Every time someone turns around, there seems to be another startup in Silicon Valley that claims to be on the verge of a so-called “breakthrough” in fusion—with massive amounts of private funding to go along with it. (To be fair, the East Coast has its share of such startups, such as the Boston area’s “Commonwealth Fusion Systems,” which reports raising $1.8 billion in funding to commercialize fusion energy.) Is there substance to an April 23 article in MIT’s Technology Review[1] that said that “record-setting magnets built by the Plasma Science and Fusion Center and Commonwealth Fusion Systems meet the requirements for an economical, compact power plant?”

It’s unclear how much substance there is to these claims of great advances in privately funded efforts at fusion. To make sense of just why there is seemingly so much interest in and money flowing toward private efforts at fusion—at the same time that many physicists and engineers who work in the field seem only guardedly optimistic at best about these multibillion-dollar efforts—I interviewed California-based venture capitalist (and UC-Berkeley professor) Mark Coopersmith via MS Teams.

(Editor’s note: This interview has been condensed and edited for brevity and clarity.)

Dan Drollette Jr: The physicists and engineers I’ve talked to[2] seem to view the likelihood of fusion coming to fruition on a commercial scale anytime soon is … distant. By “commercial,” I mean a fusion system that generates more electricity than it consumes, at affordable cost, where researchers have overcome all the technical challenges of getting enough tritium to fuel the reaction—and solved the problem of how to deal with the embrittled reactor walls caused by the constant bombardment of neutrons.

Mark Coopersmith: It does sound challenging.

Drollette: Having said that, the same physicists and engineers say they’re cautiously upbeat that fusion could eventually find its place—maybe by 2050 or 2080, or even the year 2100. Meanwhile, non-physicists and non-engineers seem much more upbeat about fusion’s prospects. There was a survey in [the peer-reviewed scientific journal] Nature[3]which found that 65 percent of companies predicted a fusion plant would be delivering electricity by the year 2035—or just over a decade from now, instead of 76 years.

Coopersmith: I saw that same quote. And my understanding is it comes from companies currently in the business of attracting capital, talent, customers, and partners—those are the companies that said: “We expect there will be some fusion electricity feeding to the grid by 2035.”

You’ve got to realize that entrepreneurs are an incredibly optimistic group. I’m an entrepreneur myself—and also an investor—and I teach entrepreneurship at UC-Berkeley[4]. And one of the 10 traits I’ve identified in entrepreneurs is optimism. You have to be optimistic as an entrepreneur, to attract those things you need.

So it doesn’t surprise me if 65 percent of the companies that were quoted thought that within 10 or 11 years, there would be some fusion electricity trickling to the grid.

But let’s step back a bit and think about what the road to commercialization looks like for deep tech.

Drollette: Great, that’s what I’m looking for.

Coopersmith: The first step on that road is: Do we understand the science?

Now, I’m not a nuclear scientist. But I see a lot is involved in making fusion technology work—and different groups are taking different approaches.

So part of this is recognizing that there may be different routes to creating commercially viable fusion electricity—and that funding for the basic research to this point has largely come from governments, educational institutions, and research institutions, as it should be.

So that’s the first of the three steps on the road to commercialization: Do we have a handle on the science?

The second step is proof of concept: Can someone show there’s net positive electricity from a fusion reaction? It appears that’s happened in some form, although on a very small, short-term scale. And my understanding is that even then, there’s debate about how you calculate all the energy that went into making it work. But let’s assume researchers are on the cusp of net positive electricity.

Then comes the third step on the road to commercialization, the really hard part: investability.

Drollette: Which means?

Coopersmith: That’s the ability to move from proof of concept—the early, ugly, working prototype that doesn’t scale—into something that can scale and be commercially viable.

And that’s a complicated process in itself. Because it’s got to be able to truly scale-up on a global level. After all, electricity is a huge marketplace—roughly $1.6 trillion worldwide.

If looked at that way, we can break things down and say: “Where does the capital come from to hit those milestones?” And it turns out it comes in three different buckets.

