It’s been a wild few years for grid-scale batteries in Texas. At the start of the decade, the Lone Star State was home to roughly 200 megawatts of battery capacity. Today, that number has skyrocketed to seventeen gigawatts…with at least five more gigawatts expected to come online by the end of 2026.
But this rapid buildout quickly led to market saturation. In the past year alone, revenues plummeted by more than 80%, far below the threshold needed for fleet operators to stay in the black.
The future of the battery market in Texas remains murky. Will the depression continue? Or will the imminent, enormous surge of data center load growth reset the state of play and bring volatility back into the mix?
In this episode, Shayle sits down with Brandt Vermillion, U.S. market lead at Modo Energy, to unpack the mechanics of the Texas battery storage market, and consider when the tide could turn back in developers’ favor. They cover topics like:
- The factors behind the recent cratering of the market
- How longer battery durations are able to capture secondary price spikes
- The degree to which weather patterns have contributed to battery demand
- Reconciling currently suppressed revenues with the hundreds of incoming gigawatts of proposed data center demand
- Whether on-site generation for hyperscaler data centers will blunt peak demand spikes and delay the return of grid volatility
- Why off-takers are considering tolling agreements to survive the market downturn
Resources
- Catalyst: 2026 trends: Gas turbines, Texas’ load queue, and China electrifies
- Catalyst: Why C&I storage is finally taking off
- Catalyst: The rise of flexible data centers
- Open Circuit: Data centers have a Texas-sized energy problem
- Open Circuit: The off-grid data center fantasy
- Open Circuit: A global energy shock and the case for distributed power
- Open Circuit: Power is caught in AI’s doom loop
- Latitude Media: How the AI boom has impacted US battery storage so far
- Latitude Media: Branch Energy heads to PJM to sell battery capacity to data centers
Credits: Hosted by Shayle Kann. Produced and edited by Max Savage Levenson. Original music and engineering by Sean Marquand. Stephen Lacey is our executive editor.
Catalyst is brought to you by GridBeyond. GridBeyond is the world’s leading technology platform for managing distributed and flexible energy resources. Learn more by visiting GridBeyond.com.
Catalyst is brought to you by Antenna Group, the strategic communications and marketing partner behind the biggest names in energy, climate, and infrastructure. For three decades, Antenna has helped breakout companies amplify their stories, build reputation, differentiate from the competition, define new categories and accelerate growth. Learn more at antennagroup.com.
Catalyst is brought to you by EnergyHub. Peak season puts every grid to the test — and the utilities that pass are the ones that built flexible capacity before they needed it. 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.
Transcript
Shayle Kann: I’m Shayle Kann. I invest in early-stage companies at Energy Impact Partners. Welcome to Catalyst.
I’ve been wanting to do this one for a while. So, at the start of this decade, Texas, or ERCOT, the market in Texas, had about 200 megawatts of grid-scale batteries. By the end of June this year, it had 16.5 gigawatts, nearly 29 gigawatt-hours. That is an enormous growth rate. 6 gigawatts came online last year alone, and most of it was built merchant, in an energy-only market with no capacity payments to lean on. Which sets up this strange split in how people talk about it. Because on one side, you hear, and this is indeed true, that the economics for these batteries have collapsed. The average ERCOT battery earned something like 84% less in 2025 than it did in 2023. In 2023, there were 58 days where the average battery earned at least 50 cents a kilowatt. Last year, there were 3.
Ancillary services saturated. The fleet crowded into arbitrage, and then the fleet flattened the very spreads that it was chasing. This has happened before, and it will happen again. Developers are of course noticing, so new capacity entering the queue has been cut in half from its peak, and a record 13.7 gigawatts of batteries withdrew from the interconnection queue in the first half of this year. Sounds like a bloodbath.
On the other side, basically every macro signal says that Texas needs a lot more grid-scale batteries. Peak demand is going to spike. We have enormous load growth coming—just look at all the data center announcements that we see coming in Texas. Even if they get delayed a little bit by this Batch Zero process, they are coming. And then meanwhile, more than 20 gigawatts of aging coal and gas are losing money as well, and some of those are going to be shuttered. Gas turbines are sold out for years, of course, and batteries are a resource that can actually show up on time. So you have this dynamic where there’s a market that is currently losing a lot of operators money, but wherein it appears we’re going to need a lot more of it at some point in time.
