Bill Gates’s Nuclear Miracle? John Gilleland Says TerraPower Needs Discipline, Not Divine Intervention

3/23/10Follow @gthuang

John Gilleland’s first day on the job was a little different from most people’s. The nuclear physicist showed up at Intellectual Ventures in Bellevue, WA, and sat down at the conference table with his new boss, CEO Nathan Myhrvold, and another, shall we say prominent, techie.

“The guy on my left looked familiar,” Gilleland says. “It was Bill Gates.”

Gilleland had been on the job for all of three minutes when Myhrvold said jokingly, “John, you’re late on your deliverables.”

That was back in December 2006. Gilleland is now CEO of TerraPower, the spinoff from Intellectual Ventures that is focused on creating a fundamentally new kind of nuclear reactor. It’s the invention firm’s biggest research project to date, spinning out as a separate entity in the fall of 2008 with 30-some staff and untold amounts of funding from Gates and other investors. It is a project that Intellectual Ventures likes to cite as a potentially transformative, homegrown invention.

The basic idea is to create a reactor that needs only a small amount of enriched uranium to get started, and then uses depleted uranium (spent fuel) or natural, unenriched uranium to produce the nuclear-fission reactions necessary to generate power for 60 years or more without refueling. The design is called a traveling wave reactor, and the idea dates back to the early 1990s. If it works, the key benefits would be cheaper power, much more plentiful fuel, more efficient nuclear waste disposal, and less risk of nuclear proliferation.

Gates has been gushing about the project as of late. He mentioned TerraPower prominently in his talk at the TED conference in California last month, calling out the proposed reactor design as a possible “miracle” innovation in the effort to provide clean energy to more of the world’s population without increasing carbon emissions in the atmosphere. (Nuclear power provides about 20 percent of the electricity in the U.S.)

John Gilleland

Gilleland (see photo, left) has been given the keys to Gates and Myhrvold’s nuclear kingdom for good reason. Previously, he co-founded and led Archimedes Technology Group, which developed improved techniques for cleaning up nuclear weapons waste, among other things. Before that, he was chief scientist and vice president of energy programs at Bechtel, and U.S. managing director of the International Thermonuclear Experimental Reactor (ITER) program for fusion energy, and he spent 16 years at General Atomics doing fusion research.

The traveling wave reactor is certainly an intriguing idea, and one that could be a true breakthrough. But the question, skeptics say, is whether it can be made to really work—and how long that will take. The idea is that the reactor makes its own fuel and uses it as it goes along: the neutrons emitted by a small amount of enriched uranium convert depleted uranium into plutonium, which splits to produce energy and also emits more neutrons that continue to “breed” new fuel. There is no precedent for TerraPower’s particular design, and the project faces some major challenges—technical, business, and regulatory. So far the physics has only been tested in computer simulations, albeit using the most advanced supercomputers available. (It’s worth mentioning that only someone like Gates could afford to fund this and risk having it not work—which is exactly why Myhrvold sees the need for an “invention capital” industry.)

On the plus side, the environment for nuclear power development is more promising than it has been in years. President Obama recently called for a new generation of nuclear plants to be built in the U.S.; they would be the first new ones in 30 years. Companies including General Atomics, General Electric, NuScale Power, and Hyperion Power Generation have burgeoning nuclear efforts in the U.S., as does General Fusion in British Columbia, and Areva, Hitachi, and Toshiba further abroad. (Reports surfaced yesterday that TerraPower and Toshiba are in talks to collaborate on a nuclear reactor, possibly involving elements of Toshiba’s “4S” fast neutron reactor—see more on this type of design below.)

TerraPower will need international partners, and funding on the order of billions of dollars, to succeed. “I am hoping that we could get a reactor built inside of 10 years,” Myhrvold told me in August 2008. “Of course, to have it built in 10 years, we have to start designing it in three years, because it takes a couple years to design it, and then you have to build it. It’s a long process.”

I spoke with Gilleland recently about this process, the milestones his group has achieved, and its realistic prospects for revolutionizing the field of nuclear power. Here are some edited highlights from our conversation:

Xconomy: How did you originally get involved with this nuclear project?

John Gilleland: Eben Frankenberg [executive vice president at Intellectual Ventures] contacted me. They were looking to see whether a startup around nuclear would be viable. I was coming off a job with Archimedes, and had sold that company. I came up [to Bellevue] with the idea of telling them they were off-base and steering them away from the endeavor. But I never left.

X: What was the original thinking at Intellectual Ventures around nuclear power?

JG: They wanted to raise the energy standard around the world. It’s great if you can supply per capita energy levels to allow people to rise above poverty. It helps with disease. They looked around at renewables and all sorts of sources, and determined the best bet would be through nuclear power, along with the other systems. It was a necessary element.

A conclusion I came to independently was that there were areas for tremendous improvement [in nuclear power]. Modern plants are very safe, but things can be improved. We revisited ideas of the previous century with new data and new computing power. Edward Teller and Lowell Wood worked together in the ‘90s on these ideas. What is a superior system? One that has an incredibly abundant fuel supply so it’s accessible to everyone, but is safe against accidents and proliferation—that’s a key problem about nuclear, but it can be overcome. It would be wonderful to have a system that didn’t in the long run require enrichment plants, reprocessing plants. When we talked to proliferation experts at various institutes, [they said] it would be an incredible reduction in the prospect of weapons. That’s one of the constructs that I, and others, hold on to.

X: So if things go really well, you could have a version of this thing built by 2020? That’s still a ways away.

