Who benefits from the AI power bottleneck?

The bottleneck holding back AI is a scarcity of power, or so goes the story. That may be true — and plenty of reporting backs it up — but different actors in the space face varying incentives to play up or play down that narrative. So what incentives are at play, and how do they shape each player's story? In this episode, Shayle talks to Shanu Mathew, senior vice president and portfolio manager-analyst for U. S, sustainable equity at Lazard. Last month on X, he posted a breakdown of the actors — including hyperscalers, chip makers, utilities, and others – and how the different incentives they face shape how they talk about energy and AI. They cover topics like: Hyperscalers’ mixed incentives: the benefits of building their own capacity vs encouraging others to overbuild Why equipment makers, chipmakers, and land developers benefit from talking up the bottleneck to boost demand for their services How independent power producers and gas players benefit from high prices How the power-bottleneck narrative has shifted over time Resources: Latitude Media: ERCOT’s large load queue has nearly quadrupled in a single year Latitude Media: The power bottleneck is changing data center financing Latitude Media: Early-stage data centers are driving up US power demand forecasts Credits: Hosted by Shayle Kann. Produced and edited by Daniel Woldorff. Original music and engineering by Sean Marquand. Stephen Lacey is our executive editor. Catalyst is brought to you by EnergyHub. EnergyHub helps utilities build next-generation virtual power plants that unlock reliable flexibility at every level of the grid. See how EnergyHub helps unlock the power of flexibility at scale, and deliver more value through cross-DER dispatch with their leading Edge DERMS platform, by visiting energyhub.com. Catalyst is brought to you by Bloom Energy. AI data centers can’t wait years for grid power—and with Bloom Energy’s fuel cells, they don’t have to. Bloom Energy delivers affordable, always-on, ultra-reliable onsite power, built for chipmakers, hyperscalers, and data center leaders looking to power their operations at AI speed. Learn more by visiting BloomEnergy.com. Catalyst is supported by Third Way. Third Way’s new PACE study surveyed over 200 clean energy professionals to pinpoint the non-cost barriers delaying clean energy deployment today and offers practical solutions to help get projects over the finish line. Read Third Way's full report, and learn more about their PACE initiative, at ⁠www.thirdway.org/pace⁠.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  A very brief word before we start the show. We've got a survey for listeners of Catalyst and Open Circuit, and we would be so grateful if you could take a few moments to fill it out. As our audience continues to expand, it's an opportunity to understand how and why you listen to our shows, and it helps us continue bringing relevant content on the tech and markets you care about in clean energy. If you fill it out, you'll get a chance to win a $100 gift card from Amazon, and you can find it at latitudemedia.com/survey or just click the survey link in the show notes. Thank you so much.

• 0:34Speaker 01

  Latitude Media, covering the new frontiers of the energy transition.

• 0:39Shayle Kann?

  I'm Shail Khan, and this is Catalyst.

• 0:43Speaker 03

  Who has incentives to talk up the power constraints? And who may not have incentives to do that? And maybe take the opposite view where it's actually not as big of a deal? And where might economic incentives actually inform or influence our views more than is appreciated by the market?

• 0:57Shayle Kann?

  Coming up, show me the incentives, and I'll show you the behavior.

• 1:11Speaker 01

  Demand flexibility is now an essential part of the data center's speed to power playbook. But how do data centers actually curtail load? Gridbeyond is the end-to-end solution for flexible energy management. Gridbeyond's advanced DERS and VPP platform is tailored to what data centers care about most: maximizing reliability, intelligently meeting curtailment obligations, reducing costs, and managing all your distributed resources in one place. Ready to transform your distributed energy management? Get started at gridbeyond.com. Catalyst is brought to you by Antenna Group, the communications and marketing partner behind the most influential companies in energy, climate, and infrastructure. For three decades, Antenna has helped ambitious companies stand out,

• 1:53Speaker 01

  build market leadership, and become category-defining brands. From breakthrough startups to global industry leaders, their team brings expertise, relationships, and track record to accelerate your momentum. Learn more at antennagroup.com.

• 2:08Speaker 00

  The power plant for the modern grid isn't a building. It's a network. Energy Hub's derms platform connects batteries, EVs, thermostats, and commercial and industrial sites into virtual power plants that over 230 utilities dispatch as reliable grid capacity day in, day out. VPPs deploy in 60 to 100 days at 40% to 60% less cost than traditional infrastructure, and they perform like conventional generation from day one. Learn about the flexibility platform for the modern grid at energyhub.com.

• 2:45Shayle Kann?

