We’ve covered AI’s massive power appetite in depth over the past year – with good reason. It’s the driving force behind much of the change and uncertainty in the energy world right now, from the error bars around our demand for electricity to the lineup of technologies vying to meet that demand.
In this episode Shayle talks to his colleague Andy Lubershane, head of research and partner at Energy Impact Partners, about five big questions arising in this uncertain load-growth environment. They cover topics like:
- The underappreciated factors that could flip the supply crunch to oversupply, like algorithmic efficiency gains, on-device inference, and off-grid data centers
- The winners of the AI-drive power boom, including utilities and grid equipment suppliers, and the potential losers like industry that relies on cheap power
- Whether there will be a “Cambrian explosion” or consolidation of nuclear reactors designs
- The prospects for enhanced geothermal after Fervo’s Cape Station comes online
- The future of grid-enhancing technologies like advanced conductors and dynamic line ratings, and whether they will make it out of “utility pilot hell”
Resources
- Open Circuit: How do we know if we’re in an AI bubble?
- Catalyst: The US nuclear groundswell
- Catalyst: How geothermal gets built
- Latitude Media: In Georgia, stakeholders still can’t agree on data center load growth numbers
Credits: Hosted by Shayle Kann. Produced and edited by Daniel Woldorff. Original music and engineering by Sean Marquand. Stephen Lacey is our executive editor.
Catalyst is brought to you by EnergyHub. EnergyHub helps utilities build next-generation virtual power plants that unlock reliable flexibility at every level of the grid. See how EnergyHub helps unlock the power of flexibility at scale, and deliver more value through cross-DER dispatch with their leading Edge DERMS platform, by visiting energyhub.com.
Catalyst is brought to you by Bloom Energy. AI data centers can’t wait years for grid power—and with Bloom Energy’s fuel cells, they don’t have to. Bloom Energy delivers affordable, always-on, ultra-reliable onsite power, built for chipmakers, hyperscalers, and data center leaders looking to power their operations at AI speed. Learn more by visiting bloomenergy.com.
Transcript
Tag: Latitude Media: covering the new frontiers of the energy transition.
Shayle Kann: I’m Shayle Kann and this is Catalyst.
Andy Lubershane: There’s a lot more uncertainty on the demand side of the equation, largely coming from how much energy data centers are going to be consuming. That’s where you may even get order-of-magnitude levels of uncertainty. In terms of future power demand,
Shayle Kann: We definitely have order-of-magnitude level uncertainty in future power demand, like even just the forecast, take all the different prognosticators on how much load growth there’s going to be in 2035, and there are huge margins between those forecasts because nobody actually knows.
Coming up: five intriguing questions about the future of electricity with Andy Lubershane.
I am Shayle Kann. I lead the early stage venture strategy at Energy Impact Partners. Welcome. Alright, this one is fun. Andy Lubershane, my partner at EIP, our head of research, who you know and love if you listen to this podcast, is back and this time he and I just came up with a list of what we think are five interesting questions that we’re both thinking about largely as they pertain to the world of AI power load growth, and all the technologies that are benefiting from that. We have some general questions about what’s going to happen in this market and then some specific questions about technologies that may be benefiting from it but still need to prove themselves out at scale. I won’t belabor it too much. Here’s Andy.
Shayle Kann: Andy.
Andy Lubershane: Hello Shayle. I’m back.
Shayle Kann: You’re back. Alright. We’re just going to do a bunch of interesting questions that you and I have been going back and forth about all under the umbrella of the thing we keep talking about, the thing everybody’s talking about in energy world, which is the load growth in the electricity sector driven predominantly by AI. So we’re going to talk about some of that specifically and then some of the, I guess, reverberating effects that we’ve seen on technologies. I think we’ll start a little broad and then we’ll get more specific and talk about a few technologies specifically, but let’s start with a broad one.
Andy Lubershane: It’s the theme of the decade.