So the question is how long the gap between these two things, the period today when economics are suppressed versus the period when they come roaring back—how long does that period last? And who’s still standing when it does? But in the meantime, the fleet that saturated the market is still growing by another maybe 5 or 6 gigawatts this year. Every one of those gigawatts pushes the recovery a little further out, while every new data center pulls it in.
To talk through it, I brought on Brandt Vermillion. He is the US Market Lead at Modo Energy. He’s coming up next.
…
Shayle Kann: Brandt, welcome.
Brandt Vermillion: Thanks Shayle. Really excited to have a conversation today.
Shayle Kann: All right. Start by telling me how much grid-scale battery capacity we have in Texas as of today, and how has that grown over the past few years?
Brandt Vermillion: Yeah, so if we go back to the beginning of the decade, we’re talking about only a couple hundred megawatts of installed capacity of grid-scale storage in Texas. And if we fast-forward to 2026, we’re talking about 17 or 18 gigawatts that’s fully approved for commercial operations by the grid operator ERCOT. And really, actually over 20 gigawatts of capacity that’s now been—that’s at least completed construction and approaching that commercial operations stage.
Shayle Kann: Okay, so big growth. How are these batteries—like, what is the revenue model for most of the storage that gets built in Texas as of today?
Brandt Vermillion: Yeah, that’s something that’s definitely evolved over the last few years. As of today, I’d say actually some murkiness to what the revenue model ultimately looks like and how projects are getting financed. Because, you know, historically there was a, I’d say, pure merchant model would be how you’d describe it. And what that really means is you’re just betting on high prices in ancillary services or high spreads between charging and discharging prices, or what we’d call energy arbitrage, for batteries to make money. And that’s, you know, still what you’re ultimately hoping for as a battery developer, that’s kind of what you’re building your investment case on. But in the last few years, a lot of that volatility has, has ultimately gone away, and revenue opportunity for storage has—”collapsed” is a strong word, but I mean ultimately that’s kind of what it has done in the last few years.
Shayle Kann: Well yeah, so talk to me more about that. Like, historically, how much of the revenue—if you’re building a merchant BESS project in Texas, how much of the revenue would you have expected to come from ancillary services versus energy arbitrage? Like, what was the expectation? And then, as this fleet was getting built, what were people generally expecting?
Brandt Vermillion: Yeah, so in the, I’d say, early part of the decade, 2022, 2023, we had—I’d say most batteries were earning, you know, more than 90% of their revenues via ancillary services. So, that’s you receive an award to provide frequency response, or contingency response, or something like that. And ultimately, you’re just sitting on that capacity and getting paid for it.
As more and more batteries were added to the system, the amount of actual installed battery capacity started to outweigh the amount of ancillary service capacity that ERCOT procures hour by hour or day by day, right? So, what you then have is a shifting of battery capacity into the energy market, which is quite a bit deeper. ERCOT’s serving 60, 70, 80 gigawatts of load at peak on any given day, right? So, what that means is, from say 2023 when you had somewhere between 80 and 90% of revenues coming from ancillary services, to today, we’ve seen it shifting to close to 80% of revenues coming from energy arbitrage in terms of that merchant market makeup for batteries in Texas.
Shayle Kann: Which is like a story we’ve seen time and time again in batteries, right? Like, the original version of this was PJM, where all the first grid-scale batteries got deployed. Ancillary services are always like a lucrative market, but a shallow market, and so you saturate it pretty quickly and then you have to make money some other way. In the case of ERCOT, that ends up being energy arbitrage. In other markets, it might be capacity markets or something like that.
But I think this is what’s interesting, right, is like, we built out this big fleet, or we have been building out this big fleet, and meanwhile, we saturate the ancillary services market. And so now, if you’re putting a battery in for the past couple of years, as you’re saying, in Texas, basically what you’re trying to do is arbitrage low prices and high prices. Texas is the best place to do that, because it’s an energy-only market more or less, and you can have really high prices.
But as I understand it, the volatility, which is what you were trading on—and you even alluded to this—has not been static over the past few years. So what’s happening in volatility, and then what has that meant for revenue for this fleet of merchant generators—or sorry, merchant batteries?
