Ep. 018 - Stop Saying Half of 2026 US Datacenter Capacity Is Canceled (Datacenter, Energy) | Jeremie Eliahou Ontiveros, Reyk Knuhtsen, Ellie Holbrook, Jordan Nanos

“Yeah, I can start with the why. I think Bloomberg started this, right, with this big headline of like half of US data center capacity is delayed. Everyone else started piling in and citing the same number. And you know that most of these articles point to the same underlying source. Right, which is a report that's available out there. And, you know, when you look at the report, the report says there was 12 gigawatts of data center capacity scheduled to go online in the US in 2026, and only 5 is under construction. Right. And I think, you know, for us, when we saw this, we were like, why, why is everyone talking about this? Like, it's just, it's just not possible. And, you know, I think that the funny thing is that, like, you don't even need to, like, do anything fancy. You can just disprove this data so easily. Like, I don't know, man, Amazon announced publicly they built 4 gigawatts in 2025. Is that going up or not in 2026? Obviously, it's going up. Amazon alone is probably going to add, like, you know, 5 gigawatts plus. So that's basically all they think is going to be going live, like one company. You know, obviously, you got all the hyperscalers. CoreWeave, you know, is gonna add a gigawatt in 2026. All of that is under construction. So you're telling me CoreWeave's 1 gigawatt under construction is like, you know, a fifth of the market? Like, come on, there's a massive issue in the denominator. Like, you're just wrong. Just don't publish that. So yeah, that's the high-level take, right? If you wanna add more deets, but for me, I think that's the, that was the trigger. It's just looking at the underlying source. You're like, oh guys, geez, you're just off. It's just wrong.”

“It was interesting. Maybe let me ask a different one because I'm thinking of it now listening to you talk. So At the very beginning of this article, we were commenting on the public research that has driven all of these articles in the media about how half of 2026 US data center capacity is canceled. But that chart that I put on screen and roughly what you're saying about the hyperscalers building for the AI labs says that, you know, even if you just take 2 of those 4 hyperscalers, we've got more than 5 gigawatts under construction. And then there's the whole rest of the market, the other 2 hyperscalers, and then everybody else. So can you opine on where the gap is in terms of people trying to do research on data center capacity under construction and just missing entire gigawatts worth of capacity? That's a real thing for next year. Yeah.”

“Yeah, I can start with the why. I think Bloomberg started this, right, with this big headline of like half of US data center capacity is delayed. Everyone else started piling in and citing the same number. And you know that most of these articles point to the same underlying source. Right, which is a report that's available out there. And, you know, when you look at the report, the report says there was 12 gigawatts of data center capacity scheduled to go online in the US in 2026, and only 5 is under construction. Right. And I think, you know, for us, when we saw this, we were like, why, why is everyone talking about this? Like, it's just, it's just not possible. And, you know, I think that the funny thing is that, like, you don't even need to, like, do anything fancy. You can just disprove this data so easily. Like, I don't know, man, Amazon announced publicly they built 4 gigawatts in 2025. Is that going up or not in 2026? Obviously, it's going up. Amazon alone is probably going to add, like, you know, 5 gigawatts plus. So that's basically all they think is going to be going live, like one company. You know, obviously, you got all the hyperscalers. CoreWeave, you know, is gonna add a gigawatt in 2026. All of that is under construction. So you're telling me CoreWeave's 1 gigawatt under construction is like, you know, a fifth of the market? Like, come on, there's a massive issue in the denominator. Like, you're just wrong. Just don't publish that. So yeah, that's the high-level take, right? If you wanna add more deets, but for me, I think that's the, that was the trigger. It's just looking at the underlying source. You're like, oh guys, geez, you're just off. It's just wrong.”

“Well, I, I, I would disagree. I think there's a pretty substantial amount of cancellations. Just what we keep saying is that these cancellations are, are early stage projects, right? Okay. I'm not sure if you have this one on the screen, but like there's this map of the US where we show like the large load requests. Right, which is like slightly different, but kind of goes to the same point. As of today, you have over a terawatt of data center load that is requested by operators in the US alone, right? This chart is as of 6 months ago. It's more than doubled now. So we have over a terawatt. Obviously, you know, the whole US system right now, peak load, 750 gigawatts. You're not gonna, you know, double it right now in sort of, you know, just a couple years. It's just not possible. It's not a reality. So obviously this is fake. There's a lot of early stage projects and it makes sense essentially, you know, I think the data center market started booming towards the end of 2023. That's when early folks started to realize, hey, there's going to be a big constraint on this. The first big deal started to get signed at that moment. And, you know, since then there's been like this massive search for power. And that means that, you know, everyone was trying to find where is power available. And that leads to some behaviors where, because everyone is doing it, then if you want to be successful, you also yourself have to be aggressive and plan multiple options. And that is actually nothing new. Like hyperscalers have always had multiple options when evaluating projects. It's even more critical in times of constraint. And so you would see sometimes for one final investment decision, see hyperscalers maybe have 10 different options, right? And so that would be 10 projects where they sort of talk to the local counties, you know, they talk to like the utilities and so on and so forth. And these are not all realistic and they're sort of just testing the field and be like, okay, which county can enable me to build my large-scale project? What do I have to workforce or do I have the supply chains? You know, where do I, where can I build it? So anyways, there's an oversupply of very early stage projects and, you know, our point and that was sort of Rake's brilliant phrasing, the cloud coded projects is that I just love this. This is amazing. That's the kind of stuff that AI is, I guess, not very good at filtering, right? So when you have humans in the loop, yes, it's pretty obvious that some of these announcements are way too aggressive. We gave a bunch of examples examples in the article of, you know, folks where, you know, they announce a 10-gigawatt project as they say, uh, the first tranche of 500 meg is gonna be available next year. And then you click on the website, you see contact us, nothing more. And then you start digging into whatever permits and whatnot, you see nothing. Uh, so at some point you're just like, okay, these people probably have a lot of netted Texas or something like that, but they don't yet have a real project, right? Uh, any human judgment would sort of filter that, but I guess for AI it's still pretty hard these days. Uh, so if you wanna build a forecast industry, you have to base it on realistic forecasts, you know, or semi-analysis. Extensive triangulation has always been our playbook, not just data centers, but also chips, because obviously this has downstream implications on Nvidia, upstream on the AI labs, their revenue, all of that stuff connects to each other. That's, you know, the semi-analysis flywheel covering every single one of these industries and building a cohesive view. You know, others don't do that and easily struggle to find the ability to put the site up.”

“Colossus 2, bro. Colossus 2. Fast, efficient, scale. Yale, super high revenue, W. Ellie, what's yours?”