  I'm Shail Khan. I lead the early-stage venture strategy at Energy Impact Partners. Welcome. All right, so this episode is a reminder to always monitor incentives. More specifically, we're talking about the narrative and certainly the reality today around power scarcity as the rate limiter on the growth of AI. Everyone is talking about it. Actually, I saw a chart recently that showed the frequency of mentions of the word power or energy on S and P 500 earnings calls over time. And you can imagine the shape of this chart. It had been pretty steady, actually trending slightly downward over time until 2025, when the line like literally went vertical. There's basically no question in my mind that power is actually a bottleneck for data center growth today

• 3:28Shayle Kann?

  and perhaps therefore for AI growth. That's sort of a truism. But it could also become a comfortable scapegoat, depending on where things trend in AI world. And because it's convenient and because it's widely discussed, it's important to think one level below the surface about what you hear from various players in the market. That's why I really loved this recent post on X initially from Shahnou Matthew that we're discussing today. Shahnou has been on this pod before to talk about public market stuff because he's a senior vice president and portfolio manager for US Sustainable Equity at Lazar. But he recently went online and went meticulously through basically every player in the AI slash energy world

• 4:12Shayle Kann?

  and laid out their incentives. Should they drum up concern about power shortages? Should they downplay them? Should they capitalize on them? This conversation for me is foundational and should act sort of as a reference and a reminder, for me anyway, anytime I hear something new about energy and AI from one of the major players in the market. So let's go through it. Here's Shahnou. Shahnu, welcome back.

• 4:37Speaker 03

  Hey, Shell. Thanks for having me on again.

• 4:39Shayle Kann?

  All right. So what got you started thinking about the incentives of all the different actors involved in data center power nexus? Like, what sparked the thinking?

• 4:49Speaker 03

  It's a good question. I feel like everyone has been thinking about this for the last two years, or at least hopefully so I don't feel as lonely. But I think the premise for this tweet or thought that I had was that, you know, there's general understanding that there's like, you know, a few ground truths. And one's like, AI is a foundational technology that's going to persist in society. Two is, we need a lot more energy infrastructure. The, you know, magnitude and duration of the cycle is of the debate, but it introduces a duration mismatch where technology is often done on cycles that are, you know, one to two years, whereas energy is typically multi-year or even multi-decadal, depending on the type of asset. And what I was trying to gear in is, you know, psychologically, humans, organizations have incentives and incentives

• 5:34Speaker 03

  drive behaviors, as everyone knows. And how could we take that, you know, lens and apply it to the power debate? Because so many times I have folks coming to me from earnings calls, or from events, or conferences, and saying, I heard this from X party, or this from Y party. And that's completely at odds with what they heard from another player down the stack. And so what we were trying to do was a clean sheet exercise was looking up and down the supply stack for AI. Who has incentives to talk up the power constraints? And who may not have incentives to do that? And maybe take the opposite view where it's actually not as big of a deal. And are there economic incentives that may inform or influence their behaviors? The goal wasn't to call out companies being misleading or anything like that. It's just namely to acknowledge where might, you know, economic incentives actually inform

• 6:17Speaker 03

  or influence their views more than is appreciated by the market.

• 6:21Shayle Kann?

  Yeah, I read the original post. And it was like, in some ways, it's kind of obvious. But at the same time, it was sort of like a lightning bolt to me. I was like, oh, yeah, like, I hear this stuff constantly as well, from every corner. Everybody is talking about the power bottleneck. And it is important to examine the incentives that drive anybody's behavior, or the narrative that anybody's telling about it. Not because there isn't a power bottleneck. I think you and I probably agree, like, there is, for sure.

• 6:47Speaker 03

  100%.

• 6:48Shayle Kann?

  The magnitude, the duration, and certainly it probably won't always be one. And so, you know, as this goes through whatever cycle it's going to go through, we need to be paying attention to like, who is saying what? And, and examining the incentives that drive that. So I, I thought it was super interesting. And what I wanted to do with you is kind of talk through some of them. What we are hearing from these various parties and then the incentives that they have to say what they are saying, or maybe to say what they're not saying. But let's start by categorizing, like walk me through at the high level. Like there are a bunch of, there are a couple of groupings of actors that have similar incentives. So how do you think about the groupings?

