This was supposed to be the year when clean energy rode the AI wave onto Wall Street. Instead, the pipeline for energy IPOs is dwindling. What happened?
A combination of high interest rates, uncertain AI revenue growth, investor concerns over capital expenditures, and local friction slowing projects are all putting pressure on AI companies — and the energy providers serving them.
Companies that planned to go public this year — Holtec, The Nuclear Company, SB Energy, and Form Energy — have all shelved their plans. This year’s breakout geothermal listing, Fervo, has seen its stock price drop 70% since May. And IPPs have taken big hits to their valuations.
This week, we dig into what the uneasy market means for power development.
We’ll look at shifting market conditions for the AI trade, the reasons for delayed IPOs, alternative pathways for exits, and who may come out ahead. We’ll end with a look at some “wild cards” that may shift the picture, including China’s AI strategy, the role of distributed data centers, and power pricing.
Open Circuit is brought to you by EnergyHub. EnergyHub works with more than 230 utilities to coordinate over 2.6 million DERs and more than 3.6 gigawatts of dispatchable flexibility through a single platform designed to perform when it counts most. See what that looks like at EnergyHub.com.
Open Circuit is brought to you by Giving Green. Giving Green researches the climate nonprofits best positioned for the next big climate win, and sends 100% of every dollar donated to the Giving Green Fund to them. Make your first gift before 2027, and it’s matched up to $500. Go to GivingGreen.earth/Circuit or select Open Circuit at checkout.
Open Circuit is brought to you by Everyday Electric. Everyday Electric simplifies savings at home, helping support a more reliable, affordable grid. Operating the country’s largest residential VPP, it coordinates 6+ GW of capacity across connected products. Learn more at everydayelectric.com.
Transcript:
Stephen Lacey: This was supposed to be the year clean energy rode the AI wave onto Wall Street. Instead, the wave might be breaking sooner than expected. Companies across the AI industry are seeing sharp stock price drops as investors worry about infrastructure spending. And it’s hitting energy companies. Holtec, The Nuclear Company, SB Energy and Form Energy have all shelved plans to go public. And Fervo, this year’s breakout geothermal listing, is down 70% from its first day of trading. What is happening?
Three numbers tell the story. $8.5 trillion. That’s how much revenue AI data center owners need to bring in by 2030 to cover their spending, their debt, and a reasonable return. Gartner says they’re going to come up nearly $5 trillion short. Then 5%. That’s the level the 10-year Treasury yield broke in September, the highest in three years. That makes every dollar of the AI build out more expensive to borrow and every pre-revenue energy company harder to value.
And finally, 71%. That’s the share of Americans who oppose data centers in their communities according to Gallup. The backlash is stalling projects and forcing investors to rethink how fast the companies serving data centers can grow. So if the AI boom is getting more expensive and less popular, what happens to the energy companies that bet on it? That’s coming up on Open Circuit from Latitude Media.
Welcome to the show. I’m Stephen Lacey, the executive editor of Latitude Media. Caroline Golin is with us. She’s the chief growth and policy officer at NRG. Hi, Caroline.
Caroline Golin: Hello. Hello. Glad to be back.
Stephen Lacey: Jigar Shah is the co-managing partner of Multiplier. Did you have fun in Austin, Jigar? When I saw you at Flex Summit, did you do any two-stepping?
Jigar Shah: I mean, I crushed it. You and I went to the same dinner.
Stephen Lacey: That was a very long table. I sometimes have a hard time.
Jigar Shah: I did not think that the structure of that dinner was good. I found that I was able to speak with the three people next to me and nobody else.
Caroline Golin: Maybe good for everybody else. I don’t know.
Jigar Shah: They did that on purpose. I get it.
Stephen Lacey: All right, so it’s a big week. Let’s get into the public markets, which are seeing shifting headwinds for clean energy companies looking to IPO. We had this combination of high interest rates, skepticism over AI revenues, slower than expected development due to local resistance and worries over the debt load now driving development. It’s all making investors wary, and we’re seeing this play out across the market. So many companies that are serving or in close proximity to AI are seeing big hits in their stock prices. And at the same time, we have Anthropic, which has a planned $1.5 trillion IPO that’s sucking up all the attention away from other companies that are planning to go public. So as a result, there are a bunch of companies in this space that were expected to benefit from all this activity. They’re delaying IPOs. They’re reportedly getting signals from investors that their appetites have diminished, and the list is Holtec, The Nuclear Company, SB Energy, and Form Energy.
So what is going on here? Caroline, what’s your read on how the narrative in the market has shifted?
Caroline Golin: Yeah, I mean, I think it’s pretty straightforward, but it’s interesting that it’s happening right now and I will be interested to see if it shifts after midterms. I think that the market gave the AI race a really, really long leash, maybe a longer leash than it should have, and they were investing and they were valuing based on potential and they were really comfortable with companies being fairly well leveraged. And now the politics have kicked up, capital costs have gotten higher. Turns out it’s hard to build power plants, takes longer than you think, and some of them are burning out and people are breaking things left and right. And so you have to be able to demonstrate an actual return and you have to be able to demonstrate more than just, I think, early stage progress to get the wins that I think people were able to get maybe a year, year and a half ago based on an LOI or based on some type of potential partnership.
I think the leash has gotten tighter and I think it should, frankly. Unfortunately, I’m not sure that the companies that are driving the demand here are being as judicious with the way they’re placing capital. So we’ll see. But I am interested how the midterms will change. I think a lot of people are waiting to see if the politics will get worse, it’ll get better, and I personally don’t think they’ve changed irrespective of whether there’s a blue wave or not.
