Carly Anderson — Dr. Carly Anderson, founder and managing director of timescale.vc (Timescale Ventures) (as introduced on the show)
Carly Anderson: You can't understate the importance of hyperscalers and data center builders to opening up these new markets and creating the conditions for a nuclear renaissance. 'Cause you really do need somebody who's willing to sign a 20-year PPA for $100 per megawatt hour. And that didn't exist until a couple of years ago.
Introduction of Dr. Carly Anderson
Erik Townsend: That was Dr. Carly Anderson. I'm Erik Townsend, and this is MacroVoices. The free weekly podcast targeting professional finance and sophisticated private investors. Episode 548 was produced earlier in August 2026. We've got a special treat lined up for you this Labor Day weekend. Dr. Carly Anderson is one of the smartest and most interesting investors I know. She's a PhD energy geek whose day job is being a venture capitalist specializing in early-stage investments in the energy-related technologies that have the potential to change the world. She's an an early investor in Aalo Atomics and a frequent source of new and exciting energy-related investment ideas. We'll discuss the formative nuclear renaissance in both nuclear fission and nuclear fusion, turbine technologies, laser enrichment of uranium, and much more in this interview.
Joining me now is Dr. Carly Anderson, who is the founder and managing director of timescale.vc. Carly has a PhD in chemical Although it doesn't sound nearly as exciting as what her PhD thesis was on, which was actually creating miniature lightning in order to heal wounds.
State of the nuclear renaissance and mass-produced components
Erik Townsend: Fascinating stuff. Carly, it's not actually all that surprising that you went from a PhD in plasma to eventually becoming my favorite champion of nuclear fusion energy. But why don't we start with something that we actually agree on, which is nuclear fission? What do you think are the most exciting things that are going on in terms of this nuclear renaissance. What do investors need to understand about it?
Carly Anderson: Thanks, Erik. It's great to be here. As I think a lot of people have noticed, hopefully, it's been a crazy exciting year in terms of ground being broken on new demonstrations for some of the smaller reactors, as well as continued progress on the regulatory front and on supply chain maturing for some of the larger designs. I know you've talked to many people on the show already about the happenings earlier this summer at Idaho National Lab. Fantastic to see so many new companies deploying technology, or at least at the first steps for showing that their nuclear cores work. In the coming 12 months, I'm sure we'll see many more milestones from some of these smaller reactor companies. Again, I know we're both very excited about this space and being able to come down that cost curve and rapidly iterate on some of these new designs.
On the larger end of the spectrum, I think there's a lot of stuff that people don't hear about in the headlines around restarting, obviously, some of the older nuclear plants, potentially one of the fastest ways to get gigawatts of nuclear power back on the grid. There's also some work going on to increase the, the power production at existing nuclear plants. And that's an area that I think is super exciting because it's also a way to bring more power online fast to meet the growing demand that we keep hearing about from the electricity sector to power data centers and all kinds of new infrastructure in the US.
Erik Townsend: Let's talk about what it takes to do that, deliver the goal, which is more power, data centers, blah, blah, blah, so forth. Okay, but conventional nuclear power plants take at least a decade to build and sometimes longer than that to get permitted and sorted out and so forth. So what are the bottlenecks? And more importantly, what do you think are the technologies or trends or changes or whatever you want to call it that's going to allow us to not have to take upwards of a decade in order to deliver more energy?
Carly Anderson: Great. Well, I, one of the key things that I think we're both very excited about is mass-producing as many components for these reactors as we can. This is something that should shrink the lead time for everything from the reactor vessels to the pumps that move the fluid around inside these power plants, to the instruments, to the safety mechanisms, to days, if not weeks, If not months.
If you look at the time it takes to build a standard nuclear power plant, and standard is that kind of 1,000-megawatt chunk, you have to, by the time you get through site works and then concrete pouring and then shipping these massive, massive metal containers to site, I mean, Vogtle was way, way overschedule and over budget. I think the best people in the world are installing these in 5 years.
That's not as fast. That's not where we want to be, right? We're looking at designs on the smaller end where you could deploy in less than 2 years, ideally on months. A typical data center is going— one building maybe goes up in 9 months at this point, and you really want your power generation systems to be able to match that timeline. So mass manufacturing, love to see that coming out of companies like Aalo Atomics and some of the other ones on the small end of things.
In terms of other bottlenecks, it was great to see Part 57 released from the NRC to create a pathway for advanced reactors that looks different from the traditional methods for light water reactors. Those regulations were set up very specifically for technologies and in a way that just isn't applicable to some of the, some of the newer designs. It's exciting to see some of the fuel supply chain mature, which we can talk more about, and the power generation side of things and like all the other things besides and outside that, what we call the the nuclear island, which is like where the magic happens and the fuel produces heat.
