In late 2007, less than 10 years into the company’s existence, Google came out swinging on the clean energy front. To a fanfare of plaudits up and down Silicon Valley and well beyond, it declared “RE<C” as its goal: make renewable energy cheaper than coal. The company invested tens of millions of dollars into R&D efforts from concentrated solar power to hydrothermal drilling. Four years later, those efforts had been scrapped.
It would be all too easy to see this as an admission of failure—big tech playing in an arena it knew nothing about, with the hubris that Silicon Valley is known for. But something else was going on. Google’s shift in strategy was a reflection of the growing success of the solar sector. Google realized its energies were better directed toward massively scaling up existing renewable technologies that had plummeted in price, rather than inventing new ones.
While Google nailed the switch from R&D to deployment, it arguably still bet big on scaling up the wrong technology. In the early 2010s, the solar race looked like a tight competition between solar photovoltaic (PV) and utility-scale concentrated solar power (CSP), which uses sun-heated fluids to drive power turbines. Google quickly invested more than $1 billion in a slew of renewables companies and utilities, including big investments in CSP outfits BrightSource Energy and eSolar. A decade later, such choices aren’t looking promising, as CSP, too, has been losing out to PV’s continuing rapid cost declines.
Google is not alone in repeatedly misjudging the dropping price of solar cells over the last few decades and its impact on how we think about clean energy. Solar PV costs fell roughly by a factor of 10 in the past decade, on top of already impressive cost declines up to that point, for a total decline of around a factor of a hundred since US President Jimmy Carter unveiled solar panels on the White House in 1979. (Ronald Reagan took them down in 1986, during his second term as president.)
To put it in perspective, if gasoline had similarly dropped in price from 1979 levels, it would cost pennies a gallon today. Gasoline, of course, is a commodity, with prices fluctuating for a number of technological, economic, and political reasons. Solar PV prices are also driven by all these factors, but over the years, technology has clearly dominated. (This year, prices for solar PV modules have increased by around 18% because of a temporary crunch in the silicon supply chain.)
In its latest annual World Energy Outlook, the International Energy Agency declared solar PV to be “the cheapest source of electricity in history” for sunny locales with a low cost of financing. These two qualifications are important. Sun is obvious—solar is always going to be cheaper in Phoenix, Arizona, than in New York City—but the report concluded that solar is now cheaper than coal and natural gas in many places.
Financing is key to why this is true. Solar PV and other renewables such as wind have low or close-to-zero operating expenses—upfront costs have always been the big hurdle, and financing has been a big reason why. Thanks in part to various government policies, solar investment has become much less risky over the last decade or so, freeing up cheap money.
As a result, solar PV deployment has increased rapidly; it’s now the fastest-growing source of electricity globally, and figures to be for some time to come. It’s starting from a low base of installed capacity, however, far behind coal, gas, hydro, nuclear—even wind, which has been cheap for longer. And therein lies one of the biggest problems for solar PV. It might be the cheapest form of electricity for many, but that on its own doesn’t make the clean-energy transition nearly quick enough.
We need ever further technological advances. Why stop at grid parity, the point where it’s as cheap to build and operate solar PV as to supply electricity via fossil energy sources? Why not 10% cheaper? Why not strive to slash costs by another factor of 10 within a decade? Such drops are needed because the hallowed grid-parity goal is misleading—the real question is at what point utilities will actually abandon existing coal plants and switch to solar, rather than merely avoid adding new coal capacity. Solar needs to be so cheap it makes financial sense to build new solar capacity and shutter working coal and gas plants still making money for their owners.
All that calls for policy to both push existing solar technology and support R&D in new technologies. The entire package includes technology research, development, demonstration, deployment, and diffusion. Every step along this chain deserves direct government support, keeping in mind that it also gets increasingly more expensive the further down the chain one moves.
How to get cheaper
To better optimize investments to get to even cheaper solar, it’s worthwhile to understand what factors have driven down the cost of renewable power over the last few decades.