Stephen Lacey: Yeah. Jigar, what’s at the other end of that leash tugging? Is it interest rates? Is it earnings to spend ratio? Is it pre-revenue companies having a harder time in the market? What do you think it is?
Jigar Shah: Well, it’s all of those things, right? I mean, for those of us who have been on this Open Circuit journey, I think we have told everybody to short these stocks for over a year and folks were like, “You’re an idiot. Why would you do that?”
Stephen Lacey: Wait, short what stocks? Are you talking about energy stocks or the data center?
Jigar Shah: Well, I think it depends. Each one of the names has a different story, but I just think that for a long time people are like, “Oh no, Jigar, it’s going to be so easy to run data centers off grid.” I mean, there’s still dumb people out there who think that they can run data centers off grid. And so those people who sold technology into those projects are still trying to convince people that you can run data centers off grid. I still don’t think we’ve won that argument, but if those stock prices went up and now all of those projects are delayed, crazy delayed. It’s not just six months delayed, it’s a year delayed. And now they’re trucking in natural gas to those projects because the pipelines are delayed. And now when you think about just how laughable it is, I feel like I’m literally watching a Charlie Brown show with a football and people are like, “Oh, that was not predictable at all.
I don’t think that anyone thought that we could see that coming.” So now that it’s happened, they’re like, “What should we do with the stock price?” Well, obviously if the compute’s not online, well then you can’t make money off of it. That’s how this works. You can lie to yourself as much as you want, but at some point if the compute is not generating tokens and you’re not getting revenue for those tokens, then there are consequences to that. And so that’s what’s happening on those sets of stocks. Then when you think about the Fervos of the world, look, I’m not going to badmouth Fervo or anybody else. The whole thing that I wanted Fervo to do was to raise $2.2 billion. They succeeded at doing that. Now they’ve got four years to prove whether they can use that $2.2 billion to make sure that all of their technology works and it can be replicated to five more projects and all these other things.
I don’t care what their stock price does because they’ve already raised $2.2 billion and now they’re going to have a chance to succeed. I can’t say the same thing for some of the other names. If you look at a lot of the other geothermal names, they didn’t go out and go public. They don’t have $2.2 billion. They have maybe $200 million. So we’ll see when they run out of cash and whether the markets are open and are willing to give them more runway when they run out of cash. When you go to the Holtec side of things, Holtec has not yet turned on Palisades, although it’s supposed to be turning on any day now, and I’m hopeful and I’m super excited for when that occurs. And then you’re going to have to build a 300 megawatt reactor there. They’ve filed for two new Holtec 300s at that site, and we’ll see where that goes.
But I think that there was some trepidation when the week was that they were going to go public, that folks were going to buy into the story that the hyperscalers were going to pay 15 cents a kilowatt hour for all that power. And so we’ll see where the contracts land.
Caroline Golin: I mean, I think on Fervo, I don’t think the lower valuation is something to freak out about. I mean, I think that across the entire industry, stocks are down, and that’s largely because you have to show more durable checkpoints on construction and on meeting your milestones. And I don’t think that’s a bad thing. I look at Fervo and I’m like, okay, if you can do what you said you were going to do, I think that will probably drive more investment in the company. But I think there was a race and an excitement. And I think the other thing that’s happening here is that we had a flood of new entrants on the nuclear side, and there was a lot of political support, still is, on the nuclear side. And I just think that it was a little bit of the dog that caught the car on the big AI build out.
And so there was a lot of interest, and we’ve talked about this before. And then there was a bit of a course correction because there’s also a bit of sanity going on now. And I think you see this all the time in markets. I just don’t think you see it with such demonstrative impact of capital that’s on the line. Everyone gets spun up about something, you get smaller companies that get pulled along for the ride. Oh, you have a deal with one of the five companies that’s 50% of the overall market right now. Okay, you’re good to go. But now there’s just a little more sanity on it and there should be.
I think there’s a lot of nuance in some of the stuff around SpaceX and Amazon and everything that’s going on there and Anthropic. But generally speaking, I think lower valuation is a result of three things. Market prices are low across the country, so there’s that. Hitting your construction timeline is people are slipping because of politics and because of supply chain, which we all knew, but no one could guarantee. And so there’s that. And the last one is that I actually, and we don’t talk about this as much, but there’s some general skepticism that the demand for the chips that are sitting on barges in ports –
Stephen Lacey: Yeah, in warehouses.
Jigar Shah: In warehouses at Amazon.
Caroline Golin: It’s there. And I think that the narrative has always been, well, they’re sitting there because there isn’t enough power. But I think there is some recognition that power is part of it, but also demand is potentially not the hockey stick that everyone once said. And so this is a course correction. And I actually, I think I’m glad to see it because if it didn’t happen, then I think we do run the risk of some massive crash in volatility. Yeah.
Stephen Lacey: Yeah. And I would add one to that last point, which is if you look at Anthropic’s S-1, the bulk of their revenue relies on two customers. 2%. And if you look at AI… No, two main customers. Yeah, I forgot the percentage of revenue, but they’re heavily dependent on two big customers. And then if you look at the bulk of AI revenues across the industry, you have a very small number of companies that represent the bulk of AI spend. And so it’s extremely concentrated and there’s not really the proof yet that this is durable in terms of revenue across a wide spectrum of industries.
