Nick Kruse 0:00
Yeah, I think if you looked from the development or the preference side, everyone would say, Hey, give me the world. I would just plop all my datacenters right on top of downtown Dallas, like all of them would go there. We'll just keep stacking them up forever. I Nick,
EB 0:32
he's a chemical engineer turned energy entrepreneur, and he sits at the nexus of power and compute. Nick, welcome to baxtels data center.
Nick Kruse 0:41
Meet me room podcast. Great. Great to be here. Eric.
EB 0:45
So before we get to before we dig in, I want to learn about you in 60 seconds. I'll ask you a series of quick questions, and please provide the first answer that comes to mind, Ready when you are all right. So do you prefer late nights or early mornings?
Nick Kruse 0:59
Late nights, definitely. Total night owl,
EB 1:01
awesome, sweet or salty,
Nick Kruse 1:03
100%sweet,
EB 1:04
introvert or extrovert.
Nick Kruse 1:06
I'm a strong introvert, but I've been working on my extra extrovert skills pretty, pretty hard.
EB 1:13
So I like it. Favorite food,
Nick Kruse 1:17
ooh, that's a tough one, tough one. I'm probably gonna go with buffalo chicken,
EB 1:23
favorite sport.
Nick Kruse 1:24
I'm not, actually not a huge sport person, but I will sit down and watch the Super Bowl every year. So I'll call that the highest
EB 1:31
with my fiance, I say, you know that she likes sport ball like or I'm gonna go watch sport ball now, because she's not a sports fan at all. And so football to her, yeah, I'd sympathize there. Okay, so what was your first job? And this is before the industry, like high school ish
Nick Kruse 1:48
or Yeah, yeah. So my first job was at a place called graffitis pizzeria, and actually spun and tossed pizzas for, you know, probably about six or seven years, basically all throughout college, until I got what I'll call my first real job, which still was an industry, but I was doing some immuno immunological research for the CDC.
EB 2:07
Oh, wow, cool. That that there's a bunch of questions right there that we um
Nick Kruse 2:15
fiction or not fiction, um fiction, usually sci fi, but I do like some fantasy as well, too. Secret talent, I wouldn't say it's a secret, but I'm a huge musician in my personal time, so I play guitar, I play drums. Music's been in the family for a long time, so I did pick up that gene, luckily.
EB 2:36
All right, and so let's move into some work questions, or work ish, questions, meeting or email. What do you prefer?
Nick Kruse 2:43
Ooh, if it's short email, if it's a long discussion, meeting and where the line gets blurred, there is where a lot of corporate frustration can come from.
EB 2:53
Right? What is your pet peeve during conference calls?
Nick Kruse 2:56
Definitely, definitely people that struggle with the tech side. I if I can never hear, can you see my screen again? I would be a very happy
EB 3:05
Yes. Or they share the wrong screen, or whatever, exactly. What is the marketing buzzword you'd ban for life? If you had,
Nick Kruse 3:16
I would say synergies. Like, if I could never hear about synergies again, that'd be great,
EB 3:23
good and then cabinet or rack cabinet, cabinet, yep, diesel Gen set or gas turbine.
Nick Kruse 3:32
Ooh, tough question. I would say it's actually between gas Gen set and gas turbine, and we'll get into that. Why later, and I'll say it's 5050, split. I love them both.
EB 3:41
Yeah, turbine or turbine? Is there a app? Is there a
Nick Kruse 3:46
I've heard it both ways. I'm not sure if there's an officially correct way, or maybe it's one of those British, English, American English type things,
EB 3:53
Potato, potato. Yeah, gas, hedge or spot market.
Nick Kruse 3:59
Ooh, I prefer taking a little risk. So I'll say, spot market.
EB 4:05
You're an advocate of prices going down over time. Perhaps I am.
Nick Kruse 4:10
I think the United States is sitting atop a ton of gas, and I think we're going to be well supplied into the future
EB 4:19
gas pipeline or on site? LNG,
Nick Kruse 4:22
I'll say gas pipeline, but if you want ultimate reliability, you do
Nick Kruse 4:26
both. Yeah, true.
EB 4:30
So where do you get your industry news from, whether in the energy and industry or the datacenter industry?
Nick Kruse 4:36
And I'll say, I'll get it from a couple places. You know, I'll give the more professional answer, which is Bloomberg, but then I'll give the more personal answer, which is Reddit. So, ah, Reddit. Okay, cool. Love me some
EB 4:48
Reddit. Yep, who would play you in the Netflix series, datacenter wars?
Nick Kruse 4:54
What actor? Ooh, that's that's a good question. I'm gonna go with, even though he's you. Passed away not too long ago. Alan Rickman, love me some. Alan Rickman,
EB 5:04
good, good. Okay, so let's move in to the regular part of the podcast. And you know you are in the gas generation business, and we'll get into what GPC does. But what can you describe for those who are unfamiliar, what, what does behind the meter mean? Sure, power industry,
Nick Kruse 5:25
yeah. So you know the meter in this context is your power meter, right? And this is kind of, you know where your line stops and where the utility begins. And when we're talking about behind the meter, what we're talking about is a solution that sits ahead of the utility. So if you were connected to a power plant directly, you would call yourself, hey, I'm behind the meter, because I don't go through utility meter. Would
EB 5:48
behind the meter? Is this? Is this something that's a permanent solution or a bridging solution? You know? How do you think about it? Or is it depend on the client's thoughts on that matter?