• 7:27Speaker 03

  Yeah, at a high level, I'll try to keep it simple. But, you know, again, there's various degrees of nuance required. But at a high level, you know, let's start with the hyperscalers who are the folks that are driving a lot of this capex and investment cycle, you know, these are, think of it as like the cloud service providers, the Amazons, the Googles, the, you know, Microsoft's other world, as well as including Meta, which is another hyperscaler, even though they don't have a legacy cloud business. That drives, you know, investment spend on GPUs. And that's a kind of the equipment, like you have the technology hardware, if you will, that includes, you know, GPUs, CPUs, custom silicon offerings. That moves upstream into like the supporting equipment, which includes electrical and cooling equipment. And then if you kind of go outside the data center, then you start to get into, you know,

• 8:12Speaker 03

  who supplies that power equipment, as well as the overall actual power of the facility. And moving upstream there is like, you know, who actually builds power, which is like utilities, as well as, you know, the labor and the EPCs and engineering that goes into those types of facilities. So if you kind of just go from who's spending the capital and follow that down the, the stream, that's a general way to think about some of these bigger pockets. And so happy to jump into any of those.

• 8:33Shayle Kann?

  All right, so let's, let's start with the hyperscalers. They're obviously at the epicenter of the data center build out. How do you think about their incentive when they talk about power?

• 8:43Speaker 03

  Definitely. And I think this is the most interesting from a, from a game theory standpoint. So just to back up for the, for the audience that may not be entirely clear, right? The hyperscalers have really strong incumbent businesses. You know, oftentimes depending on what you're talking about, cloud or digital advertising. And they think The spend on AI allows them, you know, a better product opportunity in the future such they can accelerate growth of cloud or advertising offerings. Or it'll enhance our product offerings, which, you know, will lead to better returns over time. And so you have two options to go out and get more capacity, right? One, they build their own capacity, but then two, they also can go out and procure capacity on the external market. And that comes from NeoClouds or co-locators or things that we can get into. But if you think about it from their perspective, right?

• 9:25Speaker 03

  Building your own capacity is typically what they prefer because they have control and radical integration. When they have the chance to, because they, you know, they have these, they have teams internally, they have the expertise. But when you go back to that core issue, right, is that I think everyone agrees that we need more energy. I think where the uncertainty comes in is how much energy do we need? And if you think about it from the perspective of them, if I have to go out and build all this infrastructure and this infrastructure is, you know, beyond the next few years, I need to, you know, potentially, potentially be comfortable underwriting it for the next 10, 20. That introduces a natural mismatch, right, where it's like, okay, do I want to be the long-term risk taker of the assets that I'm underwriting? I might be comfortable doing that to an extent, but beyond that, do I perhaps let the market figure out the, the excess capacity?

• 10:06Speaker 03

  So if you're a hyperscaler, the argument four would be if I really think that this is like the greatest demand super cycle and I'm severely constrained. I might try to go out and build as much as possible and then procure all the extra. The, you know, the skeptic or the pessimistic view would be they see line of sight into the demand they need. And that's all they're going to be comfortable building to and the rest I'll let the market sort out and such. So if you think about that from a power standpoint, my incentive in that scenario would be to talk up power constraints, right? Saying that there's not nearly enough of it, we need to build so much, like this is the greatest super cycle. Because that will incent speculators, developers, other companies like NeoClouds and stuff that want to compete for leases from hyperscalers to go out and build infrastructure. And this one's interesting because Satya Nadella, for example, the CEO of Microsoft, kind of told you this actually on a podcast earlier this year.

• 10:50Speaker 03

  I think it was late last year, he was on Dark Question, he talked about when he was asked, why not commit more resources to building more assets? And he ultimately said, well, I'm seeing a lot of activity from developers and co-locators and NeoClouds. And I think that I might be able to procure cheap supply in the future. Because who benefits if there is an overbuild or too much build and conditions loosen? That means cheaper leases for them and that might be cheaper than actually going out and building your own CapEx. And so that's kind of the game theory, one way or the other, in terms of why build or why would I, might I just wait to see how the chips may fall?

• 11:21Shayle Kann?

  They're complicated, I think, because you can see it going either way. On one hand, they clearly have an incentive to say that there is this massive bottleneck in the form of power and power infrastructure, because that, hopefully, from their perspective incentivizes the world to build more of that stuff so that they can build more data centers. So they should be shouting from the rooftops, this is the bottleneck. Also, by the way, it's convenient that, you know, it may, it may indeed be true, but it's also convenient for them to be like, the bottleneck is definitely power. That means it's definitely not demand, right, for their services. And again, that may be true today, but they could say that even if it weren't true, and that would be helpful to them. So to me, there's like a very clear incentive for them to say this, this power bottleneck

• 12:04Shayle Kann?