Shayle Kann: It is the theme of the decade. It’s certainly the theme of this decade so far. Well actually the first question is maybe it stems from that, which is sort of a question of will it be the theme of the next decade or the end of this decade? So here’s the question, right? I think undeniably we are currently in a state of under supply. If the supply demand balance between available electricity capacity at large scale at least and supply is mismatched and that’s why we have these interconnection and that’s why we have the long wait times and that’s why gas turbines are sold out and all this kind of stuff that we talked about and that is definitely the state of affairs today. I don’t think anybody’s going to debate that. Here’s the interesting question, when could supply demand mismatch flip back the other way and if it happened, what would cause it? In other words, when could we enter a state where, oh, wait a second, there was a lot of overbuild and suddenly there are a bunch of empty shells of data centers and or electricity load growth underperforms relative to expectations, et cetera. Do you have a view on when that might happen and why?
Andy Lubershane: It is so difficult for me to try to answer this question. It’s very difficult to imagine that balance flipping the other way. It’s possible for me to imagine the supply demand mismatch becoming alleviated over the next five years or that the supply does not grow as much as we currently anticipate, and so demand is able to keep up better than we’ve been expecting in the next five years. For it to flip back the other way is practically impossible. I feel like my past 18 months, every conversation I’ve had has been talking about all the reasons why we are in the scenario we’re in, which is that supply is falling short of demand by such a wide margin or appears to be falling short of demand by such a wide margin. But stretching my imagination, I think, I guess there’s two ways that could happen. It’s two ways that we could alleviate this imbalance and maybe flip back.
One would be that on the demand side we’re very wrong or the variables causing demand to grow so rapidly change substantially. And the other would be that the variables on the supply side change substantially. I think demand is more likely because there’s a lot more uncertainty on the demand side of the equation and that uncertainty is largely coming from how much energy data centers are going to be consuming. And there’s multiple reasons for that uncertainty, but I think that’s where you may even get order of magnitude, levels of uncertainty. In terms of future power demand.
Shayle Kann: We definitely have order of magnitude level uncertainty in future power demand. Even just the forecast– take all the different prognosticators on how much load growth there’s going to be, how much power demand from AI there’s going to be in 2035, and there are huge margins between those forecasts because nobody actually knows. And like you said, there’s a bunch of different dynamics that could drive it in one direction or another. There’s actually one that I think is interesting sort of question mark on future demand. David Cahn from Sequoia, who’s a friend of mine, has been putting out this kind of series of posts. He started with the 600 billion question maybe a year ago, which was like at that point 600 billion in CapEx announcements had been made in data centers. And he’s like, that has to pay itself back somehow through actual demand in the economy for the services from that AI.
And that’s the number, that’s the bogey that they have to hit. And now the number is who knows how much more. And he’s been doing a series since then that’s sort of looking at where’s all the money for the CapEx coming from, who’s actually paying for it and taking the demand risk. And there’s definitely some risk that there is more investment being made into data centers, then there is real economic demand for those services. But it’s not clear to me how long it would take for that to trickle through to, oh wait, load growth isn’t going to be as large. Part of this is I think that inertia on both sides of this, maybe it’s the opposite of inertia. The momentum on both sides of this is going to mean that if the train stops, it’s going to take a while to stop.
Andy Lubershane: I think that’s right. I wrote a post on Steel for Fuel about this little while ago called “Why does nobody know how much energy AI will consume?” And I think that’s one of the big variables that you just pointed out, which is just the demand for AI is still uncertain. There’s a big error bar in terms of how much individuals and corporations are going to be consuming ai, which is related to how much better are these models going to get? How quickly are enterprises going to figure out how to apply them in their businesses as well as you pointed out just the profitability of the industry. And I do feel like that is a variable that currently feels like it has a tremendous amount of momentum. It feels like almost nothing could stop the train of investment of tens of billions of dollars, hundreds of billions of investment that’s going into AI research and AI data center construction right now.