Caroline Golin: Yeah. And it may get there, but I just don’t think it was going to get there in three years. That’s the thing is everyone’s like, oh, the world is going to change.
Jigar Shah: Caroline, this time was supposed to be different.
Caroline Golin: Yeah. What was the last time?
Stephen Lacey: Yeah, what was the last time?
Jigar Shah: Well, the last time was telecom fiber and the time before that was something else. I’m just saying that every news article was like, we read all that history and this time is different because this is such a game changing technology and it’s going to put everyone out of a job in four weeks.
Caroline Golin: Well, it is a game changing technology, but I just think that it also requires, and I’m going to say something maybe a little controversial, it requires a literacy around the technology and a use case around the technology that wasn’t pre-built before it disrupted. So there’s that. Coming from the corporate side, it’s like you actually have to take time to figure out where this is going to be best used and that can slow integration in a way that maybe I don’t think people anticipated.
Stephen Lacey: Yeah. And I would push back on that a little bit, Jigar. I mean, I think there’s a question about how companies can fill this revenue gap, but there’s demand for the computing inside these data centers. I mean, it’s not like the railroads, these small companies in the railroad era that just built railroads to nowhere. These are the most highly capitalized companies in the world that are building data centers to meet actual demand. So there’s definitely a different dynamic.
Jigar Shah: I mean, potato potato, right? I mean, all I’m saying is that on this side of the fence, you’ve got people who want to build only 1,000 megawatt data centers, which are literally building the city of Cincinnati in every single place. And then on this side, you have people that want to build five megawatt data centers at substations. Utilities are saying, “Here’s where we have excess capacity. We would love for you to fill it because then we could reduce costs for everybody.” And the data center companies are saying, “We hate you. We would like to build the city of Cincinnati in every place and make your life a living hell.” So if it was so hard for them to get compute, why wouldn’t they work with the utility companies and say, “Yeah, we’ll take those five megawatt sites and actually put compute in all those long lines so that we can actually increase utilization.” And maybe they are.
Nvidia made that big announcement this week and Crusoe made a big announcement on their five megawatt unit, et cetera. But I’m just saying we are in a situation right now where it’s very obvious that on the one hand, tokens are going up in price, to your point, because we’re short compute. And on the other hand, there are people saying, “Here’s some really great solutions.” And the folks who are buying it are saying, “No, we know better than you. We are so much smarter than you.” And the stock market is inspiring that they actually are.
Caroline Golin: I mean, they’ve also sunk billions of dollars into what they thought was better, right? And so they don’t have the –
Jigar Shah: Without any knowledge whatsoever by which to say that it’s better, just with their pure masculine energy.
Caroline Golin: Masculine, of course.
Jigar Shah: So I don’t care. They could lose half their stock price and I would cheer because I’m so sick and tired of them.
Caroline Golin: But you should care because if they do –
Jigar Shah: Why should I care about the masculine energy? The whole market
Stephen Lacey: Comes down.
Jigar Shah: Well, so I’m fine with the feminine energy coming back.
Caroline Golin: This is why I still –
Stephen Lacey: Let it loose. Let it loose.
Caroline Golin: But there is no… So you look at SpaceX, still revenues, positive, still lots of deployment in AI. Market prices down. Stock prices down.
Jigar Shah: They reduced their CapEx. They’re the first company to actually reduce CapEx. They restated and reduced their CapEx, right?
Caroline Golin: But that shows discipline, does it not, or no?
Jigar Shah: I mean, that shows that they’re finally starting to listen to the non-masculine energy. I just think that at some point you have to look at the data and say, for these companies who are the most profitable companies in history to have negative cash flows, at some point you’re like, “What are you doing, man? Seriously.” Meta just announced that they went into negative cash flows. I mean, even Tesla, the only reason they made money really this last quarter was because they had a one-time increase in tariff refunds of $700 million.
Stephen Lacey: Yeah, but their revenue has slowed as well.
Caroline Golin: Yeah.
Stephen Lacey: Their revenue slowed.
Caroline Golin: But it’s across the board. It is across the board in the energy space.
Stephen Lacey: Yeah.
Caroline Golin: Cashflow revenue is down.
Stephen Lacey: Who’s most exposed? So obviously the neoclouds are the ones that are taking the biggest hit. Who else do you think is most exposed here right now?
Caroline Golin: On which side of the aisle?
Stephen Lacey: On the data center infrastructure side.
Caroline Golin: I tend to agree with Jigar on this. I think the development companies that went all in on the gigafactories are probably most exposed.
Stephen Lacey: Interesting. And then what about the energy side?
Caroline Golin: Well, on the energy side, anyone who didn’t take a DOE loan, probably.
Stephen Lacey: Why is that? Why is that?
Caroline Golin: Well, it was a joke. Well, kind of not a joke. I mean, I think everyone’s exposed. I mean, look at PJM, look at ERCOT. I mean, market prices are down. Places that were sure bets are paused. I mean, everyone’s exposed. I think the companies, and I think what you’re seeing is that everyone is saying, “Okay, how do we effect discipline then?” We all got a little crazy. We all went to the party. We all drank the AI Kool-Aid, but I think a lot of energy companies, and I talked to a lot of the supplier community and even the smaller IPP space are all recognizing we’d like to be here for 30, 50 more years. So how much of our balance sheet are we willing to leverage to do this? And I think a lot of investors are saying, “Please don’t leverage a lot of your balance sheet to do that.”