Brandt Vermillion: Yeah, so really it’s a very similar story to ancillary services, right? I mean, it’s not as easy of a comparison to draw in terms of if you have a certain amount of demand and a certain amount of supply, but it’s effectively the same thing where the more battery capacity that gets deployed, the more competitive it is to ultimately, in the case of ancillary services, get the awards, in the case of energy, actually get dispatched to discharge your energy that you’ve stored up. And so that means you have to offer at increasingly competitive prices against your competitors, your other batteries that are participating in the market, and that drives the price down.
So, same mechanism as ancillary service saturation, but as more and more batteries have then looked to energy arbitrage as the primary source of how they’re going to earn their revenue in a merchant market like ERCOT, that just means that, you know, competition for getting that dispatch to discharge is increasing, and therefore you kind of have the suppression of volatility and what you might call cannibalization of revenues that we’ve kind of seen in the last couple of years.
And just to put numbers on it, in 2023 when you had a really hot summer, kind of coinciding with ancillary services not being quite saturated yet, batteries were earning, you know, nearly $200 per installed kilowatt in Texas. And in the last 12 months or so, that’s fallen to, you know, around $30 per installed kilowatt or a little bit less.
Shayle Kann: Have you done the math to figure out what these systems generally need in order to hit their hurdles and pay back? Like, is $30 okay, it’s not as lucrative as it was before, but like that’s a financeable price? Or is $30 underwater?
Brandt Vermillion: $30 would be underwater based on our understanding of where CapEx and just general financial models stand today. I think the rule of thumb we’ve kind of heard for say like a two-hour system getting installed in ERCOT is you’re probably looking for roughly $100 per installed kilowatt, maybe a little bit less. This is a bit of a shifting target right now on the cost side as well, too, right? So, that can I’d say be a bit of a fungible number to some extent, but regardless, $30 is going to be less than what a project developer is looking for in terms of financing their project. So, I’m sure we’ll get into it in a second, but that makes for a difficult decision framework, right, when you’re looking at ERCOT as a market where there’s lots of potential load growth, but how do you potentially scale that against—or I guess reconcile that against where the revenues sit today?
Shayle Kann: Yeah, and that seems like that is the—that is the big question. I guess before we get to it though, I mean, that implies that there is a big fleet, gigawatts, maybe tens of gigawatts of batteries that are underwater on their economics operating in ERCOT right now. Have we seen the ramifications of that at all? Have we seen M&A of assets or anything that you might expect to see in a market where there are a bunch of folks who financed merchant batteries who are losing money today? Or is it just because it’s like a recent phenomenon where they’ve only been losing money for the past couple of years, and you know, these are 10-year life assets and so they’re hoping it’s going to come back?
Brandt Vermillion: Yeah, I think there’s a mix of different things. I think we’ve seen maybe some of that M&A activity that you would expect, but probably maybe not quite as much as if you just looked at it at its face value, you would think. And there’s a few reasons for that, right? You have some of this installed capacity that has been around for not just one or two years, but three or four years. And maybe in that first year of operations, ancillary services were still kind of printing pretty solid returns, or even in the energy side of the market in say 2024, which was I’d say maybe a decent year if not a very good year, you were able to earn, you know, passable revenues to keep that project above water and kind of keep yourself above your debt service for the time being.
There’s a mix of projects like that. I think there’s also lots of the developers that are participating in ERCOT are, you know, big balance sheet IPPs or potentially even, you know, more traditional utilities or things like that, right, that have the ability to potentially, I’d say, weather the storm, right, of a couple of lower years of revenues, and maybe they’ve self-financed their projects or something like that. So, it’s not a, I’d say, a full market consolidation, but that being said, there is certainly some M&A activity, and some of these projects are changing hands.
Shayle Kann: It seems like there are multiple factors at play here that I—for me at least, it’s kind of tough to tease out and separate from each other. One is, as you’ve described, batteries—especially if they are all the same two-hour duration or whatever, which we should talk about duration—but, um, you know, the more you add, the less value the marginal one has. So there’s an extent to which you could say, “Okay, the reduction in volatility and thus revenue that we’ve seen from merchant batteries in Texas over the past couple years is a function of building more batteries, the fact that we’ve built 16 gigawatts or whatever it is.”