But there are a trickle of stories now in mainstream press and in business press, the Wall Street Journal, et cetera, that mentioned the word bubble. And it isn’t entirely clear how sustainable this level of investment is and economic bubbles are the sorts of things that can pop quickly and surprisingly they have in the past. And so I’m not taking a position right now personally on how bubbly the behavior is out there in the market, but that is clearly a risk. That’s one vector of risk in energy demand growth world is that something causes investors to fairly quickly lose confidence in what can feel like an economic bubble of sorts. And by the way, that would not just be devastating for the energy sector. In some ways that’s pretty much propping up the entire US economy at this point.
Shayle Kann: It is. I think it’s unlikely that’s going to happen. I mean, well let me say first of all, remember DeepSeek people thought for a minute that that was going to be that, right? It was dramatically more energy efficient per flop and oh hey, maybe this electricity load growth isn’t real. And then two minutes later we were right back where we were. I can see–
Andy Lubershane: Well, I will say, sorry to interrupt. I was going to say DeepSeek was sort of a second degree related to a second variable that I think is causing all this uncertainty in AI driven energy demand, which is not the demand for AI services or the profitability of the AI sector. It’s algorithmic evolution. It’s the fact that we might continue to invest in AI and there might be a tremendous amount of demand for ai, but AI models become much more efficient on both the training side and the inference side such that you can consume the same amount of AI just with 10 times lower or a hundred times lower energy input. We are seeing that happen in real time. AI algorithms are becoming more efficient. They have been for the past 10 years, just a question of whether the pace of efficiency improvement changes dramatically or not. And for a moment DeepSeek made it feel like, oh my goodness, maybe we’re in a whole different paradigm when it comes to AI algorithmic efficiency.
Shayle Kann: Well yeah, ultimately it was a Jevons Paradox challenge. It was like a, “hey, it’s way more efficient, so we’re just going to do more and maybe there’s some limit to that, but I dunno where it is.” The two things you could imagine being big new shocks to the system that kind of blow up the demand story would be on device inference widespread on device inference. So it turns out, let’s just imagine that actually maybe training of new models needs to be done in these big centralized data centers, but we’re going to need fewer and fewer of them as we start to asymptote in terms of their capabilities. And then meanwhile, inference moves way, way to the edge and becomes super compressed and you can put your little GPU on your phone and run a model on it. And so inference actually doesn’t drive the next big wave of AI data centers. That’s one thing I could imagine. The second thing I could imagine in a totally different direction is that you get widespread off-grid data centers, which there’s been lots of speculation and talk about the possibility of taking stuff off grid. You don’t really see it happening yet, but if I’m imagining what are major disruptions, those are the two that feel like at scale they could meaningfully change the picture of the supply demand balance.
Andy Lubershane: I think that’s the one major supply side disruption you can see, right? Because there is the one for off-grid data centers. Exactly. There’s not order-of-magnitude uncertainty in terms of how much power capacity we can build on the grid. That’s something that we can add a certain number of tens of gigawatts through the remainder of the decade, but we’re not adding hundreds of gigawatts. That just is not possible. So the one way you could potentially step up by maybe an order of magnitude, the amount of new data center load you serve is by bypassing the grid and developing data centers in the American Southwest where you have plentiful land to build data centers themselves, massive solar facilities, batteries, et cetera, and a little bit of backup gas generation such that you can build with the same level of reliability and confidence you have and then get a fiber connection rather than have to build a whole lot of new transmission. It’s way easier to imagine building lots of new fiber to a concentrated region of the country where you have really good solar resources than it is to imagine building massive amounts of new interregional transmission lines for electric power.
Shayle Kann: So maybe to wrap this one up, I’ll force us each into just like a yes no here, 2030, let’s pick five years from now, has the supply demand balance meaningfully shifted back in the other direction or we still in this very supply constrained market?
Andy Lubershane: I bet: No.
