Comparing ERCOT and TSP load forecasts

This is Part 1 of my conversation with Nat. Part 2 will be out next Wednesday. The Video is Live on YouTube

Texas has plenty of energy stories. The harder part is finding the signal through the noise of endless filings. That is why Nat Bullard’s new venture, Halcyon, matters. It turns piles of ERCOT and PUC dockets into something manageable. It’s AI that makes dense regulatory filings sortable and understandable.

Meanwhile, Nat’s annual presentation has become must-read material for anyone trying to understand what’s happening in the world of energy: the market forces, capital shifts, and technologies driving global change. He looks at macro trends like EV adoption, clean-tech capital flows, global cost curves.

We talked about one of the main takeaways from this year’s edition: China’s dominance in EV manufacturing and exports

The Texas Energy and Power Newsletter is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

We also talked a lot about a recent study Halcyon did with GridLab to quantify the rapidly escalating costs of gas turbines. I asked what his biggest takeaway from compiling that research:

“The biggest one for me being at a $2,500 a kilowatt cost for a combined cycle gas plant. What other competing options are there for providing power?”

At that cost, there are many better options, including re-using older turbines as many of the data center developers are doing. Storage that can be sited quickly. Demand flexibility can meet much of the need. Solar is still going onto the grid at about a gigawatt per month. Targeted wires upgrades and fast local capacity can bridge the gap while long-lead assets work their way through the process.

I left this conversation thinking about how information itself is infrastructure. If we can see what’s changing — if the opaque can become more transparent — we can make better, more informed decisions.

Please share this episode with someone who will get something out of it and subscribe to keep the conversation going.

Timestamps:

* 00:00 – Introduction

* 02:00 – Guest welcome, Nat Bullard

* 03:00 – How Halcyon can find needles in large regulatory haystacks

* 06:00 – Relevance to the Oncor rate case

* 09:00 – Tracking changes in gas plant costs and across filings

* 10:00 – ERCOT vs. utility load growth projections

* 14:00 – Queue reality vs capacity

* 16:00 – Regulators’ role to align incentives and process

* 18:30 – Nat’s annual presentation

* 20:00 – How Nat builds his presentation

* 23:30 – China is the dominant EV manufacturer

* 26:00 – Will America ever import Chinese EVs?

* 27:00 – American manufacturing partnerships

* 31:00 – Can Tesla keep up?

* 33:00 – Permeability between EVs and power sector

* 34:30 – Gas turbine supply limitations and costs

* 38:51 – Closing note, turbine ask

Resources:

Guest & Company

* Nat Bullard - LinkedIn

* Halcyon - Company Website + LinkedIn

* Nat Bullard’s Famous 200-Slide Presentation

* Nat’s NY Climate Week Presentation (From Disparity to Data)

* GridLab White Paper with Halcyon on Natural Gas Turbines

Industry News & Podcasts

* Oncor Rate Case

* Odd Lots Stargate Podcast

Studies & Policy Documents

* Texas Senate Bill 6

* PUC Load Forecasting Docket 58480

* PUC Financial Assurance Docket 58481

Doug’s Platforms

* LinkedIn

* YouTube

* X (Twitter)

Transcript:

Doug Lewin (00:05.42)

Welcome to the Energy Capital podcast. I’m your host, Doug Lewin. My guest this week is Nat Bullard. Nat is and has been for a long time one of the smartest guys on energy issues. I know I have been reading his stuff long before I knew him. He does a now famous slideshow, highly anticipated by folks in the energy space that comes out every January. The last one, as many of the ones before, 200 slides long. We talked a little bit about his process for pulling all that together. He was at New Energy Finance before it was acquired by Bloomberg, wrote a newsletter that was read by hundreds of thousands of folks. He is more currently the co-founder of Halcyon, an AI-assisted research and information platform focused on energy transition. I have started using it a lot. It is great for those of us that are tracking regulatory dockets, sometimes with their many hundreds or even thousands of filings within a single docket. It is a great way to quickly summarize some of those documents. Highly recommend folks check that out. We will have a link to that as well as his famous slideshow presentation in the show notes.

This conversation did go long. I love talking to Nat. And so we split this one into two. In this first one, you’ll hear a whole lot of discussion of China. He’s based in Singapore, including a whole lot on EVs and the future of the automotive industry, a whole lot more. I think you’re really going to enjoy this and then come back next week to listen to the other one. Also, please don’t forget to give us a five-star review wherever you listen to the podcast and also to subscribe to the Texas Energy and Power newsletter and to the Energy Capital podcast. This is a free episode. It is not free to produce. Becoming a paid member at Substack really helps support our activity. So please do that today if you haven’t already. And with nothing further, please enjoy the Energy Capital podcast with Nat.

Nat Bullard (02:05.634)

Doug Lewin, thank you for having me. What a treat. I was just listening to you on the treadmill yesterday, and here we are in the flesh, so to speak.

Doug Lewin (02:14.868)

It is, I can’t tell you Nat how gratifying it was to get an email from you. I can’t remember when you sent it a couple months ago, six months ago or something saying, really enjoying the newsletter. I said to my wife, I’m like, I got a note from Nat Bullard today. Of course she’s not in the energy world. So she’s like, who? And I told her, you’re a big deal. She needs to know who you are. Nat, for anybody in the audience that doesn’t know, Nat does this, I think everybody probably will, but this amazing annual presentation. I’m sometimes asked by people, Hey, you seem to know a thing or two. Like what are some of the sources you learn from? And you’re one I always name, especially for folks that kind of want to get a global perspective. So we’re going to talk about that. What you’re doing with the presentation, some of the major takeaways, what you’re looking at for next year. But first you’ve got this new venture, which sounds super exciting to me as somebody who spends a lot of time reading regulatory filings and trying to dig through regulatory filings, ERCOT, NPRRs and all that kind of stuff. Let everybody know what you’re up to with Halcyon before we jump into the rest.

Nat Bullard (03:19.662)

With pleasure, Doug. So I spent a long time in the energy and the decarbonization research business. I was 15 years at the startup and then a company that got acquired by Bloomberg, New Energy Finance that became Bloomberg New Energy Finance became BNEF. And when I finished and took some time out on my own, I realized that I still have a real yen for the information business and the information in and around energy and changes in the system. And I also realized that it was a moment where technology allows us to do things that are different. And there are teams of people that are willing to deploy it and interested in building something new around it that viewed all of the stuff that you just described, you know, ERCOT, NPRR, or thousands of large generator interconnection agreements as a feature input for large language model and AI-driven capabilities, as opposed to a bug of people having to just read stuff by hand, so to speak, over and over and over again. And so about two years ago now, my lead investor, Andrew Beebe at Obvious Ventures, connected me with my co-founders, Bruce Falk and Alex Shuris, who are both veterans of building and operating companies and information systems at real scale, like really, really big. Billion dollar company revenue lines. Who only knows how big an information system in terms of that size, but real platforms. And we said to each other, like, what can we do with today’s new technology to answer questions and solve information challenges that we know exist?

And what we settled on is that actually it’s all of that stuff. It’s everything that exists in an unstructured, as the saying would go, fashion that is all text, that is all public, at least at the initial phases, but is, as we joke about internally, a denial of service attack on getting your work done. You want to read everything in ERCOT? It’s public. Go for it. Here you go. Here’s a fire hose of tens of thousands of pages a week per institution. And there’s 50 of them plus one if you include the District of Columbia. Start including all the other things you need to include. The federal layers, the other state layers. We’re talking about on the order of probably hundreds of thousands of pages a week of information. And within those hills, there’s gold. It’s just that it’s not often conveyed in a way that is readable by machine in our old classical way of thinking about it. But with today’s technology, with AI in particular, you know, we move from computers that can compute, that can add and subtract and multiply to machines that can read and can allow us to start extracting information that we really need. If you want to find the price of something, if you want to find a change to a forecast, if you want to find how did this number turn into that number. It often exists. It’s just that you would find it in say the third response to a request for information in this part of a proceeding in that state without any change log.

Doug Lewin (06:22.048)

Nat, I am so into that right now. So Oncor is in the middle of its rate case and there’s exactly that going on. Environmental Defense Fund and Google and a coalition of cities in their third and fourth requests for information. And within those long questions and long replies, there’s so much good information, but man, it is hard to like download every document and scan through there for what’s actually important. This is one of those use cases for AI. When people say, what are you actually going to use it for? This is one of the first things I say is like, my goodness. It makes what is just a, like you said, denial of service attack. Perfect. Like it’s just an impossible thing to do. This makes it possible.

Nat Bullard (07:06.286)

It does make it possible with a couple of other wrinkles on it. One is that it’s all in one place. So what tends to happen is you have people that specialize for good reason in a market, a jurisdiction, a service territory, an interconnection, whatever it is. And they tend to know all of that. But their ability to compare across jurisdiction, service territories, whatever it might be, is essentially nil. They don’t have the literal or figurative information architecture to do that. That’s something we built. You want to go search Texas, you can do it just the way you search Oklahoma, you can do it just the way you search Arkansas. It’s all in one place. It’s all there. You can draw comparisons across these jurisdictions. You can also, and this is something we began building in the last three or four months, abstract it up a layer, if you will, into something that has structure. So our way of thinking about it is when we have search and query capability, which we do, that people are welcome to try out, when you begin to have repeated queries from different parties about the same thing, that’s an indication that you should begin systematizing and programmatizing this into something that has a subscription basis. So that’s why the first thing that we built was, just go in and tell me what’s actually happening at the verified, costed, priced, disclosed level in the universe of building gas plants in the US.

You know, when everybody’s there trying to basically debug what somebody said on stage at CERAWeek, we can get to the number. I realized we have sources for that, right? We have places that we can go and find that. And every month we run that as a set of routines. We run that as a data science project that then manifests into an actual output that has all the numbers in it. And it’s got more than the numbers, it’s got the text updates on like this date has been pushed back a little bit or so and so is making an objection or whatnot. And it’s all put into a fashion that matches the other use case, which is not discovery, not I’m curious about what’s going on, or I want to find what I need to know more about, but rather I know exactly what I’m looking for, just give it to me. So we’re sort of attacking information in these two planes. One is sort of open with free structure for people to come in and search. The other is defined and deliberate and delivered for them on a cadence. Where the AI and LLM element comes in on that is just speed. Like this is the kind of stuff that back in the day you might’ve been able to do quarterly because you’re trying to go find out what’s new, trying to find out what’s update, trying to in your head, maintain a change log of it. Now we do it monthly at the moment in the future. The goal will be that this is essentially a continuum, right? That this lives and breathes essentially in terms of what’s been updated and what’s been changing, what’s happening right now.

Doug Lewin (09:49.518)

Yeah. And you mentioned the tracker for gas plants. We’re going to talk about that as we get into there. But what I first wanted to talk to you about was the slide deck that you did, which if folks haven’t looked at it, of course, there’s the annual one in January. If you haven’t looked at that, definitely go look at that. Just a wealth of knowledge there. And you did, I don’t know if it’s a mid-year update or just like a separate thing. It’s kind of, I think a separate thing. Was it discontinuity, disparity, data, if I’ve got that right?

Nat Bullard (10:15.512)

You got it.

Doug Lewin (10:18.988)

You did it at New York Climate Week and you had a slide in there as you do in your annual slide deck. You had several about Texas. You had a couple in this one about Texas too. And it was something that is the topic of a whole lot of conversations in Texas, which is the wide disparity between what ERCOT is projecting for load growth and what the transmission system providers, the utilities are putting out there. So first of all, talk about that slide and what you found in your research. And again, how hopefully it’s going to become much easier to kind of track these things going forward.

Nat Bullard (10:51.832)

This is one of my favorites, and it’s one of those things that’s sort of hiding in plain sight, if you will. To its credit, ERCOT has great information availability. Mapping that information can sometimes be a little bit tough, but it’s pretty good about updating stuff frequently and publicly. One of the recent things that ERCOT did was show how the transmission service providers in Texas, and then the grid operator itself, is thinking about future load, what’s going to be added to the system. And what I love about this is that the sort of the general shape of these charts is roughly the same until you get to the data center element of it. And ERCOT itself says we expect about 22 gigawatts of new load by 2030 from data centers, which in and of itself is a very, very big number. Let’s be clear. The TSPs on the other hand say we expect 78 gigawatts. Now, one of these things is not like the other one. A nearly 4X larger forecast for data centers and transmission service providers is noteworthy. It also seems to be perhaps a bit of a stretch. What is the current summer peak right now in ERCOT? 85, 87?

Doug Lewin (12:00.174)

85.5.

Nat Bullard (12:03.446)

Okay, I was close enough. This assumption would say we’re going to reach 90% of the current summer peak load just from data centers by 2030. Now, when we put it like that, that seems to be aggressive, but it’s a good lesson in how the TSPs, transmission service providers versus the grid operator, think about this. Because as your previous guest, Travis Kavulla, pointed out, there are incentives in place for those who build infrastructure to want to build more to accommodate more growth. Indeed. Yes. There are equally incentives on the part of ERCOT to keep things, as they used to say, to least cost and best fit, to best adequately reflect the reality, but also to do so at the lowest cost and with the highest reliability. And what I enjoy about this is that these are both official in their way. It’s not that like one of these is like the shadow version of it. And it’s not that like one of these is like a rebuttal. These are both essentially official ways of looking at the same information and making the same, drawing the same assumptions, but also with different purposes in mind. And yes, to your point, there is all to play for in essentially taking this one and making it look like that one. In the long run, they do have to collapse into one frame that would be reality.

And the question is, which one of these things is going to have to do either the expansion or the contraction? It is possible that ERCOT’s forecast is too low. I would wager that the transmission service provider’s forecast is probably too high. But that’s because they’re given the incentive to assume that much growth and to try to build according for it. And what we need to do in the future is to do a lot of things. One, we need to follow these things just as the data evolve. And again, ERCOT’s very good about doing that. Watch what actually has been approved for energization, change over time. Watch what’s going into the study. Watch what’s going through approval processes. The other, and this is a bigger informational challenge that I look forward to digging into in the future, is what I would call deduping. A lot of the things that are in the TSP assumption are probably assets that are being either planned in multiple parts of the state, or say the same finite amount of compute is being planned in multiple places, or it’s being planned across jurisdictions, which is an even more complex thing to try to de-duplicate. It’s a development business after all. And I think your audience of all will know exactly what I mean, is that this is what developers do. It’s a feature of good development to go out and keep multiple places live at one moment on the assumption that not all of them will come through.

Doug Lewin (14:41.73)

Yeah. You know, it’s interesting, and this is one that will be a place to look for more information and folks can go and try out Halcyon to see what it brings up for you. But there’s 58480, I believe is the PUC control number. I’m doing this by memory. You can go look at the show notes and I’ll have the correct one in there. But I believe it’s 58480 is the load forecasting docket that is open at the PUC right now. The legislature passed Senate Bill 6 to kind of deal with this because during the session, they were getting testimony from ERCOT. And some of ERCOT’s own, right, they have this kind of range of projections anywhere from 22 up to, you know, I think they have one projection that puts it like at 138 by 2031 or something. So there’s even within ERCOT’s own projections, there’s sort of this range. And then they were getting information from the transmission system providers. It was going to be much higher to your point.

So they said, Hey, we need to actually like figure this out. We need to know how much of this is real. So there’s a load forecasting docket. There’s another one, which is I believe 58481, which is actually a financial assurance. They’re going to make the project developers and data centers actually put some money down to actually have an official study and then show some kind of documentation is still going to be worked out through rules that it’s actually a real project, some kind of spend somewhere along the way. And they have to report to your point Nat, if they are investigating multiple sites. So I think what we’re going to start to see is that queue start to come down to more like realistic levels, still rapid growth, right? But probably not doubling in the space of three to five years.

Nat Bullard (16:26.518)

And what you describe is the regulatory institutions doing their essential work. And I mean this like I’m delighted to see this happening. This is something that actually is not within the gift of companies themselves to do. Markets respond to incentives and the incentive here is created by the regulatory body to say, go ahead and do this. And it’s interesting. If you think about this as a leading indicator of where other markets and jurisdictions might go is the forced disclosure of duplication is fascinating. And that’s probably something that I can only imagine happening first in Texas because of its size and breadth. That could happen in some other big states. It probably scales according to the physical size of the state and the size of the electricity system, which means guess what’s going to be first?

It’s going to be Texas. But a great effort that, again, I would hope to see picked up by other regulators elsewhere over time. And in the process, something that we just need to drive through information channels as much as possible to do and bless you for going directly to the docket number for people who are looking it up. This is one thing I’ve had fun explaining. I’ve had fun explaining to some of my colleagues. I was like, when you reach final boss level of people who do this for a living, they’re going to go to a prompt and just punch in a number. They’re going to go to a state and a number. And that’s how they’re going to think. And that’ll be amazing when we get to that point. We have people working on both planes. One come in and tell me something about X. Another is an incredibly detailed and nested question pointed directly at a docket. I will give you a for instance here, while catching up on our Discord chat with the engineers, they’re like, somebody is firing in some crazy informed queries right now. I was like, that would be Doug. I looked at the querying log, and I was like, that would be Doug. So more of that, please. Okay, many such examples, please, from all of you out there listening.

Doug Lewin (18:22.766)

Definitely. No, I love that you guys have created this tool. I’m looking forward to using it. Cause yeah, like we’re just talking about it takes a lot of time for these things. Okay. So switching gears just slightly. Again, I am just a huge fan of your annual presentation. I obviously spent a lot of time with it in January when it came out. I’ve referred back to it several times since, but I actually took a good chunk of the day, went through it all again. It was just sort of, again, just blown away by the breadth and depth, but really the breadth of this 200 slide presentation that you put together every year.

Can you just talk a little bit about, I guess this is kind of a two-parter, my audience will be used to this, I usually ask like five part questions, Nat. But a couple different things here, one kind of your process for putting it together, because I think it’s again, just kind of, you’re obviously not putting it together the week before it comes out, you must be gathering this stuff all the time and like kind of fascinated with your process. Interested if there’s anything that stands out in your head as like, just sort of a major takeaway because I would imagine, because you have so much in there about what’s going on around the world, some of which is just really stunning, this stuff about Vietnam and just things that are happening around the world that we aren’t necessarily hearing about in American news media or European or wherever listeners might be. And then of course, I asked the question I think any Nat Bullard fan would want to ask, what’s coming in the next one? What can we look for? It’s only a couple of months away from the next iteration. So can you give us a sneak preview?

Nat Bullard (19:57.55)

Sure. Thank you for that nested two-part question. I’m going to take the first part first. The way that I do this process is probably not going to be the most satisfying answer for people. I keep a giant running file of screen caps, snips, tables, links, anything that I have seen come across as catching the eye, a point of takeoff, a point of inflection, something where there has been a big headline or a change and I deliberately apply no taxonomy to it at first. So when I sort of hit wrap on my intake process last year, I had 385 slides that I took down to 200. And in the process, I probably took it down to about 170 and then built it back up with things that I know I needed to fill in. So I’m just going out there and I’m doing this live. It’s one of my favorite things to do, but it’s also like a subroutine. It’s not in real time, particularly active. Sometimes I’ll sit down on like a Saturday or a Sunday morning and I will start assembling things into sort of a more full frame. A lot of them I’m just flagging to go check on later. So, you know, where there’s higher frequency data, if it’s interesting in April, I should probably go check and make sure that it’s still valid in November. And so I put it all together into one place and then starting right about now, I start to kind of layer the themes on top of it. And so that’s sort of 80% observational. These are the things that I’m seeing. And then it’s about 20% aspirational in terms of what do I want to have come in and fill in the rest of this? And it’s funny. It’s probably because they’re the last slides in the deck.

Nobody actually looks at them. But the things that I enjoy most are my slides that are data driven, but sort of asking a question or a look ahead for us at the very end. When I do a retro on this, those are the ones that I tend to be the happiest about because I’ve raised the question further up the ladder. They don’t tend to be the ones that people come in and ask me about in the process. But I’m in the flow of just assembling sort of with no fear and with no favor the things that I find interesting from a pretty broad brush. So I get a lot of things from socials. I get a lot of things sent my way, by referral. I routinely check certain sources to see what’s interesting. And I try to learn as much from smart conversations with other people that might direct me to go out and find something. That’s the aspirational part. You should go look at this. You should take a closer look at that. And I’m always looking for novelty in the truest sense of the word. Is this new? Is this something that might not be well known? And that, I think, is where people one way or another entrust me with information. That’s the part where I hope to have the most value is, let me bring you something that you didn’t know. You know, there are things that are well priced. You know trends, you know, we know the cost of the solar module. That’s well priced. Less well priced, trade flows of Chinese solar modules into Africa. So that’s an example of where we can move from well priced to less well priced, so to speak, informationally. Or thinking too, in terms of adjacency, what are markets or technologies or business models that potentially become more relevant over time? So that’s the part one. Now part two, if I recall correctly, is like, what am I looking ahead to? What am I thinking about? Is that right?

Doug Lewin (23:25.09)

Yeah, I’m also just wondering some of those that kind of stand out that like people, particularly even like astute American power observers might, you know, look at your slide deck and be like, Whoa, man, if there’s nothing that pops to mind, we could also just talk about what’s coming next.

Nat Bullard (23:40.782)

There is so much. One is the Chinese auto complex, which is how I refer to it in the last deck. Yeah. There are only 11 countries in the world that have a full stack of doing an auto business, basically every capability from soup to nuts. China is now first among them. China makes something like 40% of the world’s vehicles.

Doug Lewin (24:00.726)

And that is, I know this from your slide, I think like 20 years ago, it was like three or four percent.

Nat Bullard (24:05.494)

Yeah, it was at the beginning of the century. It was less than 5%. I want to say it is the world’s biggest maker of all vehicles. It’s the world’s biggest maker of cars. It’s the world’s biggest exporter of cars. It’s the world’s biggest maker of EVs. It’s the world’s biggest exporter of EVs. This trend is not reversing anytime soon. And it is resonating outward to pretty much every other large auto market in the world, with the exception of the United States. And I would also say Japan. And this is something that I bring up because there are many, many things folded into this. It is trade flow. It is trade balances. It is national competitiveness. It is national champion industries. Automobility has been a classic way to move a country’s industrial capability upwards and with it, cohorts of an economy into the middle class. What happens if that’s no longer going to be the case? Or is China going to be the way Japan was of manufacturing in Kentucky, in North Carolina? You know, are we going to find that kind of path that’s driven by a combination of policy and incentives, but also market access?

Or is China just going to dominate? I’ll give you an anecdote. I live here in Southeast Asia. I was recently on a family trip in Thailand. The drive from Suvarnabhumi Airport into downtown Bangkok is quite long. And the rest of my family was sound asleep. I was not, because I was just counting billboards. I want to count all the car billboards between the airport and downtown. And I did not see a single billboard on this nearly hour long drive that was for a non-Chinese company. And of those, only a couple that were not for an EV. And one of them that was not for an EV was for a, like essentially an extended range hybrid, so small internal combustion, very big battery car. Kind of like the inverse of the plug-in hybrids that we have in the US. This is one thing that I think is really important because a lot gets wrapped into this. And it is for many reasons kind of outside the frame of what we think about in the United States. We have a huge auto business. We sell a lot of autos to ourselves. We do export, but not nearly to the scale that like Japan or Germany does relative to output. And we’ve optimized our full stack of automobility for America essentially, and for other countries that want to buy cars in the fashion that we do. China does everything else. And that’s something that I think is largely invisible, but I will put a question to you and to your listeners. Like when’s the first time you’re going to see an extended range Chinese electric pickup on somebody’s ranch in Texas? You know, when is the street going to find a way that one of those things is going to show up to someone’s like, I’ve got 400 miles of battery range and I can run 220V power in the bed.

Doug Lewin (26:53.742)

I think we’re a long, long way away from that. To me, Nat, the more interesting, that’s an interesting question, but one I wouldn’t even hazard a guess for. I think the more interesting question is when will we see some sort of Chinese manufacturing? And that’s probably a long, long way away. I sort of lost the thread. There was that CATL factory that was going to be co-located with, was it Ford or GM in Michigan? And then that got squashed, I think.

Nat Bullard (27:22.822)

I’ll give, that was the case. I’m not that positive. Yeah.

Doug Lewin (27:25.368)

That may be if America wants to export to the world, you’re almost at this point, gonna have to do it with China? Like this kind of stuff to the point of why I think your presentation is so good, because it does lead you to thoughts that otherwise you’re not really thinking about, right? You’re just like you said, like we have this American manufacturing industry, it’s mostly here, it’s sort of like internal. So we just sort of stay protectionist and we just sell to ourselves. That’s probably what happens, honestly, for the next 10, 15 years. Probably the most likely outcome.

Nat Bullard (27:58.222)

It is. It leads me to ask a couple of things, which is like how well optimized will our vehicles be for different future paths, right? For exporting to somewhere else where people want smaller, longer range EVs, for instance. How much are we exposed to the technological frontier? You know, it was Jim Farley at Ford who had imported one of these vehicles, Xiaomi’s vehicles, sorry, in order to test and he was like, this is a great car. You know, it is.

Doug Lewin (28:23.082)

And it was probably like $20,000, had like entertainment screens all over it. I mean, I don’t think we’re probably gonna be able to compete with that. And that is the thing that like that whole incident, just terrible, right? The Georgia plant, the Hyundai plant with, you know, Korean workers being put in chains and stuff like that was a path for the United States and maybe still is if there’s some amazing diplomatic work done somewhere, but like Korean technology and like an American and South Korean obviously partnership could have maybe put us in a way to compete, but now that’s looking less likely too.

Nat Bullard (29:03.092)

It’s important that you mention that because this is an industry in which it’s more than just setting up a distributorship. This isn’t paint or televisions even. This is an industry wherein you need to have a great deal of integration almost at a country level and certainly at a country immigration policy level in order to do this kind of work. And is the willingness going to be there? This is something that the US encountered in the 1980s with Japan, with Japanese companies setting up. And it took time and it took a sense of durability for it to happen as well. Yes. I think you raised the right question, but like what other parts of that supply chain could you have companies setting up and doing before it’s a full vertically integrated factory stack? And then the other question is even when that’s done, at what point is it going to really be worth the tariff to just import something? Because it’s at a price point, the quality is there, the market is there.

You know, I give Singapore as a fascinating auto market because it is tiny. Astonishingly expensive to get a car, well disclosed and visible and with data updated every month. And in December, 2021 BYD had 0.1% market share here in Singapore. It is now the biggest single brand retailer of cars in Singapore. So in less than four years, bigger than Toyota. And that tells you something. My building is full of them and increasingly other Chinese brands are showing up as well. But it’s interesting to watch a market that, look, does not have an embedded auto industry. It only has importers and distributors. Wherein with that kind of uncovered level of competition, you see the interest in stuff like this happening. And you see it differentially elsewhere in, you know, in Sub-Saharan Africa, in parts of the Middle East, certainly in Southeast Asia, even in Australia, you know, the interest from these companies coming in with a competitive product at a competitive price and seeing where that meets the market.

Doug Lewin (30:57.582)

I want to jump back to power stuff, but before we do that, this is just a really interesting thread to pull on. We’ve been talking about electric vehicles for seven, eight, nine minutes, whatever. We haven’t mentioned Tesla. We’ve been talking about American manufacturers, right? Like Tesla obviously is still one of the largest manufacturers in the world, but you do have some slides about they seem to be kind of peaking in the last year or two. Now there’s a lot that is bound up obviously in politics and the, you know, controversies that Elon has gotten himself into. You know, Tesla is still this huge question mark because the, I think the cars are still great. They’re moving towards autonomous vehicles. He’s making all these MegaPacks. There’s all these connections to the grid. You know, it’s the Energy Capital podcast. I focus on Texas, a Texas headquartered company doing some lithium mining and processing in Texas. So hard to have that discussion, I think, without at least bringing up Tesla, but I don’t know, it raises a lot more questions. It makes me scratch my head because I don’t know, I don’t know what the trajectory of that company is.

Nat Bullard (31:54.828)

You know what? There’s many, many of these questions and I actually look back to only one, which is how new is the lineup? You know, we have new trims for some of the existing models, but we don’t have a brand new mass market model in quite some time. Yeah. The Cybertruck is brand new. It’s not mass market. No, the Model 3 and Y are relatively venerable at this point. They do have updates and because they are software driven, they have a way of, you know, sort of improving themselves over time, but it’s different than the cycle time of some of the other manufacturers in the US and certainly the manufacturers elsewhere in terms of how frequently they’re updating brand new models, not just trims, but like totally new models, new chassis, new everything. And that’s my only real question is like, how do we think about it from a brand perspective? This is the trickiest thing to do. You know, every other car company, you have to talk about it like a brand. Tesla has sort of always been more of like an identity and with all these other things wrapped into it. I try to keep it very spare and just think about it from a car perspective. You know, as a car company, which is still what it is primarily, how are those cars being received in the market? And with only updated trims of late, and especially with the sunsetting of the EV credit, how is this going to play in a market that is increasingly supplied with lots of other options?

Doug Lewin (33:08.332)

Yeah. And of course, they’re looking more and more to the power sector. MegaPack is becoming a bigger and bigger part of the company, which may be a good segue to get back into the power sector. I do think I love that whole discussion because to a certain extent to our detriment, we think of these things as being really distinct. Like here’s the auto industry and here’s the power industry. But these days, as transportation becomes electrified, those lines are quite blurred and we’re kind of talking about the same thing when we talk about this stuff. So it’s getting interesting, right?

Nat Bullard (33:38.336)

It is. I’ll give you one last thought on that and then we can move it forward, which is that I’m going to ask you a question. So I did this annual presentation. What percent of my viewers of it are in the US? Do you think?

Doug Lewin (33:50.35)

I would guess 50%.

Nat Bullard (33:54.094)

43% last year. There’s a big world out there. There are lots of people from the US who want to look outside. From the outside, look into the US. And there’s a lot of people that want to look essentially mathematically in chords on a circle compared to the way that you and I are having a conversation right now. They’re in Brazil, and they’re very interested in thinking about what’s happening in Australia or vice versa.

Doug Lewin (34:16.075)

Yeah. No, I think a lot of times, obviously Americans, I do this myself. S**t, as a Texan, I do this a lot. Just think like this is where things are happening. It’s one of the things I love about your presentation. There is a great big world out there. And I do try to write about that at the newsletter too, right? Especially China and what has happened there. All right. So let’s get into some of the stuff that was in this latest slide deck, Discontinuity, Disparity, Data. And I believe you did the report right with GridLab about the prices of natural gas. So one of the things you had in there is this great slide that kind of shows the prices combined cycle gas turbines, basically averaging over $2,000 per kilowatt. Then you’ve got combustion turbines, much less somewhere in the $1,200 to $1,400. Now, one of the things we’re seeing that, and I’m wondering how much you’re tracking this or even can track this. I’m hearing more and more about this just from talking to people, but even the podcast Odd Lots last week got into this a little bit. Their guest put a drone over the Stargate facility.

Nat Bullard (35:09.746)

Stargate, yeah.

Doug Lewin (35:13.55)

They had seen the types of gas turbines that were there, and they’re pretty much like used, I don’t know what vintage or era, but they were like smaller, I assume, aeroderivatives, like Caterpillar turbines or something like that. I’ll add one more bit of data before I ask the question. Mitsubishi announced they’re gonna double their capacity of gas turbines. Do you see the gas turbine constraint kind of lifting anytime soon or do you think we’re really looking at like all the way through like 2035 these very elevated prices for new gas or used gas as it were?

Nat Bullard (35:52.34)

So I’ll start at the back. Do we see this easing? No, the demand is way in excess of supply and even Mitsubishi’s commitment. You don’t snap your fingers and magic up like rotor casting. Okay. Like, you know, these are some of the most complicated things in the world to possibly make with lots of lead time and in an industry that is rightly, I think, judicious about its expansion plans, having done this in years past, in eras past.

Doug Lewin (36:24.43)

20 years ago, right? And having a big old bubble burst.

Nat Bullard (36:27.788)

Wildly oversupplying a market for quite a long time. So that’s to my old saw here about, you know, markets responding to incentives. The incentives are complex and companies will react accordingly. The type of turbines that are being moved are fascinating. Yeah, like things that are aeroderivatives that have traditionally in these companies been sold by the oil and gas group. They were used offshore. They were used for field production. They might’ve been put behind the fence for LNG or something like that. Now being used as primary power production. There’s a move to buy used plant and move it around. I mean, that’s, you know, sure. Why not? That’s going to be of finite catchment in the long run, I think. And as far as price point goes, it’s funny because I put this slide up and the first comment I got was somebody’s like, I’m building those things and my prices are like three times higher. I’m like, show me your prices. And he’s like, I’m not doing that. I’m like, all right then.

Doug Lewin (37:24.654)

I’m not putting you in my slideshow.

Nat Bullard (37:26.222)

Yeah. Well, but also part of the purpose of building it this way, the reason that we wanted to work with GridLab on that white paper and the reason that our clients at Halcyon want this information is the rigor with which it’s sourced, which is I know that y’all are paying higher prices, but what I have here are the things that are in the regulatory disclosure. They have what’s called a CPCN or certificate of public convenience and necessity or the equivalent. Therefore, they kick into a process where we have visibility on cost, volume, time, conditions, all that sort of stuff. And that’s what we’re recording here. And also the full data set serves a broader purpose. A lot of people use it for physical trading. So they’re trying to say like with this asset coming online here and with my assumptions on asset utilization, what does that mean for demand at this node? You know, what does that mean for demand on this strip, that kind of stuff? So yeah, the numbers that we have here are lower than what people are probably getting quoted. In fact, I know they’re lower than what people are getting quoted.

But they are the revealed cost. Again, back around to my earlier comment, this is why we run this thing pretty much continuously to continue to get more sunlight. And in the long run, having come from a background where we did this quite usefully, would be to do a survey capability of it. I’d love to be able to do bids and asks here and get a sense in the long run of where people are at. We’re not quite there. It takes, having built this sort of thing before, it takes a little bit of time before you can get there. But if you’re out there listening and you’d be willing to start sharing bids and asks with me in terms of like when you’re getting quoted for combustion turbines, I’d be happy to talk. I want things that are only official. I don’t want anecdotes and I don’t want anything that’s non-verified. So that’s the purpose of doing this. But yeah, the numbers are striking to say the least. And there’s just, there’s a lot of future uncertainty about it. I mean, the biggest one for me being at a $2,500 a kilowatt cost for a combined cycle gas plant. What other competing options are there for providing power?

Doug Lewin (39:27.096)

So that kind of leads me to where I wanted to go next…

Thanks for tuning in to the Energy Capital Podcast. If you got something out of this conversation, please share the podcast with a friend, family member or colleague and subscribe to the newsletter at douglewin.com. That’s where you’ll find all the stories where I break down the biggest things happening in Texas energy, national energy policy, markets, technology, policy, it’s all there. You could also follow along at LinkedIn. You can find me there and at Twitter, Doug Lewin Energy, as well as YouTube, Doug Lewin Energy. Please follow me in all the places. Big thanks to Nathan Peavey, our producer, for making these episodes sound so crystal clear and good and to Ari Lewin for writing the music. Until next time, please stay curious and stay engaged. Let’s keep building a better energy future. Thanks for listening.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.douglewin.com/subscribe