Jigar Shah: I was drinking NA beer, to be clear. Not Kool-Aid. I was not drinking Kool-Aid. I just think that this was so obvious in coming and all of the people who are like, “I’m going to take old ship engines and power a data center with it. I’m going to rip off these engines off of old airplanes sitting in the desert in Phoenix and run data centers off of it.” Good riddance to all of them.
Stephen Lacey: But they’re still doing that, aren’t they? I don’t think that’s totally gone away.
Jigar Shah: Yeah, they are. But good riddance to their investors, good riddance to everybody who thought that basically destroying the planet to be able to power AI data centers faster by two seconds was a good idea. I hope they all lose their shirt.
Caroline Golin: Well, folks, that’s it.
Jigar Shah: No, but I’m just saying, look, I mean, there could have been… I get the fact that people were talking too much about climate change, and then the Trump administration came in and said, “All of you guys are talking about climate change too much.” But it’s not like the IPCC didn’t just come out and say that we’re definitely way past 1.5 degrees. Now we’re at 2.8 probably. We’re going to have to look at geoengineering. And then on this side, the guy’s like, “Oh, let me take the dirtiest possible power plant and figure out how to run a data center off grid in the middle of Utah.” And now they’ve got to truck in natural gas because the pipeline didn’t come in on time. And I’m sorry, I’m not crying for you.
Stephen Lacey: I want to talk about some of the delayed IPOs, and so many of these companies are just looking at the market and saying, “This isn’t the right time,” but there’s a whole range of companies that are benefiting from the development. Hundreds of billions of dollars in development are getting frozen, but there’s still a lot of development happening, and that’s really good for companies like Voltus, which is seeing gigawatt scale demand now for VPPs. It’s great for the battery suppliers. There’s just a crazy amount of batteries. The electronics component suppliers that are developing the electrical architecture, this is still a really good market for a lot of the mid-size companies along this industry.
Caroline Golin: Yeah. I mean, I completely agree with that. We’re being more disciplined. And frankly, so I think that a lot of the companies that have hesitated to go public, I think there is also nuance there. I mean, some of them are hesitating to go public because they realize they can’t meet the scrutiny that exists today that didn’t exist a year ago. That’s not necessarily a bad thing for the market. And I think some of them are not going public because there is a general sort of damned if you do, damned if you don’t, if you’re attached to specific hyperscaler load in this moment. And so I think that the former is good. I think the latter could play itself. People are banking on it playing itself out within the next couple of months, and I’m not sure that that’s going to happen.
Jigar Shah: Yeah. I think when you look at the Voltus type deals, this is a no regrets approach because if the AI data center load growth doesn’t show up, using the grid that we already have more efficiently doesn’t lose money. Being able to flex load with the same level of dexterity that we flex supply still makes the whole grid operate better. And so that’s a no regrets approach. And so yes, the ideas that are no regrets approach is like battery storage or load flex and all that stuff. Yeah, they should keep going regardless of what happens because it just makes everything run better. But my point is that when you look at the filings at NIPSCO in Indiana this week, and they’re now trying to rate base $3,500 a kW, to put that in perspective, that’s $400,000 a megawatt year for the capacity. The PJM at the rate cap is at $140,000 a megawatt year.
So they are overpaying for gas by 3X, and for what purpose? That’s going to get rate-based for 30 years. And if the AI data center load doesn’t show up, then what? Microsoft or somebody else pays for 80% of their capacity payment for a turbine that doesn’t run? I mean, this is the height of insanity in terms of how people think about what we should do as we move towards a decarbonized grid.
Caroline Golin: I don’t think the PJM cap or the market is reflective of what it costs to build new gas though.
Jigar Shah: Okay, but then give me a number, right? Because four years earlier, NIPSCO and their IRP put in $1,100 a kW for that same exact gas plant. Same thing’s true for Duke. Duke’s at $3,000 a kW now. They were at $850 a kW four years ago. At what point do you say, “Huh –
Caroline Golin: This is just too expensive.
Jigar Shah: Maybe the old way that we were doing things is not the right way to do it in the future.” I just think that the reason people are so pissed off, and you look at poll number after poll number and how governors have to be anti-utility to even get elected these days is because I just think that people are just sort of vibing it right now. And I was like, “There’s math that you can do around what the easiest way is to be able to meet this moment, and for you to do the exact opposite should get the vitriol that I’m spewing.”
Stephen Lacey: So you’ve laid out some buckets of reasons there, Caroline. Do you want to apply that to any of the companies? I mean, so SB Energy seems like the kind of company that’s facing… This is a renewable developer that is now doing integrated data center development, these mega campuses. They reportedly got pushback on the $50 billion valuation they were seeking. Form Energy felt like the market was not going to support its price goal. How would you bucketize some of these different companies?
Caroline Golin: Yeah. I think some of the other nuance that’s going on here is that similar to the cost of building new natural gas, the cost of raising capital and the cost of meeting projections from a year ago today went up by 20% across the board. So if you were preparing an IPO, say a year ago to go public this year, what’s actually going to bake the cake is higher than it was, and it’s gone up dramatically. I think if you haven’t within that year shown demonstrative checkpoints in your advancement, then you’re sort of caught in between these two bookends and that would make me question whether to go public or not. I think that there are a lot of factors here that are, and not all of them are working at the same time on every single company, on every single issue. I do think you can be a phenomenal technology and you can still have very successful development and perhaps need to be a little more conservative in the way you think about going public or what you need to show.