Nick Kruse 5:58
Yeah, I'd say there's kind of two tax with that. One is kind of where the industry's at. And then one is, you know, what, what GPC really believes the direction everyone's going to move in? So, you know, I'll kind of start with where the industry is at is, you know, this crunch in power has really forced datacenters and other large loads to be really thoughtful. You know, what are we going to do about power in the near term? But a lot still see this as a problem that's just going to resolve itself the next 345, years. And you know, from the GPC side, being industry veterans on the power side, like we see a fundamental shift here, power is going to become a compelling theme, and getting enough power is going to be a lasting and staying problem. So when you know we're thinking about behind the meter trends, we see this as a long term trend, and don't see this as a bridge. And I think a lot of people will get there eventually, and it's kind of for two reasons. You know, one reason is just the ability to construct new power generation is going to be more constrained than the ability to construct new load. So forever, generation is going to be chasing loads, so long as the demand remains. But kind of, kind of, the other tack to that too is that really, from a fundamentals first standpoint, it makes tons of sense that when you have a large load, and you know, I'll call a large load, anything more than 50 megawatts. When you have a larger than 50 megawatt load, you should really colocate the generation. You know, why go through hundreds of miles of transmission wires? You know, substations up and down, losing, you know, percentages of efficiency at every step, not being able to have that control over the power. And then, you know, distributing that across the nation, as opposed to simply building it right next door. You know, if you're a 500 KW load, yeah, it doesn't really make sense. No one builds 500 KW plants. But if you're a 500 megawatt load, you know, that's a great size for a power plant. It would make, you know, all the sense in the world, from a fundamental standpoint, to have that co located
EB 8:04
true and in the same way that you want to avoid transmission from a, you know, going from the generation to the load, which is the data center. In this case, from a data center perspective, you also want to be close as possible. You want, you want to avoid your transmission. And that's kind of getting, in this case, it's network transmission. And so you want to be as close as possible to kind of network interconnect points. So sometimes where you can generate power, or where the power is generated, where that power plant is, is often kind of in the middle of nowhere, more someplace more rural, or at the edge of town or whatnot, you know. And so I think here's the conundrum, right? Is to kind of where, where do you place this, either behind the meter, or, kind of, how do you, how do you kind of both marry, uh, kind of putting the load by the generation, but also put the the load by where they can interconnect, from a network perspective, which is often Texas, is often in Dallas, right? You know, that's where, you know, the Dallas inform, that sort of thing. That's, this is where a lot of the interconnection goes through in Texas.
Nick Kruse 9:08
Yeah, yeah, that that's kind of the problem statement, right? Is like, you know, given the power is constrained, right? You want to be able to construct your data center, but you still need the connectivity and latency, you know, how do you marry those two worlds together? And you know, we'll see sometimes announcements, oh, there's, you know, a huge datacenter that's going to be out in West Texas, or something like that. And you know that can, that can work for certain workloads, right? You know, your AI training applications where latency is not critical, where you really just need to run GPUs all day long make sure you got power but you know, when we think about datacenters being 2030, year assets with significant investment behind them, you know, you don't want to restrict them to just a single use case. You know, from the power plant and gas side, I'll say those power plants can be a little sneakier than you think, and there's a lot. Closer and more, more in the backyard than a lot of people will realize. So, you know, when you think about the Dallas area, right? You know, if you're willing to go 3040, 50 miles outside of the city, which still good from a latency standpoint, you've got all those fiber trunk lines right on i 35 I 45 you can locate a power plant out there and a data center, and get best of both worlds. You can, you know, have all the power you want. You can have the connectivity you want. And, you know, we're seeing trends where people are starting to look in that direction for just that reason.
EB 10:32
Interesting. Yeah, and then maybe describe a little bit about what your company does, you know, give you a chance to kind of articulate what GPC
Nick Kruse 10:42
does. Yeah, perfect. So, you know, in the 32nd elevator pitch, I'll say, you know, GPC is looking to be the premier behind the meter, independent power producer for datacenters. You know, we're exclusively data center focused. We bring a ton of industry experience on the power generation and CO generation side, really offering power as a service and a turnkey solution for the datacenters. We're here to bring all the pieces together. You know, the data center can focus on its core competency, which is compute and making sure you've got connectivity and serving your end customers, and we can do what we do best, which is building power plants and running and operating those assets. So we're here. We're focused on natural gas applications, because we really view thermal base load power as the, you know, optimum, most reliable solution for the data center. And you know, anything 50 megawatts to a gigawatt plus is right in our wheelhouse.
EB 11:39
So from that perspective, do you, how is your relationship with the data center? Is it? Do you own the land next to it, and you lease out the land? Do you, do you have a PPA only with the end user, and then they purchase the land next door to your generation? Or do you, kind of, do they own the land and then they and then you place your gear on their piece, or maybe all the above, I don't
Nick Kruse 12:01
know. Yeah, we'll definitely say it can be, you know, customer specific. And we have all sorts of structures, you know, kind of tackle the PPA question. So that is our preferred contractual relationship, again, kind of being, you know, an alternative to the utility. We want to look like the utility. So we're a socket in the wall you plug into you get a PPA from us that looks very similar to what you would get from a utility. So we can either PPA directly with the datacenter, you know, owner operator. We can have the PPA be held by the end tenant. It could be a combination of the above. And then, when you're thinking about location, you know, we would look to colocate on the same site. Now, whether we own our small subsection of that site or we lease it from the data center, all of those structures can work. We would really say that's data center preference, and we're flexible within all those options,
EB 12:54
okay? And then when you're doing the generation, does the data center still kind of have similar architectures. Meaning, do they still have maintain either? Do they get a grid backup? Do they also have generators? You know, in other words, what is their kind of typical What are you seeing that their typical backup scenario is? Yeah,
Nick Kruse 13:16
yeah. That's a really good question. Because really, when you're thinking about this, you got to think about reliability at like, a holistic level, right? And, you know, from the GPC standpoint, we can offer reliability that is superior to the grid. From the datacenter side, though they're still going to have their backup generation. There could be options where we could oversize the power plant, build in additional redundancy, that could allow you to eschew that, but we really think, at least, you know, until behind the meter technology, everyone's super comfortable with it. It's got 10 or 15 years of operating history. People are still going to want their backup diesel Gen sets. And that's that's been pretty consistent feedback, and that makes all the sense in the world to us, like, let's have that second layer there to protect
EB 13:58
could the, you know, I'm speaking of a non power guy here. But could the grid, you know, could they get, could that datacenter elect not to have generator backup and just use the grid as backup? And I think that there's liability, not liability, but some things fraught with that, but just as an idea, right? You know, is that a feasible solution? Yeah,
Nick Kruse 14:20
that's definitely a feasible solution. And you could even do all three, right? You could have on site generation. You could have grid as a supplement, grid as a backup, or grid, as you know, something that's 5050 with the on site generation. And then you can have your diesel backups on top of that. And you can even get into interesting thoughts, where it's you know, if the on site generation is 50% and the grid is 50% Well, do you size your diesels for 100% right? Because, like, what are the odds you would lose grid and on site at the same time? Those are completely independent things. So now, from the datacenter standpoint, maybe it makes sense we do diesel backup for 50% of the load, or something. We. Can be a material CapEx saving. So there's a lot of interesting ways to optimize once you think through it through the lens that on site generation can be a complement to the grid, if that's the path you want to
EB 15:12
go right. And then how do you think about I guess, what is your typical size of projects you work on? And in addition, what, what are you doing, turbines or reciprocal engines? You know, what size of these, these units are you typically, is your preference to use?
