Data Centers Explained: A Deep Dive with a Developer

♪ Listen to the heart beat ♪ Chevy is called the heartbeat of America for a reason, with SUVs made to move with your rhythm. The versatile equinox tackles your entire day. The spacious traverse fits your crew and your whole weekend, and tracks bring style with value you can count on. All infused with tech that has your back, so your drive always hits the right cord. Chevrolet. Together, let's drive. Croger is your destination for a Labor Day weekend savings. Find quality you can trust at a value you will love with a wide assortment of products for more exclusive brands. Fire up the grill with cook-out classics like burgers and breads, and don't forget delicious produce like fresh juicy melons, or beat the heat with frozen treats while chilling poolside. However, you're planning to send off summer, Croger makes it easy to enjoy high-quality fresh food at affordable prices. Croger. Fresh for everyone. Well, good afternoon, everybody. Welcome back to another episode of the Industrial Real Estate Show. I am pleased to be joined by an old friend of mine, Colin Plich, who used to be active in the Real Estate development market in my city. And since then, he's going on to develop data centers, mostly in the US, and we're going to talk everything data center today. First, thanks so much for joining me in bursting. In bursting podcasting up here. Thanks, Chad. It's great. I mean, I don't do a lot of podcasts, that's for sure. But so, yeah, I was excited to kind of sit out with you and catch up in a minute of 'em. But yeah, yeah, great to have you here. So, Jim was joking about you to see your background first. So, worked at a very large developer traditional warehouses and industrial properties. We did that until 2015 or so. 2015, and I, yes, born and raised here in Edmonton. Yeah, worked for part of the county for some time. And then worked with Remington, established because you know, established industrial developer here in Edmonton. And then I'm left in 2015 to start my own development company, and then transition that in 2017, 2018 to, to a company called Point One Data Centers, which is the company that I'm a partner in today in present CEO. So, you don't actually do much work in town anymore. All of your work would be where we live at that. I was also in silver from day one that, you know, we'd sort of be one of the earlier groups to the sort of the Western Canada data center market. And I still think there's opportunity for that. But in terms of the obviously Western Canada, I'll put it as a major data center market today. And so, where we, you know, where we went, we wanted to go to a tier one market. So, to the biggest market in the world and establishing ourselves there. And so, that's really where we do Eastern, Eastern US's primary area for our business, and specifically Virginia, for the Virginia Southern Virginia and North Carolina. Why is that area of the hub? Because I think at one point that 50% of all data centers was in the small geographic area, why is that? Yeah, I mean, even in 2017, there was a few what we call tier one markets, and they'll pin the globe in a few tier one markets and so the tall markets. And North of Virginia is the biggest market. And was the biggest market still as the biggest market? Why is it the biggest market? I think it goes back to pre.com bubble, you know. You know, it's, and then the network code was established there just outside of Washington DC. You know, not only with investments of the federal government, but the early telcos and the early internet companies, you know, established a network now there. And in just momentum, of data centers, telco data centers, really early internet data centers wanting to be close to that network node. So it's really a network story. And then once that, and then there's momentum that comes with that area, you want to be all of the centers today want to be near that node with us compute and that network node. And so today it still stands as one of the most. Connected. A dense network who's in the world. And so, hopefully you want a few data center operator or a company cloud computing company. And your customers want to be close to their customers. You really want to be close to a network node so that you're. I make the analogy of like a put the inner change. Like imagine an interchange was 50 or 100, you know, streets and freeways coming into it. Have your distribution, you know, an industrial distribution company. It makes sense for you financially to be as close to that node as possible. So that you can go where you want to. And so that's kind of the effect of a dense network node. And so those network nodes sort of opt up in the in the top tier markets. And then our market has continued to mature. The data center industry has continued to mature into tier two and tier three markets. And so it's not as much of a network story today as it was, but it's still a primary component. But that's that's the, you know, I'm not a network expert, but that's the primary story around Northern Virginia. It's a decent Washington DC suburb. So there's land and there's a others up. There's an ability to grow. It's not super urban. It is now, but it wasn't at the time. But it was really an art story. I remember when you left the development company and then started on your own. And then you went into the data center space. And at the time, it was very brilliant to the process. And I remember somebody even saying, I think Colin went into the call center of business. And I said a few. Yeah, I said, no, it sits data centers. But that's how early into the into the stage you were when you first started. And I want to jump into and everything. Because I think there's a lot of important topics here is that first, the data center industry has exploded as everybody knows. So now it's gone from being in Virginia to now every municipality is talking about data centers. Anybody that's sitting on 100 acres of land just thinks that it can be a data center site now. And there's so much more complexity to that. And also it's just a size has has gone up. But I remember someone was saying that 10 years ago, a 20 megawatt data center was big. And now you could see one gigawatt and larger data centers are at least being proposed. So so much has changed. Let's start the beginning. Now what makes a site? So the Virginia one makes sense. Internet connectivity. They want to be around the hub. But let's just say somebody. Let's say in Montana, they want to open up a data center in Montana. Someone has 100 acres of land. What does that land need to have in actually in order for it to be compatible for a data center? Yeah, I think it's I think a little mistake some people make is that it's like one or two just golden rules. Right? I get well, as you've got acres and you've got power. And oh, I've got network connection. I'm in the data center business. I think it's a lot of complicated than that. And so I guess really fundamentally. When when we first started in the business, power was more abundant, more available, less constrained than it is today. And so, you know, where do you wherever a data center wanted to be? Power in general available. And just like if you're doing any attempt at the moment, you assume the amount of electricity you need for that helmet is there utility can provide. That was the case throughout. I think we're in until kind of 2021, 2022, where especially in northern Virginia and other tier one markets. Power constraints are not. And so what that did is that started that that started a geographic shift for where data centers could go or want to go. Because