All of these components have to come together, and we're seeing that happen a lot faster at that smaller 1 to 20 megawatt scale.
Markets for advanced nuclear: defense, remote sites, data centers and global competition
Erik Townsend: Carly, let's talk about how this nuclear renaissance market is coming together in terms of where the opportunity is. The way I'm seeing this, what we're seeing in terms of grid-connected energy is mostly discussion about more light water reactors, good old-fashioned Eisenhower Power era technology, nuclear reactors designed in the 1950s, build more of those things. That's the thing to do. Advanced nuclear, the stuff that I tend to get more excited about, costs more. At least in the beginning, it'll cost more. And that means somebody who can afford to pay up for it has to be the customer. And there's not a whole lot of those. It seems like what we're seeing is advanced nuclear is really finding 3 niche markets, which is the military market where people have a Pentagon-sized budget for whatever special thing and they can just afford to do whatever they want.
You've got out in the middle of nowhere, the market that is for microreactors that is now dominated by large diesel generator sets that run remote mining villages and so forth. There's a nuclear angle there. They can afford to, right now, if what you're spending money on is every week you got to deliver another flight goes up to the middle of the Arctic Circle someplace just to deliver the next load of diesel fuel for your generators to run whatever you're running up there. If you can replace that with a nuclear reactor, you don't have to make that weekly fuel flight, and it's a huge cost savings. There's a market there. And then the last one seems to be data centers for AI.
Would you agree those are the markets? Do you see any other markets? And which of those do you think is going to be most exciting? And which technologies do you think are going to drive
Carly Anderson: Great, Erik. I think that that's a perfect rundown of 3 key areas for early nuclear deployment. And I'll touch on the first 2 very quickly. Defense is a— has been a fantastic enabler for so many technologies coming out of American labs. I think across our portfolio, they've been an early customer able to help buy down technical risk for things in the autonomous transportation space. Think drone delivery, in terms of new materials, in terms of satellite capabilities, that is a fantastic market.
It is limited. I like to say global markets are bigger. And at some point, we're in a place where there's only so many bases, you run into export controls. And, you know, while this is a very important tool for— these are very important markets, not just for scaling, but for national security. From an interest standpoint, at some point you start bumping up against barriers if you are in those markets alone. They're also very well suited to very small reactors. So again, these are the microreactors. These are less than 5 megawatts, often 1 megawatt, maybe even 200 kilowatts. Those, they map also very well onto remote communities. So think about mining installations, think about Alaska. And again, like In those places, these are very important technologies. There often aren't alternatives.
But if you think about the one, you know, the trillion-dollar broader energy market, those technologies aren't going to participate in a significant way. You know, you need 1,000 1-megawatt turbines to equal one AP1000. And you do run into cost challenges when things are that small. You get a lot of benefit. from mass manufacturing, but there is also benefit, particularly when you get to turbine systems and engineering systems, to having things be at a slightly larger size.
So I said that was gonna be short. It took a little bit longer to get here than I wanted to, but getting into the data center market, I think, I mean, that is where the real opportunity is from a, from a return standpoint and from a change the world standpoint, in my view. If you think about, and I think you can't understate the importance of what the entrance of hyperscalers and data center builders has been to opening up these new markets for nuclear and creating the conditions for a nuclear renaissance. Because you really do need somebody who's willing to sign a 20-year PPA for $100 per megawatt hour. And that didn't exist until a couple of years ago.
I mean, if you think about what had to happen before that, individual utilities, energy utilities in the US are relatively small. Some of them operate 2 nuclear plants, some of them operate 4 nuclear plants. They're hyperlocal. What you can do as a hyperscaler is you can help create those order books for a dozen plants or more, and you can create, you provide a secure offtake, you make the project financeable. And so that market from like a size and like opportunity also to then start exporting to other countries is huge.
Just to comment on that for a second, because I think that's been something that's frustrating a lot of us in the industry and has for a long time, is watching like what's available to the broader global market for nuclear. You know, today there are about 80 reactors under construction around the world. I think about 60 of them are Chinese design. There's something like 30 Chinese plants getting built around the world right now.
They are going to double their amount of installed capacity within the next 5 years. They're installing 5 to 10 gigawatts of nuclear power every year and will do that for the next 5 years. They have a 10-year plan. If you look at the Russian designs, India is building most of its nuclear capacity out in partnership with the Russian state-owned organizations, right? There's a split where An Indian company is, is in charge of the project and a GC is Russian. But if you're not in the US, your options right now are China or Russia. The South Koreans haven't been participating in a meaningful way.
How do we get to the point where the US has a reactor model that we have proven that we can build on time and on budget? The timeline for starting today and getting an AP1000 online and getting the confidence behind that Is, I hate to say it, I think it's 10 years.