Nick Kruse 15:31
Yeah, great question. So when we think about just general project size be very data center driven, where we kind of feel the sweet spot for natural gas generation is in that we'll call it 50 megawatt. On the low end, up to a gigawatt, you know, a little below 50 megawatts, we think it's hard to make the economics pencil out. And above a gigawatt, getting a gigawatt of gas is a substantial amount of gas supply. It can be done. But the locations where that can be done start becoming more selective. But we really feel within that 50 megawatt to one gigawatt range, gas as a primary power is an elegant solution. And we've been seeing a lot of projects, particularly in the 200 to 500 megawatt range, that have really been resonating with the equipment that we think is ideal for this. And switching over to your equipment. Question, you know, we like both gas turbines as well as gas engines. We're not technology agnostic. We care very much about the technology, but we are very purpose built, and certain locations tend to favor turbines over engines. Things like elevation or temperature can drive you to one technology or another. So when we're thinking about this, we're really thinking, you know, what is the optimum best solution to deploy? And within that space, we like the arrow derivative gas turbines. These are kind of in the 20 to 50 megawatt range. And then we like the medium speed and high speed reciprocating engines with medium speeds, kind of 10 to 13 megawatts, and the high speeds more like five megawatts.
EB 17:07
You know, I think that anecdotally, I heard that GE Verona, you know, a turbine, you know, manufacturer, you know, one of the three out there, they're, you know, if they don't sell again to traditional utilities, they'll be just fine from the demand from datacenters, right? You know, at least for the foreseeable future. And so this brings up questions around supply chain, you know, What? What? What? What do you see as supply chain issues in this because I it sounds like these turbines or engines are taking longer and longer to receive. So I guess, what are your What are your What are you seeing in the marketplace in terms of supply chain timelines?
Nick Kruse 17:46
Yeah, so I think he hit the nail on the head. Supply chains are definitely becoming challenged, and it's just kind of the reality that it's easier to prop up some data center buildings than it is to build some turbines, and the manufacturing behind that is harder, harder to scale. So we're always going to see generation lag, new demands, so long as the demand is kind of at the level we've been been seeing for the past couple of years. So really, being very thoughtful about supply chain, developing deep relationships with the supplier, which is something that GPC has done and is continuing to do is really critical, because, you know, if you can't get power, you don't have a project. So from our view, really think being thoughtful about, you know, what are the right pieces of equipment to put in, but also what is possible to be put in and and again, this is why we're somewhat favoring some of the more medium sized units, as they tend to have better availability. Lead times are a little bit better. And from a manufacturing standpoint, those processes are easier to scale up as well too. You know, you're kind of, we'll call it 500 megawatt single turbines. You know that those are going to be challenged to expand production capabilities on whereas your more mid sized ones will be ramped up. And the interesting thing about the air derivative turbines, as the name might imply, is that they came from the airline industry, right? So there's kind of a, you know, already developed supply chain that services two industries, both the power Gen as well as the airline industries. So, you know, depending on where the demand curve is for new airliners, you know, you might be able to re requisition some of that supply and pull it over to the data center side. And you know certainly when, when prices start getting pressure, the engines are going to go where, you know, whoever is paying the best,
EB 19:34
right? So in terms of the kind of from order, you know, from a customer order to kind of get standing up, say, 100 megawatt data center, like, what is the timeline? Not data center, but the generation for that piece. Like, how long does that take?
Nick Kruse 19:49
Yeah, yeah. So, you know, depends on technology. Engines tend to be faster. That's one of their advantages. If you were doing 100 megawatts of engines, you could conceivably be online as fast as. 18 or 20 months, right? Which is significantly ahead of a lot of interconnection queues. On the turbine side, you're probably looking more 24 to 28 months, which is still, you know, well ahead of a lot of interconnection queue timeframes. And
EB 20:13
you probably heard, I guess, was it just in the last month, Xai, did they purchase a foreign power plant, and they're shipping it, the turbines back.
