U.S. Markets closed

Nuclear Fusion Could Rescue the Planet From Climate Catastrophe

Jon Asmundsson and Will Wade

(Bloomberg Markets) -- About two dozen private companies around the world are working to harness a transformative energy technology that could rescue the planet from climate catastrophe. One is using space in an old factory that’s home to a mothballed U.S. Department of Energy-funded research machine in Cambridge, Mass. Another is housed in an industrial building behind a Costco outside Vancouver. A third is down the street from a self-storage facility in the foothills of Orange County, Calif.

The companies are working on commercializing fusion.

Fusion’s promise is huge. It would be the most energy-dense form of power: A liter of fusion fuel is equivalent to 55,000 barrels of oil. In its most common form, that fuel would come from a practically inexhaustible source: water. In fact, 2 cubic kilometers of seawater could in theory provide energy equivalent to all the oil reserves on Earth. “It’s ubiquitous, inherently safe, zero-carbon energy—at a scale that can fuel the planet,” says Matt Miller, president of Stellar Energy Foundation Inc., a nonprofit that promotes the development of fusion power. “Now that’s worth working on.”

It was only about 100 years ago that people came to understand that fusion was the process powering the sun. Shortly thereafter, scientists began trying to re-create it. From tabletop experiments, fusion quickly developed into Big Science. Since 1953 the U.S. government has devoted more than $30 billion to fusion research, including basic science and weapons-related work, according to data from Fusion Power Associates, another nonprofit. European countries, Russia, China, and Japan have also made huge investments in pursuit of the holy grail of energy.

Since the 1950s, however, expectations that researchers were on the verge of breakthroughs have repeatedly come up short. What’s different now is that advances in technology are bringing fusion within reach.

Turning theory into practical devices is being enabled by advances in supercomputing and complex modeling, says Steven Cowley, director of the Princeton Plasma Physics Laboratory and former head of the U.K. Atomic Energy Authority. Fusion used to be defined as “the perfect way to make energy except for one thing: We don’t know how to do it,” Cowley says. “But we do.”


So what is fusion again? The idea is deceptively simple: Smash two atoms together so they fuse into a single heavier element and release energy. It’s the opposite of fission, the process used in today’s nuclear power plants and the bombs dropped on Hiroshima and Nagasaki.

QuickTake: Nuclear Fusion

In fission, a large, unstable nucleus is split into smaller elements, releasing energy. Fusion, by contrast, starts with light atoms. Take two hydrogen nuclei, for example. Ordinarily, their positive charges repel each other. But apply enough heat and pressure, and they might get close enough for the attraction of the extremely short-range but powerful nuclear force to kick in, joining them into a single helium nucleus. When that happens, the mass of the newly formed nucleus ends up slightly less than the sum of the two hydrogen nuclei. And that difference in mass gets released as energy, in accordance with Albert Einstein’s famous equation E=mc2. Simple. Stars do it. The sun does it. It’s the basic energy process of the universe.

Early efforts to harness it, though, gave fusion a reputation for hype and disappointment. After World War II, an Austrian scientist who’d worked in Germany ended up in Argentina, where he persuaded dictator Juan Perón to fund his fusion experiments. On an island in a remote Andean lake, the scientist, Ronald Richter, set up an elaborate facility. In February 1951, he detected what appeared to be heat from a thermonuclear reaction in his reactor. The next month, Perón announced at a press conference that Argentina had harnessed the atom to create unlimited energy. A subsequent investigation found that a glitch in Richter’s instruments led to his mistaken heat reading. Richter was discredited.

Long Road

Fusion’s history is studded with disappointments as well as advances

1920British astronomer Arthur Eddington’s “The Internal Constitution of the Stars” posits that stars including the sun are powered by the fusion of hydrogen.

1938Nuclear physicist Hans Bethe describes the fusion reactions that create the energy emitted by stars, for which he later wins the Nobel Prize.

1951Juan Perón (far right) and scientist Ronald Richter (second from right) announce that Argentina has developed fusion energy.