If we're lucky, we can bring it into 7. If we are starting with some of these smaller microreactor technologies where like you and me are like, we're hoping to see one online in 2028, you know, maybe 2029 if the timing slips.
But once you have a plant like that, once you have a reference facility, you can build an order book much, much more easily and you can start iterating doing the things you need to do to bring the cost down. And I think you can get to a much more cost-effective reactor in 2020, you know, 2030, early 2030s, very, very soon.
And that creates a pathway for larger reactors and scaling up from there based on the learnings earlier on.
Factory mass production versus speed to a reference facility
Erik Townsend: Let's talk about what's most important in order for this nuclear renaissance to really succeed and change the world, as you put it. I'm very passionate about changing the world and making it a better place. So I think the big, the killer thing here is factory mass production. And I think it's really important, Carly, to distinguish the word mass production. I'm not, you know, a lot of people in the nuclear business are talking about factory production. It's more efficient to do it in a factory offsite someplace as opposed to building it onsite. I'm not talking about that. I'm talking about like assembly lines with Hopefully robotics that are very quickly, very efficiently, very precisely and repeatably with very careful quality control assembling new nuclear reactors at a completely different pace than we've ever considered before.
So I think that's the single most important thing. And frankly, I don't think anybody got it until the last year and a half or so. Now we're seeing companies like Aalo in the larger space, in the microreactor space, there's several companies. Radiant is one. You probably know more than I do, and please tell our listeners what the others are. But we're seeing people, I think, finally catch on to the idea that we need to do for the nuclear reactor what Henry Ford did for the automobile. I think that's the most important thing that gets us to eventually nuclear energy costs less than fossil fuel energy. Would you agree with that? And if so, what are the other bottlenecks? What are the things that we need to know about in order to get to the point where we can really sees this vision?
Carly Anderson: I think that's a huge component, Erik. I mean, I think I come down more on speed to reference facility. I really, really want to see these reference facilities up as soon as possible. And then after getting a reference facility online, it's what is the cost of your reactor at scale? And I think we're both aligned that by leaning into that mass-manufactured approach and building units that are small in space, And can be shipped to site very easily, and where teams can iterate and learn quickly, I think you're going to reach the cost you need to be competitive with diesel gensets first, and then with other forms of generation later, much more quickly.
Nuclear fuel supply chain: conversion and laser enrichment
Carly Anderson: There are other bottlenecks as well. Nuclear fuel is one that gets, I think, a lot of airtime, as it should for good reason. Very much a fan. And I think one of the things about HALEU that both of us recognized and immediately resonated with us Was the ability to use fuel that is available today, like what's called low enriched uranium. It is the same fuel that's used in every nuclear reactor—the ninety to hundred nuclear reactors that are operating in the U.S. around the country right now. That said, looking forward across multiple parts of the nuclear fuel supply chain, there needs to be significant updating and additional capacity built out.
When you're making nuclear fuel, there's a number of steps. Right? Once you've mined the uranium, you have to get it into a form where you can do enrichment. That step is a bit, it's a bit challenging from a chemistry perspective. It's also very concentrated. There are like 5 facilities in the world where we do that uranium conversion today.
One of them is in the US. That plant is very old. And looking forward, a lot of that capacity is currently in Russia. There are executive orders saying that we will get off that conversion supply By 2028. There is currently not an alternative, and multiple companies are running to fill that gap.
Erik Townsend: Let's talk specifically about what technologies are going to change the market and make it better. So for something like enrichment, it was originally done with this technique called gaseous diffusion. Then they moved on and eventually got to centrifuges, which is the current technology. It looks like laser comes next. So let's go through whether it be conversion or enrichment and so forth. What are the technologies in nuclear energy that are maybe candidates for upgrades?
Carly Anderson: Great. Yeah. So just to, as broad strokes, what the fuel cycle looks like right now is you mine it, you convert it to a form that you can enrich it into something, you mine it, you do conversion, you enrich it, you deconvert it, you fabricate it into fuel, and then you ship it off to your reactor.
On the enrichment side of things, like you said, there's been some shift from, you know, the OG gas diffusion technology to gas centrifuges. Now lasers have emerged on the scene. I would like to believe that lasers are the future. It's not a foregone conclusion yet. General Matter has— is working very, very hard on updating the more traditional gas centrifuge technology. Right next to them in Paducah, Kentucky, is, I think, one of your favorites, GLE, who are working on doing that similar separation. So separating the uranium-235 from 238 by exciting only one of them with lasers.
That looks potentially very promising in terms of the number of like individual units that you would need. Like lasers can be much, much more efficient about only hitting one isotope, not 2.