Nick Kruse 20:27
A lot of creative solutions like that, right? You know, gas turbines, gas engines, very well established technologies. Lots of installations across the world. You know, a lot of places that have either had changing energy needs, or newer technologies have come out, or, you know, even newer turbines have come out, and now they're phasing out the old ones. So the question is, what do you do with this old equipment, and a lot of that's going to get breathed new life with the power generation crunch that we have, because what's accessible is going to be what's used. So it
EB 21:00
sounds like, when I looked at your website, it looks like you guys do some site selection as well. Is it? Is it, are you guys choosing the site, or is that the data center user, or that your data center customer choosing the site, or a little bit of both? Or do you help from that perspective, I guess, what are? What is your role in that piece? And have more questions on that piece, but yeah,
Nick Kruse 21:21
yeah, no, that's a good area to touch on. And the answer is, it's really both. I would say our preference is generally to go to sites that the datacenters know they want, but that's just because, you know, we're not data center experts, so we don't want to pretend that we know the best connectivity, the best location, and every company has got their own goals on where they want to grow, right? You know, someone might say, Oh, we want more of a Dallas presence. Or someone might say, Oh, we want to go and go to Columbus and Ohio for whatever reason. So, you know, we love to go where the data center wants to go. But we also have sites that we're either partnered with with land developers, or some that we have on our own, that we're happy to share and say, you know, this is a place where gas generation makes sense, certainly on the GPC side, what we're looking at is, you know, where, where can you deploy in a reasonable time frame and be cost effective a gas fired power plant? And, you know, there's a couple lenses you got to look through there. A lot of them around gas supply. A lot of them around air permitting equipments, a little more agnostic on location. But, you know, that's kind of how we think about this.
EB 22:26
And is there, from a site selection perspective, what if, if you could choose a site? What? What site? What is? What are the characteristics that make the best site? From a you know, self Gen perspective, I imagine it's access to gas.
Nick Kruse 22:44
But, yeah, gas is a big one for sure. You know, I'd probably think about it from three lenses, and I'll order them in importance. So the first one is air permitting EPA non attainment zones are, might as well be no go zones for power plants. There's some exceptions to that, the Houston markets, one where you can get offset credits or other ways to work around the non attainment status, but Dallas is not one of those. So you know, kind of first screening criteria is, let's stay out of non attainment if at all possible. And then second screening criteria is, where can you get the gas right? So you want to be on some gas pipelines. And, you know, there's a bit of a danger there for people who aren't familiar with gas supplies, you might go to, you know, some maps. For instance, you could go to the RCS maps in Texas and be like, oh, there's, you know, two gas lines right in the backyard. Great, we've got gas. Well, the answer is, just because there's a gas pipeline doesn't mean there's gas for you to buy at that point. So really understanding the gas supply market is critical. And us at GPC, being 30 plus year veterans, tons of internal gas expertise, knowledge and relationships that we've cultivated, know where there's gas available and where can you can actually build these type of assets. So you know, those kind of cover options number one, or criteria number one and two, which is permitting, and then gas supply, and then kind of the third one is, I'll call it climate and constructability. Definitely. Temperate climates are really nice. You know, Ohio's got a great climate. 110 degree days are always difficult to design around, but it can be done. And that's why that's kind of number three. And I'd say it's a distant third to number one and two. So really, permitting and gas supply are the big lenses you want to think
EB 24:30
about. So it sounds like not too hot, not too cold,
Nick Kruse 24:33
yeah, not too hot, not too cold, just right, right? Although
EB 24:37
a lot of the natural gas are in markets that are too hot. It seems like Texas gets pretty hot, or Oklahoma, I suppose might get pretty hot too. But it sounds like if you go up to Ohio, Pennsylvania, maybe Western Virginia would be more temperate from that perspective.
Nick Kruse 24:53
Or, yeah, that's That's correct. And you know, there's a lot of gas in Texas, there's a lot of gas in Oklahoma, there's a lot of gas in New Mexico. Ago. What's actually interesting is there's a lot of gas in the Northeast as well. The Marcellus gas producing basin is one of the most prolific gas basins. Sits on top of, you know, call it Pennsylvania, West Virginia, kind of that area ish. So all those adjacent areas have, you know, a lot of opportunity for gas supplies well too. But again, just because there's a pipeline doesn't mean there's gas. So it's, you know, very, very location specific.
EB 25:25
What are your thoughts on, from a self generation perspective, on, you know, kind of the Dakotas, or kind of like Wyoming, there's, there's a fair amount of gas up in Wyoming, maybe Colorado, but up in Wyoming and the Dakotas, and even if you go up into Alberta or Saskatchewan, there's, there's gas up there, but then that sounds like it gets too cold. Is that something that you can is it feasible to operate in those, those markets?
Nick Kruse 25:49
Yeah, definitely feasible to operate. There's a lot of gas fired generation out there. You know, from a permitting and gas supply standpoint, those are going to be very favorable markets. You know, what I'd ask is, you know what data centers want to be out there. We see some that have some interest, right? But they tend to lean towards, you know, the big population centers, and that struggles a little bit in the Dakotas, but those are really good areas for gas generation. So, you know, if datacenters want to move out a little bit, that can make a lot of sense, you'll get some of the best power prices you can think of. You'll get some of the most favorable terms on gas supply in those areas. So from an economic standpoint, it can make a ton of sense,
EB 26:27
right? So it sounds like cold can be dealt with. You probably just have to spend a little more CapEx to to insulate the pipes. Is speaking with a non gas guy here. But you know is that the case that you just have to spend a little bit more to operate in extreme temperatures, either hot or cold.
Nick Kruse 26:43
Yeah, that's correct. I mean, cold, cold has its ways to be dealt with turbines, you know, kind of like certain temperatures on the inside or the inlet side. And if it gets too cold, you can run into icing issues or things like that. But those are operational issues that can be designed around. The one that's a little more challenging, at least for turbines, is heat affects air density, and air density affects performance. So you may have actually heard that, you know, some planes can't take off in 110 or 115 degree weather, like they just physically can't, and that's because the air is so hot it gets thinner, you don't get as much power output. Turbines kind of suffer from that problem, which is why, when we think about technology, you know, it could be engines, could be turbines. And in those hot climates, engines are oftentimes a really compelling alternative, because they don't suffer from what we call the high temperature D rate. So they can be a more elegant solution.