I think it still remains that a general today if land and power were extremely abundant and then just take off the shelf wherever you want. Most data centers were all being, you know, few nodes close to the most dense network place. But that's just not possible, right? Because there's power constraints land constraints development supply limitations. So there's that is driven to man to markets where we're in that tradition market. So you use a Montana example. I think. Yeah. But there's no there's no city or place in Montana. That would be considered tier one, two or three market now. And so I think. The job rule is that is a site shafts to be developable, right? It has to have basic utilities, you know, water sewer and. And. And. And network at them. But it has to be zone municipality wants to have to have it as an industrial types type development. And today it's got to have. It's got to have a scaled power. It's got to have the right network characteristics, right? And so this is why you're not seeing scale data centers with cloud compute data centers going to Montana or. Not to say there isn't some smaller Teleco Teleco data centers or owner user. You know, data centers in in that state. I'm sure there are just like there are new centers out there, but when we're talking about cloud compute. High compute today's big demand drivers and data centers. You know, they're. They're not necessarily not going to be attracted to a place that Montana just because it has power. And that would connect you and maybe a cheaper power than that you can get in Silicon Valley because. If you were to find that same land of power characteristics in a tier one to market that would generally be always more attractive to an end user than. then sort of a place like Montana. So I don't know if that answers your question. It's really not a, it's not a, but I think what people make a mistake about is, if I've got a plan, I've got power, I've got zoning, and I see there's some internet around, there's some network around that I've got to do the center side. I think there's many more layers to the demand story around data centers for where customers need to be, what type of compute is it? Because they were, you were at news customer, where they're customers, who are they serving, all characteristics is that business. So I think, I think it's really hard to, that's where we're going to see play out here is a lot of people making investments and bringing power to the market is what we need, generally, the industry generally needs, but there's going to be winners on those, of course, to get any industry in terms of where data centers will ultimately go. So I still, I still harken back to office or industrial type, that built to suit, a big built to suit an officer and industrial. Not, they have a game plan for where they want to be, they might RFP market, they've got their own characteristics for where they want you to be in terms of which transportation hubs they need to be close to or, where their workers are, they've got their own set of characteristics that need to be seen, and so you can take that same analogy and apply it to, the Microsofts and the aericles and the meadows and even the AI companies, they all have their own sort of characteristics for where they want to be, so you can think about it, when you back up, you think about how complex that is, you know, it's similar to saying, okay, well, there's a huge office tent in the market, I can develop them, the cheapest office building and that's bridge, or a small town in Northern Alberta, you're just not going to be able to attract to that place for similar reasons, it's going to be hard to attract in data center too, a really little place. That I'll circle back to some other questions on that, but that leads me to a topic that I wanted to give you thoughts on as what happens to the industry, is there a potential for overbuilt? So there's speculative builds, perhaps even the technology just changes. So instead of needing racks and racks of servers, perhaps it can be shrunk down to a smaller size, and at that point, perhaps these buildings become overbuilt. What happens to these buildings, which are traditionally going to be a lot more expensive to build than just a box warehouse space, and they're not conducive for any longer warehouse or manufacturing space? - Two per question, let's instead of overwhelming you with two questions or one. So the first one, can these buildings be repurposed at some point down the world into what cost? - Yeah, I think, I mean, when I first got into the industry, repurposing was really important to dynamic because primarily capital markets, right, banks in 2017, 2018 were way less banks understanding the industry's harder to get traditional construction debt started to get investment. There's less investment, there's less pension funds aren't touching it generally. It's more of a niche type market and can be in the late teens. Now, obviously it's a lot more supported by capital markets. I mean, as we all know, so it's oddly enough less of a story today because of the demand. So I think generally, every investor I've ever talked to has talked about what if we have to repurpose this thing and what if market shifts changes? I mean, that certainly is a talent. And I think that definitely plays into valuations for long term data centers. You know, I think one of the benefits that we have in our industry is the average tenant in our industry is fortune, if it's not fortune, 10, it's fortune, 5 months. So we have great credit tenants in our industry, but and I industry's growing tremendously, but just doesn't have a long-term track record. So it is always a question. Okay, well, what if we need to repurpose this building? What happens in 10 years? Well, we just dumbed out, right? It's trying to predict technology in 10 years. It's hard to know. So I think from my perspective, my personal opinion is that if we believe that there's a need for more compute tomorrow than today for 10 years and today, I think data centers, a data center that you go that has a scalable power and is operation. You may need to swap out equipment. You certainly in the swap out in services and that sort of thing, but you need to swap out switch care. You need to densify your technologies, changes, keep you prepared for that because data centers built five years ago are way different than the data centers we're building today. And so you look at that going forward. The general administration generally knows that there's going to be a lot different, a lot of different types and styles of development in the future that there is today. So I think repurposing will be important, but at the same time, my personal opinion again is that as you've got a great location and with great power and you've got to stick some in and of a building, you're going to have, there was traditionally a very sticky tennis, tennis don't necessarily have not been left, there's been at least turn over, but there's a very few tennis that move around and a certain consolidating. So tennis are really sticky as they're, 'cause they're growing so fast, hard for them to, it's hard for them to give anything up because they're having a hard time getting supply. So at this time, we don't see a lot of tennis at all, but we see a lot of, we see a lot of resurfacing. If anything, maybe there's say all these backs of enterprise data centers that get a good amount of investment and retooling. But yeah, so I think blonde-witted answer to your question. I think there's certainly a certainly considerations for what could this building be used for, but based on today's designs, there's