The first bucket I already mentioned: For decades, the funding for the basic research—the core understanding—has come in from the government, universities, endowments, and so forth.

The second bucket is a mix of a lot of public money and a little private money, all used to find out whether researchers can actually convert the R&D into something that works. At that early stage, you just start seeing some capital come in—mostly private/public hybrids that include government grants and things like that. It’s only later that we get specifically private capital.

So looked at that way, it appears there’s some private funding going on, in the past five years or so.

Now, once it gets to the point where people can truly say “Okay, we have something that works,” then it becomes commercially viable in a timeframe that’s reasonable to venture capital. And that’s when you’ll start seeing a whole lot of private money flowing into fusion energy, given the huge size of the marketplace and the opportunity.

Drollette: What stage is it at now?

Coopersmith: We can get a better sense of that by looking at what kind of venture capital is starting to flow in and who’s behind it: Who are the people and companies funding these elements?

It looks like they could be split into three different groups. Number one is corporate venture capital companies like Hitachi, GE, Google, and others like them. The key reason they’re investing in these technologies is because they want an entire sector to advance—that way, they can sell more equipment to new business models.

Intel was famous for doing this via Intel Capital—they invested in hundreds or even thousands of companies, with the idea that doing so would help accelerate the use of more and faster chips.[5]

Drollette: So people would have to swap out their computers more often and buy more Intel chips?

Coopersmith: Right. And when we look at what’s going with corporate venture capital—the names I mentioned—they’re looking at how they can become a player later. It’s not explicitly return-based for the short term.

So, that’s one kind of venture capital that would flow into fusion at this point.

A second kind is traditional venture capital and traditional venture capital funds. There’s a huge amount of venture capital spent annually—about $250 billion of venture capital in the US alone. And the way venture capital funds work is that they each have a finite fund life, typically seven to 10 years.

Some of them are longer—especially in sectors like pharmaceuticals, which take years from basic lab research to something tested and approved and able to be purchased on the shelf—but in general, investors expect to see progress and get a return on their investment within 7 to 10 years. So, if we were to start up a fund today, we’d need to see viability by 2035 at the latest. That means a company would need to launch a first fusion reactor that’s trickling electricity to the grid by then.

Drollette: It must be more than coincidence that 10 years from now works out to the end of 2034 or beginning of 2035. That explains why I keep seeing the year 2035 mentioned for companies predicting first commercial fusion—it’s driven by the needs of their funders?

Coopersmith: And that’s typical. Some may be a little longer, but yes—you don’t want to have the pot of gold so far out it’s not investable.

The economic value of that one first fusion reactor all by itself—the actual physical plant—may not be that great, but the economic value of its parent company might be massive given the size of the bet.

So there will be a few VCs who believe they can make fusion happen in that time frame—and they’ll start to make big bets on it, because the prize is so huge.

And then there’s the third bucket of investors, who have already participated in funding rounds at some fusion startups. These are people like [Microsoft’s] Bill Gates, [Open AI’s] Sam Altman, [Paypal founder] Peter Thiel, and [Amazon’s] Jeff Bezos—they’ve all invested in fusion.

And they have a couple of things going that the rest of us don’t.

First, they have billions of personal dollars on hand—so they can take part of that money out and afford to lose it. But if one of their bets on fusion pans out—even years down the road—it will deliver a 1,000 or a 10,000 percent return on investment, a “super return.”

So they place a few bets. Or lots of bets.

And people like Gates and Bezos are indeed doing this. They’re taking some of their fortunes to fund deep-tech research that may have benefits decades away.

It’s similar to what they’re doing on the charitable side. They’re either giving some of it away or using it for the benefit of humankind; their bottom line is not explicitly delivering a venture-scale-size return in a particular time frame.

Using that kind of logic, fusion is a perfect fit for somebody like Gates or Altman. For that matter, Jeff Bezos has said: “Give me a 10 percent chance on a 100-to-1 payoff, and I’ll take that bet every time.” In fact, I have used that quote in the talks I give.