“Well, I, I, I would disagree. I think there's a pretty substantial amount of cancellations. Just what we keep saying is that these cancellations are, are early stage projects, right? Okay. I'm not sure if you have this one on the screen, but like there's this map of the US where we show like the large load requests. Right, which is like slightly different, but kind of goes to the same point. As of today, you have over a terawatt of data center load that is requested by operators in the US alone, right? This chart is as of 6 months ago. It's more than doubled now. So we have over a terawatt. Obviously, you know, the whole US system right now, peak load, 750 gigawatts. You're not gonna, you know, double it right now in sort of, you know, just a couple years. It's just not possible. It's not a reality. So obviously this is fake. There's a lot of early stage projects and it makes sense essentially, you know, I think the data center market started booming towards the end of 2023. That's when early folks started to realize, hey, there's going to be a big constraint on this. The first big deal started to get signed at that moment. And, you know, since then there's been like this massive search for power. And that means that, you know, everyone was trying to find where is power available. And that leads to some behaviors where, because everyone is doing it, then if you want to be successful, you also yourself have to be aggressive and plan multiple options. And that is actually nothing new. Like hyperscalers have always had multiple options when evaluating projects. It's even more critical in times of constraint. And so you would see sometimes for one final investment decision, see hyperscalers maybe have 10 different options, right? And so that would be 10 projects where they sort of talk to the local counties, you know, they talk to like the utilities and so on and so forth. And these are not all realistic and they're sort of just testing the field and be like, okay, which county can enable me to build my large-scale project? What do I have to workforce or do I have the supply chains? You know, where do I, where can I build it? So anyways, there's an oversupply of very early stage projects and, you know, our point and that was sort of Rake's brilliant phrasing, the cloud coded projects is that I just love this. This is amazing. That's the kind of stuff that AI is, I guess, not very good at filtering, right? So when you have humans in the loop, yes, it's pretty obvious that some of these announcements are way too aggressive. We gave a bunch of examples examples in the article of, you know, folks where, you know, they announce a 10-gigawatt project as they say, uh, the first tranche of 500 meg is gonna be available next year. And then you click on the website, you see contact us, nothing more. And then you start digging into whatever permits and whatnot, you see nothing. Uh, so at some point you're just like, okay, these people probably have a lot of netted Texas or something like that, but they don't yet have a real project, right? Uh, any human judgment would sort of filter that, but I guess for AI it's still pretty hard these days. Uh, so if you wanna build a forecast industry, you have to base it on realistic forecasts, you know, or semi-analysis. Extensive triangulation has always been our playbook, not just data centers, but also chips, because obviously this has downstream implications on Nvidia, upstream on the AI labs, their revenue, all of that stuff connects to each other. That's, you know, the semi-analysis flywheel covering every single one of these industries and building a cohesive view. You know, others don't do that and easily struggle to find the ability to put the site up.”

“Excellent question, man. Excellent question. So I, I guess the first thing is, okay, this, the, the, this, this forecast sort of depicts our analysis mostly of the US grid., right? And looking at, looking at the gap, uh, context, we're not adding enough generation on the grid to meet that demand that is gonna, you know, be in the tens of gigawatts per year and just keeps increasing every single year based on all of the signals that we keep seeing, right? Um, so now is it realistic to assume there's gonna be 40 gigawatts of new behind-the-meter data centers added by 2028? You know, you have to look at what is being planned right now. Um, you know, I can tell you in the last 2 weeks, 3 gigawatts of data center deals were signed, uh, for behind-the-meter purposes, right? So the Deals are happening. Site selection I think happened beforehand. I think you saw a massive move in part of '24, mostly '25, where folks really started to have like access to a gas pipeline as one of the main site selection criteria. I think you started seeing sort of diversions where maybe some of the traditional data center operators were more focused on let's be in a tier 1 market like Northern Virginia, you know, let's find grid-connected sites because we need the 5 nines. And you saw sort of more newer operators, maybe more, you know, AI-built, that sort of try to foresee this trend. I think Crusoe is a great example. You know, they've been quite ahead of the curve on this and they've been able to, you know, sign massive deals in, you know, like Abilene's 672 megawatts with Microsoft announced publicly by Crusoe in Q1 that, you know, that's behind the meter, for example. Yeah. So anyways, I think we're getting there in terms of is the supply ready for it? Because there's like many, many developers that have secured sites. In terms of manufacturers, Ellie, I think I can probably do an extremely detailed rundown. In terms of probability of delays, I think the probability of delays is lower with behind the meter because of the way these grid constraints play out is that as a developer, you know, you talk to utility, they have no binding obligation to abide by the schedule that they provided you. And what keeps happening time and time again is is, you know, with the anecdotes we hear from developers say, I thought I was gonna get 500 megawatts by, you know, 2027. I had a handshake with the utility or whatever. And then, you know, a couple months later they tell me, actually, sorry bro, I can't do it. It's gonna be, you know, 100 by 2028, your 500 by 2032 because I gotta do bigger network upgrades because hey, actually I didn't consider in my analysis that this other guy also wants it. You don't have enough generation coming and so on and so forth. And so yeah, I think utilities are sort of realizing they're faced with more delays than they thought. And so in many cases, if your strategy is only grid, you're extremely likely to be disappointed. If anything, that's more so on the, on the power side. Behind the meter, I guess, adds a new risk, which is more the execution layer. Like, can you build a power plant on time? Can you get the permitting? That adds a new set of complexities. We're gonna see some high-profile delays. Obviously the New Mexico one, based on our analysis, is the highest of profiles, right? Basically $100 billion deal for OpenAI with Oracle. So the highest type of delays that you can have. Hope they're going to solve it on time, wish them the best. I think you're going to see more of that, but I think in terms of raw volumes, because there's like so many options now for developers, you're going to see a lot of success stories as well. And Colossus 1 and 2 are proof that you can do this at scale now. I don't know if they're the best examples because of the issues they have with regards to permitting, but they're demonstrating that it can be done and there's others doing it. Crusoe is another example.”

“Yeah, makes sense. Maybe we could talk about behind the meter now. So you guys in the second article published some numbers on this, the first y-axis on one of your charts in a little while there, Jeremy, which is pretty cool. So when we talked about the size and scale of some of these projects, I guess the behind the meter net additions when compared to like the available grid capacity is a significant difference. And you specifically forecasted over 40 gigawatts of net additions of data center capacity behind the meter by 2028. That number is effectively like it rounds to zero right now, right? Like there's a few behind the meter projects, but people are just getting going, ordering turbines, the supply chain's ramping up. What does it actually take to get there? And do you, do you expect that behind-the-meter projects are more at risk of cancellation based on some of the reasons Ellie just described, or less at risk of cancellation because they don't depend on the grid?”