Shayle Kann: Yeah, I bet. No as well. Okay, so next question is going to dovetail off of this. Let’s talk about winners and losers. We’ll keep this one pretty quick. You got to pick one winner from this whole AI electricity load growth thing and you got to pick one loser and let’s try to make it not the totally obvious winners and losers.
Andy Lubershane: Oh man, I was go with a boring one.
Shayle Kann: No, pick whoever you want. I dunno what’s obvious to you, but okay, pick one winner. Who wins in this?
Andy Lubershane: I mean anybody selling basic power system equipment, right? If you’re making transformers or switchgear or conductor or turbines of almost any kind or engines of almost any kind, if you’re making a way of producing, generating electricity or delivering it, you’re probably doing pretty well right now and I think we’ll be for at least five years to come as we just talked about.
Shayle Kann: Yeah.
Andy Lubershane: It’s a boring one, but yes, that’s one category.
Shayle Kann: I mean I was going to go even more boring and say utilities, which is just an extension of the same thing that you’ve been saying. Yeah, but that’s the right in the supply constrained market. You want to be on the supply side. That’s kind obvious. Well what about a loser though?
Andy Lubershane: Yeah, that’s a trickier question. I mean I think we’ve talked about this on the pod before you and I, or at least you and I have talked about it in some other context, but companies that are depending on low cost electricity for electrification,
Shayle Kann: Electrification by the way, or just depending on low cost electricity period, I’ve been thinking about if you wanted to cite a new aluminum smelter, which is already electrified, you don’t need, you’re not electrifying an industrial process that wasn’t before. If you’re trying to put, I mean, which by the way, we have tariffs on aluminum now. We should be producing more aluminum in the United States. Probably the single thing, by the way, this was going to be my answer too, but I think it’s like an underappreciated problem here, which is if you’re a large industrial electricity load and you want to cite a new plant, you are way at the back of the queue right now. One, you’re going to pay more and, two, siting is going to get really, really hard.
Andy Lubershane: I mean electricity consumers generally unfortunately I think are going to be losers from this boom in demand and it’s easy to sort of blame data centers for that. It’s not the fault of any individual data center development. And there are in fact ways you could see how in any given utility service territory adding a new data center, if it’s done well and you have the right contract in place with a credit worthy data center operator and they’re paying for their fair share of system upgrades and then some that you could actually reduce costs for all the other rate payers from a single data center. The problem is all this pressure collectively on the power system and on the supply demand balance for every element of that system is causing prices to go up all over the place. So any amount of growth is more expensive than it used to be. And in addition to that, we’re encountering all of this growth at a time when the core system needs upgrading and hardening and all that to boot, right?
Shayle Kann: And I think yes, the price is one problem and the sighting is a second problem and they’re both challenging, but I think the citing one might actually even be worse for large loads just because if you’re a hundred megawatts, that’s the load that you need and you’re trying to find a site that can host a hundred megawatts, there’s very little chance that one of the 100 data center real estate developers has not already tried to find that site and buy it and there’s competition for it. It’s just really difficult to do. And then your willingness to pay is going to be lower probably because probably whatever you’re doing is some industrial process that’s lower profitability, at least conceptually profitable relative to data centers where the money is flowing freely. Right? So it is really challenging.
Andy Lubershane: –someday theoretical profitability. Exactly. Yeah,
Shayle Kann: Exactly.
Andy Lubershane: I mean, I will say this is where we get back to the concept of off-grid large power facilities. Large loads cited off grid, right? I mean, we have not seen it start to happen yet, but there’s a strong theoretical case to be made that if you have an industrial facility in particular where you don’t need to be cited in any particular location, so long as there’s people that can work there, so long as you are near enough to highways or rail access, other modes of transportation to bring your goods to market and people to come work at your factory, there should be places now, I think there was a great paper late last year from Scale Microgrids and Stripe analyzing the opportunity for mostly solar powered microgrids in the US southwest. And there’s plenty of room to build those sorts of facilities. And actually manufacturing facilities probably lend themselves to doing that even better than data centers because they really are much less latency sensitive, I mean, so again, so long as you have a route to get your stuff to market, then it doesn’t matter so much where you’re cited.