1952The first test of a hydrogen bomb, code-named Ivy Mike, uses a fission explosion to ignite a fusion reaction in deuterium fuel. The 10-megaton blast leaves a big crater on Enewetak atoll.

1958ZETA excitement and disappointment as U.K. researchers announce they’ve likely created a controlled fusion reaction, but later retract.

1964A fusion demonstration at Progressland at the World’s Fair in New York.

1969In an example of cooperation, the U.K. brings laser equipment to the Soviet Union to measure the temperature of the T-3 tokamak, confirming 10 million C plasma.

1982Tokamak Fusion Test Reactor, or TFTR, starts at the Princeton Plasma Physics Laboratory. It sets a record plasma temperature of 510 million C.

1985The Soviet Union proposes international collaboration on fusion at the Geneva summit of Mikhail Gorbachev and Ronald Reagan, which leads to the start of ITER.

1989Chemists Martin Fleischmann and Stanley Pons’s cold fusion experiment can’t be replicated.

1997The Joint European Torus, or JET, sets a record with a fusion output of 16.1 megawatts, equivalent to about 67% of the input energy, a Q of 0.67.

2019Construction of ITER, an international fusion demonstration project, in the south of France is 60% complete. When turned on, ITER is expected to produce 10 times the energy it consumes, a Q of 10.


Many physicists were skeptical of the initial report, but news of the apparent breakthrough spurred research in the U.S., the U.K., and the Soviet Union. At Princeton, a top-secret U.S. government project aimed at working on the H-bomb started researching fusion technology. In 1951 scientists there began developing a device called a stellarator that would use magnetic fields to confine superheated plasma. The effort, code-named Project Matterhorn, was eventually declassified and became the Princeton Plasma Physics Laboratory.

In the U.K., work on a machine called Zeta, which “pinched” fusion fuel by running a huge current through it, led to another premature announcement of the dawn of the fusion age, in 1958. It turned out that strange instabilities in the fuel were what led researchers to mistakenly think they were seeing evidence of fusion.

The Argentine news also fast-tracked work on an idea developed by Soviet physicist Andrei Sakharov, a dissident and Nobel Peace Prize winner: confining fusion fuel in a doughnut-shaped configuration with a machine called a tokamak.

Since the 1960s, when government labs and universities around the world began constructing tokamaks in earnest, more than 200 working machines have been built. A key sign of progress in the fusion field is the chart of the so-called triple product, a measure of reactor performance. Plot this number—how hot, how dense, and how well-insulated the systems are—against a timeline, and it looks a lot like Moore’s law, the famous doubling of computing power every two years. But fusion’s improvement is even faster. “Tokamaks have beat Moore’s law,” says Bob Mumgaard, chief executive officer of Commonwealth Fusion Systems, which was spun out of MIT.

So why does it matter how hot a fusion system gets? Consider the sun. Our local star has a lot of plus-size gravity to apply to the fusion process. Its interior brings the pressure of a mass equivalent to about 333,000 Earths and a temperature of about 15 million C (27 million F). That’s the kind of forge in which fusion happens.

On Earth, with so much less gravity, you need higher temperatures: 100 million C, for example. So the first step to get there is to heat a gas and turn it into a plasma, says Michl Binderbauer, CEO of TAE Technologies Inc., based in Foothill Ranch, Calif. “That happens through adding more energy, so at some point the ions and electrons that make up the atoms fall apart into a soup of charges,” he says. “That’s the state that actually most of the universe is in—what we call a plasma.”

Almost all of the visible stuff in the universe is plasma. “We’re living probably in one of the few specks of the universe where there’s no plasma in our immediate surroundings other than lightning or something,” Binderbauer explains. What’s more, in the 1950s, when instabilities and other “funky behavior” in plasma turned out to make fusion much harder than expected, Mumgaard says, it led to the development of an entire discipline, plasma physics. The field has in turn contributed advances in medicine and in manufacturing semiconductors.