JG: In nuclear terms, that’s speed of light. But for Nathan and Bill, you should have seen them—10 years?! For them [coming from software], six months is normal. Fortunately, they’re very patient men.

X: Bill Gates is a very vocal supporter of TerraPower. How directly is he involved in the company?

JG: I get e-mails and questions from Bill on a monthly basis. Our quarterly updates last between one and 12 hours. We also have intermediate meetings, and take trips around the world to look at plants. He asks penetrating questions about the neutronics calculations [for instance], how you do the program mapping to follow the daughter products. He came in once with a 10-inch-thick book labeled “nuclear power.” It’s a nontrivial amount of time he’s spent—long hours and hard questioning.

X: What kinds of specific questions and feedback has Gates provided?

JG: He will remind us that the economics of the thing must be there. It must be competitive economically. Even if it is much lower proliferation risk, or there’s fuel forever, if you can’t afford it, it probably won’t happen. The safest position is for it to be less expensive than any other nuclear process, and less than or equal to natural coal. We have a time constraint: raising the standard of living is a key to wellness, but the climate change timescale and the time it takes to change the energy infrastructure are of comparable scale. But in that context, you must find something that is affordable. [Gates] reminds us of how the world works every once in a while, which is good for us.

X: So how do you actually make this thing operational by 2020?

JG: We have to find a place to build a prototype. We are discussing this with various institutions. We need to build a reactor in [the range of] a few hundred megawatts electric. It needs to be that large to demonstrate this reactor can live on depleted uranium fuel, that this wave action in the core exists and we understand it all correctly. The remarkable thing we found out is that the technologies basically exist to put this reactor core in the Fast Flux Test Facility in eastern Washington, and in Idaho. France, China, India, Russia, and Japan have built [fast neutron] reactors of this type—it was that type of reactor that can uniquely support running this new kind of [traveling wave] reactor core.

There are problems. We have to discover which metal is the best one to clad the fuel and have structures inside the reactor. Fundamental measurements have to be made along the way to optimize the reactor. There’s no doubt the reactor will work, but we don’t know until we’ve done more R&D.

X: What lessons do you bring with you from your time at Archimedes, Bechtel, and ITER, in terms of leadership?

JG: There are different phases. Whether its fusion or new renewables, when it’s a new project, you’ve erased the grease board. You have to bring a sense of vision and behave more like a movie director than a manager. You say, this is what we want, and you let them perform. The art is, when do you bring in discipline? Now we have to stop and build [a reactor], and change the nature of the organization. This is where you have the lighting manager in the plan, and then have a schedule. Engineers need to meet the schedule. This is where perfect is the enemy of the very good.

As I learned at ITER, it’s also culturally based. People are born with similar brains, but they learn a lot about how to think, and it will vary from culture to culture. Running a U.S.-Japan collaboration, you could sense the way the Americans, Japanese, Russians, and others would approach the problem. The Japanese were prone to go all the way to the result in consensus, and then come back and consider what to do. The U.S. would start down a beautiful road in the landscape they had planned, but then they’d see beautiful flowers on the left and they’d take an immediate turn and explore the new flowers. You have to do it just right.

At TerraPower, we’re pretty much a U.S. organization. We have different professors with different personalities and priorities, contracts with national labs and businesses, all with different attitudes and views. We’re in the transition between the initial creative phase and really nailing down what we’ll build by 2020.

X: What is the biggest remaining challenge?

JG: To see it through institutionally, we need to make sure we have the patience to push through the development all the way to operations. Then there’s the technical challenge. Some of the testing we need to do needs to be done in other countries. We don’t have a fast [neutron] reactor operating in this country. Most energy technologies benefit from superior materials—radiation resistance, strength, ability to take temperature. The best way to learn how to do something is to build one. I would like to see the U.S. build a fast reactor to enhance our ability to study the materials. We’re absolutely thrilled there’s an embodiment [a big enough core in fast reactors] that looks just like what we need to build. But we have to optimize it. That’s our technical challenge in creating a new path to fission power by 2020. We need to build a machine that looks the same but has some differences in the size of the vessel and so forth. That’s why I’m still around after three and a half years.

X: Are there lessons from the recent failure of the Pebble Bed Modular Reactor in South Africa (which had been touted as a nuclear silver bullet and was nearing construction)?

JG: Yes and no. The technology and goal are different. The lesson is you have to decide you’re going to follow through and you’ve got something different about it. I’m not an insider on why the gas reactor has come and gone. At the time of Pebble Bed, the world wasn’t building reactors all over the place. That’s relatively new. The Chinese, I believe, have a variant on it, at Tsinghua University, and they’re beginning to revive it.

X: If you could ask the God of Physics one question, what would it be?

JG: What is dark energy? [In astrophysics, this is the mysterious stuff that seems to be responsible for the accelerating expansion of the universe---Eds.]

X: So you must be pretty confident that TerraPower is going to work. You don’t need the God of Physics for that?

JG: I think he’s given us enough information, and we have to be very clever to work out his puzzle. We’re there but for some knowledge about some particular piece of metal [for instance]. I’ll turn the question around and say, we ought to thank him because he gave us the toolkit and the data—the physics is well understood. This project will just take a lot of discipline, not divine intervention.

Gregory T. Huang is Xconomy's Deputy Editor, National IT Editor, and the Editor of Xconomy Boston. You can e-mail him at or call him at 617-252-7323. Follow @gthuang

By posting a comment, you agree to our terms and conditions.