“Excellent question, man. Excellent question. So I, I guess the first thing is, okay, this, the, the, this, this forecast sort of depicts our analysis mostly of the US grid., right? And looking at, looking at the gap, uh, context, we're not adding enough generation on the grid to meet that demand that is gonna, you know, be in the tens of gigawatts per year and just keeps increasing every single year based on all of the signals that we keep seeing, right? Um, so now is it realistic to assume there's gonna be 40 gigawatts of new behind-the-meter data centers added by 2028? You know, you have to look at what is being planned right now. Um, you know, I can tell you in the last 2 weeks, 3 gigawatts of data center deals were signed, uh, for behind-the-meter purposes, right? So the Deals are happening. Site selection I think happened beforehand. I think you saw a massive move in part of '24, mostly '25, where folks really started to have like access to a gas pipeline as one of the main site selection criteria. I think you started seeing sort of diversions where maybe some of the traditional data center operators were more focused on let's be in a tier 1 market like Northern Virginia, you know, let's find grid-connected sites because we need the 5 nines. And you saw sort of more newer operators, maybe more, you know, AI-built, that sort of try to foresee this trend. I think Crusoe is a great example. You know, they've been quite ahead of the curve on this and they've been able to, you know, sign massive deals in, you know, like Abilene's 672 megawatts with Microsoft announced publicly by Crusoe in Q1 that, you know, that's behind the meter, for example. Yeah. So anyways, I think we're getting there in terms of is the supply ready for it? Because there's like many, many developers that have secured sites. In terms of manufacturers, Ellie, I think I can probably do an extremely detailed rundown. In terms of probability of delays, I think the probability of delays is lower with behind the meter because of the way these grid constraints play out is that as a developer, you know, you talk to utility, they have no binding obligation to abide by the schedule that they provided you. And what keeps happening time and time again is is, you know, with the anecdotes we hear from developers say, I thought I was gonna get 500 megawatts by, you know, 2027. I had a handshake with the utility or whatever. And then, you know, a couple months later they tell me, actually, sorry bro, I can't do it. It's gonna be, you know, 100 by 2028, your 500 by 2032 because I gotta do bigger network upgrades because hey, actually I didn't consider in my analysis that this other guy also wants it. You don't have enough generation coming and so on and so forth. And so yeah, I think utilities are sort of realizing they're faced with more delays than they thought. And so in many cases, if your strategy is only grid, you're extremely likely to be disappointed. If anything, that's more so on the, on the power side. Behind the meter, I guess, adds a new risk, which is more the execution layer. Like, can you build a power plant on time? Can you get the permitting? That adds a new set of complexities. We're gonna see some high-profile delays. Obviously the New Mexico one, based on our analysis, is the highest of profiles, right? Basically $100 billion deal for OpenAI with Oracle. So the highest type of delays that you can have. Hope they're going to solve it on time, wish them the best. I think you're going to see more of that, but I think in terms of raw volumes, because there's like so many options now for developers, you're going to see a lot of success stories as well. And Colossus 1 and 2 are proof that you can do this at scale now. I don't know if they're the best examples because of the issues they have with regards to permitting, but they're demonstrating that it can be done and there's others doing it. Crusoe is another example.”

“Excellent question, man. Excellent question. So I, I guess the first thing is, okay, this, the, the, this, this forecast sort of depicts our analysis mostly of the US grid., right? And looking at, looking at the gap, uh, context, we're not adding enough generation on the grid to meet that demand that is gonna, you know, be in the tens of gigawatts per year and just keeps increasing every single year based on all of the signals that we keep seeing, right? Um, so now is it realistic to assume there's gonna be 40 gigawatts of new behind-the-meter data centers added by 2028? You know, you have to look at what is being planned right now. Um, you know, I can tell you in the last 2 weeks, 3 gigawatts of data center deals were signed, uh, for behind-the-meter purposes, right? So the Deals are happening. Site selection I think happened beforehand. I think you saw a massive move in part of '24, mostly '25, where folks really started to have like access to a gas pipeline as one of the main site selection criteria. I think you started seeing sort of diversions where maybe some of the traditional data center operators were more focused on let's be in a tier 1 market like Northern Virginia, you know, let's find grid-connected sites because we need the 5 nines. And you saw sort of more newer operators, maybe more, you know, AI-built, that sort of try to foresee this trend. I think Crusoe is a great example. You know, they've been quite ahead of the curve on this and they've been able to, you know, sign massive deals in, you know, like Abilene's 672 megawatts with Microsoft announced publicly by Crusoe in Q1 that, you know, that's behind the meter, for example. Yeah. So anyways, I think we're getting there in terms of is the supply ready for it? Because there's like many, many developers that have secured sites. In terms of manufacturers, Ellie, I think I can probably do an extremely detailed rundown. In terms of probability of delays, I think the probability of delays is lower with behind the meter because of the way these grid constraints play out is that as a developer, you know, you talk to utility, they have no binding obligation to abide by the schedule that they provided you. And what keeps happening time and time again is is, you know, with the anecdotes we hear from developers say, I thought I was gonna get 500 megawatts by, you know, 2027. I had a handshake with the utility or whatever. And then, you know, a couple months later they tell me, actually, sorry bro, I can't do it. It's gonna be, you know, 100 by 2028, your 500 by 2032 because I gotta do bigger network upgrades because hey, actually I didn't consider in my analysis that this other guy also wants it. You don't have enough generation coming and so on and so forth. And so yeah, I think utilities are sort of realizing they're faced with more delays than they thought. And so in many cases, if your strategy is only grid, you're extremely likely to be disappointed. If anything, that's more so on the, on the power side. Behind the meter, I guess, adds a new risk, which is more the execution layer. Like, can you build a power plant on time? Can you get the permitting? That adds a new set of complexities. We're gonna see some high-profile delays. Obviously the New Mexico one, based on our analysis, is the highest of profiles, right? Basically $100 billion deal for OpenAI with Oracle. So the highest type of delays that you can have. Hope they're going to solve it on time, wish them the best. I think you're going to see more of that, but I think in terms of raw volumes, because there's like so many options now for developers, you're going to see a lot of success stories as well. And Colossus 1 and 2 are proof that you can do this at scale now. I don't know if they're the best examples because of the issues they have with regards to permitting, but they're demonstrating that it can be done and there's others doing it. Crusoe is another example.”

“Yeah, I would say it varies a lot depending on what, what signed actually, what actually is signed. There, there's a bunch of deals out there. The bulk of the volumes these days would be turnkey leases where let's say a company like, I don't know, Digital Realty, QTS, built a data center for Microsoft. You know, they signed up for lease, Turkey lease, where sort of QTS takes on everything and Microsoft just rents. You have powerchairs, which are a bit different, which is sort of a lower, lower bound on the developer and sort of more capex from the tenant. You have just power deals, right, where it's basically a PPA. Let's say Oracle with Voltagrid in the Chaco Fort County, Texas site, essentially a PPA. There's no data center involved. Right. So depending on the deals, the timelines can vary. Rule of thumb would be say deal signed, 18 months you have capacity, but you have the first tranche and then depending on how fast you sort of can build it. But I think it's pretty clear that these days the expectation is that the ramp from sort of first capacity, first phase to full ramp is expected to be, you know, faster and faster. And the point we make in that article is that from a buyer's perspective, behind the meter is now becoming much more attractive than grid because you're in control of your destiny. Money, right? Like, hey, you know you're gonna have gigawatts by X amount of time. Now obviously it has to be permitted and so on and so forth, but at least from a power standpoint, you know exactly what you're gonna have on site and when, provided that there's, you know, delays from the suppliers and so on and so forth. So the idea is, hey, if you have all of this power, then you also need to build a data center 'cause otherwise it's useless, right? So the expectation is just, is that, that the ramping up these 3 gigawatts is not gonna take 5 years. It's gonna be, be much faster than this, right? And generally what you observe on the marketplace is that when there's a gigawatt-scale deal being signed, let's say, uh, you know, Oracle with, uh, Stack in New Mexico, with Vantage in Texas, you see that financing for the whole project takes place shortly after, right? Like, you— we saw, like, you know, um, DigitalBridge raise money for Vantage, uh, just maybe a month after, um, after the lease was signed. And we're talking about, you know, $25 billion of financing. So the whole— the money's already secured, right? Once you have that deal, you secure all the money and then you just try to build as fast as possible. It's not going to take 5 years. It's going to be much older than that.”

