Elon Musk's plan to power the United States entirely on solar has one key flaw

Elon Musk has an idea: move the entire United States onto solar energy, using a 100-by-100 mile patch of land. The only problem is, you'd probably need a bit more than that.

The idea, a long-time favorite of Musk's, resurfaced in December 2019 when a Twitter user asked Musk about an old quote from Bill Gates. The Microsoft co-founder declared in 2011 that solar was "cute," but the answer lies in solutions like nuclear. Musk declared that Gates was "def wrong," declared that solar power offers one gigawatt per square kilometer, and cited a 2015 blog post from University College London written by energy research associate Andrew Smith.

"I was rather surprised, and indeed touched, to be cited in an argument between Elon Musk and Bill Gates," Smith tells Inverse. "It was quite unexpected to this working-class Grimsby lad to be called as a witness on Twitter to resolve an argument between billionaires."

The post, which seemed pretty logical in its approach, concluded that Musk was correct in his assertion. It worked on the assumption that Musk was calling for a system measuring 10,000 square kilometers, as at that point Musk had presented a map of his proposal without measurements. The United States used 3,725 Terawatt-hours over the course of one year in 2013, equating to 425 gigawatts. It then assumed an average photovoltaic yield of 0.24 gigawatts per square kilometer, based on the best panel on the market. It also looked at data for Amarillo, Texas and found panels can expect on average around 21 percent.

So then 10,000 square kilometers, multiplied by 0.24 gigawatts per square kilometer, multiplied by 0.21 gives us just over 500 gigawatts. Musk later clarified that his proposal suggested a site measuring 10,000 square miles, or 25,900 square kilometers, but the UCL analysis suggested that a smaller patch of land was sufficient.

"The science is pretty solid here: Musk is right, and Gates is wrong: the renewable energy resource that we could harness, far exceeds our electricity demands, and indeed our total demand for all energy services, not just those met presently by electricity," Smith says.

But one Inverse reader, who believes Gates is right to tout nuclear, wrote in to criticize Musk's plan. The reader, who claims to design commercial solar installations for a living, noted that the panel size is just one part of the equation:

A 1-megawatt facility using 345-watt panels takes about 4 to 4.5 acres, including aisles for maintenance, security fencing, inverters, and transformers, but no storage batteries. We’ve built several of these, as well as hundreds of smaller setups. And it produces its one megawatt only at high noon on the spring and autumn equinoxes, if the sky is perfectly clear – no clouds, no rain, no dust on the panels.

This approximately corresponds to what we know about public installations:

• The Topaz solar farm in San Luis Obispo County, California, which ranks as one of the largest solar farms in the United States, offers 550 megawatts and covers 4,500 acres. That gives around 8.2 acres per megawatt.
• The Solar Star projects in Rosamund, California, combine to offer 579 megawatts over 3,200 acres. That's 5.5 acres per megawatt.
• India's Kamuthi solar power station offers 648 megawatts over 2,500 acres. That ranks at just over four acres per megawatt.

So it seems the reader's figures are believable. Smith explained that means, in terms of how much space this giant plant would require, the system would require around 16 times more space than the panels themselves.

So how big would this giant plant really be? Somewhere around 160,000 square kilometers, based off the UCL analysis. That's about enough to almost entirely cover Syria with 25,000 square kilometers to spare.

That sounds like a lot of land, but Smith argues it is deceptively high. It equates to around 1.6 percent of the total land of the United States. This is land that in almost all cases could be used for multiple purposes. Agrivoltaics is an increasingly popular use for solar panel land, where crops grow in the shade underneath.

Smith previously explored the idea of switching the world to renewables in a 2015 blog post. It found that countries would only need to use a small fraction of their land to help the world fully transition. In this future scenario, 90.7 percent of the world's population in 2050 would be in countries using less than five percent of their land for renewables. The data, provided by the Delucchi & Jacobson 100% Solutions Project, found that the United States in particular would only need to use 0.69 percent of its total land.

Of course, all of this is an abstract thought experiment rather than a real proposal. Whether solar requires 10,000 or 100,000 square kilometers, the point is to illustrate that the technology exists to facilitate a transition.

"I consider Musk’s visualization to be merely a representation of the area of panels required to produce the amount of power needed on an annual average basis," Smith says. "It’s no deeper than that. So I don’t understand these attempts by some people to say 'it wouldn’t work in the real world, for these reasons'. To the best of my knowledge, no one is saying it should be built and used as the USA’s sole electricity supply in the real world. That’s missing the point completely."

It's unclear whether transitioning entirely to solar even makes sense. For the time being, Smith explains, the United States has a variety of alternative renewables sources like wind and hydro. Similarly, the United Kingdom has an abundance of offshore wind resources, to the point where it's been described as the "Saudi Arabia of wind energy." Solar, in its current state, would need to solve more issues around the cost and logistics of long-term energy storage before the world could contemplate going all-in with solar.

"And that’s ok," Smith says. "We don’t need to. We have already have pretty much all the scalable solutions we need to decarbonize the world’s energy supply, using a mix of clean sources – solar, wind, hydro, geothermal, bio-energy, and so on."

Is Musk's solar idea a tad misleading? Perhaps, but the point it raises still remains.

Update 01/31 6 a.m. Eastern time: An earlier version of this story suggested Musk proposed a 10,000 square kilometer installation. The article has been amended to reflect the fact that this figure is from UCL's analysis.