I’m not sure that that’s a bad thing. Do I think because people are waiting to go public that that means they’re going to go bankrupt? No, not at all. I just think that it means they’re going to have to show more due diligence on what they’ve done and they just have to be able to show a little more feasibility in their construction and that’s okay. I mean, the reality is that is the scrutiny we held this market to for decades and then we lost it for a little while. I’ve said this before on the podcast, there were times at Google where I was like, “We are kind of making and breaking things just on LOIs, and I’m not sure that’s the right way to invest.” So I think we are returning to the level of scrutiny that was there. The other thing, of course, that’s happening across all of this is IRA dollars disappeared.
And so if you’ve been planning this for years, then you’ve got all these different forces working against you. I just think it’s a higher level of scrutiny. I don’t think it means this industry’s going away. I don’t think it means it’s a bad investment. I don’t think it means these technologies won’t win out. I just think they’re having to be a little more conservative, and perhaps them being more conservative is going to force the overall… What were you drinking? I was drinking Kool-Aid. I don’t know. You drinking –
Stephen Lacey: Non-alcoholic.
Jigar Shah: NA beer. What the hell?
Caroline Golin: A non-alcoholic beer?
Jigar Shah: I think to Caroline’s point –
Caroline Golin: Oh, I was thinking PBRs.
Jigar Shah: I mean, to Caroline’s point, I think some of these companies have not hit their milestones. I love Form Energy, but they do not have a very large project up and running at scale. That’s collecting data. It’s just what it is. And I’m not badmouthing them. I’m not saying that they’re going to fail –
Caroline Golin: No, I think they’re great.
Jigar Shah: But they need to have one of those up and running. And the one at GRE is not yet up and running. But I also think that a lot of the investors in Fervo stock should know what they’re getting themselves into. I mean, it is a new way of making geothermal. I think it is well studied and DOE and others have put a lot of money into it, but now they have to get all of their techniques right. And so if you don’t want that risk profile, you shouldn’t own the stock. It’s not a guarantee. They haven’t had profitable quarters for the last five years before they went public. They represented themselves exactly the way they should represent themselves, which I think is fine. On the nuclear front, same thing. Look, I love X-energy, but they still haven’t built their first reactor yet. I mean, they’ve got a partnership with Dow.
They’ve got a couple of other partnerships with Amazon. They’re trying to get stuff done. And so they’re public. It’s going to be a very volatile stock because they haven’t built their first reactor that’s commercial, that’s NRC approved. Kairos obviously has taken a different route where they’ve got a partnership with Google, but they haven’t gone public and they’ve been trying to raise money from private sources, different direction on their part.
Caroline Golin: And I wouldn’t say that… Again, I think that the more conservative nature is not reflective of an overall erosion in power demand. I think it’s more reflective of a very excited investment community that woke up to the realities of building power in this country. I still think it’s going to get built. I still think these technologies, are all of them going to win? No, of course not. And they shouldn’t, but should they be held to a bar that ensures that they are durable? Yes. And frankly, the bar of durability sort of slipped for a little while. And we’ve talked about some of those companies and Jigar had to apologize to them for what he said afterwards.
Jigar Shah: Did I?
Caroline Golin: I think you did. But that bar slipped.
Jigar Shah: It depends on who we’re talking about.
Caroline Golin: That bar slipped. Now the bar is going back up. I don’t think that that’s a bad thing.
Stephen Lacey: Yeah, that all feels right to me. Jigar, part of what you do with Multiplier is to help companies think about their exit strategy. What are you telling companies about the pathway to going public?
Jigar Shah: Well, last year I told them the market is fully open and if you want to go into the public markets, you have a short window by which to do it.
Stephen Lacey: And that window’s closed?
Jigar Shah: I think that window is on pause for sure. You never know whether something’s closed or open until you look through the rear view mirror, but right now it certainly looks paused. And I think it depends on the company and what the characteristics are. But I do think that as Caroline suggested, we still need more power on the grid. And so for those folks, like if Voltus is able to meet this Google contract in the next 12 months, which I suspect they can, then they’ll be given a chance to quadruple that contract again or 10X that contract again because we are short capacity. Every one of the hyperscalers is short 10,000 megawatts of capacity. Contracts they signed are not coming through on time and are getting indefinitely delayed. And so they need capacity and Voltus is one of the solutions for that capacity, but they’re doing deals like one commercial customer at a time.
And so a lot of the hyperscalers are nervous and are saying, “How many of those commercial customers exist? Can you really get 10,000 megawatts of flexibility or only 1,000 megawatts?” They got to prove themselves. So they now have the chance to do that. And so I think that’s true across the board, whether it’s residential battery suppliers, whether it’s the thermostat people like Everyday Electric now, or whether it’s other folks, everyone is going to get their chance to prove themselves. All of the nuclear companies that got grant funding from the US military for five Army bases also will have a chance to prove themselves and they’re going to get paid cash to build those reactors. And now they’re going to have to deliver or not deliver and it’s in their hands.
Caroline Golin: I think one of the other aspects here that I don’t have an answer to and actually think about, maybe we can bring it back on another show is that in the past two years, you’ve also seen the movement and the consolidation with the private equity platforms. And so that has presented a different approach to the way that some of the deal making happens and the way that we think about debt, the way we think about consolidating these supply issues. And I don’t know how much that has changed the way that even smaller companies think about their trajectory. If I am a mid-size development company, is my goal exit? Is my goal to get onto one of these platforms?