“Yeah, I mean, like, I think it was always to be expected. Um, last year the narrative was behind the mirror is not possible because there's these 3 manufacturers that are so slow. And, you know, everyone was sort of throwing that chart, uh, the famous chart of like, you know, how much gas orders that were in like the early 2000s where you had this massive sort of, you know, wave. And everyone was saying GEV and Siemens are so scarred of this era, uh, because they invested massively in capacity. And so they're gonna be very conservative. Right? But you know, the, that, that leaves a market opportunity. And I, I want one question we like to ask management teams at these power companies is, you know, how AI-pilled are you essentially? You know, how much do you believe, how much risk are you willing to take? There's also a function of how easy is it for you to take risk in the sense that, you know, what's your economics on building more capacity and new factories? And there's a bunch of companies that, you know, score very well. The one that we've been flagging for a while. We especially doubled down like at the end of '25 when we did our big deep dive with Blue Energy, because, you know, I think clearly the management team is very AI-pilled. I think they have economics that sort of enable them to build capacity faster than others. And the behind-the-meter conversation is really interesting because like, I think it's like, you know, company by company, they're sort of adjusting their mindset. As you adapt to the new reality. And also sort of solution by solution, you keep adjusting your expectation just based on the constraint. So what I'm saying is that initially everyone was like, okay, this is going to be bridge power. There's going to be a matter of like 1, 2 years, I run off-grid and then I'm going to have my grid come in and maybe it's going to be backup. And so you would only consider like systems that are good at backup. But that's the ultimate disadvantage for something like Bloom is that, you know, it's like not very good at backup. You have to be, you run extremely hot, I think it's 15,000 10°C, takes like 2 days, as far as I know, to go from 0 to 100. So for backup, it's, you know, really not a good system. But hey, if you have no other alternative, then maybe, you know, maybe it's that maybe you just have to go for it, right? Like, if your option is, your other option is that I'm not going to have power, then, you know, you're screwed and you're not going to be competitive in the marketplace. And, you know, other issues are like power costs. And I think when you look at the recent like SpaceX deals, I think it's pretty clear that sort of the revenue per megawatt that folks are making on the cloud side, also on the lab side, on the model side, like revenue per megawatt go up at every layer, I guess. And as that increases, it means that power costs are increasingly irrelevant. And power costs for solutions like Bloom are not that expensive anyways, with regards to, you know, what we have in Europe, for example, with the grid.”

“And the other thing is like also estimating the bit of materials I think has proven quite complicated for many of these vendors, especially as labor rates are going to the moon and you thought you were gonna have local labor, but hey, actually everyone in Texas is, you know, already occupied building data centers, so you have to call people from Denver or, you know, from Ohio and obviously it's much more expensive. And this means that this also favors solutions that are fast to install, fast to deploy because your BOM is sort of more predictable, your timelines are more predictable and both on the labor side and on the sort of full deployment side. So, you know, I think again, like analyzing timelines and ease of deployment is key and I think some solutions score extremely well. And I also think we're gonna keep seeing more and more new entrants, uh, cuz you know, everyone looks for capacity earlier and the standards are sort of dropping, right? The, the, the, there's a few new companies that entered recently that are basically coming from the automotive industry, not themselves, but like their source engines from the automotive industry. And then when you think of this, you're like, wow, automotive, like, you know, we're talking about like 100 gigawatt plus per year of production. And that's an industry everyone knows runs at very low utilization rates, you know, 50%, 40%. These factories are not doing too well. So the capacity and the incentive that these guys have to sell their engines to the data center market is also tremendously high. So, you know, that unlocks gigantic capacity.”

“I was just going to add to that, like, we're seeing the OEMs really reacting to the behind-the-meter story as well. So, you know, we're seeing obviously the, you know, the major OEMs, the established players like, you know, your Siemens, GE Vernova, all those guys expanding manufacturing capacity. But then throughout the stack, you're seeing new, new players, new entrants, new, new types of technology, and then recycled old types of technology, which we've written about in several different notes. But yeah, I think that that's, that's a really like increasing that That's a good tell for like, you know, how this market is developing at the speed. And also I think on the permitting thing, that's something which could delay. They, since we've seen, you know, companies switch to Bloom Fuel Cells in order to kind of get speed to, well, possibly speed to power because of manufacturing footprint of Bloom Fuel Cells, but also it's lower NOx and SOx emissions. So yeah, so they're, Yeah, and I think that was the Nubius New Jersey facility, for example, is, um, they switched to Bloom, uh, I think it was Bergen, uh, Bergen turbines prior. So yeah, um, in order to, yeah, to combat the, uh, permitting issues there. So we might see, you know, shifting around of technology, um, maybe relocating to more to other states like to Texas rather than in on the East Coast, um, But yeah, no, that's a, that, that, that would be kind of the, the story going on in the BTM market right now. Full, full growth.”

“Oh, um, you know, I don't know if I have too much here. I think, uh, I think we covered that the listener needs to know. I'm not going to lie. No, I don't know. I think, yeah, I think the BTM movement is going to be very, very big, but I think, I think maybe Jeremy touched on it already, but it's interesting to see the, like what we call kind of like peak turbine in 2026 where it's like there was a huge overbuy or like a huge overpurchasing of turbines in '26 where where not everybody knew what to do with them. Not everybody could figure out how to get the permits, or not everybody could figure out how to build the data center to actually use the turbines. And so you ended up seeing a lot of these turbines go kind of underutilized, or I mean, later on start to hit the secondary market, for example. But I think this is going to read pretty bearish to people at first, and people might kind of freak out on it, but long-term we're still very, very pro behind the meter. I just think it'll be interesting to watch how people interpret this though, because it's going to, it'll probably keep picking up, right? Like you'll probably see your favorite project's turbines start to hit the market at some point and you're going to freak out. And then the question is, is it over? Right? So we don't think it's over. We still quite like BTM, but I think that's going to be a fun dynamic to watch play out.”

“Yeah, I mean, like, I think it was always to be expected. Um, last year the narrative was behind the mirror is not possible because there's these 3 manufacturers that are so slow. And, you know, everyone was sort of throwing that chart, uh, the famous chart of like, you know, how much gas orders that were in like the early 2000s where you had this massive sort of, you know, wave. And everyone was saying GEV and Siemens are so scarred of this era, uh, because they invested massively in capacity. And so they're gonna be very conservative. Right? But you know, the, that, that leaves a market opportunity. And I, I want one question we like to ask management teams at these power companies is, you know, how AI-pilled are you essentially? You know, how much do you believe, how much risk are you willing to take? There's also a function of how easy is it for you to take risk in the sense that, you know, what's your economics on building more capacity and new factories? And there's a bunch of companies that, you know, score very well. The one that we've been flagging for a while. We especially doubled down like at the end of '25 when we did our big deep dive with Blue Energy, because, you know, I think clearly the management team is very AI-pilled. I think they have economics that sort of enable them to build capacity faster than others. And the behind-the-meter conversation is really interesting because like, I think it's like, you know, company by company, they're sort of adjusting their mindset. As you adapt to the new reality. And also sort of solution by solution, you keep adjusting your expectation just based on the constraint. So what I'm saying is that initially everyone was like, okay, this is going to be bridge power. There's going to be a matter of like 1, 2 years, I run off-grid and then I'm going to have my grid come in and maybe it's going to be backup. And so you would only consider like systems that are good at backup. But that's the ultimate disadvantage for something like Bloom is that, you know, it's like not very good at backup. You have to be, you run extremely hot, I think it's 15,000 10°C, takes like 2 days, as far as I know, to go from 0 to 100. So for backup, it's, you know, really not a good system. But hey, if you have no other alternative, then maybe, you know, maybe it's that maybe you just have to go for it, right? Like, if your option is, your other option is that I'm not going to have power, then, you know, you're screwed and you're not going to be competitive in the marketplace. And, you know, other issues are like power costs. And I think when you look at the recent like SpaceX deals, I think it's pretty clear that sort of the revenue per megawatt that folks are making on the cloud side, also on the lab side, on the model side, like revenue per megawatt go up at every layer, I guess. And as that increases, it means that power costs are increasingly irrelevant. And power costs for solutions like Bloom are not that expensive anyways, with regards to, you know, what we have in Europe, for example, with the grid.”