Also, you have this confounding factor of weather, wherein if you have a mild weather period, that also reduces volatility, and vice versa—if you have a super crazy, stormy period, you get these crazy spikes, and that’s what Texas has been known for. So even just stopping there and separating out like, the fact that we’re at $30 per kilowatt installed in revenue this past year as opposed to $200 before, like, can you separate out how much of that comes from the fact that we’ve built a lot of batteries versus the fact that we’ve had a mild spring or whatever it was?
Brandt Vermillion: Yeah, I’d say the majority of that shift is definitely more on the side of the fact that so many batteries have been built, and the amount of installed capacity has increased so much, and now the batteries are not just sitting in ancillary services, but also participating in energy arb, right? So that’s, I’d say, the majority of it. But yes, there is an additional layer to it that is weather-related as well.
And I wouldn’t characterize say 2024 or 2025 or even this summer as, I’d say, like a cooler summer for instance, probably more closer in the average to above-average range, depending on how you would want to measure it. It’s more just that it wasn’t extreme, right? And I think if you look back to a summer like 2023 or some of the other winter events that we’ve seen where you have an extended period of cold weather, that’s—that’s where you actually going to see that extreme weather, I guess, kind of resulting in some of that volatility.
And, you know, that does still have the potential to happen in ERCOT say this winter, and we saw a bit of an example of it in January with Storm Fern as well, where at the end of the month you had prices getting into the four-figure range. And, you know, there’s some interesting stuff to talk about in terms of how storage operations played out in an extended cold weather event as well that actually kind of contributed to that volatility as well. But, yeah, at its core, it’s the battery deployment, but the weather is definitely also of course like a compounding layer on top of that.
Shayle Kann: Yeah, I mean that the situation you described with Fern actually relates to this the duration question, right? Which is, um, I guess tell me historically, like the batteries that are getting installed in ERCOT, are they all the same duration? Is that changing?
Brandt Vermillion: Yeah, it is starting to change, and I guess just to preface, like ERCOT is a little bit unique, mostly in the sense that it doesn’t have that capacity market, which has meant that most batteries have actually been shorter duration in ERCOT than you’ve seen in a lot of other markets. So, CAISO is the market or California where the majority of other grid-scale storage deployments have happened in the last 5 or 6 years, and most batteries there are at least 4 hours in duration because of the way that effectively their capacity construct is set up.
If you compare that to ERCOT, when batteries first started getting built out, it was almost exclusively 1-hour duration batteries because there were no real rules surrounding your duration requirements in providing ancillary services, and that started to be legislated a couple of years ago in terms of figuring out exactly how much for each megawatt of a given service, how much energy you needed to store and things like that. But that was probably the initial thing that began to push battery duration longer, along with just cells and packs continuing to get cheaper.
But the shift to energy arbitrage has also done that as well, and I’d say the average duration of batteries in ERCOT has increased from, you know, close to 1 hour a few years ago to getting close to 2 hours now, and we’re actually starting to see more and more projects come online that are in that 2.5, 3-hour range and even a couple of 4-hour duration projects. So, it is lengthening, and we can kind of talk about what some of the actual market mechanisms and market outcomes that I guess kind of warrant that change look like and maybe talk through an example today or whatever, but it’s definitely changing as time goes on here.
Shayle Kann: I’m definitely interested in, you know, you alluded to this like what happens in the extended weather event situation where, and you know, I think there’s all this weird game theory that all these merchant batteries have to play about when do you discharge and when do you not and so on. But to a first order, what you imagine is that there’s some—there’s a weather event that causes demand to rise, which causes prices to rise, which causes batteries to discharge. And if all the batteries are 1-hour batteries, you know, to a first order, they all discharge in the first hour, and so prices crash, after which they’re all totally depleted and prices spike again. Have we seen that kind of thing happen historically?
Brandt Vermillion: Yeah, I’d say it’s more of a recent phenomenon, to be honest, in say the last 12 to 18 months. And some of that is like the fact that batteries are not the only variable that’s changing on the fundamental side of the grid, right? And you’re seeing more and more 24/7 consumers on the load side, right? And so that can mean that some of these, even like an evening summer period, your difficult period for the grid operator of serving the load as you’re transitioning off of 30 or 40 gigawatts of solar generation to serving this flatter load profile that we’re starting to have now, becomes more and more difficult.