Shayle Kann: Yeah, although that’s becoming increasingly true of data centers as well, at least depends on the use case, but you’re seeing them get cited all over the place now. Okay, let’s move on to talk about a few specific technologies that have been the kind of, I’d say the darlings or at least some of the darlings of this wave of AI power demand growth. So we’re going to do one interesting question on nuclear, one on geothermal, one on GETs — grid enhancing technologies. Okay, here’s the nuclear one that I think is the interesting question in front of us in nuclear, which is let’s assume there will be a nuclear renaissance in the United States. Let’s just pause it. It’s going to happen. Will it be comprised of a Cambrian explosion of a bunch of new reactor designs? A lot of these, let’s call ’em gen IV reactors, SMRs, micro reacts, all this kind of stuff.
There are dozens of venture backed companies which are gaining a lot of steam and momentum. The U.S. DOE is running this reactor pilot program with 11 of them. Will we see the deployment of 10 or more new reactor types into commercial systems and or will this nuclear renaissance basically could be comprised of AP-1000s, which is the one gen-III-plus reactor that has been deployed internationally over and over again manufactured by Westinghouse or maybe the one SMR reactor that seems to be furthest along, which is the GE Hitachi BWRX-300. That’s the one that’s going to get built in Ontario and maybe TVA territories. That’ll be the first one. So the basic question is are we going to see a ton of new reactors deployed in the market or are we mostly just going to see the one or two that have already gotten mostly through the gauntlet?
Andy Lubershane: I think the short story is there already is a nuclear renaissance happening globally, hasn’t quite caught on yet here in North America or in most of Europe, at least western Europe up. And we can see the answer playing out, which is that there’s just a few reactor designs that are getting traction and basically it’s the ones that you mentioned, especially the AP-1000 at this point and China is very much driving that and it’s actually one area of technology in which China is still buying a significant amount of technology from a western vendor. And I think that that same pattern is going to play out in the nuclear renaissance as much as it happens anywhere in the world. There just can’t be a Cambrian explosion of new reactors. The industrial logic of the nuclear industry just doesn’t lend itself to that. I think best case scenario, it is bad for the industry if you end up with four or five competing reactor designs that are relevant in any given region because really what you need for nuclear to come down the cost curve is you need economies of scale throughout the supply chain and you need to really come down the learning curve when it comes to deployment.
And I would say that the learning curve extends all the way from policymakers and regulators down to people doing construction on the site. And that’s only going to happen if you pick one or two designs basically per region and you just deploy the hell out of ’em. Maybe we’ll get one gen-IV reactor champion, one SMR champion in each region. Any more than that, I’m not sure it’s really sustainable.
Shayle Kann: Yeah, I think in the long arc of history here, the gas turbines are a decent proxy here and there is an oligopoly of gas turbine suppliers. The three big ones are GE, Mitsubishi, Siemens, and they control 70 plus percent of the market. I don’t really see why nuclear should be so different from that. And so if they’re going to end up, and I guess you could argue maybe there are different use cases for smaller versus larger, but that’s also kind of true of gas turbines as well. So I feel like to me, I don’t really see the argument why there should ultimately be many, many of them. And if there are ultimately only going to be a couple or a few, then yeah, it feels to me like you got to make the counterargument to why the ones that are furthest along can’t ultimately be the ones that are furthest along. By the way, with companies that have balance sheets behind them as opposed to all the startup reactor companies, it just feels tough. We have some public companies, right? Like Oklo is a public company now selling reactors, NuScale, a public company now trying to sell reactors.
I dunno how it doesn’t end up just being kind of a bloodbath for a bunch of those companies and then a few of them sort of make it through at the end of the day.
