and environmentally debatable practice), it’s arguable that the land could be put to better use.

      Lastly, on the rare earth materials, my research leaves me increasingly encouraged that running out of this stuff is less an issue than I had thought. I touched on this a little bit in my recent article:

Leave a Reply