  is huge and insurmountable, and we need to, you know, have a, have a mobilization, the likes of which we've never seen in the energy sector, to solve it. On the other hand, if they don't solve it, they're, you know, if the world doesn't solve it, I should say, they are probably better positioned than anybody to snatch up the scarce resource that is the available power supply in the sense that like, if they are going to be forced to build their own capacity, which many folks are doing, like, who's in a better position to write long lead orders for G Vernova turbines than the hyperscalers who have big balance sheets and can afford to pay for it and so on. So like, you know, there is an extent to which if you're scared of the Neo clouds, for example, and you're

• 12:46Shayle Kann?

  a hyperscaler, you don't necessarily entirely want the world to solve the power bottleneck for you, because it's sort of a source of competitive advantage, at least against that other class. My guess is that between those two incentives, the one that says, no, we just want the world to solve the power problem because our job is AI, that's probably the one that wins out, but it doesn't seem pure as the driven snow to me, that it's like that clear.

• 13:08Speaker 03

  You hit it on the head, right? I think it's really, really complex when you pull it back to layers. I think the one thing I would add, right, is that a lot of the motivator for spending all this CapEx is that the, you know, each dollar of capacity not online that could be served, uh, could be serving additional demand is revenue dollars that you're not making. And so that opportunity cost is so great of not having power. So To your point, you may be comfortable, especially on like a, for example, a gigawatt data center being $50 billion, you know, spending X dollars of millions of dollars procuring deposits for long lead items, turbines, electrical equipment, switch gears, circuit breakers, et cetera, just to make sure that the long lead items are actually settled to build. So if the power is there, that you can go ahead and plug the shells in immediately.

• 13:52Speaker 03

  But there is, you know, there is an ability to walk away or cancel or, you know, just lose the deposit of contracts as well. But their incentive would be to lock up as much capacity to your point as possible, right? Because they want to have the optionality. But I think they also, you know, are playing for like, you know, worst case scenario, you could walk away from a few million dollar deposits. And I'm just speaking hypothetically here versus, you know, the opportunity cost of building a, you know, multi-billion dollar data center. So there is all these like different levers they can pull, but it definitely feels that they're keeping optionality. And then just to be clear too, like they are bringing on a lot of their own capacity, right? Like Amazon brought on 3.8 gigawatts in the last 12 months. So they're going to double their footprint in the next two years. Microsoft said they're going to double as well. And they brought on two gigawatts in the last 12 months. So they are bringing on a ton of capacity as well.

• 14:36Speaker 03

  But it is just interesting to see they're also striking deals in NeoCloud. So, you know, are they, the, the bull would say that there's nowhere near enough supply and then they're forced to do that. And the bear would say, well, actually there is want the speculative risk to go to the NeoClouds. And then they can walk away from the contracts if market conditions deteriorate. And it's, I think no one really knows, you know, time will tell.

• 14:58Speaker 01

  For data centers, connecting to the grid quickly means two things. Building distributed energy and having a plan for demand flexibility. Grid Beyond helps with both. Grid Beyond's Terms and VPP platform unlocks the potential of distributed assets and dispatches flexibility into the right market at the right time. And this is more than software. Grid Beyond supports developers every step of the way. From modeling your highest ROI energy mix to orchestrating the full set of assets around a data center. The track record speaks for itself. Grid Beyond manages over 1,400 sites and more than 3.7 gigawatts of load across Europe, the UK and the US. Learn more at gridbeyond.com or click the link in the show notes. If you work in energy or climate,

• 15:42Speaker 01

  you already know the company's Antenna Group has helped build. For three decades, they've been the communications and marketing partner behind the industry's breakout companies, category creators, and market leaders. Today, their team is working across the most consequential areas of the energy economy, from data centers and grid modernization to nuclear, geothermal, storage, critical materials, and next-generation technologies. Antenna helps ambitious companies become the names everyone in the market knows, trusts, and wants to work with. They build reputation, elevate executives, drive demand, and turn momentum into market leadership. If your company is ready to lead its market, Antenna is ready to help you get there. Visit antenna group.com.

• 16:24Speaker 00

  When you harness the power of millions of DERs in a VPP and put artificial intelligence behind the dispatch, the outcomes exceed conventional generation. Energy Hub's DERMs platform has been trained on years of real dispatch data from over 2.6 million DERs, so utilities can plan with confidence. That means consistent load reduction across a four-hour event window, even during the widest, most uncertain peak periods. Utilities target, shape, and verify capacity down to the meter. Unlock flexible capacity that costs 40% to 60% less to deploy than traditional infrastructure in a fraction of the time it takes you to break ground on a peaker plant. Learn more at energyhub.com.

• 17:08Shayle Kann?