not a lot of other issues I will say. So I think the industry is ticking and the investors and couple of markets in the industry are taking them along to you. That data center is important in the long term future and always will be added. And so I, but there still is obviously, just like there is another real estate asset classes, there's a differentiation and value by the year on a data center in the middle of Northern Virginia, whether you own a long tab, that's going to be tougher to release, you know, or it's the same dynamic because it was in clear aspects. And that speaks risk in general with what you're mentioning, but I think there's disproportionate amount of risk depending on the area, as you mentioned, like it makes sense to me that if you're in a small market in Montana versus a large, vibrant market, you could repurpose that building in the vibrant market by just sheer demands for industrial buildings in general, whereas a vacant data center in Montana, what is the use for that? I could be very interesting. I've heard a lot in the last year or two where, and you've seen this now from starting 2018 where people thought you were at call centers to not, if I don't know what it is to know, people are the average investor stinking, I should do a data center because I'm here in a corner all the time. Can you get into the process of what's actually involved in making a piece of land? And let's just say, 'cause I actually came across one the other day in our market, 100 acre site, and they said it's ready for a data center. So let's assume that someone was interested in it and they said, okay, we're gonna go and build a data center. I'm guessing you've gotta find someone actually rent that to build it on spec would be huge, huge risk. But what is the process? Someone is interested in building a data center because they're so curious about the stream and they think it's a digital goal mind and they identify a site of land. What is the actual process of getting that into a data center? - Yeah, good question. I mean, it's our industry has certainly, I'll start by saying our industry has really changed really quickly, it's changed two or three times even since 2017, 2018, and it's a function of demand and geographic growth, so, but that way I would describe it 'cause the way it is today has not always been that way. It's like earlier in 2017, 2018, 19, a data center developer and calling our industry to complement providers or operators, the data would have to speculately start a data center and to be able to win a customer and even thinking even higher than that, our industry is made up of a number of the hyperskill of customers and has some quadratic enterprise building and owning their own data center, so, you know, you think about, they're all a little bit different but AWS, Meta, Microsoft, all of them both, not so much Oracle, but it's a top hyperskillers, a big portion of their capacity is self-built and self-operated, but then there's a, there's a third to a half in generalizing across them or at least, and generally, I'll say these hyperskill customers ideally want to build and own and operate for themselves, but they are huge and they're growing really fast and so they've had to rely on developers, providers, operators to floor just in time capacity or when they can, they have supply constraints and they can grow in a market like a constrained market, they're just not able to find the right piece of land that can deliver when they need that capacity and so they have to lean on our market. So, spinning that back to your question around what, how did, what is the process? It has changed. But today supply chains is challenging and demand is really high. So what that has led to is they're There's various few people that have fear-established developer and trusted provider You done it before you don't have to spec because Because the way customers are shopping now is generally 16 18 24 months sometimes longer away some of their actual demand But they love to just take it off the shelf in a few months. Yes, but that supply isn't there the the the issue with our industry now is that there are supply because trains was challenges is why and that's why we're seeing Geographic change and in a powered land sort of gold brush hell going on now So I so if you add a piece of land here, they use thought it was going to be a data center if you're in a in a in a Established market You wait, you likely don't have you won't have to spec you likely if you're If you've got something that's attractive You're to get in front of a customer You know a good ways out before you upstart standing a building But I would say if you're if you have a hundred acre site in Edmonton It in obviously no place on an established data center market I think it's if if you had the guts and the capital to do it And you actually didn't spec a building you would get a tenant it would be it would fill up But you would it would be really scary because you'd be spending a stream amount of money before you have customer support and So that's challenging to raise capital if you do and you had Ten so it's speccing a building isn't something that established developers in our space really have So specively by land yes specively, you know get it zoned acquire utility You know acquire network get it to a place where it's a nice package in a bow with a design and permits and power story Three things you need to specively put all that together as specively You know get that in front of a customer, but in terms of spectrally building and vertical building Yeah, it's we're not seeing a lot of Well, would that process be time wise to take that land Raw land and to convert that all things let's see someone did have the appetite to take on that risk and they have the deep enough pockets start to finish how long is the process to actually get a Data center vertical. Yeah, I mean, it's it's a little bit longer than traditional, you know industrial building for sure I mean, I think I liken it to kind of the time frame to build an office building You know similar similar time frame, but it's because of multiple factors So there's there's the time it just takes to actually do the horizontal development of the site and then the very good The building And obviously every size different numbers all side, but a generous site. That's ready to go horizontally And and you really data center. Yeah, it may be 16 months right and then it may be some time to commission and actually Customers going to sit out and build out and start out when the solid is to either so but that base building is you know, I didn't be 12 68 18 months to build what the other dynamic that we have is that Because of the rapid growth of our industry Suppliers of how hard time keeping up and so we have Along me to equipment or a typical equipment Supplied houses and so it has an example You know, and this is general and this is certainly broad brush generalizing But to get a to get a backup engine has a backup generator for a data center now is over two years And and there's some suppliers of but the moods of the that may be different So I don't want to broad brush too much But generally the longest lead item in our in our industry from a criminal perspective is an engine Is there? They're just really into my hand and so you have to so you think about you want to spec a building and you're gonna take you 16 lust to build You better in order to equipment that you're before that. All right, and so And then when it comes to switch gear or UPS and And transformers and breakers and all whole watch other Critical equipment that goes in the data center Both the lead