In our Big Picture page on Energy Emissions, we saw that we need to do two things:

1. Clean up electricity
2. Electrify (almost) everything

We also saw that wind power and solar power have become incredibly cheap, and that they’re going to get even cheaper. That makes cleaning up electricity much easier than it was just a few years ago. 

But wind and solar power have a limitation. They are variable, generating electricity only when the wind is blowing or the sun is shining. So how can we rely on these “variable renewables” to power our electric grid?  

 ⇒ Read this article by Amory Lovins and M.V. Ramana, explaining why this problem is not nearly as difficult as it looks.

You might already have thought of one answer. When wind and solar generate more electricity than we need, we can use batteries to store the extra electricity, and then release that electricity into the grid later, when there’s not enough generation to meet the demand. 

In fact, lithium ion batteries (the same kind of batteries that EVs use) are great for doing this during the course of a day. They can cost-effectively soak up excess solar generation in the afternoon, or excess wind generation in the middle of the night, and then discharge it in early evening, when demand for electricity is highest and wind and solar generation are lower. (Ten years ago, using batteries this way would have been prohibitively expensive. But as more and more lithium ion batteries have been manufactured for use in EVs, these batteries have traveled along a “learning curve” and become more and more affordable.)

Unfortunately, lithium ion batteries are not cost effective for storing more than a few hours of energy. But solar and wind generation vary on timescales much greater than that. In fact, they both tend to vary seasonally. There are many fewer hours of sunshine in the winter (unless you’re near the equator), and there can also be long periods with little wind. Those cold times are just when we need the most electricity (and will need even more if we succeed in electrifying everything, including building heat). To store enough wind and solar to get us through these long, dark periods, we would need enormous numbers of lithium ion batteries – enough to make our electricity cost many times more than it costs now. 

None of this is a problem for us today. Even in the energy markets with the greatest penetration of renewables, there are still plenty of fossil fuel powered plants that can be fired up when there’s not enough solar or wind power. Experts estimate that it will start to become difficult reliably to meet peak demands when variable renewables make up about 80% of electricity generation in a given market. They make up about 12% of generation globally now, so in most places we have a long way to go before we begin to really feel this problem. For the next decade, our main task is just to deploy as much wind and solar and short-term battery storage as we can build, as fast as we can. This is an even bigger job than it seems, because we need to do more than just replace existing fossil fuel generation. In order to power our homes and vehicles and factories with clean electricity, we will need to more than double the amount of electricity we now generate and distribute on the grid. (⇒ This excellent article by energy systems modeller Jesse Jenkins does a great job of explaining the enormous task before us.) But even while we deploy wind, solar, and batteries faster than we’ve ever done, we need to start working now on the solutions we will need when we reach 80% variable renewables and we want to decarbonize the remaining 20% of the grid.

Fortunately, there are many potential solutions for the final 20%, and many smart people are taking shots on goal. Some solutions are technologically feasible right now, but face steep social and political obstacles. Others are still being proven out technologically and economically. So, it’s an open question what mix of these solutions we will end up deploying. (The answer will be different in different parts of the world.)  But enough of these technologies are far enough along that we can be confident that, by the time we need them, a range of solutions will be ready for us to draw on, from the wide portfolio of technologies we’re developing now.

⇒ Watch this short talk by Jesse Jenkins, explaining the different roles that different kinds of generation will play in a 100% clean grid.

Solutions for the final 20%

Here are some of the solutions people are working on to decarbonize the last 20% of the electricity grid, once variable renewables make up 80% of power generation.

Bottom-up grid architecture to accommodate Distributed Energy Resources

The electrical grid was built as a one-way system: a small number of large power plants send electricity out over the wires to meet the demand of homes, factories, and businesses. Grid operators forecast when demand for electricity will rise or fall (for instance, because more people will turn on air conditioners on a hot afternoon), and then ramp different power plants up or down to meet that demand. 

