In ERCOT’s new interconnection process for large loads, a data center can connect sooner as a provisional controllable load resource. That is, ERCOT can curtail power above an agreed firm level. A study from Base Power and Peak Energy shows batteries near the transmission lines to a data center could stand in for most of that curtailment.
In the study, 80 megawatts of batteries would clear every transmission constraint created by a 100 MW data center near Burleson, Texas, on Oncor’s system. A battery beside a strained line relieves it about one-for-one, while curtailing a more distant data center may relieve half as much. Any battery near the line has that edge.
On this week’s Energy Capital Podcast, Joshua Rhodes talks with Chase Dowling, head of market operations for Austin battery company Base Power. The business model is Base Power retains ownership of the batteries it installs for retail customers, and homeowners subscribe to the service.Dowling puts the batteries’ reach at “80 percent, 85 percent of all of the constraints.” The rare failures that sever part of the system, he says, still need new transmission or generation.
The limits: the battery fleet needed grows faster than the load, and West Texas lacks enough homes for large fleets. No market yet pays batteries to act as transmission.
Who pays: Dowling proposes that a data center facing curtailment fund a battery fleet near its constraints, and the host homes get outage protection.
Utility fleets: for El Paso Electric and CoServ, Base builds co-branded fleets and hands control to the utility. It now installs in Chicago, in PJM.
Backup: Dowling says a portable generator can stretch a Base battery through a multi-day outage like Winter Storm Uri.
Dowling says the challenge now is getting ERCOT to model these fleets at the transmission substation, where planners can see them.
Chapters
00:00 – Chase Dowling and the new Base Power study
01:01 – What Base Power has built since 2023
02:22 – Manufacturing home batteries as a grid resource
05:10 – The homeowner value proposition
07:50 – Base Core's generator port and extended backup
11:43 – Deployment speed and what limits it
14:43 – Regulated utilities and who controls the batteries
16:28 – Lessons from Texas for PJM
18:23 – Flexibility Without Curtailment: the paper's premise
22:32 – Why location matters: 100 MW at Burleson Switch
26:52 – How 80 MW of batteries hosts 100 MW of load
28:47 – Bigger loads, the long tail, and the 80/20 rule
35:13 – Batch Zero, curtailment risk, and who pays
38:17 – Data center PR and how big this can get
Resources
People & Organizations
Company & Industry News
Base Power raises another US$1 billion in Series D financing for US residential BESS
Base Power brings cheap batteries to residents in power-starved PJM
Base Power to launch 100-MW home battery network for Texas utility
Books, Articles & Filings Discussed
Piq Energy and Base Power - Flexibility Without Curtailment
ERCOT - Aggregate Distributed Energy Resource (ADER) pilot project
ERCOT - Trending topic: new batch connection process for large electricity users
Texas Legislature - SB 6, 89th Legislature, enrolled bill text
El Paso Electric - El Paso Electric and Base Power install first home battery in El Paso as part of new reliability pilot program
Farhad Billimoria and Conleigh Byers - The value of compute load: what cloud markets reveal about the potential for flexible data centers
Related Podcasts by Energy Capital
How Load Flexibility Could Unlock Energy Abundance with Tyler Norris
How Data Centers Can Strengthen the Texas Grid with Camus Energy CEO Astrid Atkinson
Related Posts by Texas Energy & Power
Transcript
Joshua Rhodes: Hey everyone, and welcome back to another episode of the Energy Capital Podcast. I’m really excited to have Chase Dowling here from Base Power Company to talk about what the company’s been doing, as well as a new study they’ve put out looking at the impact of distributed energy resources on the ability to unlock firm load for places like large loads, data centers.
But yeah, really excited to have Chase here to talk about it. Chase has a PhD in electrical and computer engineering from UW, University of Washington. In and around there, he also worked 10 years at the Pacific Northwest National Lab, ending as a senior applied scientist before switching over to machine learning at Tesla.
Very recently, actually, as of recording this podcast, he’s a founding power systems engineer at Base Power Company. Chase, welcome to the Energy Capital Podcast.
Chase Dowling: Thanks, Josh. Sorry for the mouthful. Yeah. Ten years at DOE and then a couple years at Tesla working mainly in energy.
Joshua Rhodes: Yeah. Yeah. Well, energy’s kind of the thread that brings us all together these days, right? It’s like energy is AI, is transportation—
Chase Dowling: Is everything.
Joshua Rhodes: —is now politics too, which we won’t get into. So it’s been a while since Doug originally interviewed Zach about Base Power, and so, you know, a lot has happened since the last time we checked in. You know, you started as a pretty novel combination of retail electricity and batteries in Texas, but now you’re manufacturing your own hardware, working directly with utilities.