time for all that equipment is all a little bit different depending on your design and what I'm actually going with That generally a long lead equipment is one of the first things that you're asked by on Other than the the rezoning's and acquisition and utility and the the land speculative part But then you're starting to have to have a supply chain in place So actually standing the building is not necessarily the first thing You know doing if you're gonna spec so that makes it all really hard to spec as well So because so unless you're a really large established provider Like a stack or advantage or a Cyrus one or a digital really They can they have the balance sheet and the growth and the customer base to go into new markets and and and be more speculative But when you're sort of that second tier developer which we're trying to be it's it's Um, it's it's more challenge for sure So it makes it that much harder To establish itself in the new market say like Edmonton and say okay, I'm gonna I've been gonna bring the market here. I'm gonna bring one of my customers here I think they're a lot easier because you probably see first Those major established developers and providers They lose it to new markets Customers themselves make moves at a new markets to share you'll see the second third fourth tier Developers because it's gonna be really challenging for the owner Yeah, well sent up Let's talk about because I know that's a huge one and we mentioned 20 megawatts 10 years ago was a big number although you could still fight even the odd manufacturing building If it was a heavy manufacturer, they might have had 20 megawatts in there. So it was still possible But that was still a crazy large amount of power now you're seeing 100 megawatt gigawatt data centers What what is the ideal number right now is case by case is there a ideal Number that people would want to have where how do you decide what that actual power Would climb me just and we'll talk about how you actually get that power editor But how do you how do you determine what the ideal sizes right now? Yeah, sure. I think it's just really market-based and compute-based what type of compute is it and where is it? I think those are the sort of the So again, it's not a broad brushed rule that every customer is looking for 500 megawatts or at gigawatt It's really a powered land rush right now because of the constraints that are happening in the tier one markets So the tier one markets constraint pushing pushing Development and supply into tier two markets and other places primarily now because of utility and so I say it that way because if you're gonna go I guess I can think about it as a as a spectrum If I can develop a 20 megawatt data center which today is a very small data center If I can develop a 20 megawatt data center in the shark nor the Virginia, that's a tremendous opportunity But if I have a 20 megawatt data center opportunity in Edmonton I don't need to pick on Edmonton, but it's where we are That is the colors side of the spectrum. That's gonna be a really challenging two You're gonna have smaller customers really small deals more telco leasing or enterprises leasing from a new smaller deals Gonna be it's gonna be multi-tatt one Even that task Quite if so so it so given that perspective the other the other factor is the type of compute Traditionally up until I'm like I say this year maybe just before that It's the primary growth and data centers have been cloud compute. All right, and so the cloud compute providers They've established their compute regions Um, and they grow those regions so in those tier one tier two markets That's what their customers want to continue to grow and so they continue wanting to grow there as so cloud compute To your point them out I mean don't quote me, but I think the biggest deal that I've seen in 2022 Um, and before it was like 70 meg least to a type of scale It's the really the biggest deal and that's right in the biggest market on the biggest to the developers And now we're seeing bigger bigger leases and all these leases aren't They're not there's a few cases, I guess, but they're not 500 could go wide and one or 500 megawatt one gigawatt leases But what's attractive about that big scale is that customers have had It's really expensive to develop in one region and network does really expensive to establish per customer Have for a user And they really ideally don't want to be moving all over the place developing gear and building with air and Can try to To network architect all that. So I think what's happened is because of the constraints in big markets If we're gonna go to a smaller market like Alberta or some other place that's less established Is gonna need to be skill and that scale isn't gonna need isn't We're gonna take it down in one lease. We're going to build it all lines. It's not that. It's It's we need scale so that when we establish our stuff here That we know we can grow in Hi, so yes, the deals are getting bigger and bigger, but at the same time, you know A gigawatt in any location is any hi, and so The gigawatt has really had challenging to find in a lot of markets in the US. So that's per overseas You know some some Some deals being announced, you know, in non-traditional markets because it's from my perspective, it's customers, if they're going to go to that sort of, sort of, sort of a new geography, they need scale. So, and then, and then I guess that it's the type of computers or cloud compute, you know, 300 megawatts site, you know, they could, they could grow, I think it's something great depending on which market it is. There's machine learning demand, which, which comes from some, some customers is some more scaled and even 300 megawatts. And then there's AI compute. And now this AI compute is definitely different than cloud compute because it's a totally different network. And so, AI compute, what I've seen anyway is that it can seemingly can, and as less constraints from a network perspective is could work and go. And so, that's why you can see some of the AI training and deals being done in new places. All right, we've seen, we'll let the code, we've seen Arkansas, we've seen, you know, we've seen sort of outside of the traditional markets, these larger deals, the deep west Texas, because it's AI training demand is like now. And so, there's power available now in a geographic, that's not, that's not traditional, then, you know, you may see some sort of demand going there. So, I'm not sure like properly sort of answered your question, but I think it's really, it really depends on the type of customer that, that, that go to the space. So, we're talking about a gigawon data center today, if that's in a, if that's in a tier one market, you're going to have a lot of the really wide way of, of the types of customers that we interested in that. If that's in a market by Deventon today, it may be just AI, they may be just AI training and we're not working to see what AI infants, where it needs to be, it will kind of, not record that your it needs, but today, it's this AI training deals that are the big scale of the new geographies that, and I'm don't see, we talked about this a couple of weeks ago, about Kevin O'Leary's proposed data center, which isn't far from here, both three, four hours from here, you know, rather remote area of Alberta. It's close to a, we meet small city there, but not a big city by any stretch. And he's proposing seven and a half gigawatts of output, when that's fully built out. And first stage, I should just saw something that he put out the other day, send it, he's doing a $2 billion raise so they can do the first phase, which