“Yeah, I mean, like, I think it was always to be expected. Um, last year the narrative was behind the mirror is not possible because there's these 3 manufacturers that are so slow. And, you know, everyone was sort of throwing that chart, uh, the famous chart of like, you know, how much gas orders that were in like the early 2000s where you had this massive sort of, you know, wave. And everyone was saying GEV and Siemens are so scarred of this era, uh, because they invested massively in capacity. And so they're gonna be very conservative. Right? But you know, the, that, that leaves a market opportunity. And I, I want one question we like to ask management teams at these power companies is, you know, how AI-pilled are you essentially? You know, how much do you believe, how much risk are you willing to take? There's also a function of how easy is it for you to take risk in the sense that, you know, what's your economics on building more capacity and new factories? And there's a bunch of companies that, you know, score very well. The one that we've been flagging for a while. We especially doubled down like at the end of '25 when we did our big deep dive with Blue Energy, because, you know, I think clearly the management team is very AI-pilled. I think they have economics that sort of enable them to build capacity faster than others. And the behind-the-meter conversation is really interesting because like, I think it's like, you know, company by company, they're sort of adjusting their mindset. As you adapt to the new reality. And also sort of solution by solution, you keep adjusting your expectation just based on the constraint. So what I'm saying is that initially everyone was like, okay, this is going to be bridge power. There's going to be a matter of like 1, 2 years, I run off-grid and then I'm going to have my grid come in and maybe it's going to be backup. And so you would only consider like systems that are good at backup. But that's the ultimate disadvantage for something like Bloom is that, you know, it's like not very good at backup. You have to be, you run extremely hot, I think it's 15,000 10°C, takes like 2 days, as far as I know, to go from 0 to 100. So for backup, it's, you know, really not a good system. But hey, if you have no other alternative, then maybe, you know, maybe it's that maybe you just have to go for it, right? Like, if your option is, your other option is that I'm not going to have power, then, you know, you're screwed and you're not going to be competitive in the marketplace. And, you know, other issues are like power costs. And I think when you look at the recent like SpaceX deals, I think it's pretty clear that sort of the revenue per megawatt that folks are making on the cloud side, also on the lab side, on the model side, like revenue per megawatt go up at every layer, I guess. And as that increases, it means that power costs are increasingly irrelevant. And power costs for solutions like Bloom are not that expensive anyways, with regards to, you know, what we have in Europe, for example, with the grid.”

“Yeah, yeah, sure. So the project which is facing quite a lot of pushback locally, not just New Mexico local opposition, but also from jurisdictional and regulatory kind of characters there is Oracle's Project Jupiter in New Mexico. So yeah, they're essentially, they're trying to construct a a pipeline to feed the behind-the-meter data center in Donia Ana Province. And essentially, this pipeline hasn't been built, and the route that they are planning to— that they want to build on hasn't been approved. They haven't got an approved route. They have got a secondary approved route, Essentially, all of the ways to get gas to the site don't seem very feasible. There's a pipeline issue which they're trying to make happen, but every FERC filing that I see that comes through on the docket, they haven't, they haven't made much progress or any progress at all because essentially it's defaulted to a type of regulatory process which essentially there is no precedent for it being done sooner than essentially 2 years. And given the local opposition as well, that this is very unlikely that it'll be sped up. There are other options to getting gas to the site, like you could ship truck CNG and LNG to to a facility, as, as we point out. However, this at scale, you know, getting to— that's, I think, the only proven, um, like, delivery of, uh, CNG at scale is kind of around 200 megawatts, or like max 200 megawatts. I think maybe even that's optimistic. And there aren't many, there aren't many manufacturers, there aren't many producers in the local area, and there are not, not enough trucks to facilitate that. So essentially, yeah, there's not only a pipeline which doesn't exist and has no viable route of existing, but getting gas to the site via CNG and LNG is even harder. Um, it's just not, not viable. So yeah, no, it's, it just, I think it brought it like, it, it's, um, it demonstrates kind of an interesting, uh, bottleneck which could occur with more behind-the-meter facilities is like building gas pipeline, um, uh, infrastructure, uh, especially in territories, um, like New Mexico that's quite not, not, not, not extremely, it's not extremely friendly area towards that kind of infrastructure. Yeah, I would, that's how we basically look through the filings. We look through all of the filings. You know, we look, we already had it on our radar because they had issues with the initial turbines that they were trying to use at the site. So they ended up switching to Bloom Fuel Cells. And through kind of that research, I spotted this timeline and I was just like, this doesn't make sense. Why are there so many local, all these pushback, all these local comments? And then yeah, FERC and the other regulators are still thinking about it.”

“Excellent question, man. Excellent question. So I, I guess the first thing is, okay, this, the, the, this, this forecast sort of depicts our analysis mostly of the US grid., right? And looking at, looking at the gap, uh, context, we're not adding enough generation on the grid to meet that demand that is gonna, you know, be in the tens of gigawatts per year and just keeps increasing every single year based on all of the signals that we keep seeing, right? Um, so now is it realistic to assume there's gonna be 40 gigawatts of new behind-the-meter data centers added by 2028? You know, you have to look at what is being planned right now. Um, you know, I can tell you in the last 2 weeks, 3 gigawatts of data center deals were signed, uh, for behind-the-meter purposes, right? So the Deals are happening. Site selection I think happened beforehand. I think you saw a massive move in part of '24, mostly '25, where folks really started to have like access to a gas pipeline as one of the main site selection criteria. I think you started seeing sort of diversions where maybe some of the traditional data center operators were more focused on let's be in a tier 1 market like Northern Virginia, you know, let's find grid-connected sites because we need the 5 nines. And you saw sort of more newer operators, maybe more, you know, AI-built, that sort of try to foresee this trend. I think Crusoe is a great example. You know, they've been quite ahead of the curve on this and they've been able to, you know, sign massive deals in, you know, like Abilene's 672 megawatts with Microsoft announced publicly by Crusoe in Q1 that, you know, that's behind the meter, for example. Yeah. So anyways, I think we're getting there in terms of is the supply ready for it? Because there's like many, many developers that have secured sites. In terms of manufacturers, Ellie, I think I can probably do an extremely detailed rundown. In terms of probability of delays, I think the probability of delays is lower with behind the meter because of the way these grid constraints play out is that as a developer, you know, you talk to utility, they have no binding obligation to abide by the schedule that they provided you. And what keeps happening time and time again is is, you know, with the anecdotes we hear from developers say, I thought I was gonna get 500 megawatts by, you know, 2027. I had a handshake with the utility or whatever. And then, you know, a couple months later they tell me, actually, sorry bro, I can't do it. It's gonna be, you know, 100 by 2028, your 500 by 2032 because I gotta do bigger network upgrades because hey, actually I didn't consider in my analysis that this other guy also wants it. You don't have enough generation coming and so on and so forth. And so yeah, I think utilities are sort of realizing they're faced with more delays than they thought. And so in many cases, if your strategy is only grid, you're extremely likely to be disappointed. If anything, that's more so on the, on the power side. Behind the meter, I guess, adds a new risk, which is more the execution layer. Like, can you build a power plant on time? Can you get the permitting? That adds a new set of complexities. We're gonna see some high-profile delays. Obviously the New Mexico one, based on our analysis, is the highest of profiles, right? Basically $100 billion deal for OpenAI with Oracle. So the highest type of delays that you can have. Hope they're going to solve it on time, wish them the best. I think you're going to see more of that, but I think in terms of raw volumes, because there's like so many options now for developers, you're going to see a lot of success stories as well. And Colossus 1 and 2 are proof that you can do this at scale now. I don't know if they're the best examples because of the issues they have with regards to permitting, but they're demonstrating that it can be done and there's others doing it. Crusoe is another example.”

