Transcript: The Big — But Fixable! — Problems in America’s R&D Ecosystem

This transcript has been automatically generated.

Subscribe to “Shift Key” and find this episode on Apple Podcasts, Spotify, Amazon, or wherever you get your podcasts.

You can also add the show’s RSS feed to your podcast app to follow us directly.

Robinson Meyer:

Hello, it’s Thursday, August 6, and after a very slow 2025, there’s been a pickup in climate tech funding lately, at least for late-stage companies. Base Power, the virtual power plant and home battery startup run by Zach Dell of the Dell family, raised a billion dollars this week, which puts its valuation at $13 billion.

Robinson Meyer:

Last week, Commonwealth Fusion Systems raised a billion dollars as well from a new set of largely institutional investors, which signals a lot of confidence in their technology. And earlier this summer, there was a spate of smaller announcements. So Endurance Energy, for instance, raised more than $50 million to build underwater geothermal plants. So what’s happening here? Is climate tech back as a sector? Is it 2020 again? Or are we seeing something a little different? Is AI perhaps triggering like an electricity tech startup boom? Well, I wanted to discuss these questions with an investor who’s in more than just climate tech, but who has been in climate tech for a long time. Our guest today is Katie Rae. Katie is the CEO and managing partner of Engine Ventures, a very interesting organization which spun out of MIT a few years ago. You’re going to hear more about it in a second. She’s a board member for a number of, I would say, climate tech companies. She would say Engine Ventures portfolio companies, including Boston Metal, Commonwealth Fusion Systems, Form Energy, and Teradar. On this episode, we talk about quantum computing, whether fusion is actually going to happen this time and what it would look like if it did, what the AI boom has meant for hardware innovation in the U.S., what’s missing from the U.S. Industrial innovation ecosystem, especially as compared to China, what John Arnold gets wrong and right about the Chinese economy, and finally, whether AI is a bubble. I’m Robinson Meyer, the founding executive editor of Heatmap News, and it’s all coming up on Shift Key.

Robinson Meyer:

Katie Rae, welcome to Shift Key.

Katie Rae:

Thank you so much. Really nice to be here.

Robinson Meyer:

Can you just tell us a little bit about The Engine?

Katie Rae:

So first of all, The Engine was founded 2016 as an entity that did two really important but distinct things. One is a venture capital fund. What we call tough tech startups, really translating out of labs. You know, we started from MIT, so principally MIT, and that has expanded over the years, translating things that could have enormous societal impact and enormous economic impact, because without the other, it probably wouldn’t work, right? You need both. And then ecosystem development play. Which is like, how do you create density of entrepreneurship ideas so that you get the very best and everybody learns from each other the fastest how to grow these companies? So that was the original The Engine. Now it is two pieces. One is the Engine ventures, which I run, and the other is The Engine, which is an accelerator. It’s space, it’s services, and ecosystem development. So it’s two distinct plays now.

Robinson Meyer:

So this category of technologies, you guys call it tough tech. Sometimes you hear deep tech. What is different about investing in it from, let’s say, the classic VC playbook? And how is The Engine Ventures like built for that? What do you even kind of consider under tough tech, let’s say?

Katie Rae:

Yeah, such a great question. First of all, this is the original form of venture capital, right? Which is that you are investing into things that have substantial intellectual property that took probably a long time to develop in academic or national labs and that are true breakthroughs. So a substantial leap forward in how we know things to work, which then can develop entire industries. Right. This is a very pure sense of venture capital. Over the years, you, in the generation you come from, think of it as software investing, where there’s very little technical risk and it’s all market risk. So what tough tech is, is that you are taking on, I wouldn’t say in general scientific risk, but you are taking on the engineering risk of building something of physical instantiation. And that kind of risk is very different than software marketing risk. But it doesn’t mean it’s different forms of venture capital. One is taking the risk that a thousand other startups could do exactly the same thing and you’re going to be the one to break through and win. And the kind of venture capital in tough tech is that you really do understand the engineering and how to scale it up and how to get it to market and partner. But it is that engineering risk phase that we’ll call tough tech. And that is more capital intensive in the beginning because it’s physical. But maybe not more capital intensive over time, but they’re just different forms of venture risk. One is you’re competing with a thousand people. And the other is there are probably one or two teams in the world with the knowledge base and ability to build this thing. And so your likelihood to win if you can gather the capital and the people and the expertise is actually pretty high, but very few people could make that happen. So that’s why a place like MIT, where you have this incredible group of humans who know how to collaborate globally on the cutting edge, is an incredible place to start something like The Engine Ventures, where you have a higher likelihood of being able to form these teams and get them out to market, and therefore a higher likelihood to win.