The world we are now entering is very different from the one the grid was built for, because the homes, businesses and factories on the “edge” of the grid are no longer just fixed sources of demand that the grid must meet. Instead, they might contain a wide range of “Distributed Energy Resources” (DERs). These include rooftops solar panels, which not only meet the demand of the buildings they are on but also feed electricity into the grid; “smart,” controllable appliances – from water heaters and heat pumps to EV chargers – that can turn their demand for electricity up or down in response to signals from the grid operator; batteries (including those in EVs) that can discharge electricity into the grid when the grid needs it; and microgrids that can “island” themselves from the grid and provide their own power during a blackout.

Collectively, the millions of DERs that will soon live at the edge of the grid can go a long way to reducing the peaks in demand that are difficult for a grid with more than 80% renewables to meet. In order for that to happen, however, grip operators need to be able to coordinate and manage these millions of DERs – something that is difficult or impossible to do with today’s centralized, one-directional grid. And so a host of startups are working to reimagine and redesign the architecture of the grid, so that it can make use of the vast resource that DERs provide.

⇒ Read “Clean energy technologies threaten to overwhelm the grid. Here’s how it can adapt,” by David Roberts. It is is a crystal-clear, must-read explainer of these concepts, drawing from academic work to lay out a vision of the way the grid can work that has helped to inspire some of the entrepreneurs working in this space.

Here are some more resources on grid architecture for DERs:

• Canary Media explainer on the power grid
• David Roberts, Rooftop solar and home batteries make a clean grid vastly more affordable
• David Roberts, Meet the microgrid, the technology poised to transform electricity
• Volts podcast: Jesse Morris on building an operating system for distributed energy

• Watt It Takes podcast with Ugwem Eneyo, founder of Shyft Energy

• My Climate Journey podcast interview with Astrid Atkinson, founder of Camus Energy

Virtual Power Plants

One way to turn some of the millions of DERs into a resource for the grid (without changing the architecture of the grid as a whole) is to allow utilities or companies to sign up customers whose homes or businesses have DERs, and connect these together to form “Virtual Power Plants” (VPPs). For instance, a manufacturer of EV chargers might pay thousands of its customers to participate in a VPP. The customers will still plug their EV in as soon as they get home from work, but they agree to let the software in the charger determine which hours their car charges in overnight, so long as the car is charged by morning. Suppose that on a hot summer evening the grid is expected to need 10 megawatts of additional power. Instead of firing up a gas peaker plant to meet the additional demand, the VPP could “provide” this power by signaling all of its chargers not to turn on until a few hours later – thus reducing demand for electricity by 10 megawatts – and getting rid of the need for the peaker plant.

This is just one flavor of VPP, and one example of the ways in which VPPs can help. For more, read and listen below.

• Canary Media, Four ways virtual power plants can help the US grid keep up with demand
• Canary Media explainer on Vehicle-to-Grid charging
• Jigar Shah, VPPieces – a series of mini-blogs about Virtual Power Plants
• My Climate Journey podcast interview with Cisco DeVries, founder of OhmConnect

• US Department of Energy report, “Pathways to Commercial LIftoff: Virtual Power Plants”

Here are some startups working to enable Virtual Power Plants, or to create a flexible, bottom-up electricity grid that can take advantage of them:

Transmission

Wind and solar power are both variable, but often they are complementary: the wind tends to pick up just when solar produces less – on cloudy days and at night. Occasionally, though, in a given location, days or weeks can go by with little wind or sun. This can happen in cold climates in the depths of winter, when demand for electricity is highest. It is one of the main reasons that powering a grid with more than 80% renewables can be difficult.

You might think that this should not be such a big problem, because the wind is always blowing somewhere, and the sun is always shining somewhere. This is true, but on most grids around the world, this isn’t much help, because although the different parts of those grids are all connected, many of those connections are small wires that cannot cannot carry much electricity. On a day that is dark and still in Chicago, solar panels in Arizona and wind turbines in Oklahoma may be producing enormous amounts of power, but there’s currently no way to transmit much of that power to meet Chicago’s need.