“Yeah, yeah, sure. So the project which is facing quite a lot of pushback locally, not just New Mexico local opposition, but also from jurisdictional and regulatory kind of characters there is Oracle's Project Jupiter in New Mexico. So yeah, they're essentially, they're trying to construct a a pipeline to feed the behind-the-meter data center in Donia Ana Province. And essentially, this pipeline hasn't been built, and the route that they are planning to— that they want to build on hasn't been approved. They haven't got an approved route. They have got a secondary approved route, Essentially, all of the ways to get gas to the site don't seem very feasible. There's a pipeline issue which they're trying to make happen, but every FERC filing that I see that comes through on the docket, they haven't, they haven't made much progress or any progress at all because essentially it's defaulted to a type of regulatory process which essentially there is no precedent for it being done sooner than essentially 2 years. And given the local opposition as well, that this is very unlikely that it'll be sped up. There are other options to getting gas to the site, like you could ship truck CNG and LNG to to a facility, as, as we point out. However, this at scale, you know, getting to— that's, I think, the only proven, um, like, delivery of, uh, CNG at scale is kind of around 200 megawatts, or like max 200 megawatts. I think maybe even that's optimistic. And there aren't many, there aren't many manufacturers, there aren't many producers in the local area, and there are not, not enough trucks to facilitate that. So essentially, yeah, there's not only a pipeline which doesn't exist and has no viable route of existing, but getting gas to the site via CNG and LNG is even harder. Um, it's just not, not viable. So yeah, no, it's, it just, I think it brought it like, it, it's, um, it demonstrates kind of an interesting, uh, bottleneck which could occur with more behind-the-meter facilities is like building gas pipeline, um, uh, infrastructure, uh, especially in territories, um, like New Mexico that's quite not, not, not, not extremely, it's not extremely friendly area towards that kind of infrastructure. Yeah, I would, that's how we basically look through the filings. We look through all of the filings. You know, we look, we already had it on our radar because they had issues with the initial turbines that they were trying to use at the site. So they ended up switching to Bloom Fuel Cells. And through kind of that research, I spotted this timeline and I was just like, this doesn't make sense. Why are there so many local, all these pushback, all these local comments? And then yeah, FERC and the other regulators are still thinking about it.”

“In terms of competing with each other, well, I mean, I think when we take a look at kind of the capacity race, you're going to notice that it's mostly driven by the respective AI labs. So most of the time, like an AWS capacity buildout will be kind of fueled by Anthropic, right? Or Meta will typically be for MSL. Microsoft typically OpenAI as well. We know that there's some Anthropic deals going on there as well, but By the model. Um, in any case, these are going to be, in any case, these are going to be, yeah, like the, the drivers of the data center buildout right now. Right. Because they're the ones who have the capital able to actually build out the data center capacity that's needed for the AI labs. They're the ones who have the kind of investment grade, like investment grade, um, financing able to start these projects and get them going forward. Most projects struggle with, you know, actually getting the financing, like if you're a NeoCloud, which if you're listening to this later on, maybe you already have read the other article, but if you're a NeoCloud, it's a bit hard to get the financing to start and go ahead and like get a data center and go ahead and buy the GPUs and expend the capital for all of this, like build out, right? Because it's going to come out to probably billions of dollars. So in essence, you're mainly left with hyperscalers driving the AI build out. Out or it's going to be very, very well-funded or well-capitalized neoclouds. Like if we look at any of the kind of Blackstone-backed guys, any of the KKR-backed guys, they've got plenty of money coming in from them to support their own buildout. But in any case, I forgot your second question actually. So if you could say that again.”

“In terms of competing with each other, well, I mean, I think when we take a look at kind of the capacity race, you're going to notice that it's mostly driven by the respective AI labs. So most of the time, like an AWS capacity buildout will be kind of fueled by Anthropic, right? Or Meta will typically be for MSL. Microsoft typically OpenAI as well. We know that there's some Anthropic deals going on there as well, but By the model. Um, in any case, these are going to be, in any case, these are going to be, yeah, like the, the drivers of the data center buildout right now. Right. Because they're the ones who have the capital able to actually build out the data center capacity that's needed for the AI labs. They're the ones who have the kind of investment grade, like investment grade, um, financing able to start these projects and get them going forward. Most projects struggle with, you know, actually getting the financing, like if you're a NeoCloud, which if you're listening to this later on, maybe you already have read the other article, but if you're a NeoCloud, it's a bit hard to get the financing to start and go ahead and like get a data center and go ahead and buy the GPUs and expend the capital for all of this, like build out, right? Because it's going to come out to probably billions of dollars. So in essence, you're mainly left with hyperscalers driving the AI build out. Out or it's going to be very, very well-funded or well-capitalized neoclouds. Like if we look at any of the kind of Blackstone-backed guys, any of the KKR-backed guys, they've got plenty of money coming in from them to support their own buildout. But in any case, I forgot your second question actually. So if you could say that again.”

“Yeah, I mean, like, I think it was always to be expected. Um, last year the narrative was behind the mirror is not possible because there's these 3 manufacturers that are so slow. And, you know, everyone was sort of throwing that chart, uh, the famous chart of like, you know, how much gas orders that were in like the early 2000s where you had this massive sort of, you know, wave. And everyone was saying GEV and Siemens are so scarred of this era, uh, because they invested massively in capacity. And so they're gonna be very conservative. Right? But you know, the, that, that leaves a market opportunity. And I, I want one question we like to ask management teams at these power companies is, you know, how AI-pilled are you essentially? You know, how much do you believe, how much risk are you willing to take? There's also a function of how easy is it for you to take risk in the sense that, you know, what's your economics on building more capacity and new factories? And there's a bunch of companies that, you know, score very well. The one that we've been flagging for a while. We especially doubled down like at the end of '25 when we did our big deep dive with Blue Energy, because, you know, I think clearly the management team is very AI-pilled. I think they have economics that sort of enable them to build capacity faster than others. And the behind-the-meter conversation is really interesting because like, I think it's like, you know, company by company, they're sort of adjusting their mindset. As you adapt to the new reality. And also sort of solution by solution, you keep adjusting your expectation just based on the constraint. So what I'm saying is that initially everyone was like, okay, this is going to be bridge power. There's going to be a matter of like 1, 2 years, I run off-grid and then I'm going to have my grid come in and maybe it's going to be backup. And so you would only consider like systems that are good at backup. But that's the ultimate disadvantage for something like Bloom is that, you know, it's like not very good at backup. You have to be, you run extremely hot, I think it's 15,000 10°C, takes like 2 days, as far as I know, to go from 0 to 100. So for backup, it's, you know, really not a good system. But hey, if you have no other alternative, then maybe, you know, maybe it's that maybe you just have to go for it, right? Like, if your option is, your other option is that I'm not going to have power, then, you know, you're screwed and you're not going to be competitive in the marketplace. And, you know, other issues are like power costs. And I think when you look at the recent like SpaceX deals, I think it's pretty clear that sort of the revenue per megawatt that folks are making on the cloud side, also on the lab side, on the model side, like revenue per megawatt go up at every layer, I guess. And as that increases, it means that power costs are increasingly irrelevant. And power costs for solutions like Bloom are not that expensive anyways, with regards to, you know, what we have in Europe, for example, with the grid.”

“And the other thing is like also estimating the bit of materials I think has proven quite complicated for many of these vendors, especially as labor rates are going to the moon and you thought you were gonna have local labor, but hey, actually everyone in Texas is, you know, already occupied building data centers, so you have to call people from Denver or, you know, from Ohio and obviously it's much more expensive. And this means that this also favors solutions that are fast to install, fast to deploy because your BOM is sort of more predictable, your timelines are more predictable and both on the labor side and on the sort of full deployment side. So, you know, I think again, like analyzing timelines and ease of deployment is key and I think some solutions score extremely well. And I also think we're gonna keep seeing more and more new entrants, uh, cuz you know, everyone looks for capacity earlier and the standards are sort of dropping, right? The, the, the, there's a few new companies that entered recently that are basically coming from the automotive industry, not themselves, but like their source engines from the automotive industry. And then when you think of this, you're like, wow, automotive, like, you know, we're talking about like 100 gigawatt plus per year of production. And that's an industry everyone knows runs at very low utilization rates, you know, 50%, 40%. These factories are not doing too well. So the capacity and the incentive that these guys have to sell their engines to the data center market is also tremendously high. So, you know, that unlocks gigantic capacity.”

“Um, yeah, go ahead. So yeah, this is, uh, grid connected. Um, so the, the, the point of this, obviously like nameplate additions are overall growing and we're talking about, you know, 60 gigawatts, 50 gigawatts per year. So it's a lot. But as everyone knows, you know, the 600, uh, gigawatt of solar is not a true gigawatt for the grid because, you know, It's intermittent and only turns on at certain hours and so on and so forth, right? Pretty simple. So, you know, depends on the area. As you see, an ELCC value is adjusted for the actual capacity value it brings to the grid. Depending on the area, it can be like 10% of it, 20% of it, sometimes below 10, sometimes over 20. Depends on, you know, it's all sort of very complex system level calculation. The problem is that again, like the grid is also sort of facing a lot of these transmission issues. It's intrinsically slow. There's giant interconnection queues. So I think the question, the real question for solar in batteries is can you do it behind the meter? Like, you know, can people build these things on site? I, I, I think it's gonna happen and we're gonna have much more content on that published on our website. So stay tuned. Obviously the high-level challenge is that the, the land required to build like massive amount of solar is just tremendous. So logistically these projects are fairly complex. You know, if you're talking about, hey, I need 20,000 acres just to throw solar panels at it, it's gonna be for a gigawatt data center or something like that, then, you know, Buying all that land can get complicated. Sometimes what happens, you know, during the course of this sort of land buyout is that the landholders sort of realize, hey, this guy wants to build a, you know, $100 billion project. Maybe I'm going to sell my land 10x more expensive than what he thought I was going to get, right? So that sort of stuff happens as well. So anyways, the complexity of building these massive projects is fairly elevated, which removes the a time to power angle to some extent for solar and batteries. But as folks sort of get bigger and bigger and start planning, you know, multiple years ahead, I think there are, there's already a lot of big projects like this underway. And the other issues, obviously SMRs and nuclear, you know, no, you know, I'm not going to reinvent the wheel. This is slow stuff, right? Everyone knows building nukes takes time. So again, not a time to power. So right now you have a lot of non-binding LOIs in the market or non-binding deals that are contingent on like milestones and execution and getting the regulatory approvals and so on and so forth. I think in the 2030s, we're gonna see a gigantic diversification of energy sources to power data centers. I think for the 2020s, we're gonna be very much in the gas world.”

“Um, yeah, go ahead. So yeah, this is, uh, grid connected. Um, so the, the, the point of this, obviously like nameplate additions are overall growing and we're talking about, you know, 60 gigawatts, 50 gigawatts per year. So it's a lot. But as everyone knows, you know, the 600, uh, gigawatt of solar is not a true gigawatt for the grid because, you know, It's intermittent and only turns on at certain hours and so on and so forth, right? Pretty simple. So, you know, depends on the area. As you see, an ELCC value is adjusted for the actual capacity value it brings to the grid. Depending on the area, it can be like 10% of it, 20% of it, sometimes below 10, sometimes over 20. Depends on, you know, it's all sort of very complex system level calculation. The problem is that again, like the grid is also sort of facing a lot of these transmission issues. It's intrinsically slow. There's giant interconnection queues. So I think the question, the real question for solar in batteries is can you do it behind the meter? Like, you know, can people build these things on site? I, I, I think it's gonna happen and we're gonna have much more content on that published on our website. So stay tuned. Obviously the high-level challenge is that the, the land required to build like massive amount of solar is just tremendous. So logistically these projects are fairly complex. You know, if you're talking about, hey, I need 20,000 acres just to throw solar panels at it, it's gonna be for a gigawatt data center or something like that, then, you know, Buying all that land can get complicated. Sometimes what happens, you know, during the course of this sort of land buyout is that the landholders sort of realize, hey, this guy wants to build a, you know, $100 billion project. Maybe I'm going to sell my land 10x more expensive than what he thought I was going to get, right? So that sort of stuff happens as well. So anyways, the complexity of building these massive projects is fairly elevated, which removes the a time to power angle to some extent for solar and batteries. But as folks sort of get bigger and bigger and start planning, you know, multiple years ahead, I think there are, there's already a lot of big projects like this underway. And the other issues, obviously SMRs and nuclear, you know, no, you know, I'm not going to reinvent the wheel. This is slow stuff, right? Everyone knows building nukes takes time. So again, not a time to power. So right now you have a lot of non-binding LOIs in the market or non-binding deals that are contingent on like milestones and execution and getting the regulatory approvals and so on and so forth. I think in the 2030s, we're gonna see a gigantic diversification of energy sources to power data centers. I think for the 2020s, we're gonna be very much in the gas world.”

“No, I was just going to make a joke about Jeremy's decision making nonprofit good with Pope. That was— yeah, no. Yeah. So interesting. Yeah, it's going to be interesting to see how much of these— yeah, how much is going to be built behind the meter from the main AI labs as well. Yeah, so yeah, seeing— and I think also they're kind of very, very, you know, behind the meter is— there's one— that's one term. There are kind of various different kind of definitions of kind of moving energy, the moving kind of energy on site and not taking all of it from the grid. Um, I mean, you can co-locate your energy, you can, um, have a net metering kind of solution, which is, um, yeah, more of a kind of interactive interactive relationship with the grid. You can have fully islanded off-grid, you know, energy, energy supply for your data center. But yeah, I think that's a, that's going to be the kind of thing that, that, that is going to be interesting to observe in the future as the more of them are built, seeing what's preferred. I mean, which state, fully islanded, bit of grid connection, maybe that might hedge some, you know, interconnection in the future. Also then there's a, you know, what happens to these assets, these turbines? Once, you know, once maybe they do connect to the grid, they might end up being part of the grid themselves. So like a lot of these turbines could be used as, can be used as gas peakers and help solve, help with solving the actual like lack of power on the grid itself. I mean, obviously there's transmission to sort out, but yeah, there's, so I think it's a, it's a, it's, it's, it's, it's more symbiotic than, than, than is kind of portrayed by a lot of other outlets. I think it's more of, yeah, it's not, I don't think it's one or the other. It will be, well, I don't think it'll be one or the other forever.”