There are other processes as well. One of the companies I really like, Hexium, down in Texas, is using a different laser technology starting from lithium metal. So now you don't need the process of conversion and enrichment typically uses fluorine chemistry, which people probably know fluorine from like the dentist. But, you know, fluorine likes to interact with bone. The dentist is a good way. Usually fluorine interacts with human bones in a very, very nasty way. So by avoiding that and going through lithium metal, that removes some of the safety.
safety things that you might want to think about or worry about.
I think one of the things I'm excited about as we develop a lot of these new chemical processing technologies is there will be adjacent technologies that benefit. Rare earth element processing technology, for instance, a lot of that could be done more efficiently with fluorine chemistry. You know, again, it's not something we do a lot in the US for some good reasons— safety, environmental, et cetera—
Turbines and supercritical CO2 power cycles
Carly Anderson: but also that pushes it to other parts of the world where
Erik Townsend: Carly, the heat energy that we get from nuclear energy is generally not usable as heat. We usually have to convert it into electricity first. And the process that we've historically used to do that, which is steam turbines, are incredibly inefficient. The thermal efficiency of a steam turbine is only about 30%. So more than half the energy gets wasted. Is there any hope for fixing that?
Carly Anderson: Yeah. In terms of technologies to turn heat into electricity, this is another area we've seen an explosion of companies trying to change how we're doing this, particularly around, like, on one side, how do you manufacture turbines in a much shorter time period, but also what other types of turbines or power generation cycles are there for converting heat to electricity much more efficiently? One of the technologies that I think we're both very excited about is using supercritical CO2 as the working fluid in these turbines.
If you do that, you can go to higher pressures and you can make the turbine equipment very small. And this again plays into that mass manufacturing argument where if you can make something small, if you can use less material, you can iterate quickly, you can come down the cost curve. And with that particular cycle, you can reach very high efficiencies of 50% even, which gets you much more power. per heat source, whether that's a nuclear reactor or whether that's something else.
Fusion versus advanced fission: timing, cost and tritium
Erik Townsend: Well, Carly, I wanted to start with a fission warmup because that's where we always agree on things. Now let's move into the fun stuff. Nuclear fusion, I think the one thing we can agree on is it's got to be the coolest kind of energy ever conceived. I think you got to get to dilithium crystals on Star Trek before you can get to something that's cooler than the idea of nuclear fusion. And just for any listeners who don't understand what it is yet, nuclear fission is when Nuclear fission is when you've got a big heavy element like uranium and you split it in half. Nuclear fusion is when you've got teensy-weensy small atoms like hydrogen or helium and you squeeze them together in order to make a bigger atom. That also releases a whole bunch of nuclear energy. That's called nuclear fusion. Carly's a whole lot more bullish than I am on it.
So Carly, first of all, you win hands down that it doesn't get any cooler than that. But once we set aside the techno geek excitement about how fascinating that technology is. I like to focus on how we can actually make the world a better place. And I'm convinced that hinges entirely on making nuclear cost less than energy from fossil fuels.
I just don't understand why you and others are so focused on fusion when we've already got fission figured out. And it's clearly going to cost less or take less time using fission to get to the technology so it costs less than fossil fuels. Why do you want to go do all this really cool technology that I don't think is commercially viable?
Carly Anderson: Ouch. Wow, Erik. Okay, so you said 2 things. You don't see how it costs less and you don't see how it arrives in time. And I think, like, let's start with that later one, right? You know, we're both excited about microreactors and having some operational in 2028, maybe 2029. But most of the new nuclear reactors, at least in the US, aren't going to come online until the early to mid-2030s, at least for the larger plants.
And so by the time you bring— we installed 6 gigawatts of natural gas last year in the US. By the time you get to 6 gigawatts of new nuclear deployed, I mean, include 2 restarts, include, okay, X-energy's unit, maybe 1 TerraPower, maybe 1 of some of these larger 300-megawatt SMRs, and then include the time to ramp up to even 100 megawatts, and we do think it'll go well beyond that. You know, we're, we're all talking about the same time period here, early 2030s, and that's when we are going to see fusion power plants operational, which to me is incredibly exciting. I know, Erik, you're, you'd like to say like, why are we already breaking ground on these things?
Erik Townsend: Yeah.
Carly Anderson: But I think like a lot has changed in the fusion landscape that people aren't aware of. For one thing, a lot of the component technologies have gotten much cheaper and much better. To take a quick dive into fusion technology, short, I promise, you can confine a star on Earth with magnets, or you can shoot fuel with strong lasers and compress it into a mini star that way.
On the magnet side of things, high-temperature superconducting magnets, You'll see HTS written around, have totally revolutionized how you build electric magnets, not rare earths, but electromagnets. And that makes your fusion device go from something the size of a football stadium down to something the size of like maybe a normal residential house, maybe even smaller. This is why Commonwealth Fusion Systems is one of the leading fusion companies in the world, is they figured out how to tokamaks, a type of fusion power plant, small.