would be about a gigawatts, is that feasible? This is a project like that, which at one point he threw out would be a $70 billion project when it was all built. It's a project like that feasible given that if it is AI training, it doesn't need to be close to a major city. And if so, do we see these start popping up all over the way? Yeah, I mean, I know our business, I know, you know, what is a market we're in, I know the countries of our customers and our customers want, I say that because, you know, I don't necessarily understand the business plan up there, I don't, but at the same time, I'm not saying that I can be successful, I think that here to, to, if I was evaluating up for our business, for it's in my personal investment, I really have to have a thesis around the scale of tomorrow's compute and the continued, the continued challenge of finding enough of which was where customers want to come, because obviously that's a really rural occasion. Today, from my perspective, it's going to, you know, it's really challenging to get the right network up there, just not just a network connection, I think it's going to have to have the right, our network care protector done, just not aware of that's up there or not. So I think, for me, if I was, if I was an analyst of it, I would think that you'd have to have a long-term view and that, and I think some people, you know, have this view and it's certainly possible that the electricity demand driven by data centers and other industries will continue to grow at 25% a year as we'll tell around now and production, not literally production is only growing at 304% a year. Now there's going to be a bunch of reaction and talking about nuclear, there's power industries really trying to respond, but I, so I think if we're thinking about 29, 20, 30 and beyond, I do think there'll be a continued electricity constraints, if we're still growing, if you believe that compute is going to multiply, if the need for compute is going to multiply and that we don't necessarily, and we don't see a huge gain in AI model efficiency or compute efficiency, you know, to your sort of, to run your earlier questions like if we see, which I think the industry needs, if we see tremendous gains in efficiency of these compute algorithms, then we just won't need that skill. So, you know, I think you'd have to have that work, I think you'd have to have a really long view that demand is going to continue to grow its current pace and that the electricity production and the current itself is not going to be able to keep up. And then we'll drive, you know, this expansion, this earlier expansion into geographies, you know, like Clamper, you know, so I don't think that's where I put my business, but I, but, you know, they can prove me wrong, I would be suppressed. I want to talk more on the power side because I do think that is the biggest thing we mean concern, not just in data centers, not just in industrial, but North America wide, and I contrast that to what they're doing in China, where they're having an arms race for power generation. At one point, they're adding a new coal plant every two days, they've got the most nuclear facilities, either underdevelopment or in progress. They're trying to expand their power keep abilities at a crazy place, whereas in North America, the lat, we had one doubt, one nuclear plant come up in the last ten years, Amazon has one that becomes trying to resurrect three mile island at the Microsoft, the one that didn't have the meltdown, there's yeah, there, but aside from that, we're trying to shut down coal, we're trying to, or at least there's public opposition to nuclear. What, so I'm concerned about it, but in your position of having to deal with this on a larger scale, what does happen, what does happen if the demand continues to grow at a piece of far exceeds our bill, bill ability to supply it? Yeah, it's a fear for sure, and to preface, I am not an energy expert, you know, I'm a real estate developer at ARG, but I'm in industry every day, you know, so I think you're right on the trends, right as the trends are, you know, computers growing, you know, computers growing at 25% caggers every year, and that can level, it depends on you, I'm separated by, and cloud, that that rate, again, where you're not, you know, the generation industry seeming to me is flat flip, because they have a pretty flat demand for decades. And so now, all of a sudden, we have this ramping of power demand, you know, what we're, the industry's feeling now is a lot of energy companies, power companies, really focusing on data centers and trying to react, so one of the big dynamics will be, you know, to talk about China, obviously that's a, you know, that's a federal focus, right, in an AI compute race, right, in a data dominance in the compute space race, energy thing, yeah, coal wars in the past, it's like, and now it's like a compute war, can I say, so the U.S. is, is, is, if it was a little flat flip, it's certainly in the U.S. federal government is focused on this race as well. And so I think what the industry is advocating for, I think starting to see is, the U.S. been a predominantly looking at the red tape and the challenges that are, that the power industry has, the energy industry has, to be interconnected, you know, nuclear plants or goggles, small modular reactors, or, or, you know, trying to interconnect new, new solar, new wind energy, this is a slow process, these utilities are designed, you know, to be slow and careful, and so a dynamic fast-moving industry is not a great mix for that, is so we're seeing the federal government really trying to, they go see it more, trying to identify where they can fast forward things, where they can remove red tape and really try to allow the energy industry to, to react to the growing demand that we have. My personal opinion is that our compute continue to grow at a similar pace. I don't know if it's going to be, you know, as hot as it is today forever, but I do think that there'll be the continued more demand for electricity and supply for the coming four or five, six years, I think that's, for me, that's clear, and so I do think electricity is more expensive for me, it's going to get harder to get and find, and so that, but how long does that dynamic last rate, how long, you know, just to do, to solve nuclear reactors, adopted or do they get or do they get You know, barely not supported because people are afraid of them. I'm not sure, but I think that it's hard to predict, so therefore it's hard to predict five or six years of, but generally for what we can see, two or three or four or five years of, we're going to see it, I think it's going to get more expensive, it's going to be harder to find, and so that's where we're seeing a lot of onsite generation, we're seeing natural gas, natural gas and breach to grid opportunities, onsite generation opportunities become more accepted, more necessary, I'll say, because of the Greek strength, so not because onsite, from my perspective, onsite natural gas generation, whether it's stop Northern Alberta or whether it's in any other place, gas is generally cheap, but it's really more expensive than the grid in a lot of places, and so it doesn't, it's for where we do business in North Carolina and in Virginia, the grid's cheaper still, but in the coming years, Russia, if I still going to be the case, and so therefore, we're also seeing more and more onsite generation, unfortunately, we're seeing some of the plants, shutdown plants get delayed, because of the demand and opportunity, the same time, I think it's a demand story, and so I do personally hope that I'll