And so even in this summer, which has overall been low volatility, some of the higher-price days were more a result of, you know, as you described, kind of batteries discharging into maybe the hour in which the day-ahead price peaked, and then as batteries start to run out of state of charge, there’s almost this secondary peak an hour or two later if say you have a combination of things happening like the wind forecast was a bit higher than what you got in reality, the load forecast was a bit lower than what you got in reality, and then all of a sudden you have this setup for this kind of secondary real-time market price spike where the batteries that held back some of their state of charge or had the advantage of being say 2.5 or 3 hours in duration rather than 1.5 hours in duration were able to kind of get the best of both worlds and discharge into both of those spikes.
And I’d say for like a Storm Fern, it’s kind of, you know, similar but different, right? In the same way that it’s extended over multiple days, right, where the first day of Storm Fern there was, you know, a lot of anticipation for, I think it was like the Tuesday morning peak or something like that, where that was when the winter record load was projected to happen, and you had a bit of a load forecast miss there where actual load came in well below the forecast, but at the same time, batteries had so much stored energy and there was essentially not really much price volatility. The day-ahead price cleared a lot higher than the real-time price.
But then, you know, a day or two later you kind of had the reverse situation where load came in above the forecast, wind was coming in below the forecast, and you had an evening where batteries discharged into a price spike, and then those forecasts stayed off for the entire overnight period, and the real-time price stayed above the day-ahead price. We’re talking say $150 versus $50 or something like that, and a lot of operators of batteries were a little reticent to charge at that price because they weren’t sure if they were still going to get that bump the next morning.
And so then all of a sudden you have this situation where you’re going into the next morning without very much state of charge stored in these batteries, talking about maybe a third of the fleet’s actual capacity that had been charged up ahead of that morning, and that was the thing that actually sent prices in the real-time market into the four-figure range and actually brought on the volatility, the most amount of the volatility that we ended up seeing during Storm Fern. So, you know, battery state of charge and ultimately understanding where the fleet sits at any given point in time is such a key variable for these optimizers and operators of storage at this point.
Shayle Kann: Has there been a mechanism—are these all like purely merchant assets, or has there been a mechanism in order to get them financed, you know like a tolling arrangement or whatever, where you offload the market volatility risk? Is anybody bearing that risk other than the operator of the battery?
Brandt Vermillion: Yeah, definitely. I think if you’re trying to get a battery financed in ERCOT today, like especially if you’re going to some kind of third-party lender or financier and you’re not self-financing in any form or fashion, you’re going to be looking for some kind of off-take to help firm up those low periods of revenue, even if you have strong conviction that volatility’s coming back in 2028 or whatever it is. You kind of have to bridge the gap to whenever that period is so that you can meet your debt service in the meantime quarter over quarter, right?
And so that’s what the bank or whoever your lender ultimately is looking for is, whether it’s a toll or whether it’s some kind of revenue share, revenue floor, or even like some kind of swap agreement with a trading house or something like that, you just need something in place that’s going to mean if prices are low, you’re at least kind of getting made whole up to your debt service agreement. And that obviously is going to scale down your upside quite a bit, especially if you’re signing at a strike price that’s relatively low. You know, that’s the inherent trade-off that a developer is making and kind of has to reconcile.
And it makes it really hard to get these deals done, too, because if you’re on the off-taker side, you want to offer as long as possible and as cheap as possible, right? And if you’re the developer, you just want to bridge the gap to when you think the volatility is coming back and maintain as much upside as possible while covering that obligation. So it makes it very hard to agree on those off-take agreements, and I think, you know, tolling has definitely been a useful tool, but it’s not necessarily the predominant way of finding off-take just because there is often such a bid-ask spread between the two sides of that agreement at this point.
Shayle Kann: You mentioned a couple times the concept of, “Oh, the volatility’s coming back in 2028.” So I think this gets to the crux of the question for the future, right? We have—here’s the situation as I understand it. We have a big fleet already operating in ERCOT, continues to operate, a lot of it’s underwater. We have an enormous queue—you can tell me in a moment how big the queue is, but the interconnection queue for batteries is even bigger—and a market wherein revenue for merchant batteries has been in decline for the past couple of years, and it sounds like not in an anomalous way, not in a way that you would say, “Well, it’s because the weather was particularly mild.”