“I was just gonna, I was just gonna say that we might see kind of as into, if you can moving on, more kind of carbon, I'm starting to see more kind of carbon capture utilization storage facilities being kind of co-located with BTM sites and with data centers as well. I think that's something interesting to observe as well going forward.”

“Yeah. Yeah. Luckily Jeremy was able to explain the whole article in 3 minutes earlier, so I can kind of expand on what he said. It's a bit too efficient with this communication. Um, basically when you have a model like ours, um, we do everything kind of in a bottoms-up way, right? Where we have like each individual data center is like, for example, a row in an Excel file for us. Right. And so you can kind of get the forecast for each individual building, the tenant, if the end user is an OpenAI or an Anthropic, um, you already have that information for that building in specific, then you can go ahead and find the timeline for the building. Right? A lot of people don't, a lot of people don't have the expertise to judge timelines on these things, right? When we go ahead and do our data center research, I like to joke that Jeremy and I have probably seen, probably seen over 10,000 to 20,000 satellite images, right? So not a lot of people can say that they understand like what that single gray pixel on a brown floor means. It's like, oh, that's actually the concrete padding coming in. Um, that's going to be a good sign. Hopefully we see vertical construction, which is the darker gray pixel in the corner over there that most people wouldn't get. So the satellite images are super useful for us, but like, that's just one that's really fun to talk about. The other stuff is a bit more boring, which is where we go through every kind of permit portal for every county, city, state, uh, country in the world, right? Um, we like to talk about how much we use Claude, as you see in the article, $170K in a week. Um, the Claude part, everybody loves the Claude part anyway. Um, so we got, you know, our lovely charts on the kind of front of the page where it's, hey, $170,000 spent in a week, right? We're token maxing, you know, we're token mogging Meta or whatever you want to call it. More tokens per employee than, you know, yeah, Meta is one of the— but in any case, we tend to use Claude, and we've actually built out basically an entire harness and workflow for our agents. We love our agents that go ahead and scan and scrape every permit portal basically around the world down to state, municipality, and city and county level. And with this, we get a ton of public filing information that allows us to basically estimate timelines with that. And without going into too much more detail, that's one part of basically three in our methodology. And for the other three, you can also buy the model to figure that out. Um, so, so great.”

“Well, I, I, I would disagree. I think there's a pretty substantial amount of cancellations. Just what we keep saying is that these cancellations are, are early stage projects, right? Okay. I'm not sure if you have this one on the screen, but like there's this map of the US where we show like the large load requests. Right, which is like slightly different, but kind of goes to the same point. As of today, you have over a terawatt of data center load that is requested by operators in the US alone, right? This chart is as of 6 months ago. It's more than doubled now. So we have over a terawatt. Obviously, you know, the whole US system right now, peak load, 750 gigawatts. You're not gonna, you know, double it right now in sort of, you know, just a couple years. It's just not possible. It's not a reality. So obviously this is fake. There's a lot of early stage projects and it makes sense essentially, you know, I think the data center market started booming towards the end of 2023. That's when early folks started to realize, hey, there's going to be a big constraint on this. The first big deal started to get signed at that moment. And, you know, since then there's been like this massive search for power. And that means that, you know, everyone was trying to find where is power available. And that leads to some behaviors where, because everyone is doing it, then if you want to be successful, you also yourself have to be aggressive and plan multiple options. And that is actually nothing new. Like hyperscalers have always had multiple options when evaluating projects. It's even more critical in times of constraint. And so you would see sometimes for one final investment decision, see hyperscalers maybe have 10 different options, right? And so that would be 10 projects where they sort of talk to the local counties, you know, they talk to like the utilities and so on and so forth. And these are not all realistic and they're sort of just testing the field and be like, okay, which county can enable me to build my large-scale project? What do I have to workforce or do I have the supply chains? You know, where do I, where can I build it? So anyways, there's an oversupply of very early stage projects and, you know, our point and that was sort of Rake's brilliant phrasing, the cloud coded projects is that I just love this. This is amazing. That's the kind of stuff that AI is, I guess, not very good at filtering, right? So when you have humans in the loop, yes, it's pretty obvious that some of these announcements are way too aggressive. We gave a bunch of examples examples in the article of, you know, folks where, you know, they announce a 10-gigawatt project as they say, uh, the first tranche of 500 meg is gonna be available next year. And then you click on the website, you see contact us, nothing more. And then you start digging into whatever permits and whatnot, you see nothing. Uh, so at some point you're just like, okay, these people probably have a lot of netted Texas or something like that, but they don't yet have a real project, right? Uh, any human judgment would sort of filter that, but I guess for AI it's still pretty hard these days. Uh, so if you wanna build a forecast industry, you have to base it on realistic forecasts, you know, or semi-analysis. Extensive triangulation has always been our playbook, not just data centers, but also chips, because obviously this has downstream implications on Nvidia, upstream on the AI labs, their revenue, all of that stuff connects to each other. That's, you know, the semi-analysis flywheel covering every single one of these industries and building a cohesive view. You know, others don't do that and easily struggle to find the ability to put the site up.”

“Yeah. Yeah. Luckily Jeremy was able to explain the whole article in 3 minutes earlier, so I can kind of expand on what he said. It's a bit too efficient with this communication. Um, basically when you have a model like ours, um, we do everything kind of in a bottoms-up way, right? Where we have like each individual data center is like, for example, a row in an Excel file for us. Right. And so you can kind of get the forecast for each individual building, the tenant, if the end user is an OpenAI or an Anthropic, um, you already have that information for that building in specific, then you can go ahead and find the timeline for the building. Right? A lot of people don't, a lot of people don't have the expertise to judge timelines on these things, right? When we go ahead and do our data center research, I like to joke that Jeremy and I have probably seen, probably seen over 10,000 to 20,000 satellite images, right? So not a lot of people can say that they understand like what that single gray pixel on a brown floor means. It's like, oh, that's actually the concrete padding coming in. Um, that's going to be a good sign. Hopefully we see vertical construction, which is the darker gray pixel in the corner over there that most people wouldn't get. So the satellite images are super useful for us, but like, that's just one that's really fun to talk about. The other stuff is a bit more boring, which is where we go through every kind of permit portal for every county, city, state, uh, country in the world, right? Um, we like to talk about how much we use Claude, as you see in the article, $170K in a week. Um, the Claude part, everybody loves the Claude part anyway. Um, so we got, you know, our lovely charts on the kind of front of the page where it's, hey, $170,000 spent in a week, right? We're token maxing, you know, we're token mogging Meta or whatever you want to call it. More tokens per employee than, you know, yeah, Meta is one of the— but in any case, we tend to use Claude, and we've actually built out basically an entire harness and workflow for our agents. We love our agents that go ahead and scan and scrape every permit portal basically around the world down to state, municipality, and city and county level. And with this, we get a ton of public filing information that allows us to basically estimate timelines with that. And without going into too much more detail, that's one part of basically three in our methodology. And for the other three, you can also buy the model to figure that out. Um, so, so great.”

“Was it like 8.5 million man hours? Uh, maybe 2029 at the earliest. 20— 2030. 2030... Let me think about how many permanent magnets are being manufactured right now. That's a tough one.”