Robinson Meyer:

Are the returns better in tough tech? Because this is why VC crowded into software, right, is that the returns were amazing. And so it was like lower risk than original VC. And then you had these incredible, huge returns on a relatively capital light formation. Do you think that eventually tough tech kind of generates these software size returns or are the reasons that say in MIT or the country overall should be in this this category of more capital and intensive companies, let’s say less about eventual investment return and more about societal return?

Katie Rae:

Oh, is that the trillion dollar question? Yes, it is. So this is a complex answer, right? So everybody, and I hear this all the time from people, oh, software returns are so much better. Well, if you go look at the average return of a venture fund, it’s negative, whether you’re in software or hardware. It’s not that just because you bet on software, you win. It’s because you bet on the best entrepreneurs in the biggest markets that can move the fastest to win. And that can happen in software and in tough tech. The difference is there are probably a thousand x the funds in software that there are in tough tech so you you play the odds right and so it is always about what is your network how can you help these companies how can you give them an advantage and how do you get the best ones that put you at an advantage to have an economically viable return in venture by the way there is luck in all these say.

Robinson Meyer:

Let’s get to examples. So what are the companies, let’s say, in the Engine Ventures portfolio right now that you’re most excited about, particularly along the lines of energy and clean energy? Because this is nominally a clean energy podcast, although we wind up talking about industrial development and industrial organization as much as we talk about anything else.

Katie Rae:

Well, see, we invest across energy, advanced systems, and I would say convergence biology.

Robinson Meyer:

What is advanced systems?

Katie Rae:

All of those advanced systems, you would think about it as like in the category would be semiconductors, quantum, AI, built environment. These are systems that are at the kind of base of how anything is built.

Robinson Meyer:

Kind of general purpose technology with a solid state spice mixed in.

Katie Rae:

Yes. I’ve never described it that way, but sure. So in tough tech, you know, you are looking on the energy side at things that are new energy development, that are much more dense forms, cleaner forms, sustainable forms, things that could be built anywhere. Where we look at that, we look at carbon transition, like how do you go from heavily carbon to lower carbon sources of creating anything, whether it’s steel, cement, you name it. But then you start to bleed into advanced systems, and then it’s anything that gives you more compute for less energy, or more compute period, more transmission period. And then on the biological side, and this will get into the energy side as well. It’s like, what are the biological systems that play with energy or materials or food or health? You know, those are the areas we work in. Now, you just asked me to name my favorite children. That’s something I don’t do because I don’t think it’s effective. But I will say across all three of those zones, there are very, very interesting plays. I think because this is an energy oriented podcast, we should talk about things like fusion or we should talk about things like, you know, long term storage. I think those are very accessible answers and interesting areas to talk about.

Robinson Meyer:

Let’s jump in. You’ve anticipated my next question. So let’s talk about fusion. I think The Engine is in Commonwealth Fusion, which is the big fusion company at the moment. Talk about your investment, but also like how do you read that sector at the moment? Because we seem to be cresting another wave of excitement about fusion. We have done that before. I’m inclined to be skeptical about it. But at the same time, we crested multiple waves of, let’s say, excitement and hype about artificial intelligence. And then one of them actually broke on the beach. We’re living in the aftermath of it. So where do we stand on fusion right now?

Katie Rae:

Okay, so I think if you look from my perspective, AI, fusion, there are a lot of these kind of frontier technologies that actually follow a fairly systematic path, kind of like Moore’s Law in computing, right? They get incrementally better and make leaps over time. And eventually they leap to the point where they make sense commercially and then have a giant impact on whatever in artificial intelligence or in energy. Then all of a sudden the world changes. You’re living through that with AI right now. Like we’ve been talking about AI for a long time and all of a sudden it’s now applied. It makes sense. it’s making productivity gains, and then the excitement goes wild. Okay, awesome. Fusion has followed that path for 50 years of kind of Moore’s Law, in fact, developing faster than Moore’s Law. And if you look at all of the systematic research in fusion, you will see that we are getting to this point of net positive energy. And it’s really sustained net positive energy is the holy grail. And the reason I’m excited and the reason I invest in Commonwealth Fusion is that we’ve built more than 300 tokamaks, which is the way they’re capturing, you know, the energy of the sun. This is essentially a leap forward because of a materials change, which is how these things happen. Therefore, we think we will get to net positive energy. And we are right on the precipice of this. You know, we invested in 2018, and they have made very systematic progress to that. If you haven’t come to see Spark, you should come see Spark, because you’ll see that this is an engineering feat that is well along its path. And I think when you get to that point, you get to the point of then scaling these things, and you get to the point of improving them very, very rapidly, and then they make a ton of economic sense. So we are sitting in that zone. But in tough tech, if we didn’t invest in 2018, you can’t make that progress. It’s not like the progress just comes out of thin air. It comes out of real engineers doing real improvements across an entire system to get to the point of net positive energy. And that’s why fusion is at a great moment.

Robinson Meyer:

We hear that fusion’s right on the precipice of net positive energy. Can you put that in AI equivalent terms? Is that like inventing the transformer? Is that like, which happened in the late 20-teens at Google, is that like chat GPT public demo? Or does that put us maybe 10 years from commercialization? And it’s a very exciting technical milestone. It signals that commercialization as possible. But it’s still 10 years before we see one of these things in PJM or in CAISO or in ERCOT, you know, a big electricity market that where Americans are consuming this electricity just in the course of normal economic activity.

Katie Rae:

It’s closer than that. I think it’s, you know, early 2030s, maybe four or five years post proof point of net positive energy, which is coming soon. You know three five years like I don’t know exactly but in that zone so it’s actually pretty close but it’s not that you know one day you wake up and it’s there these are serious engineers working through how to get there but like like someone like Commonwealth Fusion has the place they’re going to build it in Virginia right with Dominion Energy like there’s a deal We know where the steel is going to go in the ground. We know where the interconnects are and how to build balance a plant. Like all of that is happening. So we are well on our way. So it’s not just that you get net positive energy and then you start dreaming up like, well, okay, now how would we put this on the grid? Like there are serious people working on that. And so I would say this is a highly economic team who’s going to go build a huge business, right? So that’s where we are. And it’s not just that it’s people in academia or in Google labs playing with things. We are really building these things. Yeah.

Robinson Meyer:

Can you just talk through a number of your, you reference long duration storage. I think Form Energy was originally an Engine Venture firm? Am I?

Katie Rae:

I’m their first investor. Yes.

Robinson Meyer:

Yeah. And so can you just talk through a few of your other clean energy, energy investments, the kind of subsectors you’re excited about at the moment and the companies that represent those sectors?

Katie Rae:

Okay, so other areas like super excited about form. They do long duration, very inexpensive batteries that essentially help. At peak and, you know, are nicely paired with solar and wind, but also with gas, right? You’re essentially stabilizing the grid. I’ve sat on the board since the beginning. I think it is an incredible team. And the opportunities have only gotten bigger for them because of how the energy grid is playing out and how much more power we think we need and so that’s super fun we also do things like transmission so we’re first investors in a company called veer and they’re doing both how do you get more power to a data center from the point of the grid into the data center and how do you do you know long transmission where you get much more energy along the same amount of area that you would already have in transmission. Super important thing. But we also do solid state transformer. I mean, you just go down the entire stack of the energy grid and we have probably made a bet in one of those areas. So that’s materials extraction all the way through kind of the application layer of AI. And I think that is an enormously exciting stack that will kind of power how not only the grid works in the future, but how we get compute and how we get the actual materials that we need in order to kind of control destiny for the economy. And to me, that’s like real venture capital.

Robinson Meyer:

So I see The Engine Ventures as coming out of this long-running critique from MIT. I think I first heard about it maybe 15 years ago at this point, that the U.S. had these key gaps in its ecosystem. You know, the technology, the raw science coming out of American labs at that point was better than anywhere else in the world. And then you had these companies that were absolutely at the peak, you know, at the apex of the value chain and were able to get an enormous amount of value out of existing products. But that we were very bad or generally we were quite bad at taking products out of the lab that were more than, say, five or 10 years from market and putting in the long term investment, putting in the long term engineering time to commercialize them. And to get these raw advances from the lab to a place where they could be commercial products. 10 years ago, it was contrasted to China. Now it’s contrasted to China that China has ecosystems that are much better able at commercializing, let’s say, early-scale technology. And the system seems much more willing to put in investment on 10 or 15-year or 20-year cycles. There’s a lot of picking winners. There’s a lot of corruption. But this Chinese system is able to commercialize long-term technology in a way that the U.S. system maybe struggles to do. The Engine Venture seems to address part of this criticism, but I guess I’m wondering, you’re in this ecosystem every day, so what is missing from the U.S. System right now?