So those set of factors on its own implies we shouldn’t build any more batteries in Texas. Like everything in the queue should fall out of the queue, and why would anybody go, you know, like build into a bloodbath? I think the counter to that is the load growth, basically, and the belief that we are going to see so much load growth in ERCOT that we’re, you know, we’re basically resetting the table and we’re going to be back to volatility again.
First of all, am I right that that is the calculus in the queue? And then we can talk about whether you think it is true.
Brandt Vermillion: Yeah, I think if you’re a developer that’s thinking about building a battery in Texas in the next couple of years, or right now today, that’s ultimately your thesis, right? Is that the current moment is transitory, these markets are cyclical. Effectively, the old adage of low prices cure low prices, high prices cure high prices. Infrastructure finance is inherently something that kind of moves somewhat slowly and with a lag. The response to the market signal takes a little while because you see high prices, and then it takes a couple of years to actually get your project through the queue. And so inherently, you kind of have this supply and demand balance that is kind of constantly overshooting each other, right?
And so I think if you’re a developer, that’s what you’re looking at, the potential for data center-driven demand growth to ultimately result in the next couple of years. Except the added factor there, right, is the potential scale and the speed at which that can happen, right, based on the amount of load growth that has the potential to happen, or at least there are projects that have stated interest in connecting to the grid.
I mean, the headline figures out there are 4 or 500 gigawatts of projects that actually want to connect, which is very much obviously inflated. But even if 10% of that connects, that is transformational for the ERCOT network that is peaking at 91 gigawatts as of today, right? And the fact that all of those projects in theory want to connect as soon as possible, right? In reality, it’s going to be a 5, 6-year period of time where ERCOT is trying to accommodate as much demand growth as it can. And all the system operators that are seeing interest in connecting new forms of demand, primarily data centers, it’s going to be an extended period.
But, yeah, I mean the interest is there to get it done in the short term. And so if you’re a developer, you’re saying, “All right, well, this year has not been very good, last year wasn’t very good, maybe next year isn’t very good as kind of this, I guess almost development flywheel starts to ramp up and system operators are figuring out how they incorporate all these large load projects.” But as things start to kind of get figured out, and you’re starting to get a few gigawatts or even a little bit more of new load coming from data centers every single year, at the same time that other people have kind of slowed down development on the supply side with whether that’s storage or gas resources that, you know—if you were trying to build a merchant gas turbine today, you’d also be looking at the same prices that a battery’s looking at and struggling to make the economics work.
So you kind of have this setup where the demand growth could very quickly outpace what is currently as it stands today, kind of an overbuilt system or oversupplied system.
Shayle Kann: Have we seen the supply side slow down? Like is there evidence of that in the interconnection queue? Have people pulled back at all, or has it just continued to boom?
Brandt Vermillion: Yeah, so it’s, I’d say we’re starting to see the first signs of it, right? In the sense that if you look at top level, like the amount of just battery project capacity that’s in the interconnection queue is kind of stalled out at the 150, 160 gigawatt mark, which is obviously a huge number. But, you know, in prior years, we were seeing that grow by dozens of gigawatts a year. And that’s essentially stopped growing.
And then downstream from there in terms of actual project throughput, so if we look at like year-over-year growth in that installed capacity number that we were talking about at the beginning of the conversation here, I said we’re around 18, 19 gigawatts of commercially operational capacity today. That’s based off of a starting point of roughly around 14 gigawatts to start this year, which means by the end of the year, I’d expect we’ll get to around 5 to 6 gigawatts of new capacity, which sounds like a huge number. But if we look back to last year, we’re looking at around 5 or 6 gigawatts of new installed capacity. And basically the pattern for every year before that was roughly almost a doubling of the fleet size.
And so we’re starting to see at least the percentage growth of total installed capacity start to slow down. So that’s one element.
And then the next element is, you know, looking into the queue itself and like how many projects are actually progressing to new stages, how many projects are stalling out, how many projects are actually withdrawing from the queue. One of the things that we tracked last year was something like 13 to 15 gigawatts of project capacity is actually withdrawn after signing an interconnection agreement, which if we went back a few years, signing an interconnection agreement would have been like a near guarantee that a project was actually going to ultimately materialise in a commercially operational project.