EB 27:38
And I want to go back to your second point in site selection, which was talking about access to gas and gas pipelines, and land owners who might say, Hey, I have a pipeline running through my property or adjacent to my property. But I guess, what are kind of the criteria there that you look for? Is it some sort of and you're, again, you're talking to non gas guy, but is there some sort of junction points where you have that's where you tap in, or is it contractually that every, every MCF or in that pipeline is already contracted for? Or maybe it's combination both?
Nick Kruse 28:14
Yeah, that can be the real risk. Because as a data center, you know, you're on 24/7, all the time, five nines, reliability, having interruptions in service is not acceptable, right? Like, that's that's kind of a problem statement. You can't have interruption in service. So when you're talking to a gas pipeline, what you're asking for is firm transport. And all the pipelines in the country have a certain amount of firm transport capacity. A lot of times that's been sold out for a long time, and this stuff gets locked up for a long period. Like, you know, pipelines might sell firm transport in 20 or 30 year contracts. So going to them, they'll say, you know, it's great, you've got demand, right? But all our firm transport was sold 15 years ago, and those contracts don't roll off for another 15 years. So really, what you have to do is go find either pipelines that have available firm transport or the entities on those pipelines that might be willing to sell you that firm transport. Now the devil is in the details there, because a lot of times those entities are the utility or they're the local distribution companies. For instance, you know the companies that you buy natural gas for your home for they're not going to release their firm transport, right? That's what they need to run their business. So really thinking about it from kind of a gas supply lens, and figuring out where that firm capacity is is really critical. And you know, it can vary from pipeline to pipeline. If you've got two pipelines that are five miles apart, one might be sold out, one might not,
EB 29:42
right, so it's a firm transport. It's probably you only understand the capacity for firm transport, or the capacity to add, you know, to contract for more firm firm transport is by talking to the utility, the gas utility, pipeline utility, yep, yep.
Nick Kruse 29:57
And, you know, much like all the other and. Entities in the data center space or a data center adjacent space, they are being inundated by requests, so being able to cut through the noise is really critical there. And that's, you know, something that GPC brings to the table. We speak their language. We know how they think. So we know what we're asking for and what's reasonable, and that allows us to, you know, find the opportunities that exist within the space by being able to navigate that world,
EB 30:25
and then the other side of the coin, perhaps beyond just understanding the capacity within the capacity to off take from that gas pipeline. You know, whether they have capacity for their firm trust to take on an extra customer in firm transport, is it also be able to tap in? So can you a pipeline runs through, they have available capacity there? Can you just add a tap, so to speak, in that area? Or do you have to tap in at a certain junction point? You know how to pipeline operators think about that?
Nick Kruse 31:00
Yeah, that's a good question, and this will be a fun little industry story. So the way you do that is actually called a hot tap, and maybe we can link a YouTube video, because it's a it'll be a two minute video, and it's a wonderful pictorial thing, but effectively, what you do is you weld to the side of the pipe, you put a valve on it, and you drill into the pipe while it's operating and put a new lateral off of it. So, you know, these major transmission lines, when they do new additions, they don't, they don't shut down like gas pipelines are highly reliable, more reliable than the power grid, and typically have up times that, you know, if they have any downtime at all, that that is an extremely rare occurrence. So new taps and things like that are a common, common thing that happen, and they they've got a way to do it online, and the hot tapping process is exactly how they do that. So the good news is, is you generally don't have to go back, you know, dozens of miles to the nearest junction or intersection point. You can usually just tap in wherever you want.
EB 31:59
I hope the guys that are doing the guys or girls that doing the hot taps are paid a lot of money, because that sounds pretty dangerous.
Nick Kruse 32:06
Yeah, a hot tap into an interstate transmission line is not cheap.
EB 32:12
I would just not want to be the one drilling in or kind of somehow like penetrating that that high capacity gas line as it's operated. It just feels like I want to be, like, a mile away from from whatever, whenever that's happening. I don't want to be good
Nick Kruse 32:26
news is, is that they are usually pretty far away, because you can do this remotely, and they have like, little machines that kind of clamp onto the pipe and do it, do it all for them. And so I guess you
EB 32:36
coming back to the site selection piece. What are the areas that are most primed to do that self generation. You're looking at self generation and pairing it with datacenters. Where are you seeing that activity? Or where do you think the activity is going to be going
Nick Kruse 32:51
to Yeah, as we mentioned, really the marriage of gas supply and permitting are kind of the big ones. So when we think, I like to think about this in terms of what are the popular data center markets? And then let's take the subset of that that make a lot of sense for on site generation. So you know, generally, most of Texas is really good for that, outside of the non attainment zones. But Austin's not a non attainment zone, so very strong market there. Big fan of anything that's in the PJM area, which is like Ohio, Pennsylvania, kind of that mid continental leaning East Coast area like, you know, north of Columbus is really strong for that, some of Western Pennsylvania, and then some of the Southwest, and we'll call it near adjacent West, make a lot of sense too. So thinking of like Phoenix and Arizona, you know, Salt Lake in Utah, out in Nevada, whether that's, you know, Las Vegas or Reno. So there's a lot of popular data center markets that are expanding rapidly where this gas solution is really elegant approach to being able to get power faster and more reliable than the grid
EB 33:58
can you for those that aren't familiar, and I would bet that many are not familiar. But can you, can you talk a little bit about what a non attainment zone means, you know, particularly as it relates to self generation, but also just operating a data center within a non attainment zone, because I think those are similar, but different things as well.