  All right, there's another grouping of actors here where my first instinct to say is that it actually is straightforward what their incentives are, which is, let's just say, OEMs of hardware to data centers. That could be, you know, you're selling equipment inside the data center, or maybe not even hardware. Let's include, like, EPCs here, for example, as well. Like, your job is to provide services or hardware into the build of data centers. My first thought would be... Everybody there is aligned in their incentive to promote the power scarcity narrative. It's just, it just furthers their, the argument that there is limitless demand for their services and we're just going to sell out as much as we possibly can. Do you think it's more complex than that?

• 17:52Speaker 03

  I think that one is pretty straightforward to your point is that, just going back to our prior point about long lead time items and the most constrained, a lot of this is equipment as well as the labor, those are the direct beneficiaries. What do they get? They get higher backlogs, they get longer duration backlogs, they get better pricing in typically industries where you don't get, you basically get middle pricing or as much as the market will bear. Whereas now you're pricing power for the first time in generations, right? For certain equipment providers, for certain labor producers. So this one is pretty straightforward and you see that across the earnings, right? You can look at the earnings transcripts of companies like Quanta, like Record Backlog, MasTec, talking about, you know, record activity in their pipeline business as well as really a strong even clean energy pipeline.

• 18:35Speaker 03

  You can go to equipment or you see something like Avertive, which has a backlog that was up 30% organically year over year. You have Eaton, which is up 20% backlog organically year over year on, you know, like billion dollar books of business. And so I think there, that's what you'll see to your point, like a very limited narrative on talking down the potential market opportunity. Because if you think about it from their point of view, right? It's just more, more business to your point, right? Equipment, as you are power scarce, you need, you know, the latest and greatest to be the most power efficient per square foot or per IT capacity, per megawatt of IT capacity. And so you always want the latest and greatest and often the highest priced equipment as well as you want to continually get the latest generation, which requires you to size up, you know, the larger equipment.

• 19:17Speaker 03

  In terms of the labor, if labor is the most constrained element, that means that you'll agree to, you know, perhaps a higher degree of pricing that you normally would, where if conditions were, you know, scarce, you may be forced to just take, or, you know, whatever conditions that your buyers are willing to take. So for example, electricians and plumbers, we're incredibly scarce there right now. I mean, there's market chatter of companies like the hyperscalers flying out electricians from various parts of the country because they don't have enough labor. So you can imagine for the CPC companies, you know, pricing power. Is great because they're the only ones that have labor and can actually go out and deploy these projects. In an environment where that doesn't look as favorable, right? You have to agree to tighter terms. You have less overall projects to work on. You have more than enough supply of labor, folks.

• 19:59Speaker 03

  That means that your market will slow down and your pricing slows down. So like that is a vicious down cycle in terms of what happens on the other side of a growth cycle, right? So this bucket generally, I agree with you, straightforward. Talk up the business opportunity. It generally is positive and supportive for their businesses.

• 20:13Shayle Kann?

  All right. So let's talk about another one that I think of as being a little bit more complex than which is, which is utilities. Like on one hand, utilities are in the same position as everybody else we've discussed, which is basically like, look, if there's a huge bottleneck in power, clearly demand for their services, in this case electrons is going to be through the roof. They're going to be able to spend more CapEx. They should get higher earnings. Like this should be good for utilities. And so of course they would want to talk up the power scarcity narrative. On the other hand, you know, folks sometimes blame the utilities for the power scarcity challenge and like they, who is there to solve it but them to some extent? So they, there's probably some limit to the extent to which they can

• 20:54Shayle Kann?

  blame the, or place the limitation of AI growth on power scarcity. So I wonder how you think about them and what you've been hearing from them when they talk in public market contexts?

• 21:05Speaker 03

  Definitely. I think this is one of the positions that is not enviable for a lot of different market players, especially those in the other buckets. Because you, you bring up a great point where on the one side, generational CapEx cycle, I can get a guaranteed ROE on my investments and I can put through more investments. On the downside is we're increasingly seeing the political issues around rising electricity rates as well as folks being, you know, against having data centers in their local environment. So I'm like, the positive side, it's like a potential business opportunity. On the negative side, I'm getting increasing amounts of pushback. And ultimately utilities are regulated, right? They need to answer to a public utility commission and get the infrastructure they need approved. And so now they're in this interesting predicament where they have to solve for all these limitations, right? They have to