So there are signs that it’s beginning to slow down. I think just anecdotally, you listen to like what participants in the industry on the other side of the off-taker or the developer side of things are starting to say, and the I guess just amount of friction when it comes to getting that off-take agreement signed and getting a project developed or financed or whatever it is just seems like it is increasing. And so ultimately, I think there are a lot of signs that we will see some of that development pull back at least for the next year or two.
Shayle Kann: So do you have a view on the timing? I mean, these markets are cyclical, and so probably what is going to happen over some period of time is volatility’s down, revenues are down, cure for low prices is low prices, and then you have the additional variable of all this load growth. And so probably it does come back, I presume. I have no idea when, though, and for developers who are trying to put assets online, that’s a pretty important question. Do you guys have a view? Is volatility roaring back in ’28 or in 2030 or who knows?
Brandt Vermillion: Yeah, so that’s one of the things that we do is we put together our own in-house production cost model to try to get a sense of essentially just balancing supply against demand and getting a feel for, “Okay, well if we make a reasonable assumption of how much demand can actually connect in the next few years on a year-by-year basis and kind of site it on the system where it’s actually going to go, how much can we actually accommodate based on how much supply we’ve seen come through historically, how much is in the queue, kind of accounting for that same thinking that we were just talking about of how difficult is it to get projects financed on the supply side right now?”
The answer ultimately results in like we’re starting to see volatility re-emerge in our model around 2029, 2030, at least in a big way to the point where you’d be easily clearing that $100 per kilowatt mark for at least a couple of years to the point where you’d probably have a fair amount of appetite for having a battery on the system at that point.
So, it’s a difficult piece of calculus to make because, you know, that’s a few years away, first of all. So you’d potentially be riding out a little bit longer of this low volatility period, but you also—it’s difficult to know. There’s just really not enough information on the demand growth side of things to know for sure like exactly how fast that demand growth is going to happen.
And if it does really start to pick up in earnest, say in the back half of next year as ERCOT’s Batch Zero process gets finalized and really starts to pick up in 2028, for instance, and that coincides with those extreme weather events that we were talking about—you get a super hot summer in 2028 or a winter weather event between December ’27 or January ’28 or something like that—then all of a sudden you could have this volatility showing up even sooner.
But that’s the—I guess I’ve been throwing around the whole idea of the term of cognitive dissonance so much, that’s the dissonance you have to deal with in terms of reconciling today against where tomorrow is potentially and even likely going. And the timing at which that happens.
Shayle Kann: It strikes me that there’s one maybe big risk to the volatility coming roaring back concept, which is the proportion of this new load that comes with behind-the-meter generation. Or storage for that matter. But either way, if you’ve got behind-the-meter capacity that can meet the full load of your data center—let’s just say, because they’re mostly data centers—then exactly at those times when the merchant battery operators in ERCOT are going to be expecting prices to spike, those are the times that anybody who has something behind the meter is going to want to fire that thing up.
So is it possible that we will see a ton of load growth, but that load growth actually won’t, like from a system perspective, contribute to peak, and then we actually won’t see the volatility show up despite the load growth being there?
Brandt Vermillion: Yeah, I think that’s definitely a factor that you have to account for in the modeling as well, and it’s something that, you know, we try to spend as much time not only just tracking the individual large load projects or data centers that we think are coming in the next couple of years, but also looking into, you know, air permits, you know, basically filings with the Public Utility Commission or the environmental agencies in Texas and around the country as to whether they’re going to have say on-site gas generation and trying to corroborate at least public statements for an individual project that is saying that it’s going to have some amount of behind-the-meter generation.
And so that does mean that, you know, nameplate load growth is not the same as observed impact on like peak demand, coincident peak demand, that you’d actually see in the market. But our expectation, especially in the front end of the data center development where projects have potentially already secured grid connections, don’t necessarily have to rely quite as much on procuring or bringing their own generation essentially, because as we’ve kind of been talking about, the system is already oversupplied and can accommodate some amount of new load growth, right?