Kind of before we get into kind of the specific study, what’s going on at Base right now? Like, what are y’all working on?
Chase Dowling: Yeah, no, it feels like a while, but it’s crazy to say founded at the end of 2023 that we’re scaling up. We’ve got 600 MWh of batteries in the ground, and we’re just now speeding up.
We’re installing all over Texas, and we’re now entering our first new markets outside of Texas, installing in Chicago and trying to get at the capacity problems in PJM. But the biggest changes maybe since you talked to Doug or since Doug talked to Zach was we’re installing in deregulated territory as a retailer, but now we have a lot of regulated utility partnerships as well, where we come in and help facilitate the construction of a large distributed peaking resource for these regulated utilities.
But that’s the story since then, is that we’ve been scaling, growing, and installing resources faster.
Joshua Rhodes: I mean, you’re manufacturing these units in Austin. I know you’re looking— you’re also building a new facility out by the airport, but, you know, you’re currently in the old, you know, Austin American-Statesman building.
I actually took a tour of the place, I wanna say, like 7 months ago, and you had, like, you know, half of a line going. But I was there last week. I mean, you had multiple lines going, robots following magnetic tape all around, and lots of folks putting stuff together. Tell me about the manufacturing process and kind of what you’re doing, you know, here in Austin.
Chase Dowling: Yeah. So this is a big piece of being a vertically integrated designer, manufacturer, developer, and operator of an energy resource and how we build our kind of like technological advantage that we’re providing to the grid. And so manufacturing is like a key piece to the advantage here. If you look at distributed resources, and this is what we do as a company, we go and we install batteries on homes, residential batteries on homes, and we can talk more about like how this model works or what the value proposition to the homeowner is.
But we’re manufacturing these home batteries to be a grid resource first. If you go out on the market today and you look at, like, what are your options, you could buy a battery from Enphase, you could buy a Powerwall. These are very modular, very flexible consumer products that carry a margin for that manufacturer.
But we’re a grid resource first. And so to be a grid resource, you need to be robust, you need to be easy to install, you need to be capable of certain functionalities that you don’t necessarily need on, like, a consumer home battery. You need to be thinking about it here, like in Texas. You may have noticed it’s hot in Texas.
Batteries don’t like super hot and super cold, and so we have designed a battery that is exceptionally good at rejecting heat, and is gonna be available when the grid needs support from that peaking resource when it’s super hot outside and load is super high. And so in order to build such an effective resource, you need to control the destiny of the hardware that you’re deploying with.
And so by manufacturing, we’re able to assure that the batteries that we put out in the world are gonna be able to stand up to the Texas summers and the Texas winters, and now the Chicago winters. It’s like an important key piece of why manufacturing is so important. And what’s really cool about it being in the Austin American-Statesman building right across the street from the offices is that, for example, I work a lot on the software that does the grid coordination and the grid participation of the aggregate resource.
But whenever we notice a problem of like, oh, an inverter maybe derates under certain conditions, we just walk across the street and say, “Hey guys, check this data streams that we’re seeing. What’s going on?” And we can fix that problem right then and there. It’s a super incredible advantage to building a really performant resource.
Joshua Rhodes: Let’s talk about the value proposition to the customer, right? ‘Cause like you’re a grid asset, but you’re chasing a totally different interconnection than, say, you know, a multi-hundred megawatt battery that has to get maybe its own substation and its own gen-tie, maybe has to do all these different kinds of studies like on the system.
So like you’re chasing interconnections that already exist kind of at the home, and so what’s your value proposition, the Base model of getting those Base batteries into those connections at homes?
Chase Dowling: This is the best question. So I like to think of Base from the homeowner’s perspective first, either in a place with retail competition or on like a vertically integrated utility, like say here in town in Austin Energy.
So if I’m a homeowner and I look at Base, what is Base doing for me? In areas of retail competition, Base Power is a retailer, and we go to the customer and say, “Hey, sign up with Base. You can get a discounted time-of-use rate on your energy costs, and for an extremely low cost to get the battery installed, you don’t pay for the battery.”
Base pays for the battery. We install it on the home, and then the customer pays a monthly subscription fee. I think it’s like right now it’s like $12 a month in certain deregulated territories. But it depends a lot on where in the country you are. But a customer pays like a subscription fee to have that battery in their home, and it’s a very large battery.
Base retains the operations of the battery to offset the customer’s cost to serve, our cost to serve that customer. But then that battery’s quite large, where our Gen 3 Ma— our, our Base Core that we’re manufacturing at the factory is a 20 kW, 40 kWh battery. You know, a lot of consumer batteries that you see on the market today, they’re like 7 kW, 14 kWh.
This is a pretty sizable battery. You’re looking at like a small EV in terms of energy storage capacity. So it’s much larger than a typical single-family home that we would install on. A typical single-family home’s peak summer load is on the order of 4 to, 4 to 5 kW when you level out like the AC’s duty cycle. It’s a really big battery.
When we put it on that home, all of the members of the Base coordinated network, whether or not you have a battery, say, if you’re a Base energy customer, we use those batteries to manage cost to serve for everyone in that coordinated network. And so the value proposition to the customer that gets that Base battery is, “Hey, I’ve got a really, really low-cost backup solution, and I’m getting cheaper energy.”
And then what we do to make money as a business is we go and make that resource part of a coordinated network to manage our costs as a supplier.
Joshua Rhodes: Nice. One of the things I noticed different in, I think you said Gen 3, but you’re calling it Base Core. Is that right?
Chase Dowling: Yeah. Yeah. Yeah. A little slip of the tongue. We— our, like, product name, Base Core. Yes.
Joshua Rhodes: Okay. It was interesting seeing a little mock-up there as we were walking in. We actually got a tour from Justin and Zach, so it was like we got the C-suite tour, which was great. But one of the things that Justin was really interested to show, it had, like, a generator port as well, and it also connected to solar, right?
So if, like, you have a multi-day outage, like, he was basically saying, he’s like, “You can buy, like, you know, a cheap couple hundred dollar generator on Amazon that will charge the battery, that will then allow the battery to kick 20 kW out if it needed to or whatever,” which was really cool. It actually sounded like the concept for, like, an extended-range EV, but you were using it for, like, the battery for, like, the house. Did I make that connection right?
Chase Dowling: This is an excellent connection. I love this feature of the Core battery. It actually exists on a lot of our prototype hardware as well. This is such an important feature, so I’ll take a step back for a moment. So there is a generator port on the battery, and why? The value proposition of the battery to the customer is home backup, is peace of mind.
Texans lived through Winter Storm Uri, and the power was out for 4 days. And no matter what you do with a battery, there’s only so much energy in it, right? It’s only gonna last you so long. You can only ration that energy out for so long. You gotta be able to put more energy in to be able to use it to manage your cycling load.
But this is such a slept-on feature of the Core. I get so excited about it. What people are trying to do, when they get a backup generator for their home and they wanna make sure that their heating and their cooling is gonna run, they gotta size that generator to meet the demand of an AC pump, for example, you know, a 10 kW AC pump, a 14 kW AC pump.
And at that kind of power range, you need to get a pad-mounted generator that’s $12,000, $15,000 to meet that peak demand to turn that AC pump over. But if you ever, like, listen to how your AC runs, it only runs for a few minutes. The duty cycle of an AC is very short. And so what we can do with the 20 kW inverter on the battery is just push that pump over for a few minutes. And so long as you can charge the buffered-up energy in the battery at the rate equivalent to, like, your hourly average consumption, f— which on a home is, like, 4 kW, then you never actually run out of energy in the battery, and the inverter gives you that, like, oomph to get that AC pump turned over when you need it.
And a 4 kW genny on Amazon’s, like, 500 bucks. And so now what I need to do, if I subscribe with a Base battery, I pay my monthly fee, I get that discounted energy rate. I also get this generator port. I can get, you know, like, an unlimited duration of backup for 500 bucks for the cost of a portable generator to plug in.
And by extension, works with solar panels. The inverter can natively accept connectivity to solar panels, and a lot of folks don’t— didn’t realize that, you know, like, solar panels alone aren’t a grid-forming tool. They can’t back up a home alone. So just getting that inverter plopped on your house, that piece of power electronics, is such a huge power-up for all kinds of peripherals that you can attach to it.
Joshua Rhodes: Yeah. I mean, I was one of those Texans who lost power for 4 days too. I’m actually currently right now, like, I’m in Nacogdoches. I’m at my, my mom’s place, and she actually got a Generac in... She lives out in the country, and she got a Generac installed, like, 2 weeks before Winter Storm Uri hit. You know, lost power, but sailed through the whole thing.
Also burned 500 gallons of propane, which ain’t cheap. It worked. But you’re right, it’s like whenever that thing was running, it can only run at a minimum level so long, right? It can only run so low, and so I know it was sitting there burning, like, extra fuel that it didn’t really have to do if it were able to cycle and kind of buffer through.
And it’s really cool that the battery can do that. And so the value prop seems to be working out. I mean, it says, like, you’re currently is, you know, energizing roughly about 2 MW a day or so. Are you still on track for, like, a gigawatt by the end of this year? Is that, is that the goal or is that the stretch goal?
Chase Dowling: We set ambitious goals, but we set goals that are achievable, and we’re getting there. We’re a little over 600 MWh in the ground today. We won’t hit on a power basis a gigawatt till next year. We’re shooting for a gigawatt hour on the storage capacity. But we’re getting there, and it’s gonna be tight.
It’s been an exciting, exciting ride. That, that installation rate, 2 MW a day energized, that puts us at the top of the range. If you take utility-scale batteries and if you take what we’re giving to regulated utilities and what we’re bringing to the bulk power system as a bulk resource, if you take utility-scale battery developers, if you take a site, a 200 MW site, they take 3 years at best to get interconnected onto the grid, and you’re waiting those full 3 years for those 200 MW to show up.
And if you, like, divide up the time it takes to get those 200 MW online, that 2 MW a day that we’re energizing puts us at the top of the storage developers in Texas. Base is one of the fastest storage developers in Texas right now, and—
Joshua Rhodes: Yeah. If I do the math right in my head, which it’s easy math, so this is dangerous, you’re doing like 200 MW every 100 days. That’s a big battery every 100 days, every 3 months. You’re, I mean, you’re moving fast. What’s the limit now to moving faster? Is it manufacturing, in s— installing, customer acquisition, distribution, ERCOT rules? Like, would you wanna move faster, or is it just time? Like literally you’re just moving as fast as you can.
Chase Dowling: I think it is working on all of the elements of that flywheel simultaneously. On any given day, one thing may be the constraint. On any given day, it might be our ability to recruit enough trained installers in a particular geography, or on another day it might be part of our supply chain for our manufacturing that might be the constraint.
But then kind of being vertically integrated, we can work on these constraints independently, parallelize, increase the installations, like the net energization speed of the resources that we add to the grid simultaneously.
Joshua Rhodes: And when you say vertically integrated, I mean, you’re, you’re all the way down from manufacturing to, like, also the people who are installing these in homes, they work for you, right?
Chase Dowling: Correct. For Base, yeah. Through to the operations of the resource, the application that the customer sees, the maintenance of the installation of the hardware in the homes. And that’s actually an important piece when we go to our regulated utility partners that are looking to improve their peaking capacity in their portfolio. You know, like a regulated utility like Austin Energy, for example, needs more peaking resources on their system.
We facilitate a lot of these functionalities for them, and it’s just plug and play. Their ability to control the resource in the wholesale market, the ability to operate it like a grid-scale battery, just day one. Having control over each of those pieces allows us to offer that up as a resource to a regulated utility as well.
Joshua Rhodes: Okay. It’s a little bit more clear in, like, the deregulated space where you’re the rep and, like, you’re also bidding in ancillary services and things like that. When you’re working with a, with a muni or a co-op or other, are they controlling the batteries? Are y’all controlling the batteries? Is the arrangement different there?
Chase Dowling: Yeah, so for a regulated utility, they’re controlling the battery, and we’re almost kind of like EPC, but a very, like, well-instrumented EPC. So a regulated utility puts out an RFP, and they need peaking capacity, and they need it soon. And what are the options on the table? I could go out and get up a gas peaker.
I could go out and get a grid-scale battery. And as you know, a lot of these are hard to come by nowadays with the construction on the bulk power system. And what we’re able to do is we come up and say, “Hey, we take this utility-scale battery, we cut it up into residential battery-sized pieces. We help you. We build up a program with you to install it on your customers’ homes, and then we can give you the keys to that fleet. And the best part about it is that we can get you that 100 MW you need in 6 months, not 3 years.”
And that’s been a very attractive proposition to utilities that are facing acute peaking capacity needs on their system.
And so it looks like from utility to utility it’s slightly different depending on the problem they’re trying to solve. Say like El Paso procuring a fleet of Base batteries to manage a constraint on a particular substation versus CoServ procuring a large fleet of batteries to become part of their peaking, uh, resource portfolio.
In either case, the customer program that’s offered is co-branded with that utility. But we just simply hand the keys over to that utility through a control surface integrating into their EMS or ADMS, and they can facilitate the operations of that resource directly.
Joshua Rhodes: Well, that makes sense. I mean, one of the things about the energy sector writ large, I always kind of say, is like, if it doesn’t work in Texas, it’s probably not gonna work anywhere else since, like, a lot of folks kind of start and work in Texas.
You cut your teeth in Texas, so what lessons are you taking from Texas to other regions? Or have you found other regions that are maybe easier than Texas? What are you seeing elsewhere as you move into other areas?
Chase Dowling: Well, one of the markets that we’re going to right now, PJM, it’s all over the headlines that there is a significant need for capacity, and that capacity cannot be built fast enough.
So we’re just super excited to be stepping into that market with a new form of resource that can address those capacity needs. The lessons that I, I think that we have been taking away is building the muscle to build a resource that has never been built before at a scale that’s never been built before to solve a new kind of capacity problem.
The transmission system, the bulk power system, we spent 100 years building this giant synchronous machine, and we got really good at building it a certain way. And now with this, like, renewed voracious demand for energy on the transmission system, we need to find capacity everywhere. We need to put more dispatchable power generation wherever we can.
And so the lessons that we’ve been taking away from Texas are like a multitude. One, it’s like, how do you build up the warehousing capability? How do you manage truck rolls and device maintenance? How do you learn to work with the permitting authority in that geography? How do you integrate with that market or that utility? And exercising this muscle that works collectively such that when you show up to that new market, in one month, you know the list of problems that you need to burn down to start effectively putting that batteries on those homes and start energizing those resources quickly.
Joshua Rhodes: Yeah, that makes a lot of sense. But let’s turn to this paper that y’all recently put out, I think the beginning of September. It’s called “Flexibility Without Curtailment.” And so walk me through kind of the high-level executive summary of this paper, and particularly the title. What does that mean?
Chase Dowling: Yeah. I definitely wanna give credit to our partners for Piq Energy and Tom Nudell. They’ve been fantastic to work with on this paper, and it— they did a lot of that heavy lifting. And definitely credit to Tom for that title. That was his baby. But the title is getting at a contract that most balancing authorities in North America are beginning to realize needs to occur. We’re building these multi-gigawatt data centers that are being added to the grid, that, like, an individual data center, if it trips off grid, just the data center turning off creates material balancing and dispatch challenges for that balancing authority that needs to ensure stable frequencies, stable power supply.
So it’s kind of becoming accepted knowledge that, like, the data center needs to facilitate the safe operations of the data center. They become a responsible counterparty to it. And a lot of people have been talking about, like, well, they need to be flexible. They can’t just show up and operate however they please.
They almost become their own kind of resource on the grid. And when people think of flexibility on the load side, generally what they think about is like, “Oh, I can turn it down.” Texas, for example, has a history of a lot of steel mills being able to turn down when there’s severe shortages, and that was kind of a novel thing to add to the grid.
You know, we built a grid over the last century that just kind of assumes that load was inelastic. You designed the coordination of generation, dispatch, and transmission around meeting whatever demand was at the time that demand occurred, and there was no expectation that demand would be this fully equipped counterparty in the balancing of the power system.
And so that curtailment proposition, if you, like, dig a little bit deeper, in order for some of these data centers to curtail, their opportunity cost to curtail is tremendous. There was a recent paper, I don’t know if you saw it, by, uh, Farhad Billimoria and one of his colleagues, that looked at the value of lost load or the opportunity cost for certain data centers to turn down, and these numbers were upwards of $40,000 a megawatt hour, in some cases six digits, hundreds of thousands of dollars a megawatt hour to turn down.
If the only option was for a data center to turn down to ensure the stability of the grid, that hundreds of thousands of dollars of opportunity cost would be felt by ratepayers one way or another. And so what could we do to ensure that that opportunity cost doesn’t get dumped onto people who, like, we are privileged to pay for 4-cent-a-kilowatt-hour energy, right?
And that would just be a tremendous cost burden on ratepayers all over the country. So the core level is simply just add more resources to the grid, increase the operating envelope, and add flexibility to the grid directly. And that’s something that we showed in the paper where using this novel program that ERCOT has, the ADER program, which we could talk about, the advantage of these residential resources, these residential batteries, is because we can install them so quickly, speed to power is the name of the game nowadays, and so, like, this is one of the fastest resources to install.
And if we can install them at meaningful scales, those batteries aggregated at target substations can help facilitate the flexible operations of that grid and help make it more robust in the presence of these very large data centers.
Joshua Rhodes: Yeah, and I think my understanding for, like, a training data center is that it just can’t, like, stop for, like, a 15-minute interval, because if you wanna move all of those weights that you’re tuning from RAM to disk, it can take, like, 4 hours to, like, stop and then 4 hours to start again.
And so, like, for a 15-minute interval, it’s like an 8-hour interruption. The IT load is not inherently as flexible. I know that there are companies working on that, like Emerald AI and other folks, but at least I believe that that is kind of how a lot of that works. And again, I probably got the, the terminology kind of all wrong there.
But I think one of the things that I was really stoked about this paper was that it really pointed out that location mattered, right? Because these new large loads that you’re putting in places, like, they put an acute stress on particular pieces of infrastructure of the grid. The whole grid growing by 100 MW is, like, that’s just in a, a rounding error.
But, like, a couple hundred megawatts at this particular place, that has real implications for the system. So I was wondering if you could talk through those physics in hopefully relatively plain English. One of the things in their paper, you know, you talk about is if you put 100 MW of new load at Burleson Switch, which is, like, a particular bus, like, on the grid in just around the Dallas area, in the Oncor region. You put 100 MW there, what happens? What does the paper show that you’re trying to do?
Chase Dowling: Whenever you put a big piece of new load or new generation on a part of the transmission system, there’s a whole list of, kind of like a checklist of problems that you care about when you connect to this big synchronous machine.
But one of ‘em is, like, is there simply enough room on the transmission lines? And the way the TDUs and the way Oncor assesses whether or not there’s enough room on these transmission lines is they look at operating conditions, different times of year, and different failure modes. Those failure modes, we call them contingencies in the industry, and then they watch them.
They form what are called constraints on the system. Constraint is like a transmission line and a contingency. A contingency is like a power plant goes out or a meteor hits the Earth, and these sit on a spectrum of probabilities, right?
Joshua Rhodes: Hopefully the latter one’s at the tail end of that probability there, but okay, I’ll let you have it.
Chase Dowling: Super fun digression about solar storms, but it gets really interesting on the kinds of things that they’re monitoring on the contingencies. The short story here is that there’s only so much room on these transmission lines, and if I add a new big load, given that the system is coupled, if the load operates the way the load wants to operate, there may not be enough room on that transmission line.
Now a lot of constraint violations on the transmission system today are quite brief. Not all, but many are less than a couple of hours. And so when you add things to the transmission system, you affect a whole bunch of transmission lines all around you, not just like the one that you’re literally connected to.
So as you saw in the paper, if I add this 100 MW load to that spot, a whole bunch of constraints pop up, like, oh, this transmission line, that transmission line, that transmission line. And what we show is that when you put it in a spot, when you hit this transmission line, it doesn’t mean that like one extra megawatt of load at this data center site is one extra megawatt on every single one of these constraints.
It could be as few as, you know, half a megawatt or a third of a megawatt. This is, in the industry it’s called a, a shift factor or like the locational impact to all of the things around you. Now, that constraint, if it’s very far away, it might happen to be in a part of Dallas where there’s a lot of homes, a suburban part of Dallas where you can put a lot of batteries, where is really close to that constraint that cannot be ameliorated unless you were to increase the size of that transmission line or install new resources.
And those resources have like a mechanical advantage by being so close. One megawatt of injection from those batteries can realize 1 MW of relief on that constraint, whereas the data center perhaps a little further away, 1 MW of curtailment on that data center only realizes, say, like half a megawatt of constraint on that transmission line.
So as you saw in the study, there was an example data center in Burleson where a 100 MW data center, all of the constraints that it creates can be ameliorated by 80 MW of strategically sited batteries. And if you s— look closely, there’s like 10 to 20 MW pockets in a number of places that are all strategically dealing with the numerous constraints that arise.
And that’s kind of the short of it, what’s happening with the connection of the resource and what is preventing that resource from being operated safely under all the contingencies that ERCOT analyzes.
Joshua Rhodes: At the risk of getting too wonky here, please don’t answer anything that uses imaginary numbers, ‘cause even me as an engineer won’t be able to understand what you double Es do.
But you brought it up. This was 100 MW, like, additional load, but you only needed roughly 80 MW of resources to, like, effectively make that load okay to host. I mean, people might think of it as like, okay, if you put in 100 MW of load, you need at least 100 MW of stuff on the system, but you’re able to do it with less. How does 80 MW of resources allow 100 MW of new load?
Chase Dowling: Yeah, yeah, yeah. So if you look at the transmission system in Texas, there’s tens of thousands of buses where you can go and, like, install electrical hardware, and there’s not tens of thousands of buses that can host 100 MW of load.
You need the land. You need the high-voltage transformers. It takes a ton of space. It actually limits, like, where on the transmission system you can put these things. But the constraints are everywhere. They’re in the middle of the city. They’re in the suburbs. They’re in, out in the rural parts of the state where the transmission lines are running.
And you can place batteries physically closer to where the constraint is occurring, depending on where the data center’s being built. And so by very simple mental model that works is, like, by virtue of just being closer to where the electrical need is, not where the load was built, you can more effectively control the problem that’s arising.
It’s like I’m in a room, and the data center’s in the room with me, and we both need to shut off the light, and the data center is on the other side of the room, has, like, a stick to turn the light, and I happen to be right next to the light. It’s really hard to reach the light switch with that long stick, and I happen to be right next to it, and I don’t need to work as hard to turn that light switch and effectuate the control.
Joshua Rhodes: Okay, so one of the things I really also appreciated about this paper there is it didn’t stop there. It’s like you next go to a 200 MW case, and instead of less than 200, you actually need more than 200. You need about 300 MW to alleviate 25 of the 30 constraints. And to get to all of them, you would need like 1,200 MW.
So it was like the problem got exponentially harder as you made the loads bigger and bigger and bigger. And so, one, I appreciate you putting that in the paper to be realistic about it. Can you talk about, like, how did it work so well for that 100 MW? What became harder as those loads got larger?
Chase Dowling: Yeah. So there’s almost like two questions in here. Why does it work for some data centers, not others? And then why is there like a long tail? Why does it to— take such a large fleet of batteries to deal with all the constraints that materialize as the load grows? So with the first question, residential batteries aggregated at a load substation are not a panacea for all the problems, right?
Like if there’s a data center being built in West Texas, there simply aren’t enough homes in West Texas to aggregate a multi-hundred megawatt fleet of batteries. It depends a lot on like the local topology of the network, the available real estate to work with to put more resources on the system. It’s case by case.
But if you look at the data center map and where a lot of these systems are being constructed, that middle of the road, the 100 to 300 MW set of data centers, they’re getting built all around the exurbs of the major metropolitan areas, and a ton of them are getting constructed. And so there’s a ton of constraints that are gonna be able to be managed where there’s proximity to big residential communities where we can stack up tons and tons of batteries.
Now, is this gonna solve a problem out west of like Abilene, where there’s very small neighborhoods and not a ton of places to put residential batteries? Probably not. So one, it depends on the topology of the network and what’s available to work with. And then on the like tail of contingencies. Now, like the number of situations that the TDUs and ERCOT look at as failure modes on the transmission system are huge.
Like NERC has like this special ranking system for all the different failure modes that the transmission system can go through to the most extreme, where like our most important transmission lines outright fail, they fall down, they break. We saw a transmission line failure like that down in the Houston area not— about a year ago after a hurricane.
But those are exceptionally rare. And to deal with those constraints, it oftentimes severs whole parts of the system. And in those extremely rare occurrences when you sever that part of the system, you need to make up for all of that lost generation on both sides. And so to fully ameliorate that particular extreme constraint, you do need a ton of generation.
But back to that title, back to that recognition that there’s a realization that these enormous loads joining the system will have to form some amount of flexibility that can be offered to the balancing authority to deal with these extreme cases. I mean, this is kind of what SB 6 was getting after, where like in the most extreme cases, you gotta be able to help the grid from collapsing.
And you don’t need to spend a billion dollars to deal with the most common, the preponderance of the various failure scenarios where a breaker opens, a power plant trips, there’s some sort of problem. These problems happen daily on the transmission system, and they cause these shocks to price that we s— we all see on the grid status dashboard of the nodal price even just yesterday.
Those are things like transmission outages occurring, and breakers opening, and power plants tripping, and things like that. Quite common. And you can deal with a reasonable sized fleet of batteries that are much, much, much, much cheaper than a data center curtailing for several hours. You can deal with 80 percent, 85 percent of all of the constraints that are monitored, all of the contingencies that they map to, the very effective resource.
And so we wanted to show, which is really important, and not a lot of grid planners like think of the grid in this almost good enough way, but you can deal with a ton of the problems with so little work, with so little cost if you’re able to aggregate up these, these small batteries and manage the system more dynamically.
But as you point out in the paper, as you get up to those extreme ends, that’s kind of, I think, what’s on people’s minds when they think of like, what was SB 6 for? How do you deal with the most extreme failures in the transmission system that virtually never occur?
Joshua Rhodes: Yeah. That sounds like the 80/20 rule strikes again, right? You can take care of 80 percent of the problem with 20 percent of the effort or the resources or things. And so, I mean, I think that’s what your paper is kind of getting at. It didn’t take that next step, but it teed it up, I think, really well in terms of like, okay, we can aggregate these resources strategically placed to alleviate, you know, part of the issue, but it doesn’t take away the need that sometimes you just need to build new transmission. Sometimes you just gotta move more bulk power around the system. Am I summarizing that correctly?
Chase Dowling: This is totally accurate. There’s like a question of like, if I had just built more big transmission, maybe that’s an alternative to this problem here. But where all of the threads are converging to take a— another step forward, if we look at what the batch process is doing in Texas, the batch process is looking at data centers that are electing to be curtailable resources, provisional curtailable resources, and they don’t wanna have to do that. Sure. And if I’m signing up to be curtailable, that means I’m gonna do it all the time. I’m gonna have a large opportunity cost to do it. And we have this resource now, this resource that ERCOT can see in their dispatch mechanisms through the ADER program that we can build in less than 6 months to the data center that’s signing up for that risk of curtailment, oh, we can deal with most of the scenarios that you will— can be confronted with when being forced to curtail.
Now, yeah, there’s those extreme emergencies where for the safety of the state’s grid, we need to turn some things off that y— the only way to deal with it is to just build a ton more extremely expensive transmission, extremely expensive generation on both sides of a failure mode.
But if I have multiple hundreds of gigawatts of data centers electing to be curtailable, they’re taking on a lot of risk there for their business, and we can come in with these resources strategically and prevent the preponderance of that risk that they’d be taking on.
Joshua Rhodes: It sounds like, and this is probably work that’ll happen in the Lege coming around, is like, so you got the Batch Zero process, you got the PCLR, and I think what most of these loads want is actually a combination of both of those, so they’re able to, like, be flexible but have the generation also interact.
It sounds like you would need that combo, which may come in the Batch 1 process. I wanted to talk about these assets, you know, being able to allow more load, like who should be paying for that? Is there a business model you have or that you could develop that would allow the cost of these strategically placed assets to be borne by the data center who wants that 100 MW? Like, how would that work out?
Chase Dowling: There’s like two pieces to it, but this is where those threads in the batch process that I was talking about combine. If I’m a data center and I’m joining the batch process, and in order for me to get connected faster, I’m gonna elect to be curtailable, virtually any alternative form of generation to curtailing at hundreds of thousands of dollars a megawatt hour is good news for me.
And so the question on my mind is like, what generation can I pay for to prevent that curtailment? What causes the curtailment and how can I prevent it? And what we’re showing is that, like, the strategic siting of these resources can prevent that curtailment, and we can build them fast and cost efficiently.
And so that sets up a natural commercial transaction where the data center can pay for an aggregate fleet of batteries at these target substations to prevent their curtailment in operations. And so in this case, the data center would pay for the batteries. And there’s kind of like this fabled win-win-win-win-win-win where the data centers that are struggling with just general public concern over their impacts to the grid can pay for a resource that directly benefits the rate base, right?
We’re gonna put these batteries on home o— homes that are, you know, at risk of outage. Now we can offer you a tool that provides outage protection while also being able to recruit all those resources and prevent the very expensive curtailment of the data center, which would eventually land back on a ratepayer’s energy line item, right?
And then outcome there is that the grid’s more resilient, reliable, and data centers paying for that capacity that can deploy very quickly. So that’s what the commercial transaction would probably look like. The piece here, kind of the challenge and what the ERCOT is proposing as the fourth phase to the ADER pilot, is having these resources, these aggregations of batteries modeled and dispatched at transmission substation level so that they can be recognized at the transmission planning phase and in the operational phase when these constraints actually materialize.
Joshua Rhodes: That’s a very wonky point, but I think that’s really important, right? Because the location of these resources matter, you gotta see them in the SSWG files, the PQs for the thing, for the grid topology files. You gotta be able to see them, you gotta be able to model them. And to just pull a little thread on kinda what you said, the part about, like, the data centers and the PR, public relations, that is not great right now.
The value prop is more about just reducing loads. Like, it— you’re actually putting infrastructure at people’s homes. It’s not just, like, nebulous, it will reduce your bill, whatever, add it all together. It’s like now there’s this box on your house that gives you power whenever the grid goes down. Running short on time, so zooming all the way out, we got a lot of data centers coming to Texas, a lot of large load.
How large can this concept realistically get? Have you looked at that? Are you able to extrapolate from what you’ve done? How big could you go if you didn’t have annoying things like supply chains and time to deal with?
Chase Dowling: Yeah. Yeah, yeah. So if we’re looking just at Texas, the question is what is the opportunity space for storage acting as an energy resource or an ancillary service resource, or now, like, additionally as a transmission resource?
And backing me into a corner, no one is building batteries in a fashion as a transmission resource today. There isn’t quite, like, a market for that yet.
Joshua Rhodes: I don’t think you can. Like, transmission can’t own generation, right?
Chase Dowling: Right. But it’s gonna take us 10 years at best to get these 765 kV lines built to improve the availability of the transmission system, make it more robust.
I think the opportunity is quite large in Texas, definitely on the order of gigawatts. But to spread the good word of Base just a little bit, if you humor me, like, I think the aspiration of the company is we see these distributed energy resources and just the proliferation of power electronics onto the distribution system, a means to help electrify the entire world more effectively.
And so we see the upsize, the volume, how big can this go, to be extremely large. We built grids designed 100 years ago, and now we have the ability to reimagine how that grid is built. And so we’re viewing the upside as quite large, quite unlimited in terms of what we can build on the system today.
Joshua Rhodes: Awesome. Love to hear it. I think this is a good place to stop. Chase Dowling, thank you for coming on the Energy Capital Podcast.
Chase Dowling: Thanks, Josh. It was great to see you.