Andy Lubershane: Yeah, I think the next five to 10 years are pretty crucial for any of those more startup-y reactor companies because there’s clearly not room for more than a few of them. So have to, this is the time to cement themselves as the maybe one or two that make it through the ringer and start to gain significant scale and totally agree with you on the comparison to the gas turbine market. You could also say there’s a parallel here in the aviation market, any really big complex piece of machinery that takes a tremendous amount of institutional knowledge, not to mention IP to build and has really high safety and regulatory concerns attached to it.
Shayle Kann: Rocket engines. Yeah, sure.
Andy Lubershane: Yeah, yeah.
Shayle Kann: Alright, that’s nuclear. Let’s do geothermal and we’ll focus on the subclass of geothermal. That’s the kind of new one which is enhanced geothermal. So enhanced geothermal has garnered a lot more attention of late, I would say in part thanks to the fact that Chris Wright secretary of energy in the US is pretty bullish on it amongst a bunch of others. And the standard bearer, the flag bearer of enhanced geothermal is Fervo. Fervo is currently in construction on their first commercial project, it’s called Cape Station. And I think their expectation is that they hope to have a hundred megawatts out of, I think it’s a 400 megawatt project online in 2026. So that’ll be the first enhanced geothermal project ever built commercial enhanced geothermal project ever built. And I think the interesting question to me is if we assume success there, let’s assume Fervo does bring a hundred megawatts online in 2026, how big a watershed moment is that for EGS?
How much does that tell us about EG S’S ability to scale globally and quickly from there? Or is it closer to, and this is not going to be fair to EGS because it is not the same situation, but just to give you the other pole of a possibility. In nuclear fusion, it was like the watershed moment was when somebody reaches energy breakeven, Q is greater than one and NIF did it, but it was on a totally uneconomic reactor. So there’s sort of a race be second to do it on a reactor design that could theoretically be commercial. So it’s not really the watershed moment that I think people thought it might be. So anyway, for Fervo with Cape Station, let’s say they succeed, what does that make you think about EGS?
Andy Lubershane: That’s a great question because I think I have complicated feelings about this one. I think the answer, you could frame the answer either way. In one sense, I think it is a watershed moment. It’s a flag that is planted for enhanced geothermal that demonstrates that it can be done, that it can be done reasonably cost effectively. I’m sure there’s a tremendous amount of learning that happened throughout the development of first the demonstration project that Fervo has built and then onward to this full scale commercial project. And we also know that there are a number of companies that are hoping to be fast followers in the enhanced geothermal space that believe that they too can leverage the existing oil and gas supply chain for hydraulic fracturing and horizontal drilling and some of the same service providers that Fervo has utilized where some of that knowledge now resides and also develop enhanced geothermal projects.
And so there’s reason to believe that Fervo is the leader and Cape Station is kind of the starting gun. On the other hand, my own view is that geothermal unfortunately is inherently going to be a slower technology to roll out even after that starting gun than something like solar was. You could argue we’re at a point Fervo is kind of initiating a geothermal market and that we’re at a point sort of like solar was at maybe in 2008 or 2009 when the very first, relatively small but still meaningful utility scale solar projects were being built. But solar especially at that point when there was so much open land near transmission interconnection access points, it was just so easy at that point, once the economics made a certain amount of sense and there was policy support for large scale solar to roll out extremely quickly and geothermal just inherently because of the fact that there still is risk in drilling and exploration for geothermal resources, that there’s a more complex supply chain that needs to be mobilized.
I think that geothermal will inherently take more time. So while Cape Station is this marquee event and it should feel like afterwards things move very quickly, I think we might be disappointed for a few years while we wait and see more projects pretty slowly move to get off the ground. So I think we’re going to want it to be this inflection point kind of event, but I think that the inflection is going to be much slower and happen over the next five years and that geothermal will really be positioned to take off more so in the 2030s than in the late 2020s. That’s my bet at the moment.