Katie Rae:

Can we put a pin in the China thing? Because I want to come back to that. So let’s remember to come back to that. So there are two gaps in our system that I think are interesting and in some ways persist and is why the Engine Ventures got founded. So the first gap, I mean, we have funded long-term technology development in our university labs for a long time, and we’ve had a lot of corporate development in the same methodology, essentially, but internal to a corporation. And you see it at GE, you see it at Microsoft, Google. I mean, I could name 100 companies that do a lot of systematic R&D. So we’ve been pushing on the edges of science in those two ways for a long time. But if you just go to the university system, because that’s not captive to one company, there are two fundamental gaps that happen. One is there’s a difference between the scientific breakthrough and it being ready to being an engineering project. And generally, you can fund things where you have a plan, you know what the science is going to actually do, you can go and build it, and you know what the market is. Those are fundable activities. The question is, because, you know, funds are judged by their IRR and multiples of return, right? And you could be judged on both or you could be judged on one. If the time length stretches out, you’ve got to have a much bigger multiple for it to make sense economically and for you to compete with other funds. So this gap and the reason The Engine Ventures was founded is if it’s big enough, you can, like, instead of having these 10-year time constraints that most venture funds had, you could stretch that to 12, 14, 16 if you had big enough outcomes, right? Because your multiples can cover for that time stretch. But if you don’t have that, then you can’t fund fusion, unless you don’t care about your economic outcomes, because you are going to have a longer time period than most venture funds could handle. So we basically made the argument, and this is kind of filling this gap of saying, if we have big outcomes, and we’re going to bet on really, really big outcomes, we could fund these things for a longer time period. Okay, so that’s gap one. But you have to have the very best technology to fund. You’ve got to have great teams. You’ve got to have things that could win. Okay. And most won’t. Like that’s just the story of how this works. But someplace like MIT does amass the kind of talent that could have these breakthrough winners, both in the technology and the humans that could actually pull that off. The second gap is what I’ll call the scale up gap. So in venture, there’s growth capital that tends to want to understand fully what are the economics of something and are you making today? I’m going to fund your growth capitals until you get to the public markets. But I understand the economics of this and there’s no technology risk. That is the second gap, because you cannot say that in tough tech. What you can say is, we have proven that you can engineer this thing with the early stage capital, and now you have to scale this up. And you have to bring the cost down, or you have to scale up manufacturing, or you have to develop the supply chain. But if you do, the economic outcomes are enormous. And by the way, in that phase, the timescales do match a lot of current venture funds. But typically, the funds don’t have the technical expertise internally to bet on things like that. And just like you said before, in software, there are easier things to bet on. And there are things that are familiar to bet on. So those two gaps have persisted in the U.S. venture ecosystem. I would say we’ve made a lot of progress on filling the excitement of the early stage gap. And, you know, we were very systematic in trying to say, like, we want people to invest with, we want to work with other people, because having that group of people that invest makes us more competitive and the best things get funded, the best things get the most money. That’s great for the US. It’s great for the world. It’s also great, you know, for the venture people who back those companies. This later stage gap, I still think is there, right? There is plenty of growth capital if you’ve proven out all the engineering, but still a lack of that later stage capital if you’re in what we call scale-up phase. If we fix that, I think the U.S. is in a very good position. And that is done by venture funds forming with self-interest to back the very best things and limited partners understanding that they can have very big wins from that. And you start to see things like SpaceX or X Energy, Fervo, like any of those IPOs, whether they go through their bumps post IPO, they’re all going to go through their bumps, but they have created, like certainly SpaceX has created huge economic value, but that was 25 years in the making of that company, right? And I think the investors in that are pretty darn happy, right? And so you prove that these things make sense when you can become an unstoppable economic force because of the breakthroughs.