A simple, elegant, and extremely cost-effective solution is to build many more long-distance, high-voltage transmission lines to connect the various parts of the grid. Study after study has shown that building transmission not only enables us to deploy more renewables with fewer worries about variability; it also saves money. (Solar power produced in Arizona is really cheap!) However, in the US, and in many parts of the world, there are political, bureaucratic, and financial obstacles to building new transmission lines, with the result that new lines can take decades to be completed. (Russel Gold’s great book Superpower tells the surprisingly gripping story of how vested interests killed a decade-long attempt to build a transmission line that would have made almost everyone better off by bringing cheap wind power from Oklahoma to the Eastern US.)  

Because of those obstacles, it’s unlikely we’ll build enough transmission to decarbonize the grid in the cheapest way possible. Nonetheless, read the resources below to see the many ingenious ways to build as much transmission as possible, as quickly as possible:

• David Roberts, series of articles on transmission (also listenable as Volts podcasts)
• John Porcari, Laura Rogers and Jiger Shah, “A transportation, infrastructure and climate priority”

Overcapacity

In Electrify, Saul Griffith proposes a solution to the variability of renewables that complements building more transmission lines: we can build so much solar and wind that even on a short, cloudy, and calm winter’s day, it will still provide enough power to meet our needs. If we build enough renewable generating capacity to meet demand at the most difficult times of the year, then during most of the year, we will have overcapacity – more than we need. This sounds wasteful!  But in fact, it is exactly what we already do with fossil fueled power. We build gas peaker plants that we turn on only during 1% of the hours in a year. The other 99% of the year, we are paying to maintain them and have them at the ready, even though they are not generating power. Building solar panels or wind turbines whose power we don’t always need is the same, except that when we don’t need their power, they’re still producing power that is essentially free. (And in fact, we can put this excess, free power to many good uses – for instance, to produce green hydrogen.)

This is an elegant solution, and it may be cost-effective. But it faces an important, real-world obstacle. In many places where developers want to site new wind and solar farms, they are meeting local opposition to this use of the land. This means that it will be hard to build the enormous amount of wind and solar we need just to meet average demand;  building enough to meet peak demand in the next couple of decades, in the faces of local opposition to siting, would be even more difficult. For more on this idea, read:

• Saul Griffith, Electrify, chapter 8.

Medium to Long Duration Storage

Lithium ion batteries (and their cousins, LFP and sodium ion batteries) are great for filling in gaps during the course of a day, when wind and solar are not generating enough power to meet demand. To build a grid that can run on more than 80% wind and solar, we will need to find cost-effective ways to store energy to fill in gaps on the scale of multiple days and even weeks. These might include batteries that use different, cheaper chemistries; but they might also include entirely different ways of storing energy, like compressing air, pumping water uphill, and using electrolysis to create hydrogen which can be turned back to electricity later. Read and listen below to get a better understanding of the shape of the problem, and some of the many solutions startups are working on.

• David Roberts, Long-duration storage can help clean up the electricity grid, but only if it’s super cheap
• Canary Media Storage Roundup
• Watt it takes podcast interview with Mateo Jaramillo, founder of Form Energy

• Watt it takes podcast interview with Ramya Swaminathan, CEO of Malta

• Canary Media story on green hydrogen storage project in Utah
• Canary Media story on Energy Dome compressed CO2 storage
• US Department of Energy Liftoff Report on Long Duration Storage and Webinar

Here are some companies working on longer duration energy storage:

Firm clean power

So far, we’ve been looking at ways to make a grid work with only wind and solar power, both of which depend on the weather. The job becomes much easier if, as a complement to these variable power sources, we have some firm, zero-carbon sources of generation in the mix as well – sources that will continue generating when we need them, for as long as we need them. There are many options. Some of these, like tidal and wave power, are not yet mature technologies, or are expensive. Others, like hydropower, are limited by factors like geography. But innovators are working to improve all of them. 