Now, heating plasma to 100 million C sounds daunting and terrifying. Wouldn’t it vaporize whatever it touches? Short answer: no. The plasma is a handful of particles in a vacuum chamber, Binderbauer says. It’s millions of times less dense than air, its state is extremely fragile, and if it touches anything it instantly cools down. TAE’s Norman machine heats plasma to 35 million degrees, says Binderbauer. If, hypothetically, he could stick his hand into the vacuum shell, he says the plasma wouldn’t burn him. “My arm will absorb all of the energy,” he says. “I won’t even turn very warm.” Fusion, unlike fission, has no risk of meltdown. “You have to protect the plasma from the surrounding environment, not the other way around,” he says.

Fusion would have one other important benefit over solar, wind, and other intermittent sources of renewable energy, says Christofer Mowry, CEO of General Fusion Inc., based in Burnaby, B.C., near Vancouver: It’s “dispatchable” power. In most of the applications anticipated for fusion, the energy created in a reaction would heat water and run a conventional steam turbine generator. Plants could be safely and conveniently situated in cities and other places power is needed, Mowry says.

One obvious downside to fusion, reflected in the field’s 70 years of history and dashed hopes for imminent breakthroughs: It’s extraordinarily difficult to bring off.

In 1983 the late Lawrence Lidsky, an associate director of what was then called MIT’s Plasma Fusion Center, wrote an article titled “The Trouble With Fusion.” Fusion, he wrote, “is a textbook example of a good problem for both scientists and engineers. Many regard it as the hardest scientific and technical problem ever tackled, yet it is nonetheless yielding to our efforts.” Still, Lidsky laid out a laundry list of problems that, he contended, made it unlikely that fusion would ever be an economically viable source of power.

More than three decades later, the problems Lidsky identified remain. Chief among them is radioactivity. To be sure, the fuel used in fusion doesn’t pose quite the same dangers as fission’s uranium and nuclear waste. To understand fusion’s radioactivity challenge requires a slightly deeper dive into the science.

To begin, a variety of different light elements can be combined in a fusion reaction. However, the fuel that’s easiest to fuse is a 50-50 combination of two isotopes of hydrogen: deuterium and tritium. D-T, as it’s called, has been the main focus of the field. Deuterium is heavy hydrogen, the stuff found in seawater. Its nucleus consists of a proton plus a neutron (in contrast to plain old hydrogen’s lonely proton). Tritium is heavy, heavy hydrogen: a proton with two neutrons. It’s radioactive, with a half-life of about 12 years. It’s also extremely rare and expensive, but it would be bred in fusion reactors.

When deuterium and tritium nuclei fuse, energy gets released as an alpha particle (a helium nucleus, which is two protons and two neutrons) and a very energetic neutron. Those neutrons are neutral, unconfined by the magnetic field holding the plasma. They crash into whatever material is facing them, which in tokamaks, for example, is called the first wall. The crash transfers heat and also knocks the atoms in the wall’s material out of place, damaging it and making it radioactive.

Daniel Jassby, a retired researcher from the Princeton Plasma Physics Lab, says the incessant barrage of neutrons from burning D-T will create a lot of radioactive waste. Replacing weakened first-wall structures will drive up costs, he says, because of the expense of installing the new components as well as the downtime in which the system won’t be selling power. What’s more, the size of the machines means fusion reactors may produce as much as 10 times more waste than conventional fission reactors, he says. And while the levels of radiation may not be as intense as those of spent uranium fuel rods, that just means the byproducts of fusion systems are dangerous for a century instead of millennia.

The true operating costs for fusion reactors may not be low enough to cover their costs, let alone compete with existing power plants, according to Jassby. “Why would anybody want this?”


Nevertheless, a certain strain of utopian idealism has always run through the fusion endeavor. It may be what prompted the 1985 agreement between U.S. President Ronald Reagan and the Soviet Union’s Mikhail Gorbachev to cooperate on building a fusion energy project. Now known as ITER, the giant, long-delayed, 35-nation cooperative project is under construction—and about 60% complete—in the south of France.