Shayle Kann: Yeah, I think that’s probably right. The big difference between solar and geothermal, obviously solar photovoltaic panels are a product. You can put them anywhere, there’ll be different ambient conditions and they’ll perform differently, but you sort of know what they are and the product is the same everywhere. Geothermal, enhanced geothermal, right? There’s all this subsurface risk that you have to mitigate and it’s all pretty site specific. So we’ve done it in oil and gas. That’s the whole concept here is that we’ve found a way to scale hydraulic fracturing across lots of different geologies and lots of different regions, but there as well, it did take quite a while from the first well, that was ever fracked, for example.
Andy Lubershane: Yeah, and again, I would say one more interesting facet of geothermal I think is that it’s relationship enhanced geothermal’s relationship with the oil and gas industry is also both a positive and a negative. On the positive front. You have this highly skilled engineering and technical field workforce that can very quickly theoretically be mobilized to support the geothermal industry. There’s hundreds of thousands of people around the country that work in oil and gas today, some of whom probably have a keen interest in making a transition into providing cleaner energy and have the skills to do it. So that’s the positive. On the negative side, you also have what at times can be an extremely lucrative industry who can pay those people for their skills. And so I actually think to some extent the timing and the pace of geothermal deployment might also depend on the alternative for that skilled workforce. So like high oil and gas prices, really rich, rich oil and gas market harder to convince drillers to go to work in your geothermal field that’s riskier takes longer to pay off, et cetera. If we have for extended period low oil and gas prices and that market’s feeling slower and less exciting, you may be able to mobilize more of the existing workforce.
Shayle Kann: All right, let’s do one last one on grid enhancing technologies. So grid enhancing technologies is an umbrella term for a bunch of different things. You can describe some of your favorites within them. Here’s the interesting question, grid enhancing technologies on, I think anytime anybody hears about them and what they can do and what they cost for the first time and then they hear about the current situation in the power sector, the obvious reaction to it is like, well, that’s a no brainer of we should do that basically everywhere that we can. It’s a cheap way to get more capacity flowing through the lines that we already have, avoiding the need to build new transmission, which is really, really difficult to get done. It has started to happen, but I think it has been frustratingly slow to a lot of people. Why is the question, I guess, first describe your favorite GETs technologies and then why have they been slow to take off in a macro sense?
Andy Lubershane: Yeah, I think the two that make the most sense to people, and again seem like no-brainers on paper at times are advanced conductor technology. So basically this electric conductors, new wires made of different materials that can carry more power using the same footprint of transmission lines. So basically more power with the same or less weight. And if you can carry more power with the same or less weight on basically the same kind of towers that you’re already carrying power on today, you can theoretically conductor an existing transmission corridor. You can swap out the old conductors for new conductors without changing the height of the towers or anything that would trigger a new permitting requirement or potentially get some of your friendly local NIMBYs involved and just strictly carry more energy over the same transmission corridor with basically no downside. Sounds great. The other category of GETs, which I think deserves a lot of the positive attention that it receives and is often also sounds at least on paper like a no brainer, is dynamic line ratings.
Where today transmission lines are rated fairly conservatively for the worst case or near worst case, environmental conditions oftentimes on a seasonally adjusted basis. So you have a rating in the summer and a rating in the winter just to make sure that no matter what at any given time, you can safely carry the amount of power over those lines that you are carrying. Dynamic line ratings allow you to monitor the lines in real time for temperature and the amount that the lines are sagging or the amount that they’re blowing back and forth in the wind so that you can rate them dynamically in real time and potentially in many cases carry more power over that line than your static conservative rating would allow.
And so yeah, again, the question is why have we not seen more rapid adoption of these categories of technology in the past? Number one I would say is because we haven’t been in the extraordinary demand growth conditions that we are today, even just three years ago. So there wasn’t just enough of an impetus in the past that the market is more conducive for grid enhancing technologies today. But secondarily, I think the reason that we haven’t seen GETs deployed as quickly as it sounds like they should be, there’s a few reasons. One is the natural conservatism of the utility industry. And what I mean by that is this is an industry that is accustomed to operating, particularly transmission and distribution assets with minimal intervention, human intervention as these are big long remote assets deployed everywhere. You don’t want to have to have operators there all the time for decades and decades at a stretch.