Geothermal

Geothermal electricity generation has been around for more than a century, but it has been limited to a very narrow range of geographies. Startups pursuing enhanced geothermal and advanced geothermal are seeking to change that. Read and listen:

• Eli Dourado, “The state of next-generation geothermal energy”
• Volts podcast interview with Tim Latimer, founder of Fervo Energy
• News Roundup on Geothermal developments
• Article on Quaise, the geothermal moonshot company
• Catalyst Podcast interview with Jamie Beard of Project InnerSpace: Could geothermal becomes a major new zero-emissions player?
• Volts Podcast interview with Wilson Ricks on the potential for “load-following” enhanced geothermal
• US Department of Energy, Pathway to Next-Generation Geothermal Power Commercial Liftoff report

Here are some companies working on a new generation of geothermal energy.

Allam cycle gas with carbon capture

Carbon Capture and Storage (CCS) projects that seek to capture and then sequester the carbon from a power plant’s exhaust stream and then sequester it permanently underground have been around for decades. None has proved economical or practical, because it is difficult and expensive to separate the CO₂, which makes up only between 3% – 15% of typical power plant’s exhaust, from the other gasses in the exhaust. Allam Cycle power plants, by contrast, separate CO₂ from gas before it is combusted, and so produce a pure stream of CO₂ that can easily be captured and stored. Read:

• Vox article on NET Power Allam Cycle gas plant with Carbon Capture

Wave and tidal power

The first power plant harnessing the power of ocean water to generate electricity was built in 1910. In the century since, there have been dozens of efforts to harness waves or tidal power, using dozens of designs. The ocean is a punishing environment for machinery, however. So, while many designs have worked, at least for a time, the cost of building and then maintaining these generators has so far stayed high; and no design has been commercialized at meaningful scale. A new generation of startups, with new technologies, are now trying to solve the problem of making wave or tidal power robust, affordable, and scalable. Listen to this podcast interview with the founder of one of them:

• My Climate Journey podcast interview with Marcus Lehmann, founder of CalWave

Here are some companies working to harness wave power or tidal power.

Hydropower

About 16% of the world’s electricity comes from rivers dammed for hydropower. In the right circumstances, this is a firm source of power – although reduced rain and snow due to climate change have forced hydropower plants to reduce generation in many parts of the world. However, building new traditional hydropower has disadvantages and limits. First, its scope to expand it is limited, because the most suitable geographies have already been developed. Second, traditional hydropower comes at an environmental cost, radically altering and usually degrading river ecosystems. And third, flooding  valleys to create new reservoirs can generate large amounts of methane (especially in parts of the world that are warm), as biomass in the flooded valley decomposes in the absence of oxygen.

Startups are developing hydropower solutions that avoid many of these disadvantages – for instance, by inventing new turbines for generators to allow existing dams to create more power while letting fish pass through safely, and by creating turbines that can be used in rivers and canals without damming them at all. Read and listen:

• Volts podcast interview with Jennifer Gerson, “What’s Going on With Hydropower?”

• Catalyst podcast interview with Gia Schneider, co-founder of Natel Energy “How to Build More Hydropower”

• Volts podcast interview with Emily Morris, founder of Emrgy, “How to make small hydro more like solar”

“Traditional” (light water) nuclear reactors

Light water nuclear reactors generate about 10% of the world’s electricity.

Throughout the twentieth century and into the twenty first, most environmentalists opposed nuclear power. This was partly because of its close, historical ties with the nuclear weapons industry, and partly because of fears of dangerous radioactive leaks – both from reactors, and from radioactive waste that they create. Many environmentalists still oppose nuclear power for these reasons, and have campaigned vigorously to shut existing power plants down. We think that this is a mistake, for two reasons.