MIT energy systems scientist Jessika Trancik and her group find that the dramatic cost declines in solar cells over the course of three decades can largely be attributed to three factors: R&D leading directly to improvements in module efficiency (how much of the sunlight is converted into electricity) and other fundamental technological advances; economies of scale attributed to the size of solar-cell manufacturing plants and the increasing volume of inputs such as silicon; and improvements achieved through learning by doing.
None of that is too surprising, but what is less obvious is that the relative contribution of each varies greatly over time. From 1980 to 2000, R&D accounted for around 60% of cost declines, with economies of scale coming in at 20%, and learning by doing a distant third at around 5%; other largely unattributable factors account for the balance. That makes sense; it was a period of impressive advances in the efficiencies of solar cells but not a time of significant manufacturing and deployment. Since then, the pendulum has swung from R&D and fundamental technological improvements toward economies of scale in manufacturing, now accounting for over 40% of cost declines. It’s worth noting, however, that research advances still account for some 40% of declines.
The lesson for future investments that aim to make solar even cheaper: there should be direct support for all three, skewed toward economies-of-scale factors. Trancik’s findings only consider the solar PV module itself. That still leaves installation, connection to the grid, and other factors that make up total system costs. These are areas that will likely be improved as technicians and companies become more experienced. While the results of subsidies for increasing solar PV installations appear to be mixed at best, policies such as feed-in tariffs, which offer favorable long-term contracts to solar PV producers, and renewable portfolio or clean energy standards, which set quantity targets for renewables, show clear results in driving overall deployment.
No free lunch
Despite the dropping price of solar, the transition to renewables will still be costly. The big question, of course, is how expensive compared with what—climate change, too, comes with costs. Cheap solar gets even more financially attractive to developers if the social and environmental costs of carbon emissions from fossil fuels are considered.
A lot here hinges on the social cost of carbon (SCC), a tally of the financial damage each metric ton of carbon dioxide emitted today causes to the economy, society, and the environment—and, by extension, how much each ton of CO2 emitted should cost. It’s a number that says a lot about the true cost of coal and other fossil fuels—and about the appropriate support for solar PV and other renewables.
The latest US SCC, calculated by the Biden administration, puts the number at around $50 for a ton of CO2 emitted now. But that is surely an underestimate. Some calculate the SCC to go over $300 per ton of CO2, after fully accounting for the future damage caused by carbon emissions and for the uncertainties about climate change.
Whichever number you settle on, it means coal, oil, and natural gas will be far more expensive if you account for the full costs of greenhouse–gas emissions. Only then will low–carbon technologies be on the same playing field as fossil fuels.
An explicit carbon price via a tax or emissions trading system should be among those steps—but it ought not stop there. For one thing, renewable portfolio or clean energy standards, too, establish a price on carbon. Current US renewable portfolio standards at the state level translate into equivalent carbon prices of around $60 to $300 per ton of CO2, well within the range of recent SCC estimates. A federal clean electricity standard, part of the Biden administration’s proposed American Jobs Plan, could be in a similar range and would be similarly justifiable on the basis of updated SCC ranges.
Such a federal clean electricity standard would be a real boon to solar PV and other renewables, but climate policy must not end with pricing carbon. It also needs to include direct subsidies for deployment and support for R&D.
The most productive policy sequence might go something like this: first drive down the cost of renewables to create an economically viable alternative to high-carbon fuels, then price carbon via a direct price, a clean electricity standard, or something similar. The combination of the two should then lead to rapidly deployment of renewables at scale. In many ways, that is precisely what has happened, and it points to the clear need for the Biden administration and others to push for a price on carbon in whichever form that might happen.
But if far cheaper solar PV is the goal, it will also be critical to increase R&D to drive further improvements in the efficiencies of solar cells and find manufacturing advances that will allow even greater savings. And it is essential to keep exploring the scientific frontiers in search of other solar materials that could one day be even more efficient and cheaper.
Solar PV is cheap, but it is not free. Paying the price to make it ever cheaper will be well worth the cost.