Nick Kruse 34:19
Yeah, and, you know, datacenters have to get air permits too. So this is kind of all around air permitting, right? And, you know, backup generation does need an air permit. Now, fortunately, backup generation doesn't run very long, so that usually makes that a pretty easy process. But effectively, the non attainment zones are areas in the country where the EPA at the federal level has said, Hey, for one of the air quality standards that we track, we're not quite there. So there's a standard called the National Ambient Air Quality Standard, or NACS, that tracks a few criteria constituents, and the big one when we think about power generation, is ozone. Typically. So areas that are ozone, non attainment, usually have. Very stiff restrictions on what new generation equipment you can install, and usually these are around things like NOx emissions or carbon monoxide emissions. So when you're thinking of a data center, you know, once you get north of about 40 or 50 megawatts, it's going to be very challenging to build generation in a non attainment zone. So now the datacenters are kind of 200 400 600 or more megawatts, right? That really kind of puts them from a generation perspective. You want to be outside the non attainment zones. And a lot of, most of the non attainment zones are on top of population area areas. And you know, some, some have changed over the years, but really citing outside of those is going to be a strong criteria.
EB 35:44
How do you think you know for, and I know this is a little bit outside your area, but you know for datacenters that are grid connected, how do you think you know being inside a non attainment zone, like, like a Denver market, you know that has own zone restrictions? How would How is that impacted? You know, for example, datacenters have diesel generators. Typically, they fire them up once or twice a month, test them for an hour or so, you know, just to make sure they're tested. But other than that, it's they're grid connected. How do you think that being within that non attainment zone impacts that particular data center?
Nick Kruse 36:19
Yeah, it'll definitely be location specific, because there's actually degrees of non attainment. There's like, serious extreme and they all have different criteria. And what's interesting in some of the things we've been hearing is some of these larger data center campuses where previous, previously non attainment was not really a consideration, because, as you mentioned, right, they run their backup generators for an hour a month. They test them a couple times a year, and that's realistically all they're expected to do. So that's all they permit for. But it's getting to the point where, if you're putting in a gigawatt data center that even the emissions for, you know, call it one day a year, if that's what you permit for, for your backup scenario, could exceed the non attainment amount. So you know, these big campuses, one gigawatt, two gigawatt that are being announced, some of those might not be able to be developed in the non attainment zones, or they're going to have to have radically different designs than they've done before, or do something else for backup power.
EB 37:15
Can you if that particular, you know, say, a one gigawatt data center within a non attainment zone. Could they do on site mitigation to, you know, or carbon capture to offset, perhaps that one day a year emission that they're doing?
Nick Kruse 37:32
Yeah, so there's definitely ways you can do that, because what the regulators will look at is, you know, what do you actually emit? So if you got a plan to not emit things or somehow abate it, then that's an acceptable solution. You know, most of the time your diesel generators, and depends on jurisdiction, and I'm not the expert here, but a lot of times they'll come with emissions equipment on them. So, you know, you can even think about, if you've ever driven a diesel vehicle, right? You got diesel exhaust fluid, which is a little catalytic converter that actually, you know, reduces the NOx emissions. Most diesel generators have something like that too. Some jurisdictions don't require it if you're below a certain amount. Some require it all the time. So, you know, there's always kind of devils in the details of like, what was the base requirement? And if that wasn't there, that's an easy knob you can turn it's like, okay, we'll put a def system on all of our diesel generators. But if you've already done that, then now you're at, well, what do we do? And you got to get either more creative, maybe have some operational restrictions, or, you know, maybe just re evaluate,
EB 38:36
right? Well, let's, let's switch up to, perhaps talking about SB six. And so Texas recently passed Texas Bill six, or SB six, which impacts the data center industry, perhaps that the self generation industry. Can you? Can you provide an overview of what that is? And yeah, at least please to start the conversation. Yeah,
Nick Kruse 38:59
yeah, happy to do that. And I'll asterisk this at the beginning and say I'm not the legal expert here, but we have read through because, you know, this is very relevant to datacenters, very relevant to behind the meter generation. And effectively, state bill six in Texas, and there's other flavors of this bill moving through other legislatures. So this is not just a Texas problem, and it's really the grid waking up to the fact that if people are plopping in gigawatt here, 200 megawatts here, that's tough for the grid to handle. Like the grid is not an infinite resource that handles all loads and just can magically poof power wherever it needs to be. In a lot of ways, people think that it sometimes is that way, and it's like, Oh, if it has capacity, then great. If it doesn't, they'll build it, and then it'll be there. But in practice, grid stability is becoming more and more fragile over time, and you can look at the event that happened in Spain. ERCOT actually had a near miss a couple years ago in the southwest area. As you get more concentrated loads that move around a lot, and the thing people love to take. Talk about AI is how spiky those loads are and how radically they can shift. That's really hard from a grid operations perspective. So state bill six is Texas's attempt to ensure grid reliability, and it's really saying, Hey, if you're a large load, you have to meet certain criteria. And it might be, you know, ramp rate restrictions. So for instance, there's the provision in there that basically says, hey, you know, if you're above 75 megawatts, you either cannot move up more than eight megawatts a minute or down more than 20 megawatts a minute. If you've got a gigawatt data center, that's like point 8% up and 2% down per minute. AI leads move way faster than that, so that's already a problem. Then they also say, it's like, Hey, okay, if you can't meet those, then you can be a controllable load resource, which effectively means the grid can disconnect you at will. So now you're at your gigawatt data center, you're like, Okay, eight megawatts up and 20 down. That doesn't work for us. You know, we need more flexibility than that. Grid will say, Okay, you're a CLR now, which means if things get unstable, your one gigawatt of utility power just went to zero. So from a datacenter perspective, like, how do you plan around that? Right? Because you know you need power, 8760, being cut off arbitrarily is not really something that works for you. So we view state bill six as something that's very strengthening of behind the meter. And again, going back to the fact that when you're these large loads, you should really colocate the generation like it makes sense. You don't want to transmit it hundreds and hundreds of miles. Having it right there allows you to have all the operational flexibility you can need, because now you're faced with the decision, okay, well, you know, if they're going to disconnect me, then I have to plan for that. We don't know how long it's going to be. So maybe our backup generators aren't adequate. We usually only rate those for about two days. We might be a non attainment. So now our backup generators that we thought we were going to run for one hour a year at the most. Now, if the utility disconnects us, maybe we're running them for five or 10 days a year. That's that that could be a challenge. And then, you know, if you were to not take the controllable load resource path, now you have all these operational restrictions. So you know, how do you marry that to the fact the, you know, large GPU clusters are naturally going to be spiky loads. So now you're either looking at load conditioning or all sorts of these other things. And the real take home here is that the challenges that datacenters are putting on the utility are going to become a data center problem to solve, like the utility is the not, is not the one silver bullet that's going to just say we can take all these load variations no matter what happens, we're just going to handle them all like that's not going to be where we go. So now that that becomes a data center problem, wouldn't you just rather solve it completely with your own behind the meter generation that is faster, potentially cheaper and more operationally flexible. So that's really how we see state bill six playing