Robinson Meyer:

Is that a financing gap or is that an industrial ecosystem gap? Because SpaceX is an interesting comparison, because they were able to piggyback on the back of, I would say, especially 20 years ago, still extremely robust and globally dominant U.S. Aerospace industrial ecosystem, where there’s a lot of manufacturing still. We still make more of a lot of you know aerospace inputs than anyone else and over the course of SpaceX’s life my understanding is they’ve brought a lot of that fabbing in-house and they’ve been able to kind of build their own internal supply chains but when you talk about the struggle or the obstacles to scaling that tough tech companies face you’re describing it as a financing problem Is it a financing problem? It may partially be a financing problem. Or is it also like an industrial ecosystem problem where there is not the same set of contract manufacturers or commodity providers or refineries or smelters that would be able to help these companies if they were in China or even if they were in Europe?

Katie Rae:

Again, complex answer to what appears to be a simple question. And my answer to all of it is yes, which is it’s a financing problem, one. Yes. It’s also when you are building an entirely new industry that will pay off economically later. You are tending to build all of the pieces over time of your supply chain, whether it’s in partnership or whether you build it yourself, whether you have to also have breakthroughs in areas of your supply chain. The answer to all those things is yes. But if you do it and you are successful, you have an enormous, enormously positive economic outcome. And so there is incentive to do that. If you understand how you get from a standing start to like, I’m going to build a fusion company to I have one of the most valuable companies in the world, you have to understand that that involves sometimes making your own supply chain, partnering deeply with maybe someone in the US or someone in Japan or someone, any number of Europe, any number of places around the world, or incenting the building of things next door. Like you are going to have to understand how to do all of those things to build these economically powerful companies like a SpaceX or the next commercial fusion power company. Like there’s no way around that. And so it’s not like software. It’s not like, oh, I thought up an idea. I spun up my Amazon cloud account and it’s on the web today, right? And any Tom, Dick or Harry could do exactly the same thing in an hour and compete with you. That’s what I’m saying. Tough tech looks different than that. But it is possible to do that in the U.S. Yes. Did SpaceX benefit from having an aerospace industry around them? Is there a reason that they put their stuff in Southern California? Absolutely. And that is true in all tough tech companies. Like why is Form Energy in West Virginia? Well, it is the heart of where steel was built. You know, there are rivers. You understand how to move things, right? It’s an iron air battery. When you make sense of where to put one of these things, you’re often doing it because of things that have existed before and talent that you can draw from on that and a community that understands how to build industrial products. That is true in many places in the U.S. And there are many ways to partner. And so you just, you can’t shy away from the complexity of that and the optionality of that on how you build it. And that is why it is a very interesting space to play in, and it’s not so easy to do a simplistic podcast on this. Like, you’ve got to dig into these things.

Robinson Meyer:

Well, what separates a technology that scales and that you look at as a good investment at the Engine Ventures from one that stays stuck in a lab?

Katie Rae:

There are multiple things. One, it’s got to be a lot better than what exists today. It can’t be incrementally better. And almost all innovation is incremental, right? So there are very few things that are true breakthroughs, because you have to have enough time, and it has to make economic sense for long enough to get your innovation to market, right? Because nobody’s going to, your competitors aren’t going to stand still. So that gap has to be big enough that it makes sense to invest. Okay, so that’s fairly easy to judge in a lot of ways because you can look at the techno-economics, there’s generally papers published on this, you can spend real time in a lab with an academic or with an entrepreneur who’s somehow gotten this innovation. And so that you can start to assess. But then you have to say, well, how big could this market ever be? Like, how big could this company ever be? And there are a lot of companies that have really the top of what they’ll ever get to is a few hundred million dollars, maybe a billion dollars after 10 or 15 years of hard work. That doesn’t pay off for an investment where you need 10 to 15 years to build a substantial business. It’s got to be a lot bigger than that. So when I was investing in software many, many years ago, I’d be like, tell me how you get to a million dollars in revenue and in what time period. You got to go many orders of magnitude above that for it to make sense in tough tech. And so it’s how do you get to your first billion in revenue? How do you get to your first $10 billion in revenue? And why is it going to be hard for somebody to compete with you is kind of these early questions we ask. And that’s not that we know the answer, but if you can’t see that answer, you probably don’t have a company you can invest in long term. And so that’s what we look for. But they don’t happen overnight, right? Like take form. Great. you’ve got an iron air battery. Over time, the cost of making those batteries is going to go down dramatically. In the beginning, it won’t. Like, that’s why the IRA got passed, is that you had to kind of incent the early manufacturing of these breakthroughs so that they could get to the point where you could scale up and scale the cost down. And that’s why that has had more staying power than people would have thought in change of administration, because those manufacturing credits are building enormous companies that create jobs and make a ton of sense over the long term.

Robinson Meyer:

We had John Arnold on the show a few weeks ago, and one of the things we talked about with him was that he went to China for the first time a few months ago.

Katie Rae:

You know, I was with him.

Robinson Meyer:

Were you with him?

Katie Rae:

I was with him. Yes. He’s so wonderful. Like he really took us through a wonderful trip.

Robinson Meyer:

So his takeaway from that trip was that he would not invest in any manufacturing in the U.S., that China was too dominant. They have too many educated workers. Their kind of mastery of certain aspects of engineering was too high. They have scale. They have dominance of certain commodity inputs that he just didn’t think there were compelling U.S. investments in manufacturing at the moment, or at least the way you framed it was that I think it didn’t leave him very interested in U.S. manufacturing. Obviously, The Engine Ventures is invested in Form, as we’ve been talking about. It’s invested than in other manufacturing companies. You were with him on that trip, so what does he get wrong in that view?

Katie Rae:

Oh, what a great question. So I spend my days with engineers and scientific leaders. I don’t know that I’m going to say he gets it wrong. I’m going to say this is how I see it, which is that in these areas of true breakthrough, manufacturing is a part of that that must be paired with the understanding of science and the understanding of engineering and then the understanding of manufacturing, but that they all go in unison. So if I was going to manufacture something that we already know how to manufacture, and I was going to make it a little better, China looks really great because of all the things you just said, John Arnold said. Um but if you’re going to manufacture something that has never been done before that is a lot better than we have i don’t think the same rules apply and i don’t think that we’re at a standstill right like the way ai is coming about to basically help manufacturing, help us understand how to manufacture things smarter, get, you know, basically enhance human capability. The way robot arms are getting smarter because of better software. I think fundamentally that paired with true breakthrough engineering and science is something that you can’t just throw up because you have, you know, 300 million migrant workers. Like it is a very heady and collaborative task at hand. And so I don’t think that there is a distinct advantage today. Now, if you say we have a solar panel that we know how it works, and we’re going to get better at manufacturing it to lower cost, yes, China has an advantage right now. But if you’re going to say we’re going to build a next-gen fusion power plant that has net positive energy, I don’t think that’s true. And so I think to me, where the economic value is created and where I invest is in companies where you have that deep collaboration between science, people who understand the science, people who understand the engineering of something new, the materials of something new, and then the manufacturing of it, that you marry those pieces. And you see that over and over again in our portfolio and in what we find interesting and investable.

Robinson Meyer:

We just talked about China a little bit. We just talked about it kind of in the context of manufacturing. Yeah. What’s the heart of the U.S.-China dynamic that you have in mind when I was asking about it earlier and you were parlaying it to later?

Katie Rae:

I mean, first of all, the very best thing for the world is that people around the world in universities and in governments are developing breakthroughs that fundamentally make human life, planetary life better. That is my worldview. That open competition is a very good thing. I’ve been to China many times starting in the 1990s. This is not a static view of one trip that I’ve ever had to China. I believe that there has been continual improvement of getting more and more into developing breakthroughs and real scientific discoveries. And that’s happening in China, that manufacturing is important. They have a very different governmental system than we have, different funding systems. I think there are areas where their funding system works incredibly well, and there are areas where our funding system works incredibly well, and there’s areas where both can work very well. You know, I think they’re a worthy collaborator, a worthy adversary. It’s a dynamic relationship. And lots of respect for different parts of what China has done. I mean, just the economic miracle of what’s happened there is astonishing. And if you look at what the U.S. continues to produce in terms of real breakthroughs, it’s also astonishing. And so I don’t want to live in sort of a non-complex truth about where we’re headed and what’s going to happen. Like, I don’t know the answer to that. But what I do know is that where I’m going to put my life energy is looking at things that have breakthroughs that could scale and could fundamentally make society better globally. So I don’t know how it plays out with China. And I’m not in the government. But I think the U.S. innovation system, where there are people like me that look at these breakthroughs and know the time periods it take and help entrepreneurs go for it is the lifeblood of our future economy and the most exciting place to be. And it’s complex and interesting and fun and difficult, but it is. Where you get the future. And there will be collaborations with and without China, with and without the UK, Europe. This is a very dynamic system. And I love that we put risk capital against these things. And I love that it’s not dictated by any government agency where we put capital, that it’s amplified, but not dictated. And I think that long term is a more dynamic and better system, but not for everything. And so that’s where I think we are.

Robinson Meyer:

You’ve been alluding to policy, and you mentioned the IRA earlier. How have the changes in Washington, say around the IRA or around any other policy, affected the startups in your portfolio?

Katie Rae:

Yeah, I think there are some where it’s enhanced, right? Like I think areas where this administration has been more interested in some of these very breakthrough areas. And I think there are areas that were hurt. Like I think the whole decarb area was hurt. And I think things in geothermal and fusion, fission were helped. And many things didn’t change, right? Because they were rational things that needed to happen. People make a lot about administrative change. But I think if you’re looking at fundamental, breakthrough, translatable science, whether you’re a Democrat or Republican or an independent, there’s a lot of commonality between what people want. And so on the edges, things change.

Robinson Meyer:

Is that true even with the changes to grant funding and, you know, how the NSF works or how OSTP assesses grants? Or is that kind of all upstream of the areas you work on? And so if that were to have an effect, you wouldn’t see it for another five years or something?

Katie Rae:

Yeah, it definitely has. I’m not saying it doesn’t have an effect. You know that I fundamentally believe in long-term government funding of Breakthrough Science because I do think it’s how we get prosperity. That’s just like fundamental. But I do believe every single administration also believes that. So it is on the edges of that. And I don’t believe that long term we’re not going to fund that in the U.S. Because I just think the evidence is so strong that that is the right thing to do. But we’re in an era of a lot of disruption on many fronts. And one of those is in grant funding. What I see is ... just as a practical matter, and I’m not really a policy expert. That’s not what I do. And I’m not a government person. But what I see right now is in the short term, people are thinking very differently about how to translate and how to get what is already being built in these academic labs out and into startups, into things that could have an economic outcome. And that’s awesome. So it’s not like we’ve seen fewer startups in the last few years. We’ve seen just as many, and in some areas more, because they don’t have the funding and they want to translate the things that are there. I don’t know if things dry up in a few years because of that. Probably the answer to that is yes, but I don’t know yet. And there’s a lot of creativity and people thinking about alternative models to funding research. I don’t think long-term that will be as successful as what we have had for the past 50 years. But I could be wrong about that. But I will tell you that many of my conversations with people from around the globe are, how do we address the shortfall in funding in this very important area? Could we do it philanthropically? How could we get things out faster? What’s the impact of AI? Could we go even faster? If we pair what we have already in these labs with AI. And those are very meaningful, very important conversations. And I think pressing the edges of that is going to get us to a new moment. And that’s where I put my time, because I think I see incredible innovation happening at that complex edge of, I’ve lost my funding here, how do I do this? And that is exciting to me. It’s not that I want people to lose funding. I definitely don’t. I think it’s a terrible idea. But where I put my time is where I think I can have an effect, and it’s in that area.

Robinson Meyer:

You’re not a policymaker. You’ve said you’re not a policymaker, but you do presumably talk to them. So what is the number one thing they should understand that you feel like they don’t understand across parties or in parties.

Katie Rae:

Let’s leave parties out of it because I do think there are a lot of people across parties who understand this. Okay. Which is that breakthrough does not happen because you dictate it to happen. Very rarely is that how it happens. It happens over long-term collaboration with very smart people that have typically been in our very high performing academic labs that have curiosities about how you could actually achieve something of importance in a specific area. And multiple labs go at it in different ways and compete to get to that breakthrough. That is a very, very strong model and that they’re not told to do it but they are doing it out of curiosity and they’re incredibly ambitious people that long-term model will produce I think the best effects so I think policymakers are You could get super practical and say, we’re only going to fund things that translate. Well, you’re not going to have the things that can translate if that’s all you’re doing. And people will say that. I don’t think they really mean that, but they say it. They should think in longer term chunks. They should think in 20, 30 year chunks of what are we going to have then that will be a breakthrough if we don’t fund these things now, rather than, oh, we’re going to only fund things that could translate in one to two years. Like that is just short sighted, and we will be lapped if that’s how we think about things. And so our academic system and tenure and long-term steady funding of those very vibrant systems is what has produced this. So I don’t know that in one term you’re going to change that system. And I think it’s very dangerous to fully shake up that system. And I don’t think that’s what’s really being proposed. I think you hear that as a headline. But I think we will go back to funding more of these fundamental things when you start to see the negative outcomes there. So don’t do that is what I would tell policymakers. That’s as simple as that.

Robinson Meyer:

I mean, actually, one theme that’s emerging here is you’re incredibly optimistic about long-term structured competition, whether it’s between national systems, whether it’s between labs, whether it’s between companies. That’s actually like a key element.

Katie Rae:

Commonality here absolutely and that it is human beings being incented to compete openly and on a level enough playing ground i fundamentally believe that’s how you get to the best outcomes and that’s just in my bones like i see it every day i believe it and i believe it because i work with people that do that, like, you know, Bob Mumgard, or I mean, I could name 50 entrepreneurs that I look at them, and what they’ve been able to achieve over, you know, five, six, eight, ten years, and what they preside over, that is because they fundamentally are going for it. And, but I can look at examples of that globally. And I think that’s how we get to our best outcomes.

Robinson Meyer:

So in energy world, of course, we pay a lot of attention to forthcoming developments like fusion, form, ion-air batteries, long-term storage. I would say at the moment, a surprising amount of our time is spent thinking about AI. And that’s because it is driving, as you know, enormous load growth on the electricity system and pushing the electricity system into a whole new size and scale and forcing the system to grow. In ways that it has not had to grow in at least 20 years in most of the country. That makes the following question extremely important to a lot of people’s thinking. And I think they’re very curious where different people come out. When you look at the AI boom, do you think it’s a bubble? And if so, how do you think about the enormous infrastructure bets being made on the back of the AI boom?

Katie Rae:

I do not believe that anybody is going to want less AI. Okay. Like, do you use it daily? Do I use it daily? Does every single person on my team, does every engineer, does everyone in a lab in an academic setting, are they using it every day? I think so. And are they going to want more? Definitely. Because it becomes more and more powerful. And so in the short term, do I believe that there is going to be less load on the system? No. Have I been betting and will I continue to bet that we, over time, will get more efficient at producing more energy in a denser system? Yes. Do I believe that our data centers will get more efficient? Yes. And we could say talk about everything in between yes i believe all those things but i don’t believe it happens overnight i don’t believe the demand vanishes and i don’t believe that we will get more efficient in one second these things happen over time and so in the middle of that you have a lot of angst about, wait, my residential prices are going up. Who pays for the externalities of this? You have people worried about jobs. It’s a lot of change very quickly, but fundamentally, I do not believe it’s going away. And yes, will a lot of things get backed that don’t work, that look silly in the end? Will we call it a bubble? Yes. Do I believe that many things are in progress to make data centers more efficient, to make energy more efficient, to bring more power and less space? Do I believe there will be demand for that for a long time? Yes. And the way the globe is restructuring right now is only putting more pressure on that. But I believe this is a wave that has been coming and will ripple through every economy as we get to a better spot where we have more abundant energy with less carbon produced. We have more compute for less carbon produced. We’re going to use it for desalinization because water is becoming less abundant. We’re going to use it for growing food. We’re going to use that power for many things beyond AI. And those forces have been coming and are not slowing down. And so, yes, there are things that feel bubblicious and will crash. And there are things that are just foundational improvements that I think some will get overbuilt and some will get underbuilt. And they’ll be it’ll be messy. These things are always messy. But to not engage and to say we’re not going to do these things will leave us in a place in our economy that I don’t think we want to be in or we want for our children. Because I think we want to be the ones pushing these edges to make things bigger and better i believe that and it’s coming so jump in

Robinson Meyer:

Well there’s so much more to talk about but i’m gonna have to leave it there Katie Rae thanks for joining us on shift key

Katie Rae:

Thanks Rob, nice to talk to you.

Robinson Meyer:

And that will do it for us this week. Before we go, though, I do want to flag one thing here on the show that we’ve also noted in the show notes. Early in our conversation, Katie says that Commonwealth Fusion has built, quote, more than 300, unquote, tokamak reactors. That was, unfortunately, an overestimate. They’ve actually built closer to more than 150 reactors. Her team contacted us after her appearance to correct the record. We want to make sure it’s corrected here on the show. Shift Key is a production of Heatmap News. Our editors are Jillian Goodman and Nicola Lauricella. Multimedia editing and audio engineering is by Jacob Lambert and by Nick Woodbury. Our music’s by Adam Kramelow. Thanks so much for listening. We’ll see you next week.