On the laser side, the way that we do lasers, if anyone has even thought or heard of, you know, the ASML hitting a droplet of tin 10,000 times a second twice, like laser, it's obvious that laser technology has changed a lot. There are many, many, many different types of laser technologies. One of my favorite is a group out of Denver called Xcimer Energy, which literally took the lasers from the Star Wars program and Finish the job. And so now you can get laser energy for an incredibly low dollars per watt value. And then if you get into solid-state diode lasers, if you get into the simulation side of things, like it used to take years to run the simulations you needed to design a new type of nuclear plant, let alone like fusion power plant. You can run those simulations in hours today, maybe less.
So there's been a ton of game-changing technology advances that are not conceptual. They've been built, they've been demonstrated. Back to mass manufacturing, Thea Energy is starting to mass manufacture very small pixel magnets that you can control with AI to basically dial in what you want your, your fusion reactor bottle to look like. Like just the level and pace of innovation in that field is incredible.
So I think we'll start seeing fusion power plants in the mid-2030s, about the same time as we'll see some of these SMR designs come out.
Erik Townsend: Why do you think that they are better? In other words, there was some pretty good arguments to be made for why fusion energy would be safer, better, less waste, and so forth than light water reactors. I get that. But if you look at what advanced fission can offer, I don't really see that much benefit of fusion over advanced fission. So what is it that makes you like it better? It seems to me it definitely costs more, right?
Do you expect it to cost less than fission?
Carly Anderson: Yeah, I expect it to cost less. Like you, you take out your entire uranium mining, conversion, enrichment, and then decommissioning storage. Like so much of that goes away when your fuel is hydrogen. Like that's a significant part of the cost right there. Otherwise, we're looking at 2, basically 2 kettle systems made of metal. And you can get a high power density out of the fusion reactors. I don't see that being necessarily much more expensive.
Erik Townsend: Hang on, Carly. The fusion reactors that I'm familiar with run on a kind of hydrogen called tritium, a specific isotope of hydrogen that currently has a market price of $30 million per kilogram. So you can't make a cheaper fuel argument here.
Carly Anderson: I disagree. I think there are a number of fusion supply chain companies that are coming up with ways to produce the amount of tritium you would need to start a fusion power plant very cheaply. One of the key things in fusion power plant design is closing the fuel cycle so that the power plant produces the tritium that it needs to then use as the fuel. That works in a part of the fusion power plant called a blanket, for lack of a better word. That's what we call it, which is basically lithium, where when your fusion happens, the thing that it makes runs into lithium, and then that falls back apart into tritium again.
So you literally, at the end of the day, end up back with tritium. So your only input into the process is deuterium, which is a different flavor of hydrogen, but a much more common one that is found in seawater. It's actually used in the Canadian CANDU nuclear plants today.
Erik Townsend: Okay. So with respect to fission, when we talked about fission, you were discussing specific reactor companies like Aalo Atomics and Radiant Energy and so forth. I noticed that when we switched to fusion, you started talking about things like Xcimer, which is an enabling technology. So do you think the plays in fusion are on the fusion energy companies, which would be the companies that are actually building the fusion reactors, or do you think it's at the enabling technology level?
Carly Anderson: I think it's both. Several of the companies that I just mentioned are building the integrated power plant. That's Commonwealth Fusion Systems, Thea Energy on the magnet side, Xcimer Energy, Focused Energy on the laser side. There are other very interesting companies, though, that are doing the enabling technologies. Some of them are making the superconducting magnet materials.
Some of those are taking those technologies and using them in other areas. For instance, one of my favorite new companies, Spinner Energy, is using the superconducting magnets to solve problems for data centers and on the grid. Because if you have a material that can move megawatts of energy in a tiny package that doesn't break the roof of your data center, that's pretty darn cool as we go to higher and higher energy needs in a very small, for each data center rack.
Silicon carbide and solid-state transformers
Erik Townsend: Silicon carbide comes to mind when you said that as the specific material science advance, which I think is really going to have a big impact on nuclear energy and the nuclear energy industry. Do you agree? And if so, how do you see this? What are the reasons that silicon carbide and what things do you see silicon carbide getting used for?
And I guess you should start by telling our listeners who are not familiar, what is silicon carbide and why should they care?
Carly Anderson: Silicon carbide. Okay. Wow. It's a solid material that you can use for a ton of things. It's gotten really cheap because of the electric vehicle industry. Thank you, Elon Musk. And now has also moved into a lot of power control devices. It's amazing because you can switch it back and forth at high voltages, so you can make your devices really small, and it has very good thermal tolerance, so you can use it in systems that are hot. That's why we like it for other things. It's also great as an insulator for fuel, but we're going to see even other materials coming out in the next couple of years. You know, these— they're sometimes called wide bandgap semiconductors, you know, big mouthful of stuff that means Materials that can make electronics smaller, cheaper, more thermally tolerant.