EB 43:06
out. How are you seeing these datacenters dealing with some of these issues? So it sounds like the grid is pushing these problems to datacenters to solve. So the variation limits that you described that for a very large data center, that seems almost impossible. I don't know what the if it's permitted, they can't variate. Or is it over an hour? It must be a decrement over time, right? You can, you can still drop your load at night versus the day, perhaps, if your meta data center as an example. But how did, how are you seeing datacenters trying to solve that problem. Are they putting in load banks, I know x AI is looking at, you know, or it uses their own, you know, Tesla, you know, power. You know power, you know, packs, if you will, right? You know, using batteries to stabilize that load. Are these the type of solutions you're seeing? Or how do you think datacenters are thinking through those, those issues. Yeah,
Nick Kruse 44:03
there's, there's a couple thoughts there, you know, one, this is a new problem, right? Everyone's kind of innovating at lightning speed here. Before the, you know, AI, boom, right? There were not so much data center load that they even really had to think about this. So this is kind of a new problem to solve. There's a lot of people that are grandfathered in and don't have to face this issue, but new developments definitely going to have to face it. Kind of what x AI is doing with batteries makes a lot of sense in certain scenarios. You know, being kind of what we'll call load conditioners, such that the grid doesn't see these massive fluctuations that's going to have to happen one way or another behind the meter is a solution to that, because you just kind of take the grid out of the picture and you solve it. And you solve it internally. But if you have the grid in the picture, then yeah, batteries or load banks can make a lot of sense. You know. The other thing, and the other reality too, is, is, you know, GPUs don't have to have these spiky loads. Like there is a software and firmware solution to this, where. Right? They're they're dropping compute cycles left and right, and that's causing all this spikiness. You could just have them right through now. That's going to cost you energy. So that's not free. That's probably a lot of money. But then you get into what's the trade off between just burning energy you don't need, what you're going to do anyways, with load banks or batteries really right? Or solving this at the rack level. And, you know, like most complex problems, it's probably going to be all of the above. Like, there's probably going to be some load conditioning you do before you hit your behind the meter solution or the grid. And there's probably going to be some rack level software and firmware changes that fix this issue. And the videos kind of put out some guidance that they're working on stuff like this, like this is clearly a problem that isn't gonna, you know, be solved purely with equipment. So I think a solution that takes all those into account is where we're gonna land,
EB 45:52
I see and then, so if SB six is driving datacenters to do some their own generation for large, anything above 75 megawatts or the be able to be killed at any time the kill the kill switch, Mary, is this driving datacenters outside of nonattainment zone areas? Do you think it? You know, are they going to be are you going to see a cluster that are outs just outside that Dallas non attainment zone? Or are developers still looking inside non attainment zones in light of SB six?
Nick Kruse 46:23
Yeah, I think if you look from the development or the preference side, everyone would say, Hey, give me the world. I would just plop all my datacenters right on top of downtown Dallas, like all of them would go there, and we'll just keep stacking them up forever and ever. That's not reality, right? And power is going to be the constraint that pushes people out. The reality is building new high voltage transmission lines into these population centers is going to be nigh impossible, like you'll be able to do some of it, but you're not just going to start throwing 500 KV lines and just going into downtown Dallas. So by necessity, people are going to be forced out. And the moment they're pushed out a little bit, and you get outside those non attainment zones, like the world kind of opens up. And you know, when you think about connectivity, you know, it's a little bit of a lemming situation, right? That your first big anchor goes right? Your Microsoft, your Amazon, they build one out there. And then suddenly, everyone's like, All right, let's go in. Everyone's going in. This is the new hot spot. So I think we're going to see that a lot, you know, using Dallas as the example, right? Cities like waco or correspondent, which are just the next city south, right? Only 5060, miles, still close, from a latency perspective, I think those places are going to see a huge boom in activity. When
EB 47:39
you talk about Lebanon's, I think about red oak, Texas, where you had the Google data center down there, and then there's tons of developers just to the east of red oak. You know that that kind of came in right after, I don't know if that's inside or outside the non attainment
Nick Kruse 47:54
zone. It's still inside, but just barely. And it is really unfortunate because, because it's like, probably 20 miles from the edge of the non attainment so the non attainment zones are usually by county, and it's still in one of the non attainment
EB 48:07
counties, so red oak to Wilmer is all within that, just barely within the non attainment
Nick Kruse 48:12
Yeah, yeah. And I have to double check that just to make sure, but if I'm thinking about the place that you're talking about, I'm pretty sure it is. Here's
EB 48:18
an interesting question, and it might be something I don't know, but what's the biggest myth about natural gas you keep hearing and or even self generation, right? You know, what is there a myth that you keep hearing and what's your here's your chance to dispel
Nick Kruse 48:33
that myth? Yeah, I'd say that probably the biggest myth that I've heard about natural gas generation is that it's illegal for you to go do, and you have to do it through the utility that is almost universally not true in the United States. I won't opine on other countries, but generally, in most areas in the country, you can operate your own independent power producer. There gets, you know, to be some hair on it if you want to use utility lines, then things get really sticky. But for being a co located behind the meter solution in almost all jurisdictions, that's something you can do. So there's a lot of people just think, oh, you can't build a power plant. That's, that's, you know, regulated utility activity. We can't do that. That's, that's just generally not
EB 49:17
makes sense. Okay? So this, I guess, brings us to the next part of our data. Of our podcast section is like the data center war story. So think of a time in your time, or maybe one or two stories in your career that you can recount, a memorial, a memorable situation in your career, energy or datacenter, anything that you can think of or want to share?
