by Bill Gates
Although it’s still in the experimental phase, fusion holds a lot of promise. Because it would run on commonly available elements like hydrogen, the fuel would be cheap and plentiful. The main type of hydrogen that’s usually used in fusion can be extracted from seawater, and there’s enough of it to meet the world’s energy needs for many thousands of years. Fusion’s waste products would be radioactive for hundreds of years, versus hundreds of thousands for waste plutonium and other elements from fission, and at a much lower level—about as dangerous as radioactive hospital waste. There’s no chain reaction to run out of control, because the fusion ceases as soon as you stop supplying fuel or switch off the device that’s containing the plasma.
In practice, though, fusion is very hard to do. There’s an old joke among nuclear scientists: “Fusion is 40 years away, and it always will be.” (Admittedly, I’m using the term “joke” loosely.) One of the big hurdles is that it takes so much energy to kick off the fusion reaction that you often end up putting more into the process than you get out of it. And, as you might imagine given the temperatures involved, it’s also a huge engineering challenge to build a reactor. None of the existing fusion reactors are designed to produce electricity that consumers could use; they’re for research purposes only.
The biggest project currently under construction, a collaboration between six countries and the European Union, is an experimental facility in southern France known as ITER (pronounced like “eater”). Construction on the project began in 2010 and is still ongoing. By the mid-2020s, ITER is expected to generate its first plasma, and to generate excess power—10 times more than it needs to operate—in the late 2030s. That would be the Kitty Hawk moment for fusion, a major accomplishment that would put us on the path to building a commercial demonstration plant.
And there are more innovations coming that could make fusion more practical. For example, I know of companies that are using high-temperature superconductors to make much stronger magnetic fields for containing the plasma. If this approach works, it would allow us to make fusion reactors far smaller and therefore cheaper and more quickly too.
But the key point is not that any one company has the single breakthrough idea we need in nuclear fission or fusion. What’s most important is that the world get serious once again about advancing the field of nuclear energy. It’s just too promising to ignore.
Offshore wind. Putting wind turbines in an ocean or other body of water has various advantages. Because many major cities are near the coast, we can generate electricity much closer to the places where it’ll be used and not run into as many transmission problems. Offshore winds generally blow more steadily, so intermittency is less of an issue too.
Despite these advantages, offshore wind currently represents only a tiny share of the world’s total capacity for generating electricity—about 0.4 percent in 2019. Most of that is in Europe, particularly in the North Sea; the United States has just 30 megawatts installed, and that’s all in one project off the coast of Rhode Island. Remember that America uses around 1,000 gigawatts, so offshore wind provides roughly 1/32,000th of the country’s electricity.
For the offshore wind industry, there’s nowhere to go but up. Companies are finding ways to make turbines bigger so each one can generate more power, and they’re solving some of the engineering challenges involved in placing large metal objects out in the ocean. As these innovations drive down the price, countries are installing more turbines; the use of offshore wind has grown at an average annual rate of 25 percent in the past three years. The U.K. is the world’s biggest user of offshore wind today, thanks to clever government subsidies that encouraged companies to invest in it. China is making big investments in offshore wind and will likely be the world’s biggest consumer of it by 2030.
The United States has considerable offshore wind available, especially in New England, Northern California and Oregon, the Gulf Coast, and the Great Lakes; in theory, we could generate 2,000 gigawatts from it—more than enough to meet our current needs. But if we’re going to take advantage of this potential, we’ll have to make it easier to put up turbines. Today, getting a permit requires you to run a bureaucratic gauntlet: You buy one of a limited number of federal leases, then go through a multiyear process to generate an environmental impact statement, then get additional state and local permits. And at each step of the way, you may be opposed (rightly or not) by beachfront property owners, the tourism industry, fishermen, and environmental groups.
Offshore wind holds a lot of promise: It’s getting cheaper and can play a key role in helping many countries decarbonize.
Geothermal. Deep underground—as close as a few hundred feet, as far down as a mile—are hot rocks that can be used to generate carbon-free electricity. We can pump water at high pressure down into the rocks, where it absorbs the heat and then comes out another hole, where it turns a turbine or generates electricity some other way.
But exploiting the heat under our feet has its downsides. Its energy density—the amount of energy we get per square meter—is quite low. In his fantastic 2009 book, Sustainable Energy—Without the Hot Air, David MacKay estimated that geothermal could meet less than 2 percent of the U.K.’s energy needs, and delivering even that much would require exploiting every square meter of the country and doing the drilling for free.
We also have to dig wells to reach it, and it’s hard to know ahead of time whether any given well is going to produce the heat we need, or for how long. Some 40 percent of all wells dug for geothermal turn out to be duds. And geothermal is available only in certain places around the world; the best spots tend to be areas with above-average volcanic activity.