It’s a very different space than we were in a handful of years ago. I mean, between KKR and BlackRock and Blackstone, I mean, now they’re consolidating these huge development platforms and they’re picking the different pieces. And I think it’s probably the way a lot of it gets done moving forward. And so now there’s another path to it and you can go public and then go there. Or you can not. And so I think that there is, I mean, you saw this Invenergy, this $3 billion stake in Invenergy from private equity. Invenergy’s not public. They didn’t go public. They did that instead. So I think that there are more paths towards profitability and I think that there should be more scrutiny on going public considering how much the market is determined by a handful of companies right now.
Jigar Shah: Well, I mean, Charles Schwab has that ad campaign that they have now that’s called Filthy Comfortable.
Stephen Lacey: I don’t know.
Jigar Shah: I’ve seen this.
Caroline Golin: I don’t watch television.
Jigar Shah: And that’s what that represents, right?
Stephen Lacey: But what is it? Tell us the ad.
Jigar Shah: Well, there’s a bunch of people who are like, “I’m very filthy comfortable.” And it’s basically if you try to go public or you try to have these big exits, then you’re trying to become filthy rich. If you go to private equity, no one becomes rich. That’s not how private equity works. They become rich.
Caroline Golin: They become rich, right. That’s right.
Jigar Shah: You work for them, you get a decent salary and decent bonuses and a bunch of promises that they never keep on the back end. And that’s what it is. But you get unlimited capital by which to do your core business and that’s fun and it’s useful because you get to do a lot of great stuff, but you never become filthy rich, you get filthy comfortable.
Caroline Golin: Yeah. But on the flip side, if your stock is dropping and your investors are pissed, similar handcuffs.
Jigar Shah: Oh, I’m not suggesting otherwise, but I’m simply saying that that’s what a lot of people are deciding between when they think about what to do with their platform. Because a lot of the private equity guys also, to them, these are all ones and zeros on a page. It’s not clear that they actually really even care about these businesses that are underlying there. When you look at what’s happened to the renewable natural gas space and a lot of the consolidation there. And so I think that one of the big challenges that we have is on the one hand we’re talking about technologies and areas that follow the law of capitalism. They save customers money, the customers want to sign contracts, they deliver solutions, et cetera. And then on the other hand, we’re talking about technologies that are on that bridge to bankability. They have not yet proven themselves and they are in need of a lot of capital to be able to get to the other side and be able to prove that it works and that they can get their costs down and they can be mass manufactured and all of those things.
And the question that we’re asking ourselves is now that the IRA is effectively dead and the Loan Programs Office is not really leaning into more difficult names, what is the right pathway for these folks to succeed? Fervo has taken the public offering pathway. Others tried to copy them. We’re talking about private equity, but I’m not sure that private equity is going to give a lot of money to companies that lose money systematically and need another five years to prove whether they’re going to be successful.
Caroline Golin: No. And so you said something that we haven’t really brought in, which I think is interesting that we haven’t brought in, which is that all of these companies started, and we’ll put the IRA money on the side for a second. So whether you’re clean tech or not, but all of these companies started with a general belief in open markets and innovation, the opportunity in this country to succeed in sort of a capitalist system. And within the past year and a half, we’ve become more populist as a country than we’ve probably been in a hundred years. And so when I call my friends in this space and I say, “What’s the biggest challenge? Politics.” It doesn’t matter if you’re Constellation or if you’re Fervo or if you’re Tesla, I mean it’s politics. And I do think if that hadn’t emerged, I question whether the discipline in the market and the conservatism would’ve transpired.
So in many ways, and we can see the silver lining here, is that democracy is working in some way in course correcting. And so I just think that those of us who work in the business of power, even though we’ve always known on the renewable side or the clean energy side that it’s all about regulatory and policy alignment, I think generally the power industry underestimated the strength of the populist pushback. I don’t think the clean tech space ever thought that building a power plant would mean that you are synonymous with the evil overlords of AI. And I think that that emergence is probably a far more weighted variable in this than we’re giving it credit for.
Stephen Lacey: I agree with that assessment entirely. Let’s hit one more angle to this, which is the IPPs. So NRG, Vistra, Talen, Constellation are all down on the year. I think there’s a combination of things happening here.
Caroline Golin: It’s been great. It’s ’cause I joined. It’s ’cause I joined.
Stephen Lacey: I mean, we’ve got a bunch of shifting market rules in PJM, low power prices in Texas, making it hard to build new power. These companies are also seen as AI derivative companies. What is the market telling us about IPPs right now, Caroline?
Caroline Golin: I mean, market prices are soft. Most of the investor bases in the IPP space I think are generally split. There’s some investors, and this is broadly, I’m not just speaking for NRG. I think there are investors who want the IPPs to go after the golden ticket. But I think a lot of investors are like, “No, we don’t want you to. We want you to just stay steady Eddies so you can…” I mean, the IPP space, for what it’s worth, has been sort of a plug and play investment community for decades. And you ride the weather, prices are up, you make money, prices are down, you don’t make money. Now there’s this new layer of signing a bilateral and getting durable cash that way. Okay, how much of your balance sheet are you willing to leverage? Constellation’s deal with Google was a direct signal, which is a great deal.
I’m very supportive of it, but it was a direct signal that those uprates were not going to clear the market. They weren’t going to clear the market, and so they needed an offtake to do it.
Jigar Shah: But that’s true for all of these technologies.
Caroline Golin: Right. Exactly. Exactly. But if you are an industry that is 90% merchant and market, that’s why your stock is going to be down. I mean, prices are low in Texas. The cap in PJM is lower than it would need to be for a lot of the IPPs to develop anything new.