And we're looking at a couple plays in that space that we think are just totally going to revolutionize power management.
Erik Townsend: Carly, is there an opportunity in silicon carbide specifically in terms of developing that material and using it for more things? And I think it would be particularly cool if somebody figured out how to 3D print it. Is there an industry for developing the silicon carbide technology, or is it really the people that make products that use it?
Carly Anderson: So there's many steps in the silicon carbide supply chain from wafering. So companies like Halo Industries have figured out how to very efficiently slice and pattern silicon carbide, both for things, you know, like electric vehicles, but also for processing chips, which is pretty interesting. You could also use it for cladding in materials because it's very, very chemically resistant. So basically use it as a protective layer around your nuclear reactor.
Erik Townsend: Okay.
Carly Anderson: One other area that's seen a lot of attention for good reason is in solid-state transformers. Back to what do we need to hook up a lot more power, including nuclear, to the grid? All those devices to like connect the power plant to the data center, one of the limiting ones is transformers. Several companies are working to make these, again, mass manufactured, cheaper, bring the lead times down to, you know, months or weeks instead of currently years. And silicon carbide is playing a big role. in allowing these devices to be low cost and small.
Beyond nuclear: grid, data-center power and demand technologies
Erik Townsend: Let's talk beyond just nuclear energy. The data center industry and just everything else that's going on in the world is going to require more and more energy. What are the other technologies that relate to energy that are kind of the hot things to know about, where you see opportunity in the market as an early-stage investor?
Carly Anderson: Thanks, Erik. Yeah, we've talked about a lot of them already. So things like turbines, things like solid-state transformers and other power management technologies. Going upstream into producing materials, rare earth element refining, mining, magnet production, those are very interesting. Getting into even commodities like aluminum refining, that's something again where most of the US aluminum refining went away.
There are great supply chains for recycling aluminum in other places. When you look at the timeline for bringing new mines online, It is a bit scary as a venture investor. So how do you shrink those timelines? That's something we're thinking a lot about for both copper, aluminum, and other materials that go into the electric grid, as well as many other manufactured parts, products, components.
Erik Townsend: Carly, let's talk about turbines, the things that convert the heat energy that's made by nuclear and many other energy processes and turn it into electricity. You mentioned earlier the supercritical CO2 turbines are the ones that replace water with carbon dioxide as the fluid that gets circulated. It makes them much more efficient, lighter, cheaper, easier to mass produce.
Where are the other opportunities that improve the efficiency of energy?
Carly Anderson: So on the turbine side of things, to mention one other use case, we talked about supercritical CO2 turbines matching potentially very well with some of the smaller nuclear devices to create a package that can be shipped and installed in a matter of Another thing that is looking more even near-term for some of these supercritical CO2 turbines is to attach onto existing equipment and make it much more efficient. That's often called a, a bottoming cycle. And so it unlocks another 10, 15, 20% of power generation from the same assets.
This is a recurring theme that people are looking at. How do you get more energy out of the assets that are already installed? As we think about how are we going to possibly meet the data center demand doubling between now and 2030. Another area where there's a lot of design work going around on for is to meet Nvidia's roadmap to getting to 800-volt distribution in data centers.
What this means practically is the amount of energy coming into these data centers is becoming absolutely like mind-blowing. You know, somebody was describing how you build these data centers and And, you know, just a 40-megawatt data center is the size of 6 Sam's Clubs. So the scale of these things is just mind-blowing. And now you think about bringing 1 megawatt, which is basically enough energy to power, give or take, 1,000 homes, to a single cabinet.
And that's just, that's just insane. And so this is why you get to, again, a lot of these power management devices at the rack level. Backing up to right outside the data center, how these data centers interact with the US electric grid is very difficult. You need basically a box that can, on the grid side, show it like a very even, nice, smooth seas.
If it turns off, it turns off gradually, like very nice, smooth line for how the energy is being demanded. On the data center side, You've got these AI training models that are going all over the place, and they're jumping from 80% down to 20%, in some cases over, you know, a second timescale.
And so we're looking at a lot of technologies that allow that match to happen in a very low-cost, efficient way. It's very interesting. And we talk to a lot of people working on how this is being standardized so that, again, you can get back to, you know, not a bespoke device, for each time you interconnect a data center to the grid, but something that you could roll out across, you know, not just one utility area, but ideally over, over most of the US.
That's a super interesting way to do things. Also looking at a lot of companies that are aggregating demand. So again, like rather than building out additional power plants, can you create these virtual power plants that can turn on and off just like a, a real power plant effectively to reduce the peak load. That's another area that's super interesting.