Although these problems mean that geothermal will contribute only modestly to the world’s power consumption, it’s still worth setting out to solve them one by one, just as we did with cars. Companies are working on various innovations that would build on the technical advances that have made oil and gas drilling so much more productive in the past few years. For example, some are developing advanced sensors that could make it easier to find promising geothermal wells. Others are using horizontal drills so they can tap these geothermal sources more safely and efficiently. It’s a great example of how technology that was originally developed for the fossil-fuel industry can actually help drive us toward zero emissions.
Storing Electricity
Batteries. I’ve spent way more time learning about batteries than I ever would’ve imagined. (I’ve also lost more money on start-up battery companies than I ever imagined.) To my surprise, despite all the limitations of lithium-ion batteries—the ones that power your laptop and mobile phone—it’s hard to improve on them. Inventors have studied all the metals we could use in batteries, and it seems unlikely that there are materials that will make for vastly better batteries than the ones we’re already building. I think we can improve them by a factor of 3, but not by a factor of 50.
Still, you can’t keep a good inventor down. I’ve met some brilliant engineers working on affordable batteries that could store enough energy for a city—what we call grid-scale batteries, as opposed to the smaller ones that run a phone or computer—and hold it long enough to get through seasonal intermittency. One inventor I admire is working on a battery that uses liquid metals instead of the solid metals employed in traditional batteries. The idea is that liquid metal lets you store and deliver much more energy very quickly—exactly the kind of thing you need when you’re trying to power an entire city. The technology has been proven in a lab, and now the team is trying to make it cheap enough to be economical and prove that it works in the field.
Others are working on something called flow batteries, which involve storing fluids in separate tanks and then generating electricity by pumping the fluids together. The bigger the tanks, the more energy you can store, and the bigger the battery, the more economical it becomes.
Pumped hydro. This is a method of storing city-sized amounts of energy, and it works like this: When electricity is ch
eap (for example, when a stiff wind is turning your turbines really fast), you pump water up a hill into a reservoir; then, when demand for power goes up, you let the water flow back down the hill, using it to spin a turbine and generate more electricity.
Pumped hydro is the biggest form of grid-scale electricity storage in the world. Unfortunately, that’s not saying much. The 10 largest facilities in the United States can store less than an hour’s worth of the country’s electricity consumption. You can probably guess why it hasn’t really taken off: To pump water up a hill, you need a big reservoir of water and, of course, a hill. Without either, you’re out of luck.
Several companies are working on alternatives. One is looking at whether you could move something other than water uphill—pebbles, for example. Another is working on a process that would do away with the hill but not the water: You pump water underground, keep it there under pressure, and then release it when you’re ready to turn a turbine. If this approach works, it would be magical, because there would be very little aboveground equipment to worry about.
Thermal storage. The notion here is that when electricity is cheap, you use it to heat up some material. Then, when you need more electricity, you use the heat to generate power via a heat engine. This can work at 50 or 60 percent efficiency, which isn’t bad. Engineers know about many materials that can stay hot for a long time without losing much energy; the most promising approach, which some scientists and companies are working on, is to store the heat in molten salt.
At TerraPower, we’re trying to figure out how to use molten salt so that (if we’re able to build a plant) we don’t have to compete with solar-generated electricity during the day. The idea would be to store heat generated during the day, then convert it to electricity at night, when cheap solar power isn’t available.
Cheap hydrogen. I hope we get some big breakthroughs in storage. But it’s also possible that some innovation will come along and make all these ideas obsolete, the way the personal computer came along and more or less made the typewriter unnecessary.
Cheap hydrogen could do that for storing electricity.
The reason is that hydrogen serves as a key ingredient in fuel cell batteries. Fuel cells get their energy from a chemical reaction between two gases—usually hydrogen and oxygen—and their only by-product is water. We could use electricity from a solar or wind farm to create hydrogen, store the hydrogen as compressed gas or in another form, and then put it in a fuel cell to generate electricity on demand. In effect, we’d be using clean electricity to create a carbon-free fuel that could be stored for years and turned back into electricity at a moment’s notice. And we would solve the location problem I mentioned earlier; although you can’t ship sunlight in a railcar, you can turn it into fuel first and then ship it any way you like.
Here’s the problem: Right now, it’s expensive to produce hydrogen without emitting carbon. It’s not as efficient as storing the electricity directly in a battery, because first you have to use electricity to make hydrogen and then later you use that hydrogen to make electricity. Taking all these steps means you lose energy along the way.
Hydrogen is also a very lightweight gas, which makes it hard to store within a reasonably sized container. It’s easier to store the gas if you pressurize it (you can squeeze more into the same-volume container), but because hydrogen molecules are so small, when they’re under pressure, they can actually migrate through metals. It’s as if your gas tank slowly leaked gas as you filled up.
Finally, the process of making hydrogen (called electrolysis) also requires various materials (known as electrolyzers) that are quite costly. In California, where cars that run on fuel cells are now available, the cost of hydrogen is equivalent to paying $5.60 a gallon for gasoline. So scientists are experimenting with cheaper materials that could serve as electrolyzers.
Other Innovations
Capturing carbon. We could keep making electricity as we do now, with natural gas and coal, but suck up the carbon dioxide before it hits the atmosphere. That’s called carbon capture and storage, and it involves installing special devices at fossil-fuel plants to absorb emissions. These “point capture” devices have existed for decades, but they’re expensive to buy and operate, they generally capture only 90 percent of the greenhouse gases involved, and power companies don’t gain anything from installing them. So very few are in use. Smart public policies could create incentives to use carbon capture, a subject we’ll return to in chapters 10 and 11.
Earlier, I mentioned a related technology called direct air capture. It involves exactly what the name implies: capturing carbon directly from the air. DAC is more flexible than point capture, because you can do it anywhere. And in all likelihood, it’ll be a crucial part of getting to zero; one study by the National Academy of Sciences found that we’ll need to be removing about 10 billion tons of carbon dioxide a year by mid-century and about 20 billion by the end of the century.
But DAC is a much bigger technical challenge than point capture, thanks to the low concentration of carbon dioxide in the air. When emissions come directly out of a coal plant, they’re highly concentrated, in the range of 10 percent carbon dioxide, but once they’re in the atmosphere, where DAC operates, they disperse widely. Pick one molecule at random out of the atmosphere and the odds that it will be carbon dioxide are just 1 in 2,500.
Companies are working on new materials that are better at absorbing carbon dioxide, which will make both point capture and DAC cheaper and more effective. In addition, today’s approaches to DAC require a lot of energy to trap the greenhouse gases, collect them, and store them safely. There’s no way to do all that work without using some energy; the laws of physics set a minimum amount on how much will be required. But the latest technology uses much more than that minimum, so there’s a lot of room for improvement.
Using less. I used to scoff at the notion that using power more efficiently would make a dent in climate change. My rationale: If you have limited resources to reduce emissions (and we do), then you’d get the biggest impact by moving to zero emissions rather than by spending a lot trying to reduce the demand for energy.
I haven’t abandoned that view entirely, but I did soften it when I realized just how much land it will take to generate lots more electricity from solar and wind. A solar farm needs between 5 and 50 times more land to generate as much electricity as an equivalent coal-powered plant, and a wind farm needs 10 times more than solar. We should do everything we can to increase the odds that we can scale up to 100 percent clean power, and that will be easier if we reduce electricity demand wherever we can. Anything that reduces the scale we need to reach is helpful.
There’s also a related approach called load shifting or demand shifting, which involves using power more consistently throughout the day. If we did it on a large scale, load shifting would represent a pretty big change in the way we think about powering our lives. Right now, we tend to generate power when we use it—for example, cranking up electric plants to run a city’s lights at night. With load shifting, though, we do the opposite: We use more electricity when it’s cheapest to generate.
For example, your water heater might be able to switch on at 4:00 p.m., when power is less in demand, instead of 7:00 p.m. Or you could plug in your electric vehicle when you get home for the day, and it would automatically wait to charge itself until 4:00 a.m., when electricity is cheap because so few people are using it. On an industrial level, energy-intensive processes like treating wastewater and making hydrogen fuels could be done at a time of day when power is easiest to come by.
If load shifting is going to have a significant impact, we’ll need some changes in policy as well as some technological advances. Utility companies will have to update the price of electricity throughout the day to account for shifts in supply and demand, for instance, and your water heater and electric car will have to be smart enough to take advantage of this price information and respond accordingly. And in extreme cases, when electricity is espec
ially hard to come by, we should have the ability to shed demand, meaning we’d ration electricity, prioritize the highest needs (say, hospitals), and shut down nonessential activities.
* * *
—
Keep in mind that although we need to pursue all these ideas, we probably don’t need all of them to pan out in order to decarbonize our power grid. Some of the ideas overlap each other. If we get a breakthrough in cheap hydrogen, for example, we might not need to worry as much about getting a magic battery.
What I can say for certain is that we need a concrete plan to develop new power grids that provide affordable zero-carbon electricity reliably, whenever we need it. If a genie offered me one wish, a single breakthrough in just one activity that drives climate change, I’d pick making electricity: It’s going to play a big role in decarbonizing other parts of the physical economy. I’ll turn to the first of these—how we make things like steel and cement—in the next chapter.
Skip Notes
*1 I’m using the word “power” a bit loosely here. Technically, “power” refers to the rate of flow of electricity, measured in watts. In this book, for the sake of readability, I’ll use the term in its more general sense, as a synonym for “electricity.”
*2 These calculations are drawn from a life cycle assessment of dams. Life cycle assessment is an interesting field that involves documenting all the greenhouse gases that a given product is responsible for, from the time it’s produced until the end of its life. These assessments are a useful way to analyze the climate impact of various technologies, but they’re pretty complicated, so in this book I will focus on direct emissions, which are easier to explain and generally lead to the same conclusions anyway.