This is high voltage infrastructure that has immediate acute human safety implications and environmental safety implications. And so adopting any new technology, this is the most conservative part of a relatively conservative industry. And I think for good reason, it’s just very difficult. There’s a lot of validation required before they’re going to put something into the transmission system that is really fundamentally new in any way. And secondarily, I think because oftentimes there is a lot more complexity in deploying GETs than it kind of sounds like in the initial pitch, the grid is this big complex integrated system and just improving the AmpaCity of one stretch of transmission for example, changes the way that the grid operates in ways that you need to study holistically. If you’re going to change that stretch of transmission line and expect to get higher throughput, you also need to upgrade the substations at both ends, for example. Sometimes that might require you to make other upgrades on the system at the same time. And so my sense is that the adoption cycle forgets is just inherently longer than frankly, I would’ve liked to have seen it be. And I am hopeful that we’re in a different paradigm now because of higher demand, but still expecting that we’re going to need to be patient with rollout of any of this technology.
Shayle Kann: Yeah, I think one thing we’ve learned in this sector is almost never is there a galvanizing a vendor moment. Suddenly everything moves extraordinarily quickly and adoption of new technology occurs overnight like it does in some other sectors. To be fair, and there are good reasons why that’s not true in electricity, but when something does start, a new technology does gain momentum, it usually has a decade plus maybe a couple of decades worth of steam that it can ride on. So the thing for me about GETs is like it feels like the momentum is building now finally in a way that it hasn’t been historically. And if so, then the expectation shouldn’t be that you get a 10x overnight increase in GETs deployments, but it should be that you are steadily deploying more and more GETs for a very long time. And that to me seems like it should be the answer. Yes, there’s risk and we have to get over the risk, but assuming that everything works as we expect it should is an obvious set of technologies to deploy the electric grid period.
Andy Lubershane: Until recently, a lot of GETs technology has been stuck in what we occasionally call it EIP utility pilot hell, which is a bad place to get stuck or if not pilot, hell then really consigned to niche use cases where you have no other option. And it’s obvious that GETs will solve a very, very specific problem in a specific location. I think we’re starting to move beyond that, but again, agree, it’s not going to be probably similar to geothermal. There’s just for different reasons. Inflection point does not mean one to 10, it means one to two and then two to three and then maybe three to five and gradually over the course of 10 to 20 years, you get to that 10x.
Shayle Kann: Alright, Andy, we’re out of time. More questions to come.
Andy Lubershane: My question for you, Shayle, did you realize that Catalyst was going to become also not just one of the top energy podcasts, but also one of the top data center and AI podcasts? Would you have made that bet to yourself a year and a half ago?
Shayle Kann: It’s interesting. I’m constantly trying to self-reflect on whether I’m spending too much time on this podcast talking about this topic and all the little nuances of it and the tentacles that it is drawing, but it is that important I think, and it is that dynamic and it is that uncertain. And so I forgive myself.
Andy Lubershane: I’m just saying the fact that you can call to mine on device inference as easily as carbon dioxide removal is a new thing for you personally and for Catalyst.
Shayle Kann: It’s true. This is what’s happening in the world, so we got to figure it out. All right, Andy, we’ll talk again soon.
Andy Lubershane: Thanks Shayle. I’ll see you.
Shayle Kann: Andy Lubershane is the head of research and a partner at Energy Impact Partners with me. This show is a production of Latitude Media. You can head over to latitudemedia.com for links to today’s topics. Latitude is supported by Prelude Ventures. This episode was produced by Daniel Woldorff. Mixing and theme song by Sean Marquand. Stephen Lacey is our executive editor. I’m Shayle Kann, and this is Catalyst.
