During Winter Storm Heather this past year, battery storage reduced power costs in Texas by $750 million. This finding was only one of many in a report released by Aurora Energy Research this May, “The Role of Battery Energy Storage Systems in the ERCOT Market.” The report is one of the most in-depth analyses to date on the impact of battery storage, which has grown exponentially in Texas over the past decade, even surpassing growth in markets with major battery subsidies.
In today’s episode, I’m joined by Olivier Beaufils, one of the report’s primary authors and the lead of the ERCOT Market team at Aurora Energy Research, a company focused on providing power market forecasting and analytics for energy markets around the world.
Olivier was gracious enough to go through the key takeaways of the report and break down its findings with me. We discussed how the study calculated system’s savings during Heather, whether these types of savings can be expected in the future, how batteries are pairing with natural gas and renewables in the ERCOT market, the potential of long-duration energy storage, and more. Olivier also detailed what happened in Texas on September 6th – wherein ERCOT called an Energy Emergency Alert level 2, the first time it did that since Winter Storm Uri – and how batteries kept the system running and Texans’ lights on.
We recorded this episode prior to Hurricane Beryl's impact on Texas. While we don’t discuss its specific impacts and aftermath here, we are preparing a series of episodes in the coming weeks that will focus on Hurricane Beryl, its effect on Houston, and strategies for improving Texas’s resilience against future hurricanes and extreme weather events. Despite not directly addressing Beryl, this episode remains pertinent to these conversations, particularly as home battery storage systems are crucial tools for community safety and resilience.
I learned a lot from this report and from the conversation. After listening, I encourage you to take a look at the report as well; it is free and publicly available. Thank you for being a listener and don’t forget to like, share, and leave a five star review wherever you listen to podcasts.
Timestamps
3:17 - About Aurora Energy Research
4:25 - Summary and key takeaways from Aurora’s Battery Energy Storage Systems in the ERCOT Market report
9:02 - Pace in battery storage growth: what spurred it
12:18 - How Aurora calculated battery savings during Winter Storm Heather and if this level of savings is typical
17:57 - Batteries as ancillary services
21:55 - Shift in scarcity hours in TX, the duck curve, and how batteries are helping solve peak demand challenges
28:09 - Findings on peaker plants, natural gas, and the need for flexible resources
33:05 - Demand growth
37:42 - What happened on September 6th and how batteries kept the system running
41:23 - Long duration energy storage
Show Notes:
Role of Battery Energy Storage Systems (BESS) in the ERCOT Market - report from Aurora Energy Research
Energy Unplugged Podcast from Aurora Energy Research
Transcript
Doug Lewin
Olivier Beaufils, Aurora Energy Research. Thank you so much for being with us and welcome to the Energy Capital Podcast.
Olivier Beaufils
Thank you for having me.
Doug Lewin
Why don't we start, Olivier, just with a quick word about Aurora Energy Research. Tell us a little bit about the firm and what y'all do.
Olivier Beaufils
Yeah, absolutely. So Aurora Energy Research, we're at this point, the largest dedicated power market analytics firm globally. We're born out of Europe and expanded sort of in all of the other markets, competitive markets around the world in the US about four years ago. And really the main thing we do is we go in depth in studying power markets, in constructing forecasts, bankable forecasts that people will use to then go finance their project to understand the impact of the changes in the energy market on the business, on the market, on consumers and the system as a whole. So that's really what we do day in, day out. And that's really how we approach this report.
Doug Lewin
Yeah, you guys have really become, I mean you have been around the world for a long time, but in the United States really have become a go-to source on all things power markets. Really appreciate the expertise y'all have and encourage folks that are looking for energy-related podcasts to check out. I believe it's called Energy Unplugged, right? Is the podcast of Aurora. I'm a regular listener. Y'all do a great job over there as well.
So Olivier, you authored a report called Role of Battery Energy Storage Systems in the ERCOT Market. This was a very interesting report, turned a lot of heads, got a lot of attention from folks in the industry. Can you just, let's just start with just a quick summary of the report and some of the key takeaways from it.
Olivier Beaufils
Yeah, absolutely. So the main reason for this report is there's been a lot of noise, a lot of discussions about last summer, about the last 12 months and the roles that batteries have played in the market in those last 12 months. Obviously we went and had one of the hottest summers of the last 10, maybe 20 years in Texas, a pretty significant winter storm as well. And so we wanted to understand, you know, what batteries had done or not done during those events and shed some light on this.
So in terms of the key findings, really, the first one is batteries are quickly playing a critical role in the system. They've grown from roughly nothing back in 2020, 2021 to 4 gigawatts at the end of last year and then close to, I think the number is around 7.7 gigawatts right now being synchronized to the grid. So just a tremendous amount of growth and from essentially being, representing nothing of the generation capacity in Texas to over 5% of that capacity right now. And so this is a technology that we need to understand and this is a technology that is providing a lot of benefits to the grid. And so we dived into exactly what that is from that initial look.
And the first thing is, on regular days, batteries are constantly providing benefits. In particular, they're providing what is called ancillary services. I'm sure most of the listeners will be familiar, but those are critical services that are needed for ERCOT to maintain reliability, maintain frequency, ensure that there is enough capacity in case of there's an event on the grid. And by providing those ancillary services efficiently, they've been able to decrease the costs. And they've also allowed some of this gas plant to shift from providing those ancillary services into the energy market. And so day-to-day batteries are very active in the market, and they're providing services that are quite critical for reliability.
But the other one, and I'm sure we'll go into more details on that in this conversation with you, Doug, is the role during some of the critical or extreme events that we've seen. So summer 2023, as everyone that was in Texas during the summer will know was extremely hot, created a lot of tightness on the grid, a lot of tension. And there were a number of days, in particular one day, September 6, where batteries kept the lights on. Without the batteries on the system, ERCOT would have had to start shedding load. And so just that basic reliability of keeping the lights on is one of the things that batteries were able to do just in the last 12 months.
But also during winter events where we saw batteries saving essentially hundreds of millions of dollars. So in our estimates, about $750 million just in two days in system costs by providing ancillary services and making the system more efficient and enabling more capacity, especially gas capacity to participate in the market. And so therefore decreased prices for consumer by having more of that capacity and more generation online in the system. So they've been providing reliability services, they've been providing system cost decrease and overall just a lot of benefits to the Texas system.
Doug Lewin
Yeah, you know, I think a lot of people when they think about storage are still, and I think this is true even of people in the industry, although it's sort of probably rapidly fading away, but still think of storage as this futuristic thing. This is something that's going to have a real big impact in a few years. It's coming. But what we're really seeing is like right now it is having a major impact. So we'll dive into all of these, but you talked about September 6th and the determinative role that battery energy storage played on that day. You talked about the $700 million in January on a couple of days in January during Winter Storm Heather.
But I do just wanna just back up and just put a finer point on this outrageous growth. And of course, we'll put a link to the study in the show notes so people can go find it and see all the great graphs and charts that are there. But as you point out in the report, there was really effectively hardly any storage at all as recently as 2020. And then just this meteoric rise all the way up to 4,000 megawatts at the end of 2023.
We're recording in mid-July, the GIS, the Generator Interconnection Status Report just came out a couple of days ago and has storage all the way up to 7.7 gigawatts sitting here in 2024. So really in the space of four years from effectively nothing to nearly eight gigawatts. It's really just been a meteoric rise. Can you talk a little bit about what that means?
I think also you just cited the battery capacity as a percent of peak load, which again, you have a great graphic in the report showing again, effectively zero in 2020 all the way up to 5% in 2023. And it looks like by the time this year is over, that'll of course depend on what the peak demand actually is and all that, but will be somewhere around 10%. Is this in line with expectations you had a few years ago? Did you kind of see this coming? Is this surprising to you? And what does this mean overall for consumers in Texas and for the market?
Olivier Beaufils
Yeah, I mean, if I had to be honest with you, I think this has been way faster than even we were predicting. And so we have quite a lot of experience with batteries. As a UK company, we started seeing a lot of the battery ramp up in the UK, which was one of the first markets to see really a lot of development in battery storage. And so we had a fairly strong understanding and a forecast of battery storage pretty early on when we started covering ERCOT. But this growth has actually surpassed everything we could imagine. And I think pretty much everything that was out there in the market that we were aware of.
So it's been much faster and much bigger in scale than what we thought. And what's really fascinating in ERCOT, that's not the case in most other markets, is that growth has been built entirely on market economics. There's no mandate for battery storage growth in Texas. There are no requirements or RFPs that are being put out. And so really it's market economics, market constructs, and obviously federal subsidies. Let's not forget that batteries are eligible for the ITC, but beyond the federal subsidies, it's really the market economics in Texas that are enabling that.
Doug Lewin
Yeah, and I didn't see it coming either. I think anybody that tells you they could see that we were gonna be at eight or nine gigawatts at the end of 2024, and it's looking more and more like 15, 16, or even higher, gigawatts in 2025. Anybody that called that in 2020, I would like to see where they called that in writing.
So I wanna come back to this, what I think is one of the major takeaways of the study is you went and looked at a couple of days in January. Again, it was Winter Storm Heather, it got very cold. It was similar not to Winter Storm Uri, but to Winter Storm Elliot, temperatures in the teens and things like that around the state, conditions were tight. They actually looked like they were gonna be tighter. So there were some interesting dynamics in the day ahead market, but you looked at both the day ahead market and the real time market and said, the savings from storage were somewhere around $700 million. You talked about this a couple minutes ago, but can you dive a little bit deeper into how you arrived at that number? Because that's a pretty eye-popping number, $700 million in savings to the market over two days. And I guess the other relevant question is, do you think that we'll see more days like that going forward, or was this just sort of a one-off anomaly in the past?
Olivier Beaufils
That's a great question. And that was part of the core analysis we did for the study, understanding the impact on costs that batteries had on system costs, right? Because this is an important distinction. Those $750 million, you know, just like the ECRS cost that was mentioned last summer, those aren't directly consumer costs, right? There are system costs and then get passed on to consumer, but via different mechanisms.
So what we looked at is what would happen essentially, if you take out that battery capacity that was offered in the market in each of those hours over the course of the storm, how would that impact the pricing? So we have all of the offer curves and the awards from ERCOT. So for those two days, for each hour in each 15 minute period. And so we looked at these curves, reconstructed the entire offer curves and looked at where the market cleared. So we had a very specific understanding of the amount of capacity that cleared and the price for each of those hours and looked at reconstructing it essentially without the battery capacity. And the key here is to understand where that battery capacity was, right? For the majority of those hours, the majority of the battery capacity, the vast majority was providing ancillary services, right? So regulation services and then what is typically called spinning reserves, In Texas, we have different names for those. They're broken down between the responsive reserve service and then the emergency, the ERCOT contingency reserve service, and then obviously non-spin.
So they were providing those ancillary services, but if they had not been providing those ancillary services, that means other units would have had to do that, right? And typically, historically in Texas, it's been gas plants, primarily peaker plants as well as regular CCGTs providing those ancillary services. So if you take out the battery capacity that was providing those services that day, that means you're effectively taking away those gas plants from the energy market. And so then you can reconstruct an offer curve to understand where the market would have cleared for each of those hours or each of those 15 minute periods in the real time market.
It turns out that the day ahead market is more impactful in those days. ERCOT was forecasting a very large amount of day-ahead demand and so the market cleared very high in day-ahead. So that, and ancillary services are also cleared in day ahead. So the day-ahead market impact was really the largest one. But if you do the analysis on the real-time market, it's a very similar number that you arrive at. And essentially when we reconstructed it for each of the hours of those two days. So really just in a 48 hour period, you added up those savings to that $750 million number of which obviously the key sort of morning hours for those two days are where the majority of the savings are because the market was clearing at that $4,000 price. So that's how we put together that analysis and really understanding the impact that battery capacity had on market clearing and pricing.
To your second question, if I can jump in into that one, is this going to happen again? I want to say yes. I mean, this is a feature of the ERCOT market. When the system gets tight, the market generates high prices, right? There's mechanisms for that in the market but we've also seen batteries offer, also generate some of those higher prices. It just creates those high price signals. And so we will see again, whether during winter storm or summer temperatures, high prices in Texas. There's no doubt on that.
One of the things that people look at is, well, but what if we build a lot of capacity? What if we see all of this new battery capacity coming online? What if we see all of the thermal capacity that's being incentivized by the Texas Energy Fund come online, then the system has a lot of margins and maybe we don't see those situations again. Well, maybe for a few years, right? But load is growing. We're all hearing about electrification, about data centers, et cetera. And so that only lasts until load catches up. And then you're in a situation where you're going to have moments in during the year where the system gets tight and those price signals are generated.
Doug Lewin
Yes. And we will likely see other events where, I mean, the storage is already there. As we were talking about earlier, it's growing rapidly. So we will likely see days where we see similarly large savings, again, to the system, not necessarily directly to consumers, but to the system. The way you just described this, right, if I'm just summarizing, you basically have storage that is able to participate in ancillary services. This frees up some gas generation that would have been tied up in ancillaries. It's now in the market. That puts some downward pressure. You had quantified, there's the three different ancillary services you looked at, RRS, Reg Up, and Reg Down. Between 50 and 75% of those three are now batteries. Of those three, we have seen steep cost declines in each of them ranging from 20 to just about 50% cost declines. But we have not seen a cost decline in non-spin, whereas you said batteries are generally not competing because it's a four hour product. So we are seeing these savings on these particular days because gas is free to be in the market, but we're also seeing savings sort of spread across the year within these ancillary services, which as people that follow the market know, those have been rising in cost as well. Do you wanna talk any more about that finding and overall, for instance, I guess on RRS, a 49% cost decline from 2020 to 2023 in the cost per megawatt hour. That seems like a pretty significant finding as well, pretty large savings to the market and ultimately to consumers as well, correct?
Olivier Beaufils
That's right. And those ancillary services aren't small in aggregate cost, right? There are hundreds of millions of dollars, even several billion dollars total, if you pull all of them together, so those add up to real numbers if you calculate the potential cost savings that batteries have provided.
And why is that, right? Batteries are particularly well suited to provide those services. If you think about regulation up and down, batteries can ramp up, ramp down very, very quickly to respond to market adjustments for the frequency. They can also ramp up within seconds or less to respond to an outage or a unit tripping and provide essentially a spinning reserve service. And so they've been just much more efficient at providing those than sort of the legacy thermal plants that we have on the system, which have a higher cost in providing the same services, right? Because they have to consider their fuel costs, their ramp up time, and if they're not already operating their startup costs.
So as all of these combine, essentially, we're seeing real savings as batteries continue to provide more and more of those services. And we expect that trend to continue. Essentially, we haven't hit the floor yet. We're at 75%, but we've seen markets where batteries are starting to provide close to 100% of those services. And there's still room for those to decline. And not only in the services we've mentioned, but also in the ECRS service, which we didn't put on this chart because it was just created in 2023. So there was no comparison point to see an impact on costs. But as more and more batteries come into market and are able to provide those, we fully expect the ECRS service to see the exact same dynamic and also come down and become less and less costly to provide.
Doug Lewin
Yeah, I mean, in a word, it kind of comes down to competition, right? If you have energy sources competing against each other, that's ultimately good for consumers. And even when storage comes to dominate a particular category, storage resources are competing then against each other, right?
And yeah, and even in the spot market, some storage developers and operators are going to want to hold out for the $3,000, $4,000, $5,000 a megawatt hour. But of course, they're risking missing the thousand and $2,000 megawatt hour days. So those competitive forces at work are really making a big difference there.
You also got into some of the big changing dynamics on the grid. I want to talk about this as well, that, you know, again, back up say to 2019 before there was really any storage to speak of on the grid. As a matter of fact, 2019 was a pretty hot summer. We saw some pretty high priced days, some pretty, what we thought were really, really high peaks at the time, 75 gigawatts seemed like a lot. Now 75 is just sort of like an average summer day or something like that. But what we would see in 2019 was a really long peak period when we were sweating both literally because it was hot and figuratively because everybody was worried, was the grid going to stay up? You know, the general public wasn't as clued into those things at the time as they were post-Uri, but for those of us that were really watching the market, there were some pretty dicey days, but it was always starting about three o'clock and really like four or five, six o'clock in that period that everybody was worried about. Obviously we've talked about the numbers with storage, but the numbers with solar are similarly eye popping. Can you talk a little bit about the shift in scarcity hours from, you know, call it like 3 to 7PM really to kind of a 7 to 9PM period and what that means for the market, how storage plays a big role in dealing with those problems?
Olivier Beaufils
I mean, as you very well explained, I think the peak load hours in Texas are really three to six or seven, right, essentially normally. And back in 2019, we had little solar on the system. We had a few gigawatts, but it just wasn't a big factor quite yet. One of the things that I haven't mentioned yet, but that's obviously a good time to do that, is the growth in solar has also been tremendous in Texas.
Texas has added about 12 gigawatts of solar in the last 12 months only. And so essentially doubled the solar installed capacity between what was installed at the beginning of 2023 and what's installed now. And actually we've more than doubled by this point because of the interconnection at the beginning of the year. So solar is also now a force in the Texas market and that's changing the dynamics because as most of your listeners will already know, what matters for dispatchable capacity and maintaining system reliability is not actually just the peak load, but it's the net load, right? So it's the peak load minus the generation that you're getting from non-dispatchable sources. So renewables, wind and solar essentially in Texas. And so before the events of solar, wind tends to generate at night primarily and less so during the day. And so it wasn't solving your problem during the day, but definitely helping during the evening hours. But solar obviously generates during the day. So it has this effect that really helps the system. \
What happens is solar obviously ramps down at sunset. So right around 7, 8pm especially as you get further out in the summer into late August, September the sun will set earlier. So by the time you hit 7pm you start seeing really an impact, but the wind hasn't ramped up yet. And so you're really in that valley where the solar is coming down, the wind's not ramped up yet, and you have a two to three hour window where suddenly your net load is becoming quite high. It's rising because you don't have the solar anymore, but you don't have the wind yet. And that's really where you need as much dispatchable generation as you can because Texans are still consuming a lot of power. As we know, it's still very, very hot at 7 PM in the summer in Texas. But suddenly you have to bring these gigawatts, right? These gigawatts on the system to maintain stability. And that's where battery storage play this really, really helpful role to the system in that they can shift some of that solar from the middle of the day that's cheap or even sometimes, even from the night before that they could charge from the wind and then they can dispatch it right during those hours that are really stressed and where typically prices are also high so they're capturing those high prices. And that's really that new dynamic on net load, right? In California, they call it the duck curve because of the solar valley during the day, but in California they don't have the wind. And so that duck curve has a very different shape. In Texas, it's very peaky where you really need all that energy during those few hours of the evening.
Doug Lewin
Particularly well suited to storage and particularly the shorter duration, couple of hours solves the vast majority of the problems because of the difference as you were describing it in the Texas duck curve. Sometimes I hear it referred to as armadillo or dead armadillo curve, or people have come up with some creative things to replace duck curve, but it still kind of looks like a duck, particularly, and again, in your report, you have some graphs that show this. As you start to get out into the 2035, 2045 kind of timeframe, you do see some steeper ramps there and storage seems particularly well suited to deal with those ramps, right?
Olivier Beaufils
Absolutely. And those ramps are becoming bigger. So not only is storage well suited, but all of the storage that's coming online is actually needed.
Doug Lewin
Yeah. And you know, the other variable to consider there as well, Olivier, is the thermal outages as well, right? We see there are times of year where… that's really what I try to help people understand is that you really have these three variables. It is the demand, which is mainly in the summertime driven by cooling load and in the winter, obviously, driven by heating load. And then you have these other factors of renewable output and thermal outages. And so when you put all three of those together, that you got that high demand, low wind, high thermal outages, that's really where you end up getting into trouble. I'll let you speak to that, but I just want to point out that this report, while it is obviously very bullish, I would say anyway, on storage and what it means for the market, has some pretty interesting findings on gas and gas peakers as well.
One, that with the rise of batteries, we haven't seen a decrease in gas or the use of gas peakers. So I'd love for you to speak to that, but there's also a graph you have that shows that Texas in 2035, according to your modeling, will need about 38 gigawatts of capacity with a load factor above 30%. So not to be not too jargony for folks that are more at a beginner phase of this stuff and want to keep up with these conversations, we're talking about power plants that would be running, you know, more than one third of the hours during the year. Cause a lot of power plants are going to just run in those hours of scarcity where prices are high and the system really needs it. And basically what you're showing is a system that is moving more and more towards valuing those resources that do operate with lower capacity factors and you quantified a need or you put a potential capacity need of 20 gigawatts of flexible gas generation and other 20 gigawatts of storage. But that would suggest, and you put it pretty clearly on your slide that among current combined cycle and coal plants in ERCOT, there's about 55 gigawatts. We probably only need about 38 of that in 2035.
So this is a very big shift in the market, not happening tomorrow or next year, but over the course of a decade, doesn't mean we need no baseload in your view, as expressed in this report. We're gonna need some, but probably less of that, but a lot more of the flexible stuff running 5, 10, 20% of the year, rather than 40 or 50%. Did I characterize that correctly? Please feel free to say if I didn't, but then also add any additional color or explanation you'd like.
Olivier Beaufils
No, that's exactly right. That's a good summary. I think the fundamental shift is that the system needs more flexible capacity. It needs less of those baseload generation. As you said, it still needs some, right? And it can take different technologies to provide that. But it doesn't need as much as it needed in the past because we have all this renewable generation coming up, because demand patterns are also evolving, people are consuming electricity at different times. And so that flexibility is really what is driving a lot of the needs.
And maybe just to go back to the use of gas generation, I think one thing that people sometimes miss is the fact that this is not a zero sum game, right? Where if I have a little bit more of something, that means I have to have a little bit less of something else. And the reason for that is because we are in a system that's evolving quickly. We're in a system that's growing in terms of demand pretty rapidly. Right? And as the system grows, it needs generally more capacity, but it also needs different types of technologies to provide that capacity. And that's really the fundamental thing that I think people need to understand is, you know, batteries aren't going to suddenly displace CCGTs. They're playing different roles. Some of the, you know, baseload power, mostly coal plants are less needed because they lack, especially coal plants, they lack that flexibility. And so as they get older, a lot of those units are retiring, right? We see all these announcements from the different utilities of coal retirements. And it's not because, you know, people just, it's not for any political reasons in Texas. It's really because those coal units aren't economical. They're not providing the services that the market really needs. And so the market's shifting to different technologies that are providing those services. So we're seeing in our modeling a continued need for gas, but gas that is more flexible. So really peaking capacity. And we're seeing that need remain in a way with the growth in battery capacity because it is playing a different role, because it is going to generate, you know, say for the longer duration events that the current battery fleet at one, two or four hour is not really well suited to fulfill.
Doug Lewin
Yeah, exactly. I think that to the greatest extent, we can get folks to think of these issues in terms of the system and not any individual favored technology or group of technology, but the system and how the system works together, the better off we're going to be.
You mentioned load growth and how this isn't a zero-sum game because we are, there's a lot of reasons why it's not a zero sum game, but one of those is rapidly rising load growth. How does that factor into your modeling? You or I don't know if Aurora has done particular work on this, but obviously there's been a ton of attention on ERCOT sort of bombshell at the regional planning group that we might go to 152 gigawatts. Are your analyses predicated on that type of growth or somewhere in between the old sort of projections of 90 gigawatts? How do you view some of these projections around rapidly rising load?
Olivier Beaufils
Yeah, that's a great question. So I mean, the very short answer is somewhere in between. We're definitely not projecting 150 gigawatts by 2028. I don't think the system can handle it if anyone was still in doubt. But I think one thing that's really important for people to understand is the numbers that was put out at that meeting, the 150 gigawatt, is not a forecast, right?
It is just a stack up of all of the load interconnection requests that the different TSP's in ERCOT are looking at for those specific years. So it's everyone that's submitted interconnection requests. And as we probably know at this point, not all of these interconnection requests actually translate into load and projects that are moving forward. So a number of those are not going to materialize.
So that means two things, right? One is there is load growth. And obviously I think that would be foolish to ignore that. We're seeing the system grow. We're seeing the need for more capacity to fulfill that growth. But at the same time, you know, there's a lot of hype around that growth sometimes. And I think part of what we do day in day out is trying to, you know, look beyond the hype as to what is actually tangible, what is likely to move forward and what's a reasonable sort of forecast to plan against because when you're making an investment in, you know, a large generation facility, you want to look at the realistic scenario. You don't want to look at the hype scenario.
Doug Lewin
Yeah, that's right. And I think also ERCOT is, as you said, it's not a forecast. They're taking these things that are coming into their interconnection queue and they're assuming in most of the cases, not all, some of the Bitcoin they're discounting, but they're generally assuming a 24-7 load. And just like the price signals are there for battery storage and for flexible gas, there are price signals to load to reduce or curtail during those few hundred hours out of the year as well. So I hear you on that.
Olivier Beaufils
I think that's a topic that I've also heard you talk about quite a bit, but I think something people don't think about is these loads are also going to look at becoming more efficient. When you think about data centers, data centers have every incentive to become more efficient and to be able to reduce their costs in terms of providing power. So there is going to be efficiency gains as well that are going to materialize as we go into the future. And hopefully we can make some of these maybe expand or incentivize some more efficiency. But I think that that's something people need to keep in mind is some of this is not going to be as much as we think it is right now.
Doug Lewin
No, I think that's right. And there's, there's the efficiency on the large side with the data centers and all of that. And there's going to be a lot, there is a lot of efficiency potential on the residential side as well, that I see more and more willingness of policy makers and utilities, stakeholders in general to, to really try to get after. So all of that will be impactful on load growth. I think that's right.
I do just wanna, as we're starting to wind down, I wanna make sure though, before we end, that we do talk a little bit more about September 6th. I will often stand on the proverbial soapbox and say we are not focusing enough. I think it was a pretty significant event, right? With frequency dropping all the way down to 59.77, the biggest drop since Winter Storm Uri.
And as a matter of fact, you know, being in an energy emergency declared by ERCOT and actually leaping right past the EEA-1, which, as bad as Uri was, they didn't do that then. They did go through the one, two, three. Luckily on September 6th, they never reached three, which of course is rolling outages, but they just bypassed one because it was such a precipitous drop. And as you talked about, as we were beginning this podcast, batteries really played a major role. Can you talk a little bit more about September 6th? What happened that day from your viewpoint and talk a little more about the role that batteries played on that day.
Olivier Beaufils
Absolutely. So, I mean, as you described very clearly, right, so in terms of the situation, right around 7:25, so call it 7:30, ERCOT declared an emergency level two because the system reserve dropped and the frequency dropped to levels, obviously, that weren't good for system stability and reliability. And so what we did is we looked at battery dispatched hour-by-hour and 15 minute-by-15 minute period across the day and understanding how much battery discharge was done. And obviously some of that was in ancillary services being called, but generally how much energy batteries actually provided to the system. And so what we saw is the highest hour of battery discharge was exactly that hour from 7 to 8 where batteries provided on average 1.4 gigawatts of energy across the hour, but actually discharged on a 15-minute basis all the way up to 2 gigawatts of energy. And I think that's something that's really important to keep in mind because then we compare that to the amount of physical reserves that were remaining in the ERCOT market for each of those hours of highest system stress, right? And so in the case of September 6, 7 to 8 PM, and when you look at this, you see that there was only 600 megawatts of reserves left before ERCOT had to declare an EEA 3, so an emergency level 3, which is, you know, it's not that they have zero reserves left, but it's the threshold at which they have to start shedding load to maintain system stability.
And so essentially, we had 2 gigawatts or 2,000 megawatts of battery dispatch during that 15 minute period when ERCOT was only left with 600 megawatts of reserves, which means if the battery dispatch had not happened on that day, ERCOT would have had to shed load because it would have been below that threshold and by quite a lot, right? It would have been below by 1.4 gigawatts with which at that point is actually really close to just system collapse.
And so to put it simply for readers, batteries kept the lights on. Without the batteries, ERCOT would have had to start shedding load. This obviously would not have been a Uri. I think that's also important to understand. We were in a summer period where ERCOT would have had to shed, say, a gigawatt and a half rolling outages across ERCOT for probably no more than an hour as the system was recovering and ramping up other capacity. But some people would have been left without power for half an hour, 15 minutes, half an hour, maybe an hour across the state. And so that was essentially the role and the reliability that batteries provided.
Doug Lewin
Yeah, absolutely. And there, you know, that two gigawatts you referred to, I believe that that was the first time, there may have been one other that summer where we were at two gigawatts. There have now been 10 times where storage has been at that level. And now once, in fact, in May, when we were having a spring day, it was particularly hot. There were a lot of thermal outages. Wind was a little bit lower. And storage hit three gigawatts that day. And I think we're going to see a lot more of those kinds of days going forward as well.
I do want to ask you one more thing. Do you have thoughts on longer duration storage? Because the vast majority of storage we have is one to two hours. And we've been talking about, and I don't want to gloss over, like one to two hour storage is extremely valuable. If that's lost on anybody this far into the conversation, then they just haven't been paying attention. So in no way am I saying it's not valuable. It is. And it'd be nice to have some longer duration storage as well, but it seems that for whatever reason, the market isn't sending the signal for that longer duration, whether it be three, four, six, eight hour. Because you can have storage in those kinds of durations. Certainly, eight can get kind of hard, but you could certainly do three or four hour storage, and we're just not seeing that much of it. Do you attribute that to, well, what do you attribute that to? And do you have thoughts on how we could attract three or four hours? And I guess I should just ask the question, do you agree with that? Would it be valuable to have that longer duration? Or maybe my assumption there is just wrong and it's just not really what the system needs and that's why we're not getting it.
Olivier Beaufils
No, I would agree with you. I think there is definitely value in having longer duration storage because you have more ability to maybe withstand slightly longer duration events. We know most of the summer events are one to two hours, but there are moments where you're, say the wind doesn't ramp up fast enough. And maybe that one to two hours extend into three, four or five hours. And you're very happy to have longer duration storage to handle those events when that happens because that's something that happens every summer.
So I think there's definitely value in having those three or four or five hour battery storage. I think the reason we're not seeing them right now is really two things. One is cost. It is quite a lot more expensive to have a four hour system versus a one to two hour system. And so we are going to need to see those costs coming down for those systems to be economic in the ERCOT market. And then the other one is that the bigger price signals right now are more a one to two hour price signal. And that's sort of the additional arbitrage you get from that third and fourth hour is just much lower than the amount of money you get from the first and second hour.
So I think it's the combination of the two. The price signal is not as strong for those third and fourth hour. It doesn't mean they're not valuable, but they're just not as valuable. And the costs on the other hand that you have to put up for those systems is too high. So that's the situation we're in right now. In our modeling, we do see for our system emerging, especially in the 2030s, naturally, even without market design changes. So it is something that becomes economic in the Texas market. It's just not something we're seeing quite at that moment.
Doug Lewin
Okay, good, yeah. I have a lot more questions, but I don't wanna go on too long here. So I think we'll leave it there, but I look forward to diving into some of the economics around storage in future episodes. Hopefully, future conversations with you, Olivier. This has been great. Is there anything that I didn't ask you that you wish that I had? Anything else you'd like to say?
Olivier Beaufils
No, I think, thanks a lot, Doug. I think we've covered it pretty extensively, honestly. And maybe one comment to say is, watch out this summer. This clearly is stacking up to be another hot summer in Texas, and we're continuing to see load growth. So I think a lot of the findings we have in this report we're likely to see being repeated, and probably even in a bigger scale with batteries providing more of that critical energy, more of those ancillary services over the summer. So I think for everyone to watch, right, we'll have, you know, August is coming up in a few weeks and I'm sure we'll have a lot more examples of that coming up.
Doug Lewin
For sure. I think that's absolutely correct. Lots more, so stay tuned everybody. And really thank you, Olivier. I think this report is a big contribution. Of course, we have a link to it in the show notes. I hope folks will check out the research and all the good stuff coming out of Aurora Energy Research. You guys are doing great stuff and appreciate the contribution you're making to the knowledge base of the industry and stakeholders. So thank you. And thanks for taking time to be on the podcast.
Olivier Beaufils
Thanks for having us and thanks for your conversation.
Share this post