Jigar Shah: We have to separate some of this stuff, right? I think if you’re talking about new assets, the IPPs, broadly speaking, were started by, let’s say AES back in the 80s, and they were using PURPA to bring on technologies. Yes, if you’re building new natural gas right now, new nuclear uprates, none of those things pencil without a long-term offtake agreement from a hyperscaler that’s guaranteeing you a return on investment at a price that pencils. So I think that makes total sense.
Caroline Golin: That’s the story, right? I don’t think there’s anything particularly interesting.
Jigar Shah: It’s not the whole story. So then if you look at Texas, I’d say that, I mean, you tell me, but I don’t think that the vast majority of batteries were deployed by the traditional IPPs. No. I think there were a lot of new companies that had a new business model with new metrics around.
Caroline Golin: Well, I think the business model was there was a six-year window to capitalize on. And if you look at returns in ERCOT for storage, it’s decreasing because essentially it’s going to run out of its ability to chase the margin from peak. And that’s a natural occurrence with a technology like storage, unless you start pairing it and start integrating it and blending it in different ways. But no, that’s absolutely right. And so prices are down and prices are soft. The other thing though in ERCOT and PJM is that… It’s interesting. I was actually talking about this with a friend the other day, which is that there was like PJM is looking at ERCOT and going, “Man, that’s the way we should have done it. There’s no capacity market there. The capacity market is what’s killing us.” And then ERCOT’s looking at PJM and would never admit this, but there’s a number of entities within ERCOT are going like, “Without a capacity signal, we’re going to erode plants.
Power plants are going to go offline if prices continue to drop.” That’s just economics. And so there’s this interesting interplay between the two. Neither of them I think have it right. And I think the truth is, and Jigar, you’ve said it a million times, it is time to have a very different view of what market structures need to look like moving forward. And I don’t think there is a perfect market out there. And I think that I will say it now, but I think there will be legislation in Texas next year that asks this. Is the market today what we need going forward? And I’m hopeful that there are people with interesting views.
Stephen Lacey: All right, so let’s speculate to end here. What are your wild cards going forward? I think we can see data center resistance as a wild card. I don’t think anybody two years ago saw just how powerful the resistance would become. Right now, a lot of people are sitting here freaking out about the cybersecurity threats of agents. Banks now have agents swimming around in their systems making transactions. They don’t know which ones are malicious. You have people who don’t know much about hacking now able to use open weight models to send teams of hackers out. I don’t know how this is going to affect markets, but you can see a world in which AI companies take a hit if there’s this massive cyber event that wreaks havoc on the financial system. So this has downstream consequences. I’m kind of curious about what’s in the back of your mind.
What are the crazy things that could happen that actually might not be that crazy at all?
Caroline Golin: Jigar buys a tweed vest. That’s the craziest.
Jigar Shah: A Buckaroo
Stephen Lacey: Jacket.
Jigar Shah: A Buckaroo jacket. You go. I’m staying away from the cowboy boots.
Caroline Golin: You go.
Stephen Lacey: If you think about the wild scenarios that people are worried about right now, do any of them feel like they are material enough?
Caroline Golin: Here’s the truth on this. This whole agent thing or I actually, and I’ll leave my personal opinions out of this. I actually think what you will see happen if something like that happens is the narrative that the only thing that can tackle AI terrorism is AI.
Stephen Lacey: So it makes the AI labs more powerful and concentrated.
Caroline Golin: Yes.
Stephen Lacey: Yeah, I buy that.
Caroline Golin: So I don’t know that that does anything to… I mean, it does more for political unrest. I think the wild card is going to be the narrative around China being months behind us goes away and they’re no longer months. They are ahead of us.
Stephen Lacey: Yeah.
Caroline Golin: And then the question becomes, well, do you need to host all this infrastructure to benefit from the integration into the economy? And I think that is not the wild card. I think that is the card that plays out within the next year, year and a half.
Stephen Lacey: Yeah. And at some point the systems are so powerful, being months behind, does it really even matter that they’re going to be so capable?
Caroline Golin: Right. Or there’s just so much differentiation and optionality in which type of framework or large language model you choose.
Stephen Lacey: Yeah. Yeah. All right, Jigar, what’s your wild scenario?
Jigar Shah: I mean, I think I go back to Caroline’s point on politics. I mean, part of the reason why the markets are so uneasy is because they don’t understand who the sheriff is in town. So whether it’s on the AI side or whether it’s on all of these CFTC oil trades that didn’t get prosecuted. I mean, today, people no longer believe that Brent, the oil price at Brent actually represents the cost of oil. Just think about what that means. When you have an actual metric that you’re basing billions of dollars of contracts against, and you’re like, “I don’t even know what that number means anymore.” And so when you think about all of these things, it leads you to this populist approach, which I don’t think people have fathomed what the pathway is to AOC becoming president.
Caroline Golin: Well, it’s not even AOC. That’s what this looks like.
Jigar Shah: Yeah.
Caroline Golin: I think that’s
Stephen Lacey: Reasonable.
Caroline Golin: Vance and Warren and AOC sort of all saying the same thing, what does that mean? But I don’t know. I do think a lot of the moratorium concerns are really more about a pause because I think governors are getting wise to the fact like, “Oh, this is a negotiation. This is more than just maybe tax. This is more like what do I want out of this?”
Jigar Shah: I mean, this is certainly what Jennifer Granholm has been saying, but I would suggest to you it’s a little bit different than that. I hear you, but I think this is more about do we need large data centers?