Fusion is one of those 100-to-1 payoffs. With a super-long time frame, they can afford to make those bets. So those are the three different types of capital, and each thinks really differently.

But to reiterate, in the traditional, core, venture capital category, there needs to be a roadmap to commercial viability and venture-scale returns in a reasonable time frame.

Drollette: So, when I see a headline that says Bill Gates invested in XYZ, that should be taken with a grain of salt? It doesn’t necessarily mean he sees a cutting-edge technology that will bear fruit tomorrow?

Coopersmith: Yes, it needs to be taken in context.

Drollette: So the world of investment capital is not a monolith—there’s different strains to it?

Coopersmith: Let’s deconstruct that for a moment. Venture capital is considered to be highly risky equity: Typically, 75 to 80 percent or more of startups fail. So the remaining 20 percent need to deliver a big return, on the order of 10, 20, or 30 times the investment to make up for the failure rate of all the others. That’s called the “power law of venture capital,” which says that most returns come from a tiny fraction of your investments—typically 80 percent of your returns come from 20 percent of the companies you’ve invested in.

To make that happen, VCs have to invest in lots of companies that they believe each have the potential to deliver a 10-, 20-, or 100-time return.

Things are not quite there yet when it comes to fusion technology companies, because there just aren’t that many companies out there which are backed by proven science.

So there needs to be a few more players before the market takes off.

For comparison, look at the number of companies that are in AI and machine-learning right now: There’s hundreds of them, with billions of dollars invested, because it looks like within a few years, there will be venture-scale returns.

Of course, there will be losers, too. But the ones that win will deliver massive returns.

Drollette: Which reminds me of something you mentioned earlier, that I wanted to circle back to. You said there’s 10 traits common to entrepreneurs?

Coopersmith: Without going through the whole list, I’d say the main things are a bias to action, the embracing of change, and a kind of “break-the-rules” mindset.

Drollette: As in “It’s easier to ask for forgiveness than to ask for permission”?

Coopersmith: Yeah, that kind of … resourcefulness.

Now, out of all those traits that entrepreneurs have, optimism is at the top. And I would hope that the leaders of these fusion companies are telling a story of optimism because—especially with deep tech—it normally takes longer for these new technologies to become commercially viable. There will be early sparks of success, but to bring them to scale is a huge challenge, especially when dealing with brand new science.

Drollette: Okay, so this brings up something that a mutual colleague, Sam West[6]—who had introduced us to each other—had mentioned earlier. As you know, West is a clinical psychologist who specializes in corporate behavior, and he said that there’s really a clash of two different cultures here. Scientists come from a whole different place—they want lots of testing, they want data, and they are cautious. They want the same result to happen many times in a row—what some call a “six sigma” level of confidence[7]—before declaring that X causes Y. It’s a whole different kind of mindset.

Coopersmith: I think that comparison is right. Venture capitalists are used to getting excited by new ideas. They have a long history of waging bets on the untested. They’re comfortable making decisions and investments proactively and preemptively with lots of uncertainty.

And that’s something VCs have in common with entrepreneurs: We’re dealing with new science, new ideas, new markets, and new technologies all the time. But having said that, as a venture capitalist, you do want to have some kind of roadmap that helps take some of that risk off the table along the way.

So, you invest according to where things are on that roadmap: When there’s a huge amount of risk, you invest a smaller amount. As the risk is reduced and the certainty of this technology is increased, you grow closer to commercialization—there’s a working technology, customers buying it, employees being able to deliver, all of that—so then companies become worth more because the risk has been reduced, and the time frame to commercialization is closer. So there’s a natural kind of relationship when you look at risk and the availability of capital, and how capital is deployed over time as valuations of these companies rise.

Drollette: I see you’ve put up a chart on the screen.

Coopersmith: Yes, this shows the amount of annual funding that’s been going into fusion-focused startups from 2019 through last year. The bars represent the amount of dollars that went into companies, and the green line is the number of deals.