• 21:50Speaker 03

  bring on as much infrastructure as possible, but also do it at the lowest cost or at least the most Efficient use of dollars such that they can't be blamed for inefficiently spending capital or artificially inflating demand. So I think utilities were, you know, the last few quarters, right? You talked, or even like the last one, two years. It was a lot more talking about the gigawatt pipeline and how much they're growing. And, you know, you have folks throwing out tens of gigawatts in their pipeline or hundreds of gigawatts. And now it's talking a little bit more about how can we bring these gigawatts on in a very effective way and then ensuring that they have the labor, they have the equipment, they have the ability to execute on this quickly. Some of these are also in regions, right, where they're impacted by regulatory impacts, such as like Texas with SB six or PJM just actually went through and they couldn't get to a resolution on

• 22:35Speaker 03

  what's the, you know, approach to large load tariffs. But you're, you're right, utilities are conflicted in the sense of that they don't want to miss out on the opportunity, right? Because if they can't bring on the power or they aren't permitted to bring on the power, someone will take their load elsewhere to a different utility. And so they're in the business of problem solving right now, whereas, you know, it will last one to two years, talk up the opportunity. Now it's in the, in the realm of can I do this effectively and cost efficiently?

• 22:59Shayle Kann?

  How do you think about the, there's, there's another category, which is like the, I'll group together like real estate owners, land owners. And then the, the like powered land developers, the developers who are going off and trying to find sites and make them powered and then sell them to a colo or a hyperscaler or whoever. What's their incentive?

• 23:18Speaker 03

  Their, their incentive is to talk up the constraints as well, because it increases the value of the asset they're owning, which is the real estate or the powered land bank. If we think about solving again, if there's a constraint, that means that anything that accelerates you through the constraint is valuable, right? Or increases in value as a constraint gets worse or it remains. And so if I have a powered land bank, that means that I can move a lot faster than someone that has just a plot of land that has no interconnection access or anything like that, right? So I can charge a more premium price to developers and new clouds, et cetera. Anyone else that would want to be further along up the development cycle. And so they're generally universally talking up the opportunity that they have and the ability to move faster. I'd say, you know, a lot of what we're seeing there is, or at least what we hear chatter wise, right?

• 24:00Speaker 03

  Is that like you have transaction values that are much higher than they've ever been, or at least in recent And cycles as folks all are chasing the similar opportunities. So pretty much any landbanks that are powered that are near plentiful energy sources, something like, you know, in the Northeast with near gas fields or in West Texas, right? You're seeing construction activity be plentiful because people think that it's a much faster path to value realization. So you're seeing that increasingly move up. I think like, you know, one tall tell signal or maybe like a leading indicator that could like start to suggest weaknesses if you start to see real estate offerings that are powered that are near, you know, tier one cities or have really good access to energy infrastructure or fiber infrastructure

• 24:39Speaker 03

  that don't get a premium value or are struggling to sell in the market. That would be an interesting indicator that perhaps the market's not as strong as possible. But at least right now, the market indicators would suggest that that value is considerable because people are still focused on time to power right now.

• 24:55Shayle Kann?

  All right. I want to move on to the chip players.

• 24:59Shayle Kann?

  They're an interesting one here. Like if you think about Nvidia's incentives, on one hand, they share the incentives that most of these other folks do of like, let's play up the power constraint in the hope that it gets solved. On the other hand, they want to be making the argument that we are just going to sell effectively infinite chips forever. Like we're just going to sell so, so, so, so many more GPUs. And if power is a real constraint on that, that's the limiting factor and it's kind of outside their control, then one would think that would affect their perceived value and their perceived growth, I would say. But, but my sense is that Nvidia is talking kind of like the hyperscalers and really the story they're telling is

• 25:42Shayle Kann?

  this is a major constraint. I mean, possibly just because it's true, but maybe that's their incentive as well.

• 25:47Speaker 03

  Yes. And I think they actually, if, if, again, this is all very hypothetical game theory, but if you are Nvidia, right? Like you are pitching that you are getting the best performance token per watt is, is, you know, what Jensen will call it. And basically arguing that they can extract the most tokens per watt of useful energy. And so they're giving you the lowest cost efficient dollar to the highest level of performance. And so that actually gives them pricing power, right? Because since we are power constrained, And that constraint is expected to get worse and power prices rise. The Nvidia chip holders actually, the TCO advantage that they have versus less efficient chips actually goes up in that scenario. Right? So they actually continue to do better as power remains constrained because it reinforces their pricing power.

• 26:30Speaker 03

  And so this introduces like the recent, like, you know, the, the interesting angle where like custom silicon, for example, for those, your TPUs are all the rage this week, even though they've been in existence for a while. There's arguments that you can find more purposeful chips for certain workloads that require less energy draws. And if, you know, you weren't as constrained on energy, you might be, or you might be willing to use some of those chips. That actually erodes your pricing power for Nvidia, right? Because the, the idea is we're power constrained. I want to get the maximum amount of knowledge for every dollar I spend. And I can do that right now with Nvidia chips that I may not be able to do with custom silicon chips. And so it actually reinforces their competitive advantage. And so they have a kind of an interesting angle here where, to your point, they're going to sell more and more chips and you need power to bring online those chips.

• 27:12Speaker 03

  But they also probably want the market to remain tight because the longer it's tight, the better their pricing power actually ends up being.

• 27:18Shayle Kann?

  So on balance, basically, everybody we've talked about so far, you know, sometimes it's nuanced, but pretty much everybody is incentivized to play up the power bottleneck. And indeed, you know, they are, it is also real, but, but they are playing it up. Who do you think is on the other side of this equation? Like who is incentivized to say that power is not such a big bottleneck?

• 27:38Speaker 03

  Yeah, I, I think this part's a little bit more interesting too, because to your point, I feel like a lot of people benefit if the power constraint remains. And, you know, that's just like a okay thing to observe. It's just facts are facts. But on the other side, you know, I'd call it a few buckets and the, the first that I think is independent power producers. And this might seem counterintuitive, but let me lay out my logic is if you're an independent power producer, that means that you have a fleet of existing assets today. Those generate money based on selling into electricity markets. And so, you know, the input to their revenue is electricity prices. If we're in scarce conditions and those continue to grow. They generate a really healthy incremental margin on that, right? Because they're not really building out new assets. You have slightly higher O and M perhaps by running your assets harder, but you're just flowing through a high electricity price.

• 28:22Speaker 03

  That is extremely profitable for them. And that's actually been the thesis for a lot of owning these assets. So these are like the Constellations, the Talons, the Energies of the world. One thing I noticed, it's interesting. If you look at the like the last two earnings calls, right? You have the folks like the CEO of Constellation talking about the fact that, you know, energy prices actually aren't supportive of building new gas plants, which may or may not be true. And that, you know, there's different items that we have or levers to pull, like flexibility or, you know, things of like Tyler Nurse's suggestion of, you know, throttling workloads when you're at peak conditions to enable more excess capacity to be built and used today. I think that's interesting that the IPVs are talking out of that because like, if you game theory it out, it basically is suggesting that perhaps they don't want a lot of new supply to enter the market, right? Because what does that do? That potentially dampens prices.

• 29:07Speaker 03

  And I just laid out a path where they can generate a lot more healthier EBITDAs if they can, you know, sell into higher or appreciating power price markets. And so I think that's someone that is probably trying to downplay the impact just because their value in the marketplace goes up as long as this constraint, you know, remains. And then, you know, another area that I think is a little bit more interesting too, it's almost a derivative of how the market works. But like LNG providers, for example, a lot of them just sign very like long-term, you know, contracts. And then the basic arbitrage in US LNG is you buy cheap domestic gas and then you sell it to other markets like Asia or Europe to be able to, you know, sell into premium markets. Their model actually gets a little bit into trouble here.

• 29:50Speaker 03

  Or the arb gets less interesting if domestic gas prices rise, right? Because you go from taking a cheap input to a rapidly inflating input that's maybe no longer cheap and that kills your arb on the other side too. So they're probably incentivized as well to talk down some of the power issues to ensure that, you know, the supply goes to them versus going to a local market that they can be served much more profitably as well. So that's like two areas at least that came to mind in terms of the other side of the trade where, you know, perhaps trying to talk down the power constraint issue so that new supply doesn't enter the market and you get to continue to, you know, raise your prices. That also, you know, Other area would be like, just think about that a lot now is the natural gas E N Ps also as well. Right. If everyone's just chasing this glut, then you probably bring on more production. That actually keeps a lid on your prices when you want prices to be rising.

• 30:33Speaker 03

  So I think the way to think about this one is, does it, who benefits from a rising input price to their business models and they're probably in the camp of talking down expectations.

• 30:43Shayle Kann?

  Right. That's kind of where, I guess of those three groups, I think the I P Ps and the, the natural gas E N P companies are kind of aligned, right? They both benefit from high prices of their commodity. So they don't want a ton of new capacity to get built. That's just good for them. The L N G folks are on the other side of that trade, right? They're buying the natural gas that gets produced and selling into another market. So they don't, they don't want prices to rise. They do want more production. They want as much drilling as possible, right? And so then the R gets bigger. So for them, again, this is complicated game theory, but like, but you know, with the L N G players, you'd think they'd want to say, oh, it's a, you know, not, they should be telling you that we're talking about the power bottleneck. We're, it's actually a natural gas bottleneck. We need to be drilling so, so much more.

• 31:29Shayle Kann?

  And if that happens, then it keeps their prices low and keeps their R big.

• 31:33Speaker 03

  So that's a really interesting point though, right? Cause like, yes, they want natural gas production to be high because that means low, but they also don't want local sources of demand to, to accelerate. Because that means you have a lot more willing buyers, in this case, gas power plants, right? So if a bunch of C C G Ts go up, then all of a sudden you have a local buyer that's willing to pay a premium price that raises the input price for them. They want to buy cheap nat gas. But all of a sudden, if you have local demand that goes, you know, incredibly higher, you're competing with another source. And you know, producer might be able to go to somewhere else versus you. And that impacts your ability to serve your contracts on a longer term time horizon. So it is, it is complicated as you get, as you start to peel the onion, the layers of the onion back, you start to see all this like, you know,

• 32:13Speaker 03

  there is general incentives like for and against it within all these pockets, but it is just like, it's hard to parse out without going through these exercises and say like, who benefits in what market? And that's the, the, the rule or the role that investors play, right? Is we're generally trying to underwrite the range of scenarios, good, bad, neutral, and then probability weight, what's most likely to happen.

• 32:34Shayle Kann?

  Maybe taking a step back, I guess, just to to wrap it up. It seems clear to me that this this narrative, this general narrative of the power bottleneck has taken hold substantially over the past, I don't know, you tell me, but 12, 18 months, something like that is when it's really taken hold. How would you describe the the sentiment around that narrative today? And like, have you seen it changing at all? Is it is the fervor growing right now? Like, where are we in the trajectory of people talking about this problem?

• 33:06Speaker 03

  Yeah, I'd characterize it at a high level. The last one or two years was really just getting our heads around the issue. What's the source of this demand? How strong is it? And that's largely driven by GPUs and say, all right, we need the latest and greatest to train our models. Scaling loss hold. That means we can justify more investment and we'll continue to, you know, perpetuate the cycle. I think what it turned into recently now is folks are getting into that ROIC question is like, all the dollars that we're spending, you know, which is on the order of half a trillion dollars now a year, is that generating positive ROIC? And that question mandates, like, are they generating revenue and dollars from it? And so that moves you upstream to, okay, like, are folks actually doing that? And then that leads to the area where there's the question of if I had more supply, could I generate more dollars?

• 33:48Speaker 03

  And would the, you know, ROIC on each of those dollars be positive? And that's the part that we're in right now is what is the real constraint here? If you talk to a lot of the, you know, providers that sell AI solutions, it's because we don't have enough powered shelves or power to to get more supply online and be able to generate more dollars. But I think that's like the real question is like, if said another way, if we had unlimited power, unlimited constraints now that were all solved, would the AI trajectory still be as strong as it is today? And that's the question that we're getting into is like, which constraints of these are solvable? And if those constraints are relieved, does the cycle, you know, perpetuate positive for that player or negative? And I think right now, as folks are trying to answer that question, and that's why you see kind of the gyrations that you do in the market every other day, right?

• 34:30Speaker 03

  Where, for example, a lot of the AI power trade or AI plumbing trade, wherever you want to call it, for like a physical infrastructure, has done tremendously well for the last few years. I think folks oftentimes think about it as just Nvidia, but the names, like the, a lot of the companies that we talked about earlier, like the The EPCs, the MEP subcontractors, the networking and optical connectivity equipment, the switch gears, the circuit breakers, the transformers, the chillers, all these people have benefited. And so now it's just about the durability of those constraints, because that's ultimately what we're trading on now is like the two, three, four years outward, can they actually grow into what the valuations have grown to?

• 35:06Shayle Kann?

  All right, Shonu, this was fun. Thank you for helping me think through all the complicated incentives that are out there, mainly to just talk about how big a problem power is. But we'll have you back on when things take a turn, as they inevitably will.

• 35:19Speaker 03

  Yeah, it'll be fun to tag along and we'll see what happens here. And the narratives are always constantly changing and markets giving new opportunities to new narratives. So let's circle back here in a year or two and see where we landed.

• 35:31Shayle Kann?

  Shonu Matthew is a Senior Vice President and Portfolio Manager for U.S. Sustainable Equity at Lazard. This show is a production of Latitude Media. You can head over to latitudemedia.com for links to today's topics. Latitude is supported by Prelude Ventures. This episode was produced by Daniel Waldorf, mixing and theme song by Sean Marquand. Stephen Lacey is our Executive Editor. I'm Shail Khan, and this is Catalyst.