When ITER achieves its first plasma, which is slated for 2025, it’s expected to hit a fusion milestone: It will produce more energy than it consumes. “There’s nobody knowledgeable in the space who doesn’t believe when they turn ITER on that it’s going to produce net energy out,” says General Fusion’s Mowry. ITER is expected to produce 500 megawatts while consuming 50. In the parlance of the field, it will have a Q>1. Specifically, since it’s expected to produce 10 times the energy put in, it would have a Q=10.

In the plasma physics community, there’s no question that fusion is viable. Now these startups are aiming to build a working—and profitable—fusion power plant, Mowry says. “Private fusion ventures are not going to work on fundamental plasma physics and fusion science,” he says. “They sit on top of that half a century of hard-won knowledge, and they’re all about commercialization.”

Here’s a snapshot of three such companies:

Commonwealth Fusion Systems, Cambridge, Mass.

TECHNOLOGY: Developing high-temperature superconducting magnets to confine plasma in a small tokamak called Sparc.FUNDING: $115 millionINVESTORS: ENI, Breakthrough Energy Ventures*, Future Ventures, Khosla Ventures, and others(* Michael Bloomberg, founder and majority owner of Bloomberg LP, which owns Bloomberg Markets, is a member of the Breakthrough Energy Coalition)

Commonwealth Fusion Systems, which was launched by professors from MIT’s Plasma Physics and Fusion Center in 2018, is looking for space. For the time being, CFS and MIT design and technical teams are working in what used to be the control room for Alcator C-Mod, an Energy Department-funded experimental tokamak on MIT’s campus. The machine, which sits in a large bay two doors away, ran a so-called high field using especially powerful magnets and set a record for plasma pressure.

CFS is seeking to make the next advance in magnetic confinement using new, commercially available high-temperature superconductors. The discovery of such materials was an advance that won the Nobel Prize in Physics in 1987.

Before high-temperature superconductors became available in the past decade, tokamak builders faced a trade-off: use a lot of power to run a high magnetic field or run a lower magnetic field in a much bigger device, like ITER, says Mumgaard of CFS. The new superconductors will enable the company to build a smaller, cheaper version of an ITER-like machine. “Two years from now, we will have that magnet done,” he says.


CFS’s subsequent step will be to build a demonstration machine called Sparc that will use the new magnet technology. Sparc will be about 12 feet tall and could fit into half a tennis court. Construction is supposed to start in 2021 and finish in 2025. A commercial version, called Arc, is expected to follow. It would be approximately twice as big, fitting into a basketball court.

CFS’s tokamak will burn D-T fuel, which means it will confront the first-wall problem. The solution, Mumgaard says, is “to build a machine so you can replace the wall very easily.” Replace it often enough, he says, and it wouldn’t get very radioactive and could be stored and then recycled. “You can choose what you put around the machine,” he says. “Right now we can go with the stuff that’s cheap and easy. And yeah, it’s activated. But in the future we can put in stuff that lasts longer.” One potential solution would be using specialized alloys that are more resistant to becoming radioactive, though the industry is still working to develop such materials.

The radioactive material from fusion reactors is drastically different from fission waste, Mumgaard adds. “It’s basically not stuff that’s biologically active,” he says, unlike the volatile gases that can escape in a fission accident. “So it’s like a completely different category. Whether or not we can explain that well to the public, you know, is one of the challenges that we have to figure out in fusion.”

Still, Mumgaard is upbeat. “Fusion is a big endeavor, and there’s a lot of excitement around it,” he says, adding that enthusiasm is coming from energy people, investors, and academics. “We’re trying to birth an industry here. And it’s a fun place to be.”

General Fusion, Burnaby, B.C.