So the deals have remained relatively consistent—it’s been relatively flat, with the exception of this big blip in 2021 and a little bit of a bounce in 2023. But the big blip was in 2021.

Now, a couple things happened that year. First, there was a huge amount of venture capital in general being deployed, and perhaps there were a few breakthroughs in some of the science around fusion technology.

And there were two big deals that took place in 2021, one to Commonwealth Fusion Systems—$1.8 billion in funding, a huge amount—and about $500 million to [Sam Altman’s Everett, Washington-based fusion research company] Helion Energy. Now, these are really big, significant deals. But if you take out those two deals—those two companies—then the line pretty much stays the same.

In the rust-colored type-font at the bottom, you can see the average deal size. So the amount of dollars per funding round that have been made since 2019 has stayed in this range of $20- to $50- million, with the exception of those two big ones. Take them out, and 2021 actually had deal sizes on the low end.

Now, this other chart shows what’s been going on in venture capital in general, globally.

And it shows that in 2021, nearly $650 billion was invested in venture capital—a huge amount that is really kind of amazing, considering we were in the midst of a pandemic—though it’s gone down since then. That also happens to be the year of the big spike, with those two large investment rounds in fusion technology.

So, we look at all that and see that while there may have been some science that was proven right around 2021—some early successes—there was this general exuberance to fund a lot of things at that time. And often when those trends converge, you have some massive funding. So that’s really what took place at that point.

Here’s one more slide to look at.

The blue line shows VC deals for fusion technologies going all the way back to 2013. We see that the amount of private capital that went into fusion technologies has stayed relatively consistent. And again, outside of that huge spike in 2021 where we saw some growth, the deal size and funding actually hasn’t changed that much. And I think that’s because the finish line continues to move out, and capital sees that.

Drollette: In other words, the people who have the capital are reading the same things that you and I read in the science press—such as the news that the publicly funded fusion research facility announced a 10-year delay to first plasma. VCs are not blind to that.

Coopersmith: And what typically happens? They go elsewhere.

I mean, there are some funds that invest across all sectors, but many funds have a particular focus on certain sectors or related sectors.

But that’s changeable. I would expect that as soon as there’s more viability and potential venture-scale returns, we’ll see some very large funds spin up around energy—and in particular around conversion to new energy sources. That hasn’t really happened yet, in part because of that time frame I talked about.

Drollette: Because it was astonishing for me to hear the pessimism about commercial fusion[8] from some of the people who were actually working in engineering and physics—and then read something about $1.8 billion being invested in just one fusion startup.

Coopersmith: Well, that’s the one big bet, right? Or, more precisely, one of the two big bets made in 2021. And we haven’t seen other big bets made since then. So, you can say that those two data points aren’t representative. Instead, the trend is relatively gradual and upward as fusion continues to be de-risked.

Drollette: Do you have any reasons why there were such big deals for Commonwealth and for Helion?

Coopersmith: I don’t have any great insights into that.

Drollette: What about energy in general—what did you find out about renewables while you were digging up this information and generating these charts? Are certain areas really taking hold?

Coopersmith: That depends. From an investment standpoint, solar had attracted a lot of capital. I’m not sure how much capital is being deployed in solar right now, in part because the economics of solar—such as the government incentives in the US and globally, or the heavily subsidized cheap solar panels that come in from China—make it difficult to evaluate solar as a profitable business at venture scale. So even as we see capabilities in wind, solar, geothermal, and others growing, their ability to attract capital is still kind of uncertain.

Drollette: We talked about a disparity between the engineers and physicists on the one side, and Silicon Valley on the other. But what about within the valley? When the investment world thinks about fusion, are they still going by the buzzwords of trying to “fail quickly” or “fail forward,” or anything like that?

Coopersmith: No, it’s different. I wrote a book called The Other F Word[9] on bouncing back from failure, and I think that at this point we’re still into the basic science of fusion. I think that there continues to be the need for government funding and support for basic research here. We’re just not seeing venture-type mechanisms coming into this arena at a large scale—not until we understand the science a little bit better.