Nick Kruse 49:44
Yeah, sure. So I'll give one kind of we'll call it project construction, and the advantages of experience, as we're going to call the theme on this one, which is, so I was on a team that was executing a project, and kind of came in at the tail end of it. So how? Had been there for most of it, and we were getting to the commissioning phase. And you know, a lot of times you do what is called a hydro test, which just means you fill all the piping and the tanks and the vessels with water, and you make sure there's no leaks. Right, smart thing to do in the chemical industry, you don't want to have. So we came in, the guys had already done that, and we were actually getting ready to do kind of final commissioning and get this unit online. And we had this one tank, and, you know, we got there, and we're like, wow, there's, there's still water in this tank from the hydro test. And so we go and drain out the tank, and then we notice, well, there's like, a leak over here, and where's that from, and there's a leak over there, where's that from. And everyone kind of starts looking at each other, and it's not a good feeling, you can, you can feel the quietness in the room. And, you know, we came to find out that they filled the tank with what was effectively swamp water. Call it fire water, but this was, you know, not clean water. And that day, I learned about a little something called microbial induced corrosion, or mic, which is actually where bacteria, and there's lots of them, will come in, and they'll eat metal, like you would think, bacteria doesn't eat metal, like that's not a thing, right? Apparently, it is a thing. And this very large stainless steel tank, all the welds got eaten out. It had more or less holes in it. They basically turned it into Swiss cheese. That was a very expensive mistake to fix. And kind of where I'll take this back to, you know, where the data center audience may care about this is that there's a lot of lessons in doing behind the meter generation or self generation. You know, some people are thinking, hey, we'll just build a team and go do this ourselves. You know, that's definitely possible. But, you know, I always like to think about where is everyone's core competency, like, what are you good at? Data Center? Guys are good at datacenters. US power people are good at power generation. And we've got lots and lots of experience building power plants such that we've already learned those lessons. Like, tuition is expensive in physical assets, and should always ask yourself, it's better to have someone pay for that learning experience. And you know our team has has learned our lessons, and you know we're here to execute flawlessly.
EB 52:09
And that story that makes sense, and that story, even though I've always considered myself to have a steel stomach, it's turning me into a germaphobe,
Nick Kruse 52:18
yeah, yeah, yeah.
EB 52:20
Microbes eating through
Nick Kruse 52:22
metal I've never heard, yeah, stainless steel specifically. So, you know, this wasn't even carbon steel. It was crazy. And you know, at the time I was a lot, lot younger in my career, that was news to me, like I had never heard of that
EB 52:33
before. Cool. So I guess that brings us to the next section. Is topics. So if you can share something that you you find useful or that maybe others might not use or know about. Is there anything that you would recommend?
Nick Kruse 52:52
Yeah, I'll say I love my 3d printer. It's a hobby more than anything. But, you know, it's pretty cool to be able to create stuff in your own house. So Big Three printer fan, and I have a lot of fun with that. I've got the bamboo lab. I think it's the P 1s model. I have to double check that, but yeah, love my 3d printer. You get to print little gadgets. If you ever have to fix anything, you can just make your own parts.
EB 53:16
What are, what are the types of, what are examples of things that
Nick Kruse 53:21
Yeah. So I'm kind of a bit of a computer nerd, right? And I usually build my own computers, and one day, like, just, kind of like a clip in there broke off. And, you know, this is a piece of hardware that no one sells it anymore, so it's like, how do you get a replacement clip? And you know, it's a 20 cent piece of plastic, but good luck going on Amazon and finding something that's going to work. So I, I have three printed a new one, and for you know, 50 cents of plastic had a solution to my problem. And you know that that was a really great situation, so paid for itself in some ways. Well, not really, but closer.
EB 53:56
It's little things that justify, you know that the big toys,
Speaker 2 54:00
exactly. I removed value there. Yeah,
EB 54:04
exactly, exactly. I mean, it's just fun to be able to do that. For example, I have this is not as fun, but it's, we have a trash can that you roll out to the curb that the trash truck comes and, like, it's all mechanical. Now, you don't have trash guys, right? They don't come and throw the trash into the truck. They come. And I think there's probably every community now, they come in, you know, put it upside down and shake it, you know, these mechanical arms and put it back. Well, it's broken off, like, what? There's two clips that hold on that the lid, and when the clips broke, and so it's been broken for the last couple of months. And I've been like, do I call the trash company to replace that? But I feel like, if I had, maybe I'm justifying myself getting a printer to replace this, to save, you know, adding this trash can, get in a whole new trash can. But I only probably need a little piece that I can the kind. The clip that holds it on, that, that trash lid on, and I, yeah, sounds like I could probably make it,
Speaker 2 55:06
yeah, yeah, no, you might get some good value there. Okay, cool.
EB 55:12
And then, where can we, where can listeners find out more about you or contact you? You know, where can we reach you?
Nick Kruse 55:20
Yeah, so our website, which is www, dot GPC infrastructure.com, and the GPC stands for gas powered compute, can reach us there. We've got a contact form, and then, of course, more than welcome, reach out to me on LinkedIn. I'm Nick crews on LinkedIn. You'll see me as one of the GPC members. You know, always welcome customer conversations, or anyone that just wants to learn more about behind the meter generation and what we do
EB 55:44
Nate, we'll put your those links into the show notes. Perfect. All right. Well, thank you for joining us.
Nick Kruse 55:51
Awesome. Thanks, Eric. It's been a great conversation. Had a good time.