So those projects that I think are coming in the next say 1 to 3 years are probably going to be leaning less on on-site generation. And then I think as we get into the later part of this decade and the early 2030s, in the kind of the time period that we’d expect the system to start getting tighter again, that’s when, you know, the planning studies on the system operator side are going to indicate that, “Hey, we can’t really get you a grid connection. You might be a 1-gigawatt data center that wants to get a 1-gigawatt grid connection. We can only service, you know, 200 megawatts in Year 1 of when you’re looking to connect,” or whatever it is, and you kind of have to progressively ramp that over time or just say, “You know what, like we’ve actually got a line to a gas turbine, we’ll just go ahead and pay the price for that combined cycle plant and make that work and ultimately consume less from the grid.”
So it is something you have to kind of be aware of, and I guess our expectation is in the long run something like between, I’d say, 50 and 70% of nameplate is going to result in actual grid draw essentially, is kind of our, I’d say, rough rule of thumb. I’d say there’s a fair amount of guesstimating and assumptions based on initial leanings and just kind of conversations within the industry that’s happening there. But, yeah, that’s part of the calculus as well.
Shayle Kann: All right, final question for you, lest we spend the entire time talking about ERCOT and Texas. To what extent is this general phenomenon of like saturating ancillary services, declining revenue, you know, entering this period of like overbuild that may or may not get swamped by load growth in the future—to what extent is that an ERCOT-specific thing versus a dynamic we’re seeing play out with obviously different markets in different locations?
Brandt Vermillion: Yeah, I think there’s definitely lots of similarities you can draw to other markets, and some of them are you have an—you can almost have a market like ERCOT or CAISO or even like the UK as kind of learnings, or markets you can learn from in terms of how the battery deployment will go and just the cycle power markets in different regions of the country will go in the next few years.
So fundamentally, as you start building out grid-scale storage capacity, they’re always going to target ancillary service markets first. Those markets are quite shallow. They’re probably going to provide the most lucrative returns you can provide. PJM is actually a really good example of this right now where they did a redesign of their regulation market last year and prices have been obscene for the regulation service in PJM for the last year or so, where if you had a battery in that market you could be kind of similar to early days in ERCOT, making your investment back in a single year.
But the limitation there is that that market is only say 6 or 700 megawatts deep. PJM’s got about 500 megawatts of batteries right now. As soon as you get to a gigawatt, 2 gigawatts of batteries, that should I’d say compress quite a bit, right? And so of course that’s very similar to what we’ve seen in ERCOT, and then at that point you start to see batteries moving into, you know, providing energy arbitrage or participating in the capacity market in the markets that actually have those things.
And ultimately, you know, batteries tend to cannibalize their own revenue opportunity regardless of if it’s going to be in a capacity market construct or in a merchant market or energy-only market construct as well.
So, yeah, I mean I think we will see the same story that has played out in ERCOT play out in other markets where, again, you know, high prices cure high prices, kind of interconnection queue woes outstanding in some of these other markets that, you know, ERCOT hasn’t necessarily faced. But, yeah, I mean I think that is ultimately what we’ll see, and then the question from there is, well, A, like what’s the next phase beyond that kind of cannibalization of your first AS and then secondly arbitrage opportunity? Is there a recovery as, you know, other variables start to shift, whether that’s load growth taking off in that region, or if it’s thermal—if it’s in a region where there isn’t very much load growth coming, say take California for instance, is it then that you start to see a lot of thermal generation retiring and a changing supply-side mix that starts to change, well, the price shape and ultimately the revenue opportunity for storage?
Shayle Kann: Brandt, this was super informative and a lot of fun. Thank you for the time.
Brandt Vermillion: Thanks so much, Shayle. It was a lot of fun, appreciate it.
Shayle Kann: Brandt Vermillion is the US Market Lead at Modo Energy.
The show is a production of Latitude Media. You can head over to latitudemedia.com for links to today’s topics.
This episode was produced by Max Savage Levenson. Mixing and theme song by Sean Marquand. Ann Bailey edits the video version of the show. Stephen Lacey is our executive editor.
All of our episodes are up on YouTube. Subscribe to Latitude Media for episodes of this show and Open Circuit. You can also find the audio version of the show anywhere you get your podcasts.
I’m Shayle Kann, and this is Catalyst.