Caroline Golin: Yeah, that’s my thing on the, is it worth it? Is the infrastructure worth it?
Jigar Shah: Right.
Stephen Lacey: I mean, the companies are saying, yes, they do. But companies are lying for their teams have so many parameters. But these models are so much bigger. There’s so many parameters. They’re not. Yes, they are. They’re not. Oh my God, the train –
Jigar Shah: Epoch AI has debunked it and said basically that because all of the data is now behind anti-fetching approaches, they’re never going to get to the weighted models that they need to use four gigawatts of compute. The current models that they’re training right now still are less than one gigawatt.
Caroline Golin: We’re going to have to pause on that one and hold because I don’t necessarily agree with that, but I think the question is still, can you become a leader in the next version of the global economy integrating AI without hosting all the damn data centers? And I think a year ago we told everyone you can’t. And I just don’t know if that’s true.
Jigar Shah: Well, but you said something differently earlier, Caroline, which I think is worth talking about, which is that –
Caroline Golin: We’re still talking? All right.
Jigar Shah: That ultimately this is about integration of the tools. It’s really like what Red Hat did to Linux. Linux was powerful, but no one knew how to use it and Red Hat figured out how to make it usable for people. I mean, yes, the models matter, but what matters more is this layer where people can help train people around how to really unlock productivity. And to me, that layer of value creation doesn’t require 1,000 megawatt data centers. And the reason it matters to me is, just to be clear, is not about whether I win this argument. It’s about if you’re a populist and you have a building that needs a million dollars worth of upgrades, putting a 500 kilowatt data center in the basement of that building and getting them to pay for the million dollars of upgrades is a pretty easy trade. I think you’re going to see that all day long.
Caroline Golin: Okay. So what you’re really arguing is that it’s all going to happen. It’s just going to happen in a more edge infrastructure.
Jigar Shah: Because that’s where the trade comes, right?
Caroline Golin: Because that’s where the negotiation can be. That’s where you can actually get something out and the communities aren’t getting anything out of the bigger ones.
Jigar Shah: Yeah. We just talked about the delta between transmission and distribution. There are so many buildings right now that are office buildings that are 50% empty. They have 3,000 amps of unused service. So you can put a data center there, and what the mayors want in return is that they meet their energy efficiency standards and the water efficiency standards and they get upgraded because a lot of these buildings are in disrepair. I just think that there’s a level of political efficiency here that people are not seeing because they’re just so obsessed with going bigger and bigger and bigger.
Stephen Lacey: I totally agree, and I think this is where China has a leg up. They actually have a strategy to make AI useful throughout the economy, throughout provinces, so that local leaders figure out how to adopt it inside government, make it useful for people.
Caroline Golin: So does the Middle East, frankly.
Stephen Lacey: Yeah. And I absolutely think that that is a huge advantage for China and for the Middle East. But I think you still need a lot of these centralized data centers for the kind of training that we’re talking about. I don’t know how many, I don’t know the exact translation between the trillions and trillions and trillions of parameters and the amount of gigawatts you need. That to me is still a very open question, but these models are getting absolutely enormous and there is going to be a need for these large mega campuses. We just can’t deny that.
Jigar Shah: Which we have 36 of them already, to be clear. So are you suggesting we need another 36 of these large campuses?
Stephen Lacey: No, I mean, I don’t really know. I don’t know.
Jigar Shah: But you should know because it’s destroying the stock market.
Caroline Golin: Well, if Stephen knew he would not be hosting this podcast.
Stephen Lacey: Yeah, exactly. Right, right.
Jigar Shah: No, but I’m just saying the reason –
Stephen Lacey: I’d be an operations officer at OpenAI or something.
Jigar Shah: Well, no, but there are a lot of operations officers in OpenAI and they don’t get listened to either, right? Because I’ve talked to many of them and they agree that inference can be small. The reason I bring this up, Stephen, is not to be pedantic about it, but just because the fact that this is whatever it is, a trillion dollars of infrastructure per year and no one has the answers to this most basic question means that people are investing capital without understanding very basic things that they should know.
Caroline Golin: I just really think that investment and developers underestimated how hard it is to build large centralized power plants.
Stephen Lacey: And now they’re finding out –
Caroline Golin: And everything goes from there. Everything just goes from there. And whether they pivot, and maybe they should, or whether they just lose a ton of money beating their head against the wall, we’ll see. But I think at the end of the day, it was just people thought that if they had a contract for turbines, then the turbines would get built.
Stephen Lacey: That is such a good place to close it, and I think that is what is driving this market correction. People are figuring out just how difficult this stuff is. So well said, a fun conversation. Caroline Golin, so good to see you. Always a treat.
Caroline Golin: Thank you, Stephen.
Stephen Lacey: Jigar, great to see you. Your sweater just made my day, made me feel cozy New England.
Jigar Shah: Well, it is a fantastic fall.
Stephen Lacey: Good to see you both. Such a joy seeing you. Open Circuit is produced by Latitude Media. The show is edited by me. Sean Marquand and Anne Bailey provide production support. You can find all of our episodes anywhere you get your podcasts, Apple, Spotify, elsewhere. Transcripts are at latitudemedia.com. And while you’re there, check out our newsletters. We’re covering all the stuff that we talk about on this show. And of course, if you like to watch podcasts, go to Latitude Media’s YouTube channel and you can subscribe and see Open Circuit and Catalyst, and we hope to see you there. Thanks so much for being here. I’m Stephen Lacey. We’ll catch you next week.