continue to find alternative methods of power production and geographies so that we're at, we're not trapped in a market, I say that we're power becomes extremely expensive and aren't again to enter, and the industry is going to make sastas still burn coal, obviously, that would mean then, and I get, so still must be challenging even right now, we build the grid can theoretically sustain today's demand, but it's still going to be very difficult to get onto the grid. I can't imagine any electrical company in any area just saying, oh yeah, we'll allocate 100 megawatts to you, and there's another one down the street, we'll just give them 100 megawatts as well, how challenging is it actually getting to those utility? It's a very, it's becoming as way more challenging than it was five years ago, and it's often than it's, I will prophesy by saying it's a very geographic specific. If you're trying to connect, if you're trying to get a new site in Northern Virginia as an example, because that's where I'm doing business, the utility is telling you, it's four to seven years, from the time that you said you, you know, I think four is either Roosevelt or Roosevelt. So by the time you say, hey, Mr. Utility, I've got a new site, I have a new demand. The time, by the time it takes to engage in an engineering, engage in them studying their grid and understanding what's going to take to serve you in the allocation, and then add to that, then the long lead time it takes to get a 230 KV breaker and a 230 KV transformer in the substation is like three years plus four years now. So it's, so it's really challenging to get that type of scale and grid service in certainly in the top markets. And again, that's why you're receiving, if you're, so the utility is not, the story's not flat across the nation or across the North America, because if you're going to a new geography, that, and utility that's less constrained, that's more excited for data centers, because some would say, what are utilities in the US? You know, no, they love data centers, and they continue to grow, but they've seen a lot of it, and they've, so they're not as, you know, they're not as exciting and, you know, there's merely been getting, they're getting hundreds of applications for data centers. And so it's not like you're going to utility that they're seeking this application for the first time or excited to support it and then doing all they can. That's, so that's a little bit easier situation to get scale power. You might sign, you know, availability and capacity in certain places, but it, so that's one of the supply challenges that we have in most markets now is that it's hard to get equipment and it's hard to get the utilities' attention, because there's so much to add. But most of that demand, I'll just say, is a lot of that demand is speculative, Chevy is called the heartbeat of America for a reason, with SUVs made to move with your rhythm. The versatile Equinox tackles your entire day, the spacious traverse fits your crew and your whole weekend, and tracks bring style with value you can count them, all infused with tech that has your back, so your drive always hits the right cord. Chevrolet, together, let's drive. Relax and let Croger Delivery handle your grocery shopping this week. We start with only the freshest items, then carefully pack your order in our refrigerated trucks, and our cool trucks keep your food fresh right to your door. Fresh groceries your way with Croger Delivery, and right now, enjoy free delivery on orders over $50. Croger, fresh for everyone. Powered land, you know, gold rush. You know, so you can imagine the utilities getting hundreds of applications from every ton to canary that owns a piece of land that wants to make it a data center. So their utilities thinking, oh my gosh, cool look at all this data center demand. How much is real customer said it's customer back to mad? And not a fraction of Fisher, you know, else. What risk is there on a developer that wants to develop something on spec, goes and gets the power. Maybe to do in a 50 megawatt site, they go and get that, but they can't get a tenant for whatever reason. Maybe it's maybe someone else built one a better one down the street. Any number of reasons is they don't get a tenant. Are they committing to that 50 megawatts of power for some period of time? That's the first changing too. I mean, when we first started getting in this trade, I think you'd probably find in most utilities around here is that if you can find that 50 megawatts of available capacity in the infrastructure that's there already, then there's not a lot of economic costs upfront, but in a place that is constrained, seeing a ton of application for the same amount of capacity that they have, they're becoming yeah, and well, it's harder to deal with. And what that means is instead of them doing a bunch of engineering for you for free, you know, you're putting down 250 grand for them to just start to see if they can serve with you. And so they're making it harder and harder for the for the Glant speculator to grab power. And again, that's in the real estate trade market. So if you're in this market, let's say I'm into this example and you're able to, you're probably I'm assuming a lot here, but you're probably not going to have to make major commitments to that scale of service. If that ever structures there, but if you're asking for it and they can't get it to you for three, two, three, four years, because they have to do a major major transmission upgrade or the substation. There's major, there's major spending nest occur that many utilities are are having forcing developers and customers to securitize that. So you're I mean, but it made major LCA's and what we call kind of a take paid to model. All right, so you're having to commit to that and our and so again, it makes it real challenging for a smaller developer to add up and in that scenario to make up his flatwork. What I love so far, what everything that you've described is that there are rules of thumb. I think that is the biggest way to put it. Yeah, is that someone just thinks, okay, yeah, I'll just go put a data center here, but there's so many complexities, be up to the utility cookups, the costs and might be involved in doing it any commitments, timelines of doing it, the type of tenant that you actually attract when you get one of the sports and tenant fortune 500 companies, or whether you piecemeal and adopt a smaller ones, there's so many nuances that go into this that there's no simple answer. So I think anybody that is considering a data center needs to be very hyper focused and understand that area as opposed to saying, well, they built this in Virginia, so I'm just going to take that model and they're wheticated in in Toronto or within Williams, Montana, it's there's so many complexities and differences between the two. I think you're right. And I was what I was saying is everything I'm saying to it, I'm very sure I'll be sure to wrong it till the years, right? Because I will leave it in the industry since 2018, it's changed three times already. Like, it's fundamentally changed demand and where data centers will go, what's the profile, what is the network market address change, what? We start off as all service contracts and that was some more, more net leases. Well, it's changing. I mean, so I didn't as much as, you know, sound like I know I'm talking about today, I guarantee people watch us two years from now, all the person run a lot because it's, so that doesn't make a challenging as well for our industry too. We've got to be really careful and use set your design. Let's say today, I'm going