First, while nuclear power can pose dangers to human health and to the environment, these dangers are dwarfed by the health and environmental harms actually caused, every day, by burning fossil fuels. The two deadliest nuclear accidents in history were the accident at Chernobyl, Ukraine in 1986, which caused about 433 people to die prematurely (including deaths from cancer), and the accident at Fukushima, Japan, in 2011, which caused or will cause about 2314 premature deaths. These were terrible disasters. But to understand how dangerous nuclear power is, we have to compare these deaths to the premature deaths caused by burning fossil fuels to generate an equivalent amount of electricity. Most deaths from fossil fuels are caused by air pollution, but they also come from water pollution and from accidents that harm workers and communities at every step of extracting, transporting, refining, and delivering these fuels to the sites where they are finally burned.

⇒Read this article from Our World in Data, which makes the comparison in meticulous detail.

Drawing on data about deaths connected to generating nuclear power all over the world, from the dawn of the nuclear age in the 1950s to the present day, researchers at Our World in Data created this chart:

The same article from Our world in Data also offers the following, helpful way of thinking about the data the chart summarizes.

Let’s consider how many deaths each source would cause for an average town of 150,000 people in the European Union, which… consumes one terawatt-hour of electricity per year. Let’s call this town ‘Euroville’.

If Euroville was completely powered by coal we’d expect at least 25 people to die prematurely every year from it. Most of these people would die from air pollution.

This is how a coal-powered Euroville would compare with towns powered entirely by each energy source:

• Coal: 25 people would die prematurely every year;
• Oil: 18 people would die prematurely every year;
• Gas: 3 people would die prematurely every year;
• Hydropower: In an average year 1 person would die;
• Wind: In an average year nobody would die. A death rate of 0.04 deaths per terawatt-hour means every 25 years a single person would die;
• Nuclear: In an average year nobody would die – only every 33 years would someone die.
• Solar: In an average year nobody would die – only every 50 years would someone die.

Of course, there could one day be a nuclear accident worse than Fukushima or Chernobyl, and this would change the data. However, no accident is likely to come anywhere near to outweighing the estimated 1.1 million to 2.5 million people who die every single year as a result of burning fossil fuels for electricity. Our understanding of the danger of nuclear is simply out of whack.

The second reason we think it is wrong to oppose nuclear power – and especially wrong to campaign to shut down existing nuclear power plants – is climate change. The death tolls from fossil fuel power estimated above do not include deaths from climate change. So, fossil fuels are far, far worse than the numbers above indicate. Nuclear power, on the other hand, emits no CO₂ or other GHGs at all.

You might think that in comparing nuclear power to fossil fuels, we’re asking the wrong question. Why don’t we just replace both fossil fuels and nuclear with renewables like solar and wind? We’ve already seen one reason: making the electrical grid run reliably with more than 80% wind and solar is a difficult problem, for which clean, firm nuclear power is one potential solution.

But even if you don’t think we should add any more nuclear power to the grid, it’s important to remember where we are right now. We’re in a race against climate change to build enough clean power sources as fast as we can, so that we can retire fossil fuel power generation. As it is, we’re not going nearly fast enough, and we’re adding CO₂ to the atmosphere every day. Closing down existing, zero-carbon nuclear plants means that we are setting ourselves even further behind in the race, because it means we need to build even more wind and solar to replace them before we can stop burning fossil fuels. And in fact, in each case where activists have succeeded in shutting down nuclear power plants – for instance, the Indian Point reactor in New York, as well as nuclear power plants all across Germany and all across Japan – the result has been an immediate increase in fossil fuel consumption to make up the shortfall. It’s a setback the climate cannot afford.

Read:

• Article on the climate imperative to keep existing nuclear plants open
• Volts podcast interview with Jigar Shah, Director of the US Dept. of Energy Loan Programs Office:

Advanced Nuclear Power and Small Modular Reactors (SMRs)

Building new nuclear power plants is expensive. They take well over a decade to complete, and in many parts of the world, they tend be subject to huge cost-overruns. Because nuclear power is clean and firm, it plays a valuable role on the grid complementing variable renewables, and so it will be valuable even if the cost of of a megawatt hour of nuclear power is many times higher than the cost of a megawatt hour of solar power. But recent nuclear power plants have been prohibitively expensive. If we cannot find a way to bring these costs down substantially, new nuclear power will play no substantial role meeting the final 20% of power demand.