Geothermal, we're going to see a lot more geothermal activity coming online. That does take time. There's a lot of subsurface engineering that has to happen. We have much better tools for characterizing the subsurface. And again, modeling, can't say enough about how modeling and simulation have totally changed how we build hardware companies.
It has been just the most massive unlock that you can— honestly, I would even go so far as to say this is what now makes deep tech an investable sector, is being able to actually simulate and design these products on a very, very short timescale.
Geothermal: drilling, siting and cost
Erik Townsend: Pardon the pun here, but let's go a little deeper on geothermal. I think geothermal is the coolest idea ever, but hang on a What you have to solve in order to solve geothermal is somebody's got to figure out how to drill holes through 400-degree granite several thousand meters below the surface.
That's a fairly hard technical problem. We're not quite there yet. Let's suppose we solved it. Okay, where would we be? Well, where we would be is similar to the fracking industry in oil. We would have the ability to send these very expensive teams of people out to very remote, expensive, hard-to-get-to places in order to go and create some geothermal wells.
Okay, how does that ever get mass produced in a factory in order to bring the cost down? I don't think it does. Whereas with nuclear, I can see that path, factory mass production. Do you see a path to geothermal that achieves that big step function of efficiency that might make geothermal competitive with mass manufacturing?
Carly Anderson: It's a great question, Erik. As an early-stage investor, I would love to just take the cop-out and say that I'm an all-of-the-above energy person.
Erik Townsend: You got a slice of all of it, just in case.
Carly Anderson: Exactly right. But I guess to double-click on some of that, you said something that I think is very important, which is 400-degree heat. And if where that appears near the surface where you can get to it now is way out in the middle of nowhere, there aren't that many sites. Fervo has bought them all up because they're very smart about how they— about their business plans. What would need to happen for geothermal to get really attractive is to be able to, like you said, cost-effectively drill very, very far underground to where you can hit 400-degree heat, not just in the middle of nowhere, but next to a— in closer proximity to a load. And so Quaise is a very interesting company that is working on technologies to do that. There are some other folks poking around there.
That solves, you know, one of several challenges for geothermal, and people are working on all of them. So you have to drill through the rock quickly, cost-effectively. You have to create enough surface area with your heat source to get enough heat out fast enough. And then you have to find an area that's— then there's a volume question, right? Where is there enough heat in like this, you know, I don't know, know, mile-square chunk underground where all that heat can come up your pipe.
So there are a number of questions to be answered. There are people working on all of them. Those projects, again, like, it's the density of power you get from an individual site is something that needs to go up quite a bit, I think, for these things to get to your 5 to 7 cents per kilowatt-hour cost target. But I think there's definitely situations where that could make I'm excited to see what happens with some of these new drilling technologies. And there's material opportunities as well. A lot of the cost of drilling is broken drill bits. One of the things that we're looking at are materials for, you know, not just these silicon carbide tools for data centers, but when you're using a really hard thermally conductive material, it also shows up as useful in mining. We love these synergies between industries where the market makes the market bigger.
You can learn faster and you can actually access new markets without having to totally rebuild a manufacturing technology.
Robotics: industrial versus humanoid
Erik Townsend: Well, Carly, this has been a fantastic interview, but before I let you go, I want to touch on one last topic, which is a little bit outside of your wheelhouse as an early-stage energy investor, but I know it's one that fascinates you as much as it fascinates me. And that's the direction of robotics. What I used to think is, okay, industrial robotics is really cool, really important, The robots on TV, movies, and so forth, the so-called humanoid robotics are probably just an amusement. But the robotics that build cars and so forth, I thought that was really it. It seems like the world wants to go the other way. Humanoid robots really are going to be a big thing and we're going to invest big in that.
Which way do you see the robotics industry going? Is it more important to stay focused on industrial robots that build nuclear reactors more quickly, or do you think that this humanoid robotics where people own their own robots and have robot assistants and so forth, is that going to be a big trend?
Carly Anderson: Well, Erik, thanks for giving that prequalification that I am not a robotics expert. I appreciate that. But it's a super interesting space and there are just a ton of people, especially where I am in San Francisco, who want to talk about robotics nonstop. For us in the industrial sectors, we pay way more attention to what's going on in form factors that aren't necessarily human. How do you get robots into spaces where humans can't be or don't want to be because the materials they're working with are toxic, it's high off the ground, those type of things? And I think that's where we, like, that's where we've been seeing the most robotics penetration is doing a lot of these dangerous, remote, Or difficult tasks. You know, one of my favorite companies, Pyka, is basically replacing—they started off replacing the pilots and crop dusters, and at this point they are covering tens of thousands of acres across South America. They've flown in Central America as well.
I think this is one of the point things to point out is that a lot of robotics use is happening outside of the U.S. I think in the U.S. we rightfully have a lot of concerns about safety. Another one of the companies that know from my past life is FORT Robotics. They're, they're in the process of SPACing. Congrats to the team. That's awesome.
Right. And their whole focus is how do you have robots and humans be able to be in the same workspace and not run into challenges? Right. Tons of work to still be done on robotic safety. You have people working on soft robotics parts, trying to do things with pneumatics instead of actuators, you know, in addition to having off buttons.
You know, in the near term, the industrial robotics space is super, super interesting. We're seeing all kinds of stuff in shipbuilding and in decommissioning and in construction. Having robots in the home, I think we're 5+ years out from robots in the home. This is my unqualified energy investor opinion on robotics, but I'll take a bet with you on that if you'd like, Erik.
Timescale Ventures and Carly Anderson's investment thesis
Erik Townsend: Well, Carly, I can't thank you enough for a terrific interview, but before I let you go, please tell our listeners a little bit more. You've got such a cool job. You do early-stage venture investing. You're a VC focused energy— I shouldn't say it. I'll let you say it because you actually know the answer and can describe your role. But I think it's just super exciting that you're focused on such interesting technologies. It really is the stuff that I think can change the course of history if we get it right. So tell me a little bit more about what you do and particularly how you got into this particular niche of investing.
Carly Anderson: Thanks, Erik. Yeah, Timescale Ventures, like you said, we're an early-stage deep tech and critical technologies fund. Energy is a huge piece of it. We do a lot of investing in things like advanced nuclear and fusion and adjacent technologies, in how do we power data centers. Industrial robotics is also a big piece of it. We do things in logistics, we do things in mobility, manufacturing, compute, and other critical technologies.
I think, why did we start this now? We're at a really incredible moment where the talent stack has matured so much. After my PhD, I helped build a startup for a couple of years, and the talent that was willing to go into startups, people hadn't built hardware companies before, right?
At this point, the people coming out of companies like SpaceX and Rivian and Tesla, as well as folks from industry who understand what startups are and that, hey, we might actually be able to do things on a faster timescale, And spoiler, this is why we named the firm Timescale Ventures, is because you need to be able to get technologies to market fast if you want to make a difference, both financially and from a human standpoint in the world today.
The capital stack has matured. There are downstream investors that we know very well. The tools to build companies have gotten extraordinarily better, from simulation to prototyping to materials to Full-on robotics components that you can buy off the shelf. And like I said, like the talent going into the sector has— we've never seen the level of engineers in hardware that are building companies right now. And so with Timescale, we're working with them from the earliest days. It's super, super exciting. You're going to hear a lot more from us in the coming months in terms of having stuff out on Substack, having stuff up on the web. And I'm excited to— To see where this goes.
Fascinating times.
Erik Townsend: And Carly, for people who might want to follow your work or learn more about what you want to do, what's the website? How do they contact you?
Carly Anderson: We'd love for them to come visit us at timescale.vc. We have a Substack up. Again, the handle is @timescalevc, and they're always welcome to reach out to myself or my co-founder, Christina O'Connor, directly on LinkedIn or through one of our mutual contacts. Love meeting people who are as passionate about this space as we are.
Outro (no postgame) and disclaimer
Erik Townsend: Folks, that's a wrap for this prerecorded episode. We don't have a postgame segment this week because Massil and Patrick are off this week. They'll be back next week with our regular show format. MacroVoices is presented for informational and entertainment purposes only. The information presented on MacroVoices should not be construed as investment advice. Always consult a licensed investment professional before making any investment decisions.
The views and opinions expressed on MacroVoices are those of the participants and do not necessarily reflect those of the show's hosts or sponsors. MacroVoices, its producers, sponsors, and hosts shall not be liable for losses resulting from investment decisions based on information or viewpoints presented on MacroVoices.
MacroVoices transcripts are machine-transcribed, then corrected and edited by AI under MacroVoices' editorial rules: names, terms and figures are checked against the audio where the text is uncertain, and this free transcript is the corrected verbatim text: it keeps what each speaker said, including fillers, repeats and false starts (the subscriber edition is the edited reading copy). Nothing is added; every edit is logged against the verbatim transcript, which MacroVoices keeps. Timestamps refer to the published audio. Errors are still possible — where it matters, listen. Nothing on MacroVoices is investment advice: the views expressed are the participants' own, and MacroVoices, its hosts and producers accept no liability for decisions made on the basis of this material.
Subscribers get the formatted edition of this episode as PDF and HTML — the research-note outline with every market call flagged, the contents with links into the text, the guest's earlier MacroVoices appearances and related episodes, the edited transcript and the glossary — and the one-page Episode Brief. Open the subscriber edition.