to build this across our sites for the next five years. You better be ready for a difference in change as a technology. It can change really fast. And so there can level up again and see a level as it is customer-dependent. So it's changing that fast when it does make it pretty challenging. So I think that's why the industry works well now because we're able to do more built-to-soup. I mean, we're not having to in the red markets, you know, I'm introspectively speculated. And then I will say at the same time, the industry does need land speculation, right? It does need is hard for the, really hard for the customers and really are treated with big developers to procure and land bank. And we talked three or four years earlier in the past trying to get power. So the industry does need developers to can a customer do need land speculators to go out there and get ready to land, get it to sound, you know, acquire the utility, put together that power story. Then you do have something of value if you're in the right situation because then some day our company, but also the more established developers and customers, again, they're not having a really art time shopping and land bank many years out. So there is a business plan still for, for a power of land speculation. And because some of their hard customers and users are developers, I will need to take that to buy that item. We'll pay it pretty penny for it and typically. And so I think, you know, I don't want to book who all land speculation is going on because the industry does. But it's, it's a dangerous game in the markets here. Yeah, it comes with risk. I mean, there's probably higher the risk, you know, higher potential for words, but also the higher the talent side risk. So yeah, very fascinating to see this. I want to jump into some of the changes that you've seen. But even before that, the big things that any site's going to need, obviously, you need to be zoned accordingly to have local government support for it. You'll need power, which we already talked about. I think the two that aren't as obvious, even the people in the industry would find it very obvious would be the water and then the internet activity. What I think, again, I think people will, in the industry, will understand why the water is required. But some people might might not understand and see used as a cool thing. And it uses a sort of water water. Can you talk about how much water, just a data center could use? Sure. And to that, there's that list. I'll add tax strategy as well, because I think it's sweet. And I was talking about that. And it had the factory image of important for a data center because of the capex size. But to your question, water, using water in a data sense, cooling in a evaporative way and in direct, direct, evaporative way, is a really efficient way to cool, to cool the data center. But it uses to put a lot of water, extreme amount of water. And so if you're in the past and then today, measuring a data center's efficiency, there's something called PUE power usage effect in this ratio is if you're using water to cool your PUE, we can be a lot in Europe, meaning your data center is going to be using electricity efficiently, as live only as possible, meaning the most of the power that's going into that data center grossly, most of them is being used by servers, right? And less of it is being used for cooling, working on another and other uses in the data center. So the smaller you can have, you're you're cooling use for power in the more attractive data center. So how efficient the data center is using a tremendous amount of water makes that makes your data center as quite a fish. It's one of the most efficient ways to to cool the data center. But it uses a tremendous amount of water. So it's not a cool rule. The data center is used water for cooling, there's a lot of many data centers, half of not more than half, when I'll use air cooled chillers and not use, you know, water, that was maybe a chiller who can set the building, but it's a closed system and they will not use. So that means your your data center is a little less efficient, you're using more electricity, but you're using virtually not wired, right? So it's not a golden rule that a data center will use a lot of a lot of data centers designed, do use a lot of water. So I say two or three years ago, we see a lot of data centers and customers prioritizing the use of water because they want to be as efficient as possible. But I think over the last two years, from my chair, customers have, you know, people, there's a lot more eyes and there's a lot more glory that they have to do on how much water they use. So there's now, there was a WUE, which is a water user defecting this like how much water is your data center using. And so when you're PWS or Microsoft or Google can purchase some of the major companies, you know, so very mindful and watch how much water they're using. There's a lot of eyes on that. And so in many cases now, you know, the customers will save their airking for swap and simply because they don't want to use that much. And then we're going to places where there is that there isn't enough water, right? Like even in the big markets in the bigger binaries, so just not enough. Like, they're using a tremendous amount. And so, unfortunately now you're going to new geographic areas where the water system's hot. You would find here the downtown Edmonton, you know, you're not going to have enough water. So you're in a force to use air cold and so on and be a little bit less efficient. So I, so it's not necessarily a golden rule that I think is trending away from using water. If you had asked me, but you know, trends got to come and go, so we'll see. So if you don't, if you have a site and you have this water connection, it has to be quite scaled. And obviously, when you get into 100 megawatts, 500 megawatts, it's almost an unimaginable amount of air at that scale. So it's, that's makes it a pet scale, makes it also in a life possible from my perspective to, to responsibly reuse water for cooling. So it's not golden. It's very interesting. So I'm guessing it used to be a fiber line that's going in there. What about redundancies? Because if the fiber line goes down, just the network go down. And I kind of want to talk about you, because you mentioned redundancies on and generated the long lead times to get those right now. And maybe that's the simple answer on power. But what about redundancies on the network connectivity? How do you trust that? Yeah. They're going to get the, again, the first rule, I think you would think about is not every customer network architecture is going to be the same. So again, you know, golden rule. But as an example, on the side of the spectrum, you're going to have applications and then and then it's extremely network sensitive. They need to be close to a network out, you know, and that specific network node. And then one millisecond of latency is too much and send a lot of cases. But on the other end of the spectrum, again, if you're tiptoe mining is certainly one example of a type of compute that seemingly doesn't need, doesn't need much more than just any kind of hidden connectivity. But in a data center space, the AI trainings, I get I'm not an expert, so I'm going to sound like one, but seemingly that AI training type of compute is lasting work sense. And so that's why we can see it going to places with, you know, less robust number can, and the other factor is cost. Like you think about it, network is like a toll road. And you're not just necessarily unless you own it here as a customer, you're not, you have to pay, you know, as a cost. And so that, and then that's a distance function as well. And so the closer you are, because you're transmitting a lot of data, the closer to your other network, the cheaper your network cost is going to be. So here you are, the more data you're transmitting can be extreme. And so, so that yeah, that's another sort of layer to the network, is it's got to be the right network characteristics, not only from the, we call it dark fiber, availability, like there's availability, so that a customer can, can contract for it to its private fiber. And in a system, and they need to be the right sort of distance characteristics, can then down to your question. It's in most cases, you have a redundant path, meaning, you may, you may be contracted to use a primary path between where your data center is and other data centers that you're in as a customer or a board of network known. But if something happens to that network, like you said, I mean, do we get a, we get an interruption, we have something happen there that needs to be redundant rate, right? And so data center, you're almost at every case contracting for redundant pathways back to where everything went. And that test rank goes to where it gets specific. So it's, again, it's really hard to apply a goal. I've done network or I don't, or I've done dark fiber and I don't, it's, you know, the right sort of network characteristics. for the right type of compute that's going to be calling it today, sir. It's not having to be done and see till a deal. And, okay, that is that simple. I mean, I won't speak for every type of compute and every customer that's having not that, certainly not an expert, but from our perspective, we, from our business product, and we would never buy a parcel land that did not have, it's just a robust, unavailable dark fiber and pre-done a network path. Is that going to be quite challenging, smaller remote areas where it's tight? Yeah, because I would, I could think that bringing in a small rural area where there might be a thousand people in the town with a village, they're not going to have a high, two reconnectivity fiber line because the handle was added. I need 50, I need 50 pairs of dark fiber, that dark fiber line might have light as that availability, but because it has multiple paths, you know, it is a message. And so, again, depending on the customer, they may need to do their own network. Some of them have their own network characteristics, they own their own network, they own preferred network containers, you know, so that's all three year graphically away they are from where they ultimately want to be or where they're transmitting to can be extremely expensive. And so it has, and that's why we haven't seen a lot of data-centered development in extremely rural areas, unless it's maybe the customer built and owned and they know exactly what they're doing and they've got it all planned out, it's just so it's hard to be a developer, a speculative investment in land or in an insight and not have a wide enough array of network characteristics so you can serve the specific customer that ends up coming to you, though, thinking about it on a different level on the industrial side, it'd be the equivalent of a distribution company that says, well, we can either locate in the middle of the city or in the city or we can go find land half an hour outside the town and we'll get land for a fifth of the price, but now all of your additional costs are then much more expensive, your transportation costs, getting labor, it comes up all the other set of challenges, so it might make sense for a company that that's looking in a small rural area, but it might also make sense for them just to be able to be located in one of the major nodes because that's exactly what it is, right? You'll see some of you may go to the point, well, look, Google in over here, look, this customer went over here, it sort of breaks the mold, well, it's a triumphant, triumphant, triumphant, it's a whole bunch of characteristics and a trinity of almost characteristics, right? So then, like I said earlier, the general, from my perspective, if we could grow, if every customer could grow in the tier one market, it's right next to the network node, they all would, well, that's what they wanted me, but generally, they could have supply constraints, so there's trade, stingling to places and networks, but end down, so that's what we're seeing. It's different than traditional real estate dynamics, it sits persuasive, that's very interesting. You mentioned tax fusion, that's one that I hadn't thought about before, so I'd love for you to expand on that. Well, you think about the cost here, it was, it was around billions of dollars and a data center, yes, the data center is extremely expensive for a developer or developing a kind of key data center, you know, maybe 12, 13 billion dollars a megawatt, you know, to develop a certain key data center, and I'm not an expert, but, you know, what I'd throw around in the industry is that if then the customer is putting two or three times that at least in value of racks and servers and build out in the data. So if it is a, if it's a 70 megawatt data center, or that's going to be around a billion of caps for a company like us, a developer, the customer is build out, it's going to be a stability in that case, so huge numbers. And so when you're going into a municipality that has a tax regime, that's, let's say, like bringing it back to industrial, that says it doesn't equip in tax. And if servers are, are considered equipped, then the tax bill, and they, a lot of the tax regimes around municipalities that were never designed around extreme scale of compute. And so, you know, customers used to let some go where they were provided incentives on the local municipalities to locate tax, tax breaks, infrastructure deals, because it doesn't, it was really extremely hard for a customer or a developer at that scale to just move into any tax regime, it makes it really expensive. And so tax regimes that have urban disabilities that have, you know, specific compute, computer taxes, and even specific depreciation schedules for those depreciation schedule, the server is very fast versus a piece of traditional, maybe industrial equipment versus the building. So, you know, it's very important to have the right and competitive tax rates and structure of that customer coming out of wise, it can be completely a non-starter for a sort of move into a certain area after it's there, because their tax pool can be incredibly high. It just adds an end to the layer of complexity, you have all of that. But instead of complexity, what do you guys are active right now in Kirkbuck, Carolina, and in Virginia? So, what do you guys have in the goal right now? What's we focused on? Yeah, over it. Since 2018, we've put together, we have five properties in total, we have four hyper-skill, probably we call hyper-skill properties that are sites that are, you know, our smallest type of scale site is under megawatts and our biggest one is potentially up to a gigawatt. And so, across in Virginia, we have about a gigawatt of a developable capacity in our portfolio, and then with Carolina, we have had gigawatt as well. And so, you know, when we first started acquiring sites, you know, in the industry of a 20 megawatt deal is a huge deal, right? And so, we were, we made a play on market adjacency and scale, and that thesis has played out, but also the utility constraints and other supply constraints, you know, we're still a company that's looking to transition from being an aggregator of sites and a developer of the development of land, getting our sