SPEAKER_00 0:00

World War III started months ago. And somehow that's good for natural gas, AI data centers, behind the meter, and at least three stocks in the AI investor portfolio. We have an unusual episode today. Eric is off, so it's just me. I'm sorry. He's leaving big shoes to fill. So I asked another industry expert to step in and share their experience. So you're about to hear a conversation between myself and Doomberg, an energy industry veteran, one of the most prescient names in this space, who's made many very productive calls in the AI trade, particularly going back to 2024, when they called behind the meter one of the more essential trends, which then led to stocks like Bloom Energy, V GE Vernova, and many others we've talked about performing very well. Now, we have talked about why this show is not a geopolitic show, but we do have to talk about it from time to time, particularly with energy, because energy is a geopolitical instrument. So here's why this conversation matters beyond that to anyone investing in AI. For the past two years, the bottleneck in AI was chips, chips, chips, chips. Today, the bottleneck is electricity. So you can buy all the NVIDIA black wells you want, but if you can't power them, they're expensive paperweights. And the numbers here are really staggering. Texas's grid operator has over 460 gigawatts of connection requests. Most of them are data centers, and that's against fewer than 10 gigawatts approved, which of course leads to behind the meter in companies like GE Vernova, and we've talked about SpaceX and their desire to bring 10 gigawatts online using local turbines. So this is not abstract for us. A big chunk of Eric's AI investor portfolio sits directly in the path of this trend. You've got Solaris Energy Infrastructure Partners, ticker symbol SEI, which rents gas turbines to data centers, including SpaceX. You have Vertiv, ticker symbol VRT, which handles power and cooling inside the data center. It's been one of the best performers in the portfolio. You have American Superconductor, AMSC, which is a bet on a grid that is only going to get more stressed, and we discuss at length in the following interview. There's uh MCOR and IES Holdings, these are electrical contractors, and then there's Regal Rec Stored, which was originally added as a robotics play, but also makes power modules going into them. So today the question is simple. If energy is the constraint on AI, who wins and who loses as what Doomberg calls World War III is already in motion. All that and more is next.

SPEAKER_02 2:24

The size of the hydrocarbon resource base underground is best modeled as being infinite.

SPEAKER_01 2:31

Period.

SPEAKER_02 2:32

Everybody gets excited about fusion. You know, we've written critically about fusion, and people scratch their heads. Like fusion is um an unneeded technology that solves problems that don't exist. Nuclear waste is literally uh the greatest job of branding that opponents to technology have ever done. Nuclear waste is not waste, it's it's merely fuel waiting to be recycled by future engineers. I think the GE Vernova trade, if you pushed us, is done. I think everybody kind of knows that story. What's your edge? This is a bubble. Bubbles can go on for a long time. Um it is simultaneously a very large bubble and an incredibly transformative technological leap. Both things are true. And there's a whole array of corporate structures that are just gonna go away. And AI is going to expose just how many useless white-collar jobs there are. We are in the middle of World War III. Um, in our view, World War III started in 2014 for reasons we can get into. Um, whether or not you believe that, it doesn't really matter. It's an effective mental model from which predictions can be made.

SPEAKER_00 3:46

Okay, so my guest today is one of the most widely read voices in energy anywhere. From X to some to Substack, they're a must-follow. They're also one of the most mysterious. So I'm joined by Doomberg, which is the anonymous team of former industry executives and commodity sector veterans. You'll know them from the Green Chicken Avatar with the tagline Chicken Little gets a terminal. Doomberg launched in 2021 and they've built one of the largest financial publications on Substack with hundreds of thousands of readers. One of the things that I really enjoy about them that makes them different is they start with the physics and the chemistry of energy, not just press releases, and then they follow the molecules to the money. If that sounds like what we're doing on the AI Investor Podcast, then you know you trust your instincts here. Uh Doomberg called the move to off-grid gas-powered data centers back in 2024, well before it showed up in the headlines. If you'd followed that sort of advice, then it might have led you to investments like GE Vernova or Bloom Energy. So obviously very prescient call then. So if you care uh about energy prices, geopolitics, and the collision of AI and data centers, you should be reading Doomberg. The good news is there's also a free newsletter and a paid tier. And listeners of the AI Investor Podcast can also access a special discounted rate by visiting 247wallstreet.com forward slash go, forward slash Doomberg. You can lock in a 17% discount today. I will put that in the show notes as well. Doomberg, thank you so much for joining the AI Investor Podcast.

SPEAKER_02 5:09

Austin, long time in the making. Great to be with you. Um, more than thrilled to be partnering with uh with 24-7. It's been fun and we're both learning a lot and looking forward to a fantastic discussion today.

SPEAKER_00 5:19

Uh me in particular, so we have joked with our listeners, you know, this is not a geopolitics show. So I'm gonna ask forgiveness of our listeners as we start with some geopolitics because I don't think you can understand energy today without starting there. And I promise for our listeners, we're gonna get it back to AI and and energy production. But, you know, Junberg, can we just start with you know the 30,000-foot view here on energy in the markets today? This year we've had a war with Iran, the Strait of Hormuz is still running below normal capacity, Brent is above $100 a barrel, we've got the diesel squeeze, yet natural gas is still around three bucks. So can you just give our listeners a state of the energy today? It's it's hard for us to understand, you know, the dynamics and the many moving pieces here that are affecting energy prices, which, you know, for us will end up affecting energy and AI compute.

SPEAKER_02 6:04

Yeah, I was gonna say I wouldn't underestimate the impact of geopolitics, particularly around energy and the AI race, which is something we have been chronicling for years. As you mentioned, we wrote about, I think the piece was called Irreconcilable Differences about how AI DNA centers would have to go off grid because the grid just wasn't designed to accommodate them. Um, and if you take a step back, the a key driving force of geopolitics is in fact the AI race. The two clear frontrunners in that race are the United States and China. The United States powers its AI uh uh training models and inference models using predominantly natural gas, either directly or via grid, which sources its electricity, you know, um, at least from a plurality measurement from natural gas. And then uh the main competitor is China, who um, for all the greenwashing and you know climate concerns, um, notwithstanding um China powers its grid based on coal. Coal is still dirt cheap. Um and that is the the the two those are the two fuels that are that are driving the AI race. We've also chronicled the uh semiconductor sanctions war between the U.S. and China. We have some pretty strong opinions on that as well, and that too is driving geopolitics. And so if you're interested in the AI space, um the types of questions you let off with are exactly the starting points that you should um you should begin with. So let's answer the question then. 30,000 view, uh, 30,000-foot view. We are in the middle of World War III. Um, in our view, World War III started in 2014 for reasons we can get into. Um, whether or not you believe that, it doesn't really matter. It's an effective mental model from which predictions can be made. Um the the war largely involves a battle over the US dollar uh in its uh primacy, its hegemony in global geopolitics. That matters for the AI race as well, because you know the US dollar is the weapon of choice for the US to stunt Chinese AI and semiconductor chip ambitions. Um and on one side of the war is the U.S. and its allies, although that alliance is fracturing, um, which is kind of what happens when you lose wars. And um, on the other side are the most sanctioned countries in the world, Russia, Iran, North Korea, and China, although less so uh compared to the other three, just because China has such a chokehold on certain key uh supply chains that are required for both the military and the AI industrial sectors. And so that's uh the sort of the starting point. Next level down then is how are those wars going? And we would say quite badly. Um NATO is being defeated in Ukraine. Trump was defeated by Iran both in June of 2025, uh, when it was a proxy battle between Israel and Iran, and then more directly in February, March of this year, when we think Trump made the disastrous decision to sink his presidency and go to war with Iran. And so um one of the things we like to say on podcasts and in our newsletter is that um one of the reasons you shouldn't start wars is because they're very hard to lose. And we don't mean tactically, they're actually quite easy to lose. Tactically, many wars are lost before they start. Um losing wars is hard mentally, it's hard emotionally. And we're seeing that play out right now in Ukraine. That war's lost. It's been dead lost for 18 months. It's been lost since Ukraine's major counteroffensive of 2023 was uh was defeated by the Russians. Everything else has been a grinding war of largely one-way attrition. And um the insistence on prosecuting a war that is lost is actually quite a cruel one because it condemns the war to continue until the last Ukrainian. And if you're a Ukrainian, that's particularly cruel. Same story in Iran. I mean, Trump should have taken the L in the spring, been plenty of time to reconstitute, um, take care of the home front ahead of the midterms, which are looking pretty ominous for the president now. That, of course, has significant impacts on energy and AI as well. And so all of it's interrelated. And we write about 90 articles a year trying to pull it all apart.

SPEAKER_00 10:23

Uh well, thank you for that level set. And really easy to understand why energy prices are so high. We've got the diesel squeeze, oil above $100 a barrel right now. Yet natural gas is still around $3 a barrel. So how can that be the case when oil and energy is seems to be broadly in crisis while natural gas is practically free? And you had mentioned that natural gas is the energy source of choice for U.S. data centers, and we talked about that with turbines and behind the meter. So, you know, is this just a lucky happenstance that the U.S. finds itself in? That might be a tailwind for data centers? I mean, how is this dynamic possible right now?

SPEAKER_02 10:58

Yeah, minor correction, um, $3 per million BTU. Thank you. Which is the rough equivalent of about $18 a barrel oil on an energy content basis. Um, this is something we have been writing about, predicting, and uh chronicling for some time. The reason why natural gas is dirt cheap in the U.S. is because oil is expensive. And that's a consequence of the fact that oil and natural gas are largely co-produced together from the same well across the big shale patches in the U.S. And in fact, natural gas is very, very regional fuel. Um, there are lots of places in the world where natural gas is prohibitively expensive, uh, including Europe. And there are places in the world where natural gas has up until very recently traded for negative prices. People will pay you to take it away. That has been true in the Permian Basin in the U.S. for much of the past 18 months. And it's still shockingly cheap there and in places like Alberta, British Columbia, where there's just too much gas, not enough egress, and um you you literally don't have anything to do with it, it stands in the way of the production of oil. So the environmental constraints around flaring and umor even just venting, which is still done in parts of the world. Um, if you bring up natural gas to the surface in the Western world, you have to do something with it. Um that thing that you do has to be economically uh responsible. And so you collect it and you clean it and you try to sell it. But if the local demand isn't there, as a gas, it's hard to handle. And so if you don't have the infrastructure to handle it, it gets given away, it gets burned, it gets it gets sold for rock bottom prices. And so because they're co-produced, when the price of oil spikes, all those drillers in the Permian and pick your favorite uh you know, uh shale patch, they start drilling huge volumes uh because they want that expensive crude. And they'll give away the natural gas to get that crude. And so um we wrote a piece on this called Generational Arbitrage, you know, um, a couple of years ago. Um you're going to see eventual engine switching, all of these arbitrages eventually close, and one of the ways it's going to close is a boom in data center construction in the US. Um, and so, in a way, the war in Iran has made natural gas cheaper than it otherwise would be simply because of co-production.

SPEAKER_00 13:24

Fascinating dynamics there, and I think that actually that dovetails nicely with something that you had said in 2024, where you described data centers as buildings uh where natural gas goes in one side and data comes out the other, and you had predicted that sitting right on top of natural gas fields would be an inevitable choice for where to build data centers. Two years later, we're seeing Chevron and Microsoft planning a 2.7 gigawatt project in the Permian. SoftBank has a nine gigawatt gas build-out in Ohio. So, how much of your call has come true? Has anything surprised you with that? And when I hear this dynamic about high oil prices resulting in you know low and affordable gas prices, obviously not a good thing for all of the other reasons that implies, but specific to the AI data center build-out, it seems, you know, how long would you expect this trend to persist?

SPEAKER_02 14:13

So the thing about natural gas, I mean, it we could do a whole show on natural gas. Um so, first of all, it has directionally come to pass in much the the way that we thought it would, the question of um of of this AI build-out and a behind the meter, but not quite as fast as we thought. One of our errors, you know, we're writing a book and one of the chapters is on being wrong. And when we when we embrace all of the uh forecasts that didn't quite play out as the way we thought they would, and when we embrace those and try to understand why, a common theme is that we're just too early. Um just because it should be that way doesn't mean it is. Um and so, yes, you're beginning to see the thing that was apparent to us when we wrote irreconcilable differences be realized by local regulators and big tech and energy producers alike. Some companies were early to the game. I could think of like EQT in the Appalachian region and so on. Um, but it it directionally it's been a good call. Um I would say that we have been surprised at the lag of that fundamental realization. Um another thing to consider, of course, yes, we're seeing all of these data centers in the Permian, but it's hot there, and half of the energy in a data center goes towards cooling, so it might even make more sense to park them in British Columbia and Alberta, and sure enough, we're seeing that uh roll out as well. Places like Iceland, you know, if you if you have the uh cheap energy and your ambient temperature is lower, that's that's a that's a win-win.

SPEAKER_00 15:43

I was going to ask if this is a major tailwind for Canadian tech, not an industry I follow well. And obviously there's other geopolitical tensions there, but it does seem that Canada seems to be unusually well positioned with pipelines, uh cooler weather, and a lot of space to build.

SPEAKER_02 15:57

Alberta is the Texas of Canada. They have a uh very industry-friendly, you know, tech-facing premier, the equivalent of the governor. Um, her name is Danielle Smith, very impressive um woman. And she um has navigated Alberta's resource extraction and development with um quite quite good skill and um has been working closely with the new Prime Minister of Canada to cut through the red tape to fully take advantage. Look, there's so much energy in Alberta and British Columbia. You asked me, you know, um might there be an overbuild of AI data centers and could we see you know the price of natural gas eventually converging and maybe, but man, there's just an enormous amount of natural gas in North America. And once the infrastructure is built to handle it, the break-even cost to bring it to surface is is de minimus. Um so there's a huge pure play natural gas field in Appalachia, you know, Ohio, Pennsylvania, um, New York, although they don't exploit it. Um the break-even cost for natural gas uh in that region, at least from the cost leaders like EQT, who I mentioned earlier, and I only know them because we're doing a deep dive on EQT for our Doom Zoom next month, um, is like two bucks a million BTU. So that's like $12 a barrel oil in basically infinite supply. I mean, the the amount of gas coming out of Appalachia alone would make it the third largest producer of gas in the world if it were a standalone country, um, behind only you know the US and Russia.

SPEAKER_00 17:33

Okay, that is that is an incredible statistic. And is one one more macro question, just because I'm fascinated by this, then I'd love to get to maybe some industry technicals here. But it when you're talking about this, we're what I'm hearing is abundant, affordable natural gas for the foreseeable future, in large part driven by high oil prices, but also just pure capacity and supply. And that the app that Appalachian stat is amazing. But I'd love to hear, you know, we heard a counter-argument from, I think it was um Matt Smith was on the Invest Like the Best podcast. And he had argued that natural gas was headed for a shortage with demand passing supply by 2028 and predicting a crisis by 2030, largely driven to LNG exports. So that seems to be diametrically opposed to your view. Um, and in that view, gas goes from about 10% to as much as 40% of hyperscaler costs, which suddenly makes AI much more expensive on a per token basis. Meanwhile, you've been writing about the abundance of natural gas in the Permian. So who is who is right? Where do you see this going for those next few years?

SPEAKER_02 18:36

I mean, Mr. Smith is not here to defend himself. Um you know, we publish what we believe to be true under our masthead, and we live and die by every every article. And um, we did a summary of that podcast appearances and his white paper, and I thought we gave it pretty fair treatment. Um his fundamental difference, I would say, between worldviews that we hold and what he holds is this belief that the U.S. will prioritize LNG exports when push comes to shove, and we just think they won't. So he's actually not making a call on a supply shortage. He is making a call based on a bottom-up analysis of what he thinks the industry could plausibly produce, and then saying demand is going to be greater than that. So, first of all, if you look at his supply predictions, they're in line with ours, and also far higher than the sort of peak shale gas pessimists that have been calling for a uh decline in U.S. gas production for years. I mean, if if the if the if the year ends in a number, they've been saying shale gas isn't going to be economic anymore, and they've been wrong for decades. One wonders how much wrong they have to be before they might reconsider their models. But um to Matthew's point, he does see substantial growth in natural gas. His argument is the combined demand of AI and LNG exports is such that um the market will tip to a shortage and prices will skyrocket. Two counterarguments to that. Um they'll just toggle the exports of LNG. Don't kid yourself, the AI is a nationally strategic fundamental priority, a bipartisan priority. But two, I would I would not underestimate the stampede of new supply that would come online at the first sign of elevated prices. Look, um if we did a thought experiment, and if the president tomorrow said there is a floor price of five dollars per million BTU under natural gas, the the utter stampede of drilling a new supply that would be rushed online to meet that incredible price would shock industry observers. Um one thing that outsiders perennially do, even somebody who studied the market as closely as Mr. Smith has, and again, we're not here to besmirch his analysis. We read it, we enjoyed it, we we wrote a polite counter to it. Um we have all day for polite debate, authentically held views, you know, politely expressed. Um one thing that analysts do, insiders and outsiders alike, is they always wildly underestimate the technical potency of the oil and gas industry. For some reason, there's this association of commodity with not technical. And unless and until you visited a deep-sea oil rig or an LNG export terminal or a you know uh state-of-the-art oil refinery, you have no idea what you're talking about. I mean, like there's thousands and thousands of PhDs and engineers and technicians and and other experts who work day and night doing this dizzying array of work that would be indistinguishable from magic as an outsider looking in, and yet they get no credit at all. You know, they're drilling miles underground and landing on the head of a pin. And just as their it's super advanced AI powered models indicated, there's hydrocarbons there. And they bring it up to the surface for free, for next to nothing. It's insane. And yet Um pessimists are always afraid we're gonna run out. Um and we just we just wholeheartedly reject that view. And in fact, um again we're writing a book on the subject and we have a whole chapter dedicated to nullifying this peak oil nonsense. I can tell you, you know, lateral thinking, um it is uh the the size of the hydrocarbon resource base underground is best modeled as being infinite period. Um all other assumptions about the size of the hydrocarbon resource available for humanity to exploit do not do as good a job of predicting subsequent events as that foundational assumption does, which is resources are never constraining. It's always something else. Find out what those constraints are, and at the first sign of high prices, those will be removed.

SPEAKER_00 22:50

Very well said. And and we have talked about a lot on the podcast here. It is it is very difficult to be a successful investor and to be a pessimist, but it's very easy to sound intelligent as a pessimist. And it's often a very tempting view, but we we share your view. Um, one of optimism is then solutions that people will overcome. And the the peak oil, you know, peak oil. I think it had a moment around March of this year where it sort of came back uh as high energy prices became a factor and people started to worry about peak oil again. But people have been worried about peak oil for decades at this point, and yet we still have energy prices that are relatively less affordable than they were decades ago. You're talking about the technical side of the energy trade. I just said I wanted to get to an industry, you know, more niche questions, but I actually have one more I want to pause on. Let's do nothing was more technical than nuclear. And you've said that the barrier to nuclear is political, not technical. Um, so I'd love to hear more about that. And you said that we should be building big reactors by the dozen, but it seems that the political opposition, and I mean, if we're having a hard time building data centers, I can only imagine the resistance to modular nuclear reactors to power those data centers. So, what role does nuclear play here? And thinking specifically from the investment perspective, nuclear was the trade for about a year as people started to anticipate energy bottlenecks. Then it turns out natural gas is affordable, and people are starting to realize the timeline on some of these nuclear reactor build-outs. So many of those investments have given back um their previously jubilant share prices. But I'm wondering now, is that an opportunity? Do you still think nuclear is a solution here, or is it simply unrealistic given the political opposition?

SPEAKER_02 24:22

I I would say we've gone sort of full cycle on optimism around nuclear. Um we first initially predicted, I think correctly, as that boom kicked off, that nuclear and its obvious potential to power data centers would begin to wipe away the political barriers that have stood in the way of a true nuclear renaissance. Um but we've become cynical. The industry um has become flooded with um fly by night, hocksters and pump and gumpsters. Um here's the challenge with nuclear. The real challenge with nuclear. The problem's already solved. Everybody knows what to do. There is no technology needed to catalyze uh a nuclear renaissance. All we have to do is build a hundred AP1000s or a hundred can-do reactors and deploy them. And in fact, settling on old, reliable, safe, proven, industrially um, you know, uh relevant designs. There's only you could pick three or four hectro in the Russian designs, they're good too, and um build entire supply chains around it, and you could be cranking out reactors every year like China is for a billion or two a gigawatt instead of you know a multiple of that here. And why part of the reason why is the venture capital world has convinced itself that all that's missing is some fancy technology for people to get excited about and put money behind. And it's just not needed, it's a solved problem. Like we are everybody gets excited about fusion. You know, we've written critically about fusion and people scratch their heads. Like fusion is um an unneeded technology that solves problems that don't exist. And so the pursuit of fusion uh merely punts the ball on fission. Like France built a fleet of nuclear reactors 50 years ago, um, and it still gets two-thirds of its electricity from nuclear. Today, those things are immortal if you maintain them. If France could do it 50 years ago, find me the technology barrier. There is no technology barrier. It's political. And now with the attention returning, we see these, you know, let's do mus small modular, let's do, you know, pick your favorite technology, let's do um thorium reactors, and you know, oh, this one produces less nuclear waste, and this one has no meltdown risk. I mean, these are fantasies. Um nuclear waste is literally uh the greatest job of branding that opponents to technology have ever done. Nuclear waste is not waste, it's it's merely fuel waiting to be recycled by future engineers. Well, 95% of the embodied energy and uranium is still there. Well, you know, uh yes, when you recycle it, you produce some you know, you produce some unwanted uh material that could be used to make bombs. Well, yes, we understand that um proliferation risk is real, but that all can be handled. That that's politics, that's not technology. Um so here we are, another hype cycle, another window for nuclear to get its act together to finally, you know, uh and it's it's going to be missed. Um natural gas is going to win.

SPEAKER_00 27:30

Then let's stick with natural gas. And I'd love to turn to turbines. So for the AI Investor Podcast, lead investor Eric Bleaker has discussed how behind the meter is absolutely essential for the data center buildup. There's permitting, local opposition, uh grid limitations, which you've written about. Um, today, if you're doing behind the meter, right, that requires natural gas and turbines. Um GE Vernova has a backlog of, I think, 116 gigawatts, which is a mind-bending number, and is taking orders for 2031. Siemens and Mitsubishi are booked out years as well. So if natural gas is as affordable as it is, are turbines the real ongoing choke point in the future? Should investors be excited about this? Or is this a case of you know, the problem with choke point, or you know, the solution of choke points is choke points and that they will get solved more quickly than people expect? How do you view the turbine environment right now? And is that a place where we should consider investing because it will be lean for years and therefore price setters, or is this a solution waiting to be solved right around the corner as capacity comes online and we should look elsewhere in the industry for opportunity?

SPEAKER_02 28:39

Yeah, look, we did a we did a piece um called Talib's Law. And um, this is from, you know, um uh Talid, of course, who is a you know an interesting character, a pretty deep thinker, brilliant man actually. Um he posted something on Twitter that uh that I saw back in the day that kind of stuck with me. And um something along the lines of um not all gluts lead to shortages, but I've never seen a shortage that doesn't lead to a glut. Um and we have a shortage of machines capable of converting natural gas to electricity today. You framed it as a shortage of turbines, I would frame it as a shortage of solutions that can convert natural gas into electricity cheap enough that the discount that natural gas presents can be captured. And so this is why you see Bloom Energy, which is not a turbine company, but is in fact a solid oxide fuel cell company sold out as well. Um we've had some mixed analysis of Bloom Energy. Um, you know, it um there's a lot of questions swirling around it, but forget about the company itself, and we don't do individual stock investments but um or recommendations, but um Bloom is a manifestation of the market crying out for solutions to convert natural gas into electricity that do not um do not require turbines. Elon Musk, of all people, have been has been making some investments in this regard. Um there is a shortage of such machines, and if you follow Talid's law, all shortages will lead to GLUT. So, what's the investor to do? I think the GE Vernova trade, if you pushed us, is done. Like everybody kind of knows that story, what's your edge? So if we were, say, participating in the private markets where we do a lot of our own investing, we would be looking for the startup that allows data centers to take natural gas and get data out in a way that does not uh find itself on the critical path of turbine production. And there's some alternatives out there that are interesting, but that's where the money will be made. Um closing of such arbitrages is where um generational wealth is created.

SPEAKER_00 30:46

So then continuing the turbine car I love that point, by the way, and I'm I'm gonna borrow that with with with attribution in the future. I do I do love that. Uh let's stick with turbines. Uh one of the investments in the most recent investments in the AI investor portfolio that Eric made is Solaris Energy, ticker symbol SEI. So the company rents gas turbines to data centers like SpaceX, so they can get power fast without waiting on the grid. You and I both agree the grid is not up to the challenge here. So this behind the meter opportunity is, you know, an opportunity. We're trying to figure out where to put capital. So is renting turbines a lasting business based on what you just said, or does it fade um as you know the capacity constraints ultimately result in a collapse?

SPEAKER_02 31:28

So, first time I've heard of the company, I'll just give you my uh impressions. Um, smart idea. Um effectively, they are trying to capture um the bit ass spread in turbine demand in a unique way, um, and good on them. And um, to the extent that speed um is rate limiting for a lot of these projects, um, they will certainly have a boomlet of revenue, whether they can service that revenue in a profitable way. Um, but all parties come to an end eventually. I would just be mindful of you know, that business model is one technology invention away from being made obsolete, which is wow, a really cheap thing that you can get in a month turns natural gas into electricity, and we figured it out. Now, you can argue that's years away, and maybe they'll take some of those profits and become innovative because they're in the industry. And I wouldn't bet against such companies, as you say. You know, um much easier to get rich uh investing as an optimist than as a pessimist for sure.

SPEAKER_00 32:32

In that sense, I wonder, and this is an investment that I own personally, is um, but in that sense, I wonder, based on how you described it, if maybe they're a little like neo clouds at the moment, where they're sort of benefiting from the pricing arbitrage of you know, renting high-end GPUs and purchasing them. So maybe companies like Solaris, you know, they're benefiting from that arbitrage today. And then, you know, TBD on how what they do with that advantageous position, do they invest it in new technology to remain competitive, or does that arbitrage window close as capacity comes online? Uh interesting and maybe you know potential parallel there with some of the neo clouds.

SPEAKER_02 33:05

Yeah, look, I can only give you my personal investment philosophy in the private world. Um, we try to get you know uh early investments into companies that we're happy to own because they cash flow. And if somebody wants to overpay for it, you hit the bid. Um and so I don't know what this company's capital allocation looks like if they're printing cash today or are they returning it to shareholders, you know, are they or they're gonna gamble and try to build some more sustainable business on the back end. But if they're paying out dividends or doing stock buybacks and you're getting remunerated as a shareholder and somebody eventually overpays for it, that's great. Um and that happens, you know, there's cycles and um you have to know when to when to sell into strength, as my friend Tony Greer at TG Macro likes to say, if if you're not if you're not gonna sell into strength, when will you sell? Um and so you don't want to, you know, are you renting this stock? Are you owning it? And again, this this is not security specific. Um, this is just how we look at investments. You know, if you get along a private company and it's cash flowing and it doesn't need more more capital to grow and it it occasionally distributes investor cash back, and somebody comes along and wants to pay you 20 times revenue for it, well, you know, it's yours. Enjoy. But I'd have the same mindset here. You know, this is a bubble. Bubbles can go on for a long time. Um, it is simultaneously a very large bubble and an incredibly transformative technological leap. Both things are true. Um, shale was a bubble, a lot of equity was destroyed, and shale radically transformed the geopolitical chessboard forever. The same story with AI. So, you know, I wouldn't be too pessimistic. I was talking to a pretty famous hedge fund manager a couple of months ago about how we believe that the demand for in for inference is infinite. And if you can't make money in the face of infinite demand, you know, I said just go get long something. You know, it's it's it's not to sound like Davy Day Trader here, but I mean, yeah.

SPEAKER_00 34:54

We uh we we share the uh view that we we will say the demand for intelligence, but in this world, you know, inference and intelligence or maybe analogous, we share the view that that is infinite, and then uh potent particularly as it becomes recursively self-improving and the power and potency of that intelligence is you know faster, cheaper, more capable, and it moves into different verticals like robotics. So we we also see infinite demand. Uh one of the uh we're talking a lot about behind the meter here and power delivery. And in the last six months, the share of projects building their own power has gone from about a third to roughly 50%. And I think it's trending even higher. Uh so co-host and lead investor Eric owns grid names like American Superconductor, electrical contractors like MCOR and IES. So I'm curious, you know, we have our view, which I've stated, but do you see the grid ever being able to catch up and support this level of demand for AI data centers, or is off-grid or behind the meter just the what it is always going to have to be going forward?

SPEAKER_02 35:59

The the purpose of the grid is not to feed AI demand. The purpose of the grid is to be the foundation of the standard of living of developed economies, um, which is the driving force behind why we wrote irreconcilable differences in the first place. Um you do not want to find out what the true clearing price of electricity is at the margin and then charge everyone in society that price. There's a reason why grids are a public good. Um you want to have automobile assembly plants and you want to have homes that can be reasonably powered and workers who can live a nice standard of living because their electricity bill doesn't take 80% of their paycheck. Um and so there's this fundamental tension between the speed and the value creation of the venture capital world and you know the benefit of the US national highway system. You know, it it you know you you wouldn't allow AI to ruin the highways and therefore interrupt you know food trucking and you know um Amazon Prime, heaven forbid, and um and so on. Like the the grid is a public good and it needs to be protected as such. And um and you know, the the whole regulatory structure is uh is built around that. You know, we like to think a lot about market structure and and you start with purpose. What's the purpose? The purpose of a grid is to um deliver reliable electricity at basically below cost, like below the marginal cost, uh, you know, to uh uh as uh in a way that uh people uh consume more of it. It it is um it is not meant to be a classic supply-demand market clearing uh um situation. And and so the pure capitalists will say that's nonsense and that's communism. Well, look, there's there none of the models are perfect. Um, but historically, the electricity grid is the foundation of high standards of living. And what is the economy other than a way to organize ourselves in order to impart a reasonably good standard of living for all of society, including a higher floor than raw capitalism would uh would typically permit.

SPEAKER_00 38:05

I like I like that point of the grid being a public good, not a tool for private industry, which would indicate that behind the meter just has to be the way these get built going forward. It seems similar to the US Postal Service, right? Which was not built for e-commerce and billions of dollars of returns and packages every day. Um, but it is still a public good and it serves a different purpose. I think in the US now we do like 750 million of e-commerce returns every day. So there has to be a different solution to that, whether it's which Amazon saw presently with their own delivery and fulfillment network and UPS and FedEx and others filling in the void there. Um I I would love to you know home stretch here. Last few questions from us. I'd love to get sent to some maybe specific recommendations and where you, uh, given your understanding of the energy market, would recommend our listeners look. So our AI portfolio owns power and cooling companies, but no oil, gas pipeline, or utility companies. And lead investor co-host Eric has floated buying gas producers as a hedge. Uh, I'm not sure if you would recommend that or you agree with that. So, just an open question to you if you could add one energy sector to an AI investor's portfolio today, what would it be? Where should we look and why?

SPEAKER_02 39:18

Yeah, we just did a deep dive last month for our premium tier uh on midstreams. So midstream companies are uh, you know, again, a different risk return profile. These are capital compounders, you know, classic, steady, 8%, 10%, 12% per year, dividend payers. There's some tax complications depending on which one you buy. You have to be careful, you don't do your own diligence. But a midstream company builds pipelines, gets paid largely by volume, not by price of the underlying commodity. Um so um, if pipelines are constraining to connect gas fields to data centers, they build them and they they collect a VIG for every molecule that passes through their pipeline. And um they're wonderful companies, they're well run, they have a little bit of leverage, they usually pay good dividends, um, Fortress balance sheets. We did a full deep dive uh on midstreams last month, including 10 interesting companies to look at. Um so you know, if if you decide to come to Dunberg and jump up to our premium tier, you'll get the full archive of all those presentations. Um but that's the place we would look. Wonderful.

SPEAKER_00 40:25

So these are the classic toll booth companies, right? As you said, they benefit from volume. So the more that flows through, the more they the more they make, they generally are income-rich investments, so they're popular with retirees. Not asking for you to give away the secret sauce here, but just some names maybe for our listeners to look at. You know, we're thinking companies like Enbridge, ticker simple ENB, uh enterprise product partners, EPD, Kinder Morgan is the big one, I think, KMI. Um there's you know, and just a uh uh an account consideration, uh because of the way these issue dividends, generally better to own them in a non-retirement account for our listeners who are interested.

SPEAKER_02 41:01

So the complication is that these can trade as um as publicly traded partnerships, broadly known as MLP master limited partnerships. Um they all not all of them do, but many of them do. You want to be very careful with those because you're not gonna get a 1099, you're gonna get a K1. And if you've done any private investing, you know the difference between the two. Um and so you do want to consult your tax advisor or or at least have a conversation with Perplexity or or pick or Claude or pick your favorite, you know, um AI model. They're usually pretty good about explaining the differences and what you should look for. Um there's nothing bad about it, it comes with some more reporting requirements, and if you if that's a cost you can bear or a complexity you can handle, then that gives you a bit of an edge because some people don't like to own these things because of that. A couple other names, um, energy transfer, um uh TC Energy up in Canada, a North American integrated play. If you want some exposure to Alberta in particular, they have a near monopoly on all the natural gas that's moved around Alberta. We wrote a piece on that uh not that long ago. Um, you know, so there's some great names. Um EQT is integrated. Um, they're a natural gas player in Appalachia. They also uh they're they're run by the Rice family. Um, you know, um Toby Rice knows how to make money, as we like to say, around the office, and um uh and his brother Danny. Um, and they um they run a pretty good shop. So they own Mountain Valley Pipeline now, and they they also they're also forward integrating into directly selling LNG, trying to capture more of the value in case Matthew Smith's predictions come true. There's lots of great companies that pay good dividends that are pretty reasonably priced. But again, these aren't going to be your 10 bangers. They're not going to be, you know, they're not growth company. But as far as like anchor, uh set it and forget it, reinvest the dividends and wake up one day and you have more money in your brokerage than you uh thought you did, though these are the types of companies that we would look at. Um but it depends where you are in life and what your risk tolerance is and the rest of your portfolio.

SPEAKER_00 43:00

Well, for many of our AI investors who might be uh used to and normalize to the idea of 100% returns in a year, that will sound absolutely pedestrian. But if you've been around the investing industry long enough, you will know that that is that is heroic and worth its weight in gold when you can find those sort of positions that you know just keep churning and cooking reliably. Uh it's not every year you can you can you know double your portfolio, nor should you expect it. Uh Doomberg, thank you so much for your time. Just one quick open question, last one for me. If people are looking to invest in the AI industry today, and there's a lot of ways you can do it, right? Chips, energy, whatever the you know, the model companies themselves, with your background in energy, is there anything that you feel like is misunderstood or overlooked that you'd love to clear the record on? You know, here's where I think you should look, or people have this narrative about this aspect of energy that's wrong today.

SPEAKER_02 43:49

Yeah, I think um based on our own dabbling in the AI startup universe, uh, there's a massive disconnect between the casual corporate person. And people who are actually familiar with not only the existing frontier, but the pace of change of that frontier. And there's a whole array of corporate structures that are just going to go away. And AI is going to expose just how many useless white-collar jobs there are. And your instinct would be: gee, let's sell AI to these companies and let them be the first to change. That's not going to work. What we think is going to happen and where we've put our own money is that entirely new business will be spooled up using AI as their foundation. In other words, boring typical business industries that are going to get steamrolled by new entries, new entrants that are well capitalized. And you're beginning to see some of the frontier shops with extra investor money looking at doing private equity roll-up type stuff where they basically just buy market channel and destroy the carcass and rebuild it, you know, with an AI hardwiring from the beginning. And then start with a team of AI experts and build around brand channel technology, demand creation, operations with AI being a core strategic input into each of those pillars. Look, we do it, we think about it with Doomberg all the time. We're a publishing company, just like you. What is our brand? What is our channel? How do we create demand? Um what's what's our technology strategy? How do we do our operations? And because we're small and we're nimble and we're quick, we have been quick to embrace um AI. Last thing I'll leave you with is a bit of a contrarian take, which it is our view that the value of authentically human is going to skyrocket um going forward. And um that foretells two things. Um maintain human authenticity if it is core to your business model, which is why we will always write our articles, and I will always um deliver the presentations for our pro tier. But also technologies that enable authenticity to be confirmed in the age of AI, human authenticity, they're going to be priceless as well.

SPEAKER_00 46:25

Well, yeah, I certainly love and endorse that view, but Doomberg, uh, you know, uh authentically human, I'm here for it, authentically chicken as well. So thank you so much for joining. Uh I I really appreciate your time and your expertise. I know our listeners will as well. So, uh, where can people learn more about you on Substack? Where should they go? Um we will include a link in the show notes if people want to join your newsletter, but if they want to follow you elsewhere, maybe on X or other channels, where can they where can they get your insights?

SPEAKER_02 46:50

So we've gone dark on X, um, mostly because we're full on Substack as our channel partner. They have an X equivalent called Notes. Um but Doomberg.com is everywhere. Uh, all things Doomberg goes there. Um, we have our annual newsletter, which you have um provided the discount link for. Um that's a great entree into the Doomberg world. And after reading enough of those articles and being you know tantalized by our monthly Pro Talks, maybe you could upgrade to our premium tier as well. We have about 500 articles in the archives and probably 50 or 60 Doom Zoom presentations now. And if you subscribe, you unlock the full archive of either the articles or the or the presentations, depending on which tier that you choose. Um, Austin, this has been great. Um happy to come back anytime. I've really enjoyed our collaboration and um best of luck going forward.

SPEAKER_00 47:35

Likewise, same to you. Um wishing you all the best out there. Until then, we'll talk to you next time.