And that’s where this conversation started. I see that as the role of research institutions, of governments, of the Department of Energy, to continue to push forward with the basic science.

Drollette: I suppose that could change, if the people who are doing this research come across some interesting spin-off technologies along the way. For example, if plasma technology advances greatly, or this research leads to new ways of detecting cancer or treating cancer come out of it, or they come up with what theoretical physicist [and former head of Argonne National Laboratory] Bob Rosner calls “self-healing metals,” then that will quickly change how the market views this area?

Coopersmith: Well, yes, this technology doesn’t exist in a vacuum. The development of the computer processing and data transmission that’s going to be required to keep these fusion reactors operating safely, for example, will yield a lot of benefits.

You mentioned self-healing metals and other side benefits, and that’s absolutely true. A lot of times, there are things that we learn from new discoveries that may not follow a linear path to an application. There may be unanticipated, alternative applications. For example, the GPS that we all use on our phones every day came out of some early technologies designed for space exploration.

So that’s the interesting thing: We don’t know when someone will create something or come across some new science that may have an application that could bear fruit closer in—and some of those might create VC bets as well.

Drollette: One last question just popped into my head. I’ve sometimes watched a comedy show called “Silicon Valley” on HBO. Have you ever seen it?

Coopersmith: Oh, yes.

Drollette: How realistic was it?

Coopersmith: I’d say it’s only one degree away from the truth. People thought it was purely a comedy, but I knew people who really were like that. It was closer to reality than most people realize—scarily close. And hilarious.

Drollette: I read that the series’ creator, Mike Judge, actually had worked for a startup in Silicon Valley as an engineer, decades ago, and that informed how he depicted it. He was having fun with it, and its excesses.

Coopersmith: For some of my classes, I make that show required watching.

Drollette: Seriously?

Coopersmith: Seriously. It was fun and funny—and it completely captured the quirkiness, the excesses, and the eccentricities of some of the players. Nailed it completely.

Endnotes

[1] See April 23, 2024 MIT Technology Review article “MIT’s superconducting magnets are ready for fusion.” https://www.technologyreview.com/2024/04/23/1090425/mits-superconducting-magnets-are-ready-for-fusion/

[2] See interview with Bob Rosner in this same issue of the Bulletin of the Atomic Scientists, November 2024

[3] See July 8, 2024 Nature article “ITER delay: What it means for nuclear fusion.” The subheading says “The world’s biggest fusion-energy experiment is likely to be beaten to its goals by other projects—but the massive reactor still has value, say scientists.”   https://www.nature.com/articles/d41586-024-02247-2#:~:text=In%20a%202023%20survey%2C%2065,by%202040”%2C%20he%20said

[4] See UC-Berkeley “Learn2Launch” https://learn2launch.berkeley.edu/faculty-and-administration

[5] For more on the strategy behind Intel Capital, see June 18, 2019 Forbes magazine article “What Corporate VCs Can Learn From the First Decade of Intel Capital”  https://www.forbes.com/sites/gilpress/2019/06/18/what-corporate-vcs-can-learn-from-the-first-decade-of-intel-capital/

[6] See January 16, 2023 Bulletin article, “Interview with Samuel West, founder of the Museum of Failure” https://thebulletin.org/premium/2023-01/interview-with-samuel-west-founder-of-the-museum-of-failure

[7] See “68-95-99.7 rule” in Wikipedia https://en.wikipedia.org/wiki/68–95–99.7_rule

[8] See MIT Technology Review article of May 10, 2023 titled “This startup says its first fusion plant is five years away. Experts doubt it” https://www.technologyreview.com/2023/05/10/1072812/this-startup-says-its-first-fusion-plant-is-five-years-away-experts-doubt-it/

[9] See Goodreads review at https://www.goodreads.com/book/show/22942547-the-other-f-word