TECHNOLOGY: Developing  magnetized-target fusion machine in which plasma is injected into a cavity surrounded by swirling molten metal and then compressed by synchronized pistons to create fusion.FUNDING: More than $100 millionINVESTORS: Bezos Expeditions, Chrysalix Venture Capital, Khazanah Nasional, and others

General Fusion, outside Vancouver, is taking a different approach to building a reactor. Founded in 2002 by plasma physicist Michel Laberge, the company dusted off a 1970s design by the U.S. Naval Research Laboratory. Called Linus, the design included features that inspired General Fusion’s concept. “It’s basically the fusion equivalent of a diesel engine,” Mowry says. General Fusion’s machine addresses the first-wall problem by facing the plasma with swirling molten lead and lithium, which absorbs the neutrons. “We inject the plasma into a spherical cavity made out of liquid metal, and then we have basically an array of lots of pistons that are synchronized to collapse that cavity down very quickly around the plasma, heating it up until it burns—just like the analogy of a diesel engine,” he says.

Today’s high-speed electronic controls made it possible to synchronize the pistons with a precision that was impossible in the 1970s, according to Mowry. “That’s an example of what we call enabling technologies,” he says. The company is getting ready to build a scale model demonstration that it aims to complete in 2025.

“Fusion’s time is really coming now,” Mowry says. Before joining General Fusion, he worked in the energy industry for 30 years, including founding a company that designed so-called small modular reactors for fission energy. Now, he says, fusion is becoming competitive with fission. “When you look at the realistic time frames for commercializing advanced gen-four fission technologies, it’s no shorter than that time frame to commercialize fusion these days,” he says.

TAE Technologies Inc., Foothill Ranch, Calif.

TECHNOLOGY: Developing beam-driven field-reversed configuration machine, which fires two plasmas into each other in a confinement vessel so that their magnetic field holds them while heated by particle beams.FUNDING: More than $600 millionINVESTORS: Goldman Sachs Group, Vulcan Capital, Venrock, and others

TAE Technologies, started in 1998, is the oldest company in the field. The late plasma physicist Norman Rostoker, who co-founded the company, took a long view, CEO Binderbauer says. Early on, Rostoker asked what fuel would be most likely to enable a viable fusion power plant—instead of what would be the easiest. He chose hydrogen and an isotope of boron, known as boron-11, because they produce no radiation during fusion and are readily available.

The catch? You have to cook the boron-11 fuel at temperatures of billions of degrees. So that’s the path TAE is taking. Such temperatures have already been achieved in particle physics experiments, according to Binderbauer. “When we talk about temperature, what it really is, it’s sort of how fast and with what energy are these particles zipping around and colliding with each other,” he says. Consider the Large Hadron Collider near Geneva and convert the experiments there into temperature units, Binderbauer says. “CERN actually has created trillions-of-degrees situations where an operator will control it, put these particles into these storage rings, and they run around there,” he says.

TAE’s current machine, which accelerates two plasmas into each other in a confinement vessel and heats them with particle beams, is called Norman. It operates in the neighborhood of 35 million C. The company’s next device, called Copernicus, is aiming for 100 million C.

Like other moonshots, the effort to harness fusion has been both inspiring and frustrating. The finish line may still be years away, but breakthroughs along the way have been sufficient to keep attracting scientists—and, more recently, investors.

And fusion could have an important place in the future energy mix. “The statistics will tell you in the next 25 years we’re going to double the amount of electrical demand and consumption,” Binderbauer says. “To me, finding baseload power that is decoupled from having to burn fossil fuels is very, very critical.”

The potential market is enormous, requiring an investment of $10 trillion or more in generating equipment by 2050. “You can build multiple very-high-value companies in a market like that,” he says. “And we will never even step on each other’s toes.”

Asmundsson is Go editor of Bloomberg Markets and Wade covers energy for Bloomberg News in New York.



To contact the authors of this story: Jon Asmundsson in New York at jasmundsson@bloomberg.netWill Wade in New York at wwade4@bloomberg.net

To contact the editor responsible for this story: Christine Harper at charper@bloomberg.net

For more articles like this, please visit us at bloomberg.com

©2019 Bloomberg L.P.