A wide range of companies are working to solve this problem, with new designs for nuclear power plants that they hope will make them cheaper to build and operate. None of these has succeeded yet, and the jury is still out whether any of them will. There is plenty of well-reasoned skepticism. Here are some resources to help give you a feel for this landscape.

• David Roberts interview with Suzy Hobbs Baker and Jessica Lovering
• “Advanced Nuclear Reactor Technology: A Company Compendium” — a good overview of the many companies developing Small Modular Reactors
• US Department of Energy: Liftoff Report on Advanced Nuclear, and Webinar
• My Climate Journey interview with Carolyn Cochran, co-founder of Oklo

Here are some companies developing Small Modular Reactors (SMRs):

Fusion

The nuclear power plants we have been talking about so far all harness the energy produced by splitting atoms apart. This is nuclear fission. But in the sun and other stars, immense energy is produced in a different way: by joining hydrogen atoms together, to create helium. This is nuclear fusion. We have created fusion reactions on earth by exploding hydrogen bombs. But we have not yet created fusion reactions that we can sustain, control, and harness to create electricity. Whereas fission requires dangerous radioactive elements like uranium or plutonium as fuel, fusion requires an isotope of hydrogen that is cheap and abundant (deuterium), and another (tritium) that is expensive and limited in supply right now, but that fusion reactions can produce more of once they get going. And whereas fission creates long-lived radioactive waste, fusion does not. For these reasons, engineers and policy-makers and have long thought of electricity from controlled fusion as a holy grail – an unending source of clean, cheap power that could usher in a new age of human prosperity.

Controlling fusion is a devilishly hard problem. To do it, we need advances in many branches of physics and engineering, from plasma dynamics to materials science. There is good reason to think that problems in these fields can all, eventually, be solved. Scientists and engineers have been pursuing them since the 1950s. The running joke, for many decades, has been that fusion power is about thirty years out – and always will be. In the last twenty years, however, dozens of startups have begun to pursue a wide variety of approaches, beyond the few already being pursued by large, government-funded laboratories. They have been enabled by real advances in computing power and superconducting magnets. There is now some reason to hope that we might develop workable fusion power plants in coming decades.

Proponents of fusion often look to it as a silver bullet that will solve our energy and climate problems. This seems to us to be a mistake. First, as Michael Liebreich argues in the article we provide below, even if we can get a working fusion pilot plant in the coming decades on the fastest, most optimistic timetable, it’s unlikely that we will be able to build enough, fast enough, at commercial scale to contribute meaningfully to meeting our energy needs by 2050. But we need to reach net zero by 2050 to prevent terrible global warming. We cannot afford to wait for fusion to help.

Second, while fusion power would be valuable for many reasons, it might not be the super-cheap power source that many suggest. It is true that the fuel would eventually be cheap. But most of the fusion designs being developed would still require very large, very complex physical plants made with expensive, special-purpose materials – not the sort of small, replicable modules, like solar panels or even wind turbines, that can be stamped out in huge quantities by factories, and that tend rapidly to descend learning curves as a result. And in order to turn the heat generated by a fusion reactor into electricity, most designs will still need to incorporate steam-driven turbines of the kind that operate in coal and gas plants. (Although at least one startup, Helion Energy, is working on a design that does not need a steam turbine.)

For a better understanding of where fusion is now and what it promises, read and listen:

• Catalyst Podcast, Is Nuclear Fusion Getting Closer?
• Canary Media roundup of fusion projects
• Michael Liebreich, Investing in Fusion – Fools Gold or Gold Mine? (An argument that, even if fusion succeeds, it probably won’t be fast enough to make a difference on climate.)

Here are a few of the many companies working to develop nuclear fusion: