Texas has spent decades building transmission to serve load growth. The pattern worked when growth rose steadily with new homes, oil and gas operations, and the gradual expansion of the state’s industrial base. It is being tested by a different kind of customer: data centers requesting interconnection at a scale that exceeds what the grid can physically deliver.
Eric Goff, founder of Goff Policy and a long-time participant in the ERCOT stakeholder process, walks through how the system is adapting. The current large load queue sits above 400 gigawatts, a number Goff describes as effectively infinite because the constraint is infrastructure, not demand.
In an interview with host Joshua Rhodes, Eric covers a lot of ground, including:
How the batch zero policy, now working its way through ERCOT governance, would replace one-off utility studies with a single, system-wide study and a constructable transmission plan.
How the decision to build 765 kV transmission compares to the 138 and 345 kV shifts of past generations.
How Senate Bill 6 gave ERCOT and utilities multiple tools to disconnect large loads before emergencies.
Why artificial intelligence hyperscalers behave differently than the crypto miners that came before them, with a value-of-lost-load above the wholesale price cap.
Whether a minimum transmission charge can protect existing rate payers as new load arrives.
Goff argues the infrastructure decisions Texas makes now will determine whether the data center build-out lowers per-unit costs for everyone or shifts them onto residential consumers.
Energy Capital Podcast is produced by ClarityForge Studios.
Timestamps
00:00 - Introduction & Eric Goff
02:36 - How Policy Gets Made in ERCOT
06:35 - What’s Actually Driving Load Growth
09:18 - Is the 435 GW Queue Real?
10:41 - Inside the Batch Zero Process
14:21 - Building Transmission for Load, Not Generation
16:56 - Large Load Flexibility and Controllable Load
18:18 - SB6 and the Power to Curtail
22:38 - The Dispatchable Campus Idea
27:27 - Does Transmission Planning Need to Change?
32:32 - Paying for Transmission: 4CP, 12CP, and a Minimum Charge
39:39 - Five Years Out: Betting on Infrastructure
Resources
People & Organizations
Company & Industry News
ERCOT files Planning Guide Revision Request 145 for Batch Zero
ERCOT Board approves $9.4B 765-kV Eastern Backbone project (RTO Insider)
PUCT approves first 765-kV transmission lines in ERCOT region
Texas lawmakers push back on 765-kV transmission plan (KXAN)
Books & Articles Discussed
Related Podcasts by Energy Capital
Transcript
Joshua Rhodes (00:04.558)
Hey everyone, and welcome to another episode of the Energy Capital Podcast. I’m really excited to have Eric Goff, the founder of Goff Policy here, to basically tell us how ERCOT works. Eric is one of the smartest people out there when it comes to kind of the current comings and goings of entering around ERCOT. Deeply involved in a lot of the policy and a lot of the procedures and a lot of other things happening. Eric founded Goff Policy in 2019. Before that, he had seven years at Citigroup.
Joshua Rhodes (00:31.726)
energy trading, market operations, earlier roles at NRG Energy and Reliant and Constellation. He’s been a longtime participant in the ERCOT stakeholder governance system. He serves as past chair of multiple ERCOT subcommittees and working groups and up until just recently served as a sole representative for Texas residential consumers at the ERCOT Technical Advisory Committee appointed by the Office of Public Utility Council. The firm has grown significantly in the past couple of years. You’ve been spreading into Western markets. You’ve been hosting
Joshua Rhodes (01:01.58)
symposiums and all kinds of things. Eric, welcome to the Energy Capital Podcast.
Eric Goff (01:07.896)
Thank you so much. It’s great to be here. I appreciate the invitation.
Joshua Rhodes (01:10.946)
Yeah, no, absolutely. So as I do with most people when I’m going through their LinkedIn before talking, I actually came across something I did not know about you is that did you really co-found compost peddlers?
Eric Goff (01:23.18)
I did with my friend Dustin Fedakko.
Joshua Rhodes (01:26.23)
Okay, so I don’t know if we’ve talked about this. Maybe we have. I was in East Austin at around that time and I saw your guys, maybe you, I don’t know, like riding these modified cargo bikes with basically blue barrels carrying compost around. Was that you?
Eric Goff (01:44.142)
That’s right. That was us. We had, we called everyone the peddler. And I did some of the compost shifts, but we had part-time and full-time peddlers and launched it before we had municipal composting in Austin. And I think that maybe the best thing to say is that we accelerated the city’s own plans to municipal composting.
Joshua Rhodes (01:47.65)
Right, right, right.
Joshua Rhodes (02:08.09)
It was just funny. just, remember one time thinking, and this was also with like PediCab folks around that time before like e-bikes really kind of made it into where they’re the most fit people in the world riding tens of miles a day carrying heavy loads. was insane.
Eric Goff (02:23.308)
Yeah. Some of the hard part too is because it’s just pure literally inertia with another maybe grid thing, but you have to stop the bike and start the bike and stop the bike and start the bike with many, many pounds of compost in that barrel.
Joshua Rhodes (02:36.814)
It was insane. All right. So I’ve already burned enough of our precious time talking about, but inertia is a good tie in. We’re going to get to the grid. So Eric, you do a lot of policy, a lot of policy consulting, a lot of policy work in ERCOT. For folks that either don’t live inside the ERCOT stakeholder world, which is the vast majority of people on this planet, how does policy actually get made in ERCOT? Where does it start and where does it continue?
Eric Goff (03:00.184)
Sure. It’s a unique system that doesn’t exist in many other locations. And it’s changed some. I’m sure we’ll get into that since the winter storm Uri happened. historically, the energy companies in Texas, going back to the 90s at least, I wasn’t in the business then, so I’ve been told, were active in working together to establish what ERCOT would be and do.
Eric Goff (03:30.33)
And they established this stakeholder process. At the time, it had a stakeholder board and it operated by having like a balance and continues to have a balanced group of buyers and sellers that many recommendations at the time decisions. And if anyone doesn’t like something, you can appeal that to the public utility commission. But the process drives consensus because you have to get along long-term with your peers.
Eric Goff (03:58.798)
Right. And so many things don’t end up being appealed to the utility commission. They just kind of get worked out. Everyone leaves a little bit dissatisfied, but it’s an effective process. Since Winter Storm URI, for good reason, people were upset about how the process was working and, you know, there’s a lot of finger pointing. And so there was significant change to the process. So now everything goes to the utility commission ultimately. And the Stankold process is more of like an advisory role to the commission.
Joshua Rhodes (04:28.064)
Okay. And so are these the groups, the technical advisory committees, correct me if I’m saying these wrongs, and these working groups like the large load working groups, is this kind of where all of this is the negotiations happen and things are made?
Eric Goff (04:40.558)
So ERCOT has offices in Austin and those meetings are technically open to the public, but they’re also broadcast online. State law has required that TAC, which is the senior stakeholder group, as well as the ERCOT board, have been accessible on the internet so anyone can follow along. during COVID, some people just kind of dialed in for fun. But the process has dozens of people in some cases,
Eric Goff (05:10.434)
for many of these large load questions, hundreds and hundreds and hundreds of people that dial in to listen. But it’s many of the same people that speak that are kind of established figures in the stakeholder process have been around for a while.
Joshua Rhodes (05:21.998)
In general, like, is the policy that comes out of these things, I mean, there’s negotiations, there’s consensus or some form of consensus or some things like that. how quickly are these groups able to move? I know there’s a lot of pressure right now on figuring a lot of things out that I want to get to, but how quickly are these groups able to move to make new policy?
Eric Goff (05:40.18)
It depends on the issue. historically, when I say historically, I’m going to typically mean like before winter storm, URI and after winter storm. But before URI, a lot of it was stakeholder driven and led. then when the new commission and new board and new CEO came in to kind of right the ship with ERCOT and the grid, a lot more became.
Joshua Rhodes (05:48.129)
Three years ago. Okay. Yeah
Eric Goff (06:09.438)
led by ERCOT staff. today it’s largely similar to how other independent system operators work, where the staff will recommend something, bring it to the stakeholders and ERCOT board and to the commission. And there’s back and forth dialogue among all three groups. I don’t think it’s more efficient than it used to be, but it is more public process.
Joshua Rhodes (06:35.054)
Okay, well there’s some trade-offs there, I guess. I want to jump into probably one of the biggest, the hottest button issues that’s being worked out right now. And this is, you joked about hundreds of people dialing in, probably straining WebEx to the point where it’s not able to function anymore. I think I’ve been one of those flies on the wall a couple times. I want to get into, okay, everyone’s talking about load growth right now. What is actually happening?
Eric Goff (06:37.752)
Right.
Joshua Rhodes (07:00.014)
Let’s level set on kind of what is actually happening before we get into kind of like what the process we’re trying to build around this.
Eric Goff (07:06.702)
Great. So ERCOT was the only North American market that had growing demand for electricity year over year for decades. And many others had stable load growth or very minimal load growth. And I think that’s driven in part by kind of state policy as well as just kind of the wide open spaces, lots of natural resources that are low ground and above ground, and there’s opportunity to grow.
Eric Goff (07:36.31)
and also a growing population because of that. And so a lot of the growth was in new home construction, right? And that kind of stuff too. So ERCOT’s been used to needing new generation, not just to replace existing generation, but to serve demand. And a lot of that was happening and continues to happen from wind and solar and recently batteries and less so with more traditional natural gas and coal.
Eric Goff (08:05.336)
although that’s been happening too. And in the past couple of years, there’s an opportunity that was created, I think, because of the Texas market design, which is a lot of cheap electricity, or as is the trendy way to say it, either it’s an abundant amount of electricity or a dominant amount of electricity. Yeah, that’s right. And so that cheap electricity was an opportunity initially for cryptocurrency miners.
Joshua Rhodes (08:25.762)
Come one, come all.
Eric Goff (08:34.068)
where their main input is energy, right? Then with scale, other data centers, and we’re seeing cloud and increasingly AI data centers that the capabilities to do machine learning have been around for a long time. But to do it at the scale and with the incredible outputs that we are getting today, it’s using essentially the same process we’ve had since at least the 80s or not before.
Eric Goff (09:02.744)
but you can do a lot more with when you have energy to apply to it.
Joshua Rhodes (09:06.562)
Yeah, no, as a lot of folks, they’re surprised to find out that like the concept, the artificial neural network that underlies the large language model, we’ve known about these since the 1970s. It’s just like we finally figured out how to shove the entire internet into it. But when you shove that much information into it, takes a lot of energy to hold all of that information, particularly in like the training side. And speaking of like a lot of energy, was trying to find the latest numbers from ERCOT. Number just go up as far as I can tell. And so I haven’t believed them for years, but I think the latest one
Joshua Rhodes (09:34.946)
that I saw out there was something like 435 gigawatts of large load in our so-called large load queue. What is your sniff test on this? mean, like how much of this is real? Like so.
Eric Goff (09:47.372)
I have a crazy way to say it that I think is clippable, but then I’ll explain it. I think right now, the amount of demand for new data centers in Texas is infinite. Okay. And that’s because there’s no way to serve the amount of load on the grid that is desired by developers. Yeah. And so it’s just above the number that the grid can serve. So it might as well be an enormous number, right?
Eric Goff (10:15.798)
And the constraint is the infrastructure side, not the demand side. So until we figure out how to serve that demand, it’ll continue to be more than we can possibly serve. And the difficult part of that is which one of those projects are going to get built and how do know when and when, how do you plan for that infrastructure, and how do you make sure that those data center developers are paying for that infrastructure and not anybody else?
Joshua Rhodes (10:41.772)
Yeah, and so that brings us into one of the processes that’s trying to be worked out right now is this whole batch zero process around, okay, which tranche of these new large loads are going to make it through. We’re setting aside transmission for these, we’re setting aside transmission capacity for these for like the first time that we’ve ever done anything like this before, as far as at least from what my understanding of what Pablo Vegas said to me the other day on the podcast.
Joshua Rhodes (11:06.744)
Can you talk me through where the batch zero process is right now? Like how much is in there? What’s it looking like? What are the milestones we’re at and where are we headed?
Eric Goff (11:14.336)
Yeah. So we are a couple of weeks away from establishing a rule for Batch Zero. And Batch Zero is the one process to connect a large load in Texas. And that’s 75 megawatts or more, which is these days a small amount of load. Right. And so the reason we’re moving to a Batch Zero process is because
Joshua Rhodes (11:31.086)
Seems quaint.
Eric Goff (11:38.018)
There’s loads of big enough that they were affecting other load studies. And so they’d have to be restudied and restudied and restudied, which wasn’t leading to many approvals, which doesn’t help with that infinite demand problem. And so many of the developers and ERCOT collectively and the utilities all kind of came to this idea that we need to have.
Eric Goff (12:00.394)
a single system wide study. And the hard part is how you transition from individualized studies to a singular study. there are other challenges we can get into, but that’s the big one right now is you can’t study all 435 gigawatts or whatever it is, right? And so how do you pick which ones to study and which ones to ignore?
Joshua Rhodes (12:22.072)
So I mean, one of the things we’ve had a generator interconnection queue for a while in a very like consistent process. Everyone knew what the steps were. We’ve recently been talking about this large load interconnection queue, like we actually had a queue in a process, but this like, actually didn’t, we just had a word we could put on there and we called it a queue and people thought it had a process. But you mentioned that before, it’s like, you can’t go from all these individual studies. So mean, like, how was it done before this batch zero process?
Eric Goff (12:46.958)
Eric Goff (12:47.418)
So the funny thing about governance is it just means making up rules in a way that we agree to make up rules, right? And so before the cryptocurrency growth that I mentioned earlier, it was just done by utilities. You requested an interconnection and they said, okay, it’s available in this timeframe. And that was generally within like the capital planning horizon and it was fine.
Eric Goff (13:15.086)
the crypto facilities could build really quickly. And so we’re requesting really tight timelines that was shorter than the transmission planning timeline that ERCOT was used to. so it couldn’t keep up. And so for that reason, ERCOT tried to create this substitute process to plan for a transmission on a one-off basis. And that didn’t really work for a variety of reasons that are now
Eric Goff (13:43.712)
In the past, it was not worth getting into, but quickly, for example, they didn’t actually create a constructable transmission plan. They just said, it’s feasible for a transmission plan to exist. And so one of the things that I’m excited about from this batch process is we are going to have a constructable transmission plan. And ultimately, I think all the problems are infrastructure problems in this energy market. There are secondary questions around like,
Eric Goff (14:12.396)
market design and incentives and those are really important and I care and many other people care a lot about those but today it’s just about constructing infrastructure.
Joshua Rhodes (14:21.324)
Yeah, I know that there’s like parallels in there with like the connect and manage that we’ve done for generators in there a while, right? Like you said, there was no transmission plan. Here’s an affirmation that it can happen or something. I’m just putting words in your mouth. Yeah. But it’s an interesting parallel there because, it seems like, correct me if I’m wrong, but like in general, we’ve often built transmission for load in ERCOT. Whereas like other grids, like in PJM, if you have a capacity construct, you’re often building transmission for generation.
Joshua Rhodes (14:50.388)
and you put the cost of those deep network upgrades like on the generator cost. Now, they’re having all kinds of issues and I’m going to talk about PJM with someone else later, or I hope I am. But like, this kind of feels like a little bit moving towards that kind of construct if we’re actually going to put together like transmission plan for these loads. But we’ve kind of done that in the past. Am I understanding that right?
Eric Goff (15:12.894)
You are historically with one exception. We built transmission to serve load growth. And there’s this belief out there because of the one exception that a lot of the transmission is to build for like renewable generators. And we did that once and it was successful in getting a bunch of wind that was, you low cost energy. But the issue with that is how we pay for transmission.
Eric Goff (15:41.484)
When we added new transmission costs from that and consumers pay for transmission costs, we didn’t add new consumers. We just added new wind generators. energy costs went down, but transmission costs went up and people noticed another bill. What could be different about it this time is we’re adding new transmission costs and we’re also adding new giant consumers that will have giant energy bills to pay for that transmission.
Eric Goff (16:11.062)
I don’t know which way it’ll go yet, but it’s entirely possible for the development of these new large loads to lead to lower energy costs and lower per unit transmission costs because you have more payers now than the increase in transmission costs. So I think that should be a goal of the state is to let’s make sure that as this change is happening.
Eric Goff (16:38.574)
We do it in a way that benefits everybody and not just the new giant loans that are coming. We hear all these complaints and well-placed fears about what does this mean for me. There’s an opportunity through policy to fix that. So we should do it.
Joshua Rhodes (16:56.482)
I mean, I think it can happen. I don’t think it’s a foregone conclusion that it’ll happen that way, but I think it’s possible. I’ve got some ideas on that, but this is podcast where I’m interviewing you, not you interview me. We get to talk about that. I wanted to pull a thread a little bit on, okay, so you brought up crypto miners a couple of times. And one of the things about, particularly like when crypto was getting big, they talked a lot about their flexibility.
Joshua Rhodes (17:18.445)
about being controllable load resources, about being able to consume energy, know, pay for infrastructure, but then like we’ll go away whenever things get tight. So one of the big debates right now around large loads is, know, how can they participate in the market? Do they need to be flexible? Do they need to be controllable load resources? There’s these things around, if you have onsite generation, can you net that generation against like your consumption to appear to be flexible on the system? Are you bringing your own generation?
Joshua Rhodes (17:47.118)
My understanding of the way that it works right now is that if you are a load and you want to be a controllable load resource, ERCOT actually wants to turn the load off. They won’t allow you to net against local generation. And for crypto, that was fine because they were like, okay, just turn off the ASICs and whatever. They’re like, we’ll just hash them back up, not too long. But with a data center, they’re not so fine with that.
Joshua Rhodes (18:12.226)
Can you summarize the current thinking around like large load flexibility in Texas? Like where are we at right now? Where might we go?
Eric Goff (18:18.71)
I can definitely do that. And I want to start with the bill that the Texas legislature passed last session with SB6. The legislators were very concerned about the growth of these new data centers. The time is very new. We weren’t nearly at the hundreds of gigawatts that we are now, but was during the session, there was a giant new load forecast that was published and it made people very nervous. So peppered throughout that bill, the Senate
Eric Goff (18:48.17)
and the house made sure that there are multiple ways that ERCOT and the utilities could turn off large loads before an emergency happens. Either by controlling the generation or requiring the installation of a disconnect device or technology, having a couple of different programs to do that. Whether you net against an existing generator to establish conditions so that ERCOT can take back control of that generator before an emergency. So,
Eric Goff (19:17.166)
Before we get into the flexibility, the state has established rules to turn off all this new load if it gets to a crisis. And if, as I believe will happen, this also results in new generation, if you turn off all that new load but have all the new generation, all that new generation can serve the existing load. And that’ll keep the traditional native customers
Eric Goff (19:42.882)
have a more reliable position than they were because this excess generation that to serve these giant data centers that are offline during the emergency. But there’s also the question separate from like that operations thing for emergencies of like, how can we create incentives for them to turn off before we get to an emergency? So we don’t have to take those like strict, you know, mandated government functions. So the crypto guys, what’s different between them and the AI guys is they’re
Eric Goff (20:12.756)
value of lost load. So like how much money do you make by consuming a megawatt hour of electricity? And for the crypto guys, was hundreds of dollars, sometimes less. And for AI, it appears as though it’s above the cap for electricity prices. And so they’re not going to turn off unless they have to. And that have to could be in two ways. One is we get to the emergency.
Eric Goff (20:41.526)
Or two is there is some electricity available for you, but not 8,760 hours of electricity for you. And so you might be willing to say, sure, cut me off sometimes if it means I can use power three quarters of the year or four fifths of the year. It doesn’t appear as though that’s a super attractive option for these hyperscalers because while they definitely want speed to power, they also want to install these
Eric Goff (21:09.836)
very, very expensive computer chips in a place where they can be used and useful. so if they can be used for four fifths of the year or be in a warehouse gathering dust, they might pick the four fifths option, right? But I think they’re hoping to find a place where they could be used like 99 % of the time. So we’ll see. The other issue is that ERCOT, like you said, wants to control the load directly. And that’s just how their systems are set up.
Eric Goff (21:39.884)
where they have separate resource entities for each thing and they want to measure what the thing is doing. I do think that there is an opportunity to use energy storage systems to provide that flexibility for these sites. I don’t know that every staffer or cat agrees with me. This is an area of active debate. what is a uninterruptible power supply versus a battery? If it’s only
Eric Goff (22:09.164)
behind the meter functions as a part of the load, never puts power under the grid. It’s just consuming electricity and it’s a consumer. So I think that is and will be allowed and is kind of the next iteration of how do we have load flexibility for these data centers. Because the energy storage systems can be the flexible part of the load so that the data center facility doesn’t have to be.
Joshua Rhodes (22:38.146)
Yeah, I mean, it seems like that would be a rather obvious step here. Correct me if I’m wrong again, but haven’t we had this issue with renewables and storage? People wanting to build batteries behind like one meter for like a solar farm and a battery. They’re required to be on two meters. Is this like the legacy of the policy we got right now?
Eric Goff (22:59.412)
It’s a legacy of the policy and the systems. So the ERCOT IT systems expect these to be different things. They want to be able to model one thing. Even within a wind farm, 90 % of the turbines have to be the same vintage and variety. If you replace it with a modern turbine, you have to create a separate resource. It’s antiquated, but this is a problem we’re going to have to solve to think about like what is a dispatchable campus.
Joshua Rhodes (23:02.84)
Okay.
Eric Goff (23:28.622)
where you’re just putting out power at the point of interconnection. We’re more than capable now of having more than one kind of thing in a dynamic model. And so there’s no reason not to do it. And as these facilities are using as much power as cities, we’re going to have to figure out a way to have a dispatchable campus approach. And that’s going to take some thinking and policymaking, but I think we need to do it. And it could lead to innovations around microgrids.
Eric Goff (23:58.103)
Micro is a funny word in this context, but it’s how we think of it.
Joshua Rhodes (24:01.696)
micro relative to what it just defined dispatchable campus for me just
Eric Goff (24:05.386)
Yeah. So a dispatchable campus might be a data center or advanced manufacturing facility or a neighborhood maybe that has batteries, DERs, solar, maybe a natural gas or SMR, who knows what the future holds that are all dispatched at like a common point of interconnection to the grid. And the campus operator
Eric Goff (24:33.814)
would be responsible for making sure that what they tell ERCOT they can do is what they actually do. Just like what we have with a generator today. It’d just be a more complicated arrangement. And ERCOT will probably still want to know, like, okay, if I’m relying on you to do something and you’re relying on a bunch of batteries, I want to know the state of charge of all those batteries behind or inside that campus and stuff like that. The technical details to work out. Another issue for the policy nerds,
Eric Goff (25:03.096)
that I know listen to the podcast is like, what do they get a nodal price or do they get a zonal price? here we go. Yeah. So stuff like that.
Joshua Rhodes (25:11.446)
Yeah.
Joshua Rhodes (25:13.028)
I thought when I was talking to, had a recent interview with the folks at Austin Energy talking about the load zone and we went through that, generators get paid where they inject power, but load pays the load weighted average of the load zone that they’re in. And if you’re buying and selling both at the same time, or it does open up a real debate on where should you be netted at.
Joshua Rhodes (25:31.446)
I know you’re in a couple other regions as well. Like how is Texas doing relative to other regions in this? I think there’s a lot of data centers trying to come here. Our numbers are huge. So we’re kind of forced to try to build this plane as it’s going. Like how are we doing relative to other places? Are other places doing anything? And we’re just the ones that are figuring it out. How’s that look?
Eric Goff (25:49.23)
Other places are doing kind of one-off things. we’ve seen some announcements of like fully integrated utilities like the Louisiana Energy Meta Deal, right? But that’s done through a bespoke arrangement, right? What happens here, because we have competitive wholesale and retail markets, lots of land, no zoning rules in the counties, and abundant
Eric Goff (26:16.736)
solar and wind and abundant natural gas, you can just kind of move quickly. You don’t need to ask nearly as many permissions as you do in other places. And then I can’t wait for you to get to your PJM conversation soon because here the people that are doing the things are the ones that pay for the things for the most part. And in PJM, when the capacity market is being set by all these new data centers, everybody pays for the new data centers. It’s a crazy, crazy system.
Eric Goff (26:44.726)
It’s not functional and they know that and so their market system in quotes is breaking.
Joshua Rhodes (26:50.144)
Yeah, I mean, we won’t get into it on this, BGM just put out a paper where they have broached the idea of getting rid of the capacity market and going to like an energy only market. may do something like, remember the LSCRO thing that we considered briefly like years ago? They
Eric Goff (27:03.746)
Yeah, they may
Joshua Rhodes (27:04.758)
do something like that to soften the blow. I don’t know.
Eric Goff (27:07.808)
Yeah. I can’t wait to watch from afar, but ultimately the reason that we’re here is just because it’s worked, you know, and it’s not perfect. We haven’t figured out the ideal energy market and you can’t go down to the corner store and buy a bag full of mega one hours. Right. So we have to have a construct. Totally. Yeah.
Joshua Rhodes (27:10.303)
Exactly.
Joshua Rhodes (27:27.564)
Yeah, totally. And we’ve touched a little bit on this. One of the big things, you said infrastructure multiple times and a lot of that is transmission. That’s one of the hardest pieces of the system to build. mean, traditionally we’ve planned transmission around reliably serving load. We tell ourselves we have an economic test for transmission, but then every once in a while we build these big tranches. We referenced the CREZ lines earlier. We’re doing it with this Permian reliability project, STEP project, you know, East and West. Even the current set of like
Joshua Rhodes (27:54.538)
know, 765 backbone was really pre AI and data centers. I mean, a lot of it was focused on serving oil and gas load out in the Permian. Does transmission planning need to change with this current set of drivers being new massive loads outstripping historical growth rate?
Eric Goff (28:09.294)
Yes, and it is changing, but it needs to change even more. So ERCOT has said multiple times that it’s time for a new era of transmission planning. That’s the phrase they use. And this 765 decision is a generational shift in how we build and plan for transmission. I’ve heard long time veterans of the industry say that their earlier generation was
Joshua Rhodes (28:13.42)
Okay, tell me about it.
Eric Goff (28:37.088)
So happy, so like two generations of engineers ago were so happy when they went from 138 to 345 because it made such a big difference. This is very much like that where it’ll enable moving a lot more energy across the street a lot more efficiently with fewer big power lines because you can move so much more energy on a high voltage line. So there’s a lot of growing opposition to these lines that you mentioned from people that don’t want to see them on Pretty Vistas.
Eric Goff (29:06.786)
The alternative is more lines, not fewer lines. So that’s one thing where ERCOT is looking to do this unified approach to transmission planning that integrates all of the variety of transmission planning they do today into one process that’ll be modeled on this batch process. But when we do that, we need to modify the way we look at how consumers benefit from these transmission lines.
Eric Goff (29:36.91)
One of the things that’s held up investment in transmission in the past is skepticism over changes to how we say that a project is economic. Because you’ve talked about connect and manage where a generator can build wherever they want and suffer or benefit from the dispatch they get. And so if we have a system where a solar generator can build in the desert in the middle of nowhere on a 69 KD line,
Eric Goff (30:05.938)
and then suffer curtailment, then consumers pay for the transmission line to help them reduce the curtailment. It’s just not something that was persuasive to many of the large companies here in Texas that pay a big share of the transmission lines. And maybe it could be economic or maybe it couldn’t, but when the people that are pushing for it or the generators that would directly benefit, there’s just intense skepticism. Is this for you or is this for me?
Eric Goff (30:34.678)
I think the way we solve that is just by saying the reason we plan transmission economically is to benefit consumers. if you can prove the consumer benefit, great. And if it so happens then that the way we save consumers money is by paying for a transmission line that increases the amount of solar output from that far away generator, great. We prove that it’s benefiting consumers.
Eric Goff (31:04.492)
and then incidentally benefiting the solar generator. So we have to do that.
Joshua Rhodes (31:09.644)
Yeah, I’ve often cast CREZ and the New Step projects. It’s not even just about the electricity sector. It’s really a total economic enabler for the state, right? It’s like the state runs on energy, oil, gas, electricity. And so it’s like the better the sandbox is, the more efficient that system is. Like it’s not just getting the cost of electricity down. It’s the cost of everything else that we use the electricity to build the thing that we sell around the world. I don’t know. I’ve been more framing it that way. I don’t know. What do you think about that?
Eric Goff (31:39.886)
I think that’s really true. Another related issue is like we’re having all this interest in co-locating generation and giant data centers in ways that will like impact communities because you need the data center near where the jobs are, which means you need the generator near where the people are. And transmission doesn’t exist at the scale to move a gigawatt of power a few hundred miles away to the data center near where the jobs are.
Eric Goff (32:08.458)
So we can’t have a financial arrangement to have like a power purchase agreement to buy a thousand megawatts of power from some random place. And if I can’t physically consume the energy at my location, and that’s why I keep coming back to infrastructure. Once we have investment in transmission, we’ll have investment in generation. It’s a necessary predicate with this scale of large loads.
Joshua Rhodes (32:32.588)
Yeah. And tying together a couple of those before we move on is like, we’ve done connect and manage for generation. It’s possible maybe to use some of that inertia to connect and manage for load as well. Like if we can get this like behind the meter netting issue thing. Cause I agree with you. I don’t think the data centers just don’t want to turn off, right? It’s just the perceived value of the tokens and the product and everything they’re making is just so high. They’re just like, no. So talk about infrastructure. We’ve talked about transmission. Do we need to plan it differently?
Joshua Rhodes (33:01.144)
But also, mean, we’re also talking about it, we’re going to pay for it differently, right? I mean, this is like just the Public Utility Commission staff just released their recommendation to go from a forced CP where we calculate how much of the transmission year owed from the four highest 15 minute intervals of ERCOT serving load in June, July, August, and September. And they’re recommending a 12 coincident peak. So all the months in the longer time, a 30 minute period, why do we need to change or do you believe we need to change how we pay for transmission? Why are we going down this path?
Eric Goff (33:31.318)
Yeah, so this can get really complicated really quickly, but essentially the way that it works today with 4CP is you look at how much people used and the peak moment of June, July, August, September compare that to how much everybody else used in that same peak moment. And then it creates a ratio of you, which everybody else. And we have assumed demand charges for small commercial and residential. We don’t measure the actual demand.
Eric Goff (34:00.152)
but it’s effectively that. And that ratio is how much you have to pay for all of the transmission costs for that year. So there’s been this incentive for large users to reduce how much they’re using during those four times, and they all have to do it momentarily, and it doesn’t substantially change the amount of spend that we have on transmission, certainly not in that year. Maybe hypothetically,
Eric Goff (34:27.872)
over time, if that happens every year, reduces the amount we need five years from now, maybe. So it’s cost shifting. And, you know, the utility commission staff and their analysis and then the analysis of the Office of Public Utility Council, who represents residential and small commercial consumers in this state, residential consumers are paying today more than they should be. So the fear that the legislature had...
Eric Goff (34:55.554)
because of how the crypto facilities were operating when they were curtailed during those time periods, is that we would spend billions of dollars on more transmission and the data centers would curtail and avoid paying for it. So the last section of SB6, the bill that is regulating the data center construction, is to evaluate for CP and consider changes to it. And so the PUC, think, is very interested in
Eric Goff (35:23.884)
the legislature is very interested in making sure that they’re protecting consumers against the existing issues that are already there, as well as this potential for it to be even worse. And so in response, many of the large data center developers or trade associations, my clients, I’ve been proposing on behalf of some of these people, a minimum charge for transmission.
Eric Goff (35:52.204)
which would be instead of based on the amount that you actually used, it’d be based on some percentage of the amount you contracted to use from the utility. If you have a thousand megawatt data center and we have like a 50 % charge, then it can be no less than 500 megawatts of charge no matter what you use in that equation. We’re working on analyses to figure out
Eric Goff (36:17.846)
what the right level of minimum charge is in order to protect all the other repairs in the state from increased costs. And, you know, hopefully we’ll have some good evaluation from that in a few weeks. That’s kind of a really pending question. And so with that context, 12 CP or 4 CP doesn’t really matter. In each one of those measurements, you’ll use the minimum transmission amount instead of the actual transmission amount. So there’s no incentive to try to avoid the cost, right?
Eric Goff (36:46.702)
They could still choose to curtail and connect and manage based on the energy price, right? We walk them into respond to energy prices, avoiding transmission charges that other people have to pay for. Yeah, I don’t know if that’s an incentive that we, right?
Joshua Rhodes (37:05.134)
get it. But I mean, recently, the peak demand and the energy price are divorced, right? It’s like, we used to, oh, 4pm on the hottest day in August is coming, like, ERCOT’s going to peak out, prices are going to be in the thousands because we have scarcity pricing and now we’ve got 35 gigs of solar and it’s just $5, $10. Like, even ERCOT’s peaking and like prices are nowhere near the cap. And so people have talked about moving it to net, but we’re trying to pay for transmission. This is when we’re actually moving the power across the system, right? It’s like,
Eric Goff (37:33.006)
Exactly. What a great problem to have is when you’re peeking out the system, energy cost is five bucks. That’s a great problem to have.
Joshua Rhodes (37:39.692)
No, it’s better than the alternative, right? But it does seem like it kind of misaligned some of those incentives. I guess we’ll see how it shakes out. I wanted to drop down into the distribution area, but I think we’re kind of running out of time. So I know you’re a huge fan of like the energy only market. Do you think that it can survive all of this? All of like low cost, like the zero marginal cost generation, hyperscale load, the increasingly sophisticated financials that are happening around it? Can the energy market still work here?
Eric Goff (38:08.79)
I think it continued to work and perform admirably because when you have zero marginal cost units that are sharing the price most of the time, that’s a great problem for consumers to have. And that’s what led to this opportunity for growth is there’s a lot of cheap energy. And when there’s not enough energy, the price gets very high. And so that’s a strong incentive to keep the existing stuff working.
Eric Goff (38:38.176)
or to not use energy during that time. And if it happens often enough, then it creates an incentive to build new generation or to better maintain existing generation. Right now, the cost of new entry for a new generation is being set by the hyperscalers, where they’re paying whatever it costs to get new generation to co-locate because they don’t have the transmission they need yet. And so we don’t have a market that’s meeting kind of the old cost of new entry, let alone the new cost of new entry.
Eric Goff (39:07.788)
But functionally, that’s okay because we’re getting the investment in generation, right? And so I’m really curious what happens next. Will we have merchant generation that can pay the price of the cost of new entry? Not until we have transmission where you can transact across it and like deliver via financial PPA, just to connect something to some local listeners. Maybe now is not the best time to be a utility buying a turbine.
Eric Goff (39:36.578)
But that’s a separate question.
Joshua Rhodes (39:39.662)
Now we’re really tying together a couple of different podcasts. All right, so last question. Looking at these load forecasts, these load forecasts are really interesting because like, if you look historically, we grew at like two and a half percent or something like that. It’s like the next five years that everything just goes crazy. And then I forecasts are hard, especially about the future, but then it basically kind of returns to like a pre now crazy, like normal. Now I don’t know that I believe any of it.
Eric Goff (39:42.466)
Ha ha ha ha ha.
Joshua Rhodes (40:07.094)
So five years from now, if you’re looking back, you’re traveling in the future, you’re looking back, what will people say about how Texas beat the odds and did it and kept the Texas miracle going?
Eric Goff (40:17.356)
It’ll be because we paid for infrastructure. And ideally, when I say we paid for infrastructure, the state allowed it to happen and made sure that the significant substantial payers of that are the new loads that are coming in. And we can totally achieve that. And we just have to try and it’s going to be a significant infrastructure challenge. To the extent we don’t build enough infrastructure, though, we’ll just get
Eric Goff (40:47.989)
fewer data centers, we’re still going to get as many data centers that the infrastructure can support. And so the real question, and I don’t think anyone knows the answer to this question yet, is like, what is that actual demand for new data center facilities? And AI, whether you like it or not, is being integrated into every part of our lives, and that is going to take energy. And so I think from a public policy perspective, it’s like, OK.
Eric Goff (41:14.508)
With this enormous investment in the grid, how can we make sure that sure we get what the AI companies want, but we also are benefiting every single use of electric grid at the same time. And that’s how we can get that miracle.
Joshua Rhodes (41:28.362)
I I think if anywhere is going to be able to do it, I think it’ll be Texas. I we built the CREZ lines to get the wind, but then that allowed us to electrify oil and gas. And so now we’re building these Permian liability projects, electrify oil and gas. That’s just going to get us more solar territory. But it’s just building that whole thing. We seem to every once in a while get it. OK, a big tranche of infrastructure now that’s bigger than any one particular project would be a good idea to let everyone to let the markets go and let the miracle continue.
Joshua Rhodes (41:55.662)
I think I agree with you on that. think if we’re able to look back in five years and say, did we beat it? It’s because we built the infrastructure that let the market do it. Eric Goff, thank you for coming on the Energy Capital Podcast.
Eric Goff (42:05.292)
My pleasure. This was a great conversation.
Joshua Rhodes (42:09.23)
Thanks for listening to the Energy Capital Podcast. If today’s conversation helped you make better sense of how the energy system actually works, share the episode with a colleague and hit follow on your podcast app. You can find us on Apple Podcasts, Spotify, and all the usual platforms. For deeper analysis and context each week, subscribe to the Texas Energy and Power at texasenergyempower.com. That’s where you’ll find every episode, every article, and our latest updates. We’re also on LinkedIn, X, and YouTube.
Joshua Rhodes (42:38.594)
where we share clips, insights, and ongoing commentary on energy policy, markets, and the grid. Before we go, a quick note. The views expressed on this podcast are my own and do not represent the official positions of the University of Texas, Ideasmiss, Austin Energy, or Columbia University. A big thanks to Nate Peevee, our producer. I’m Joshua Rhodes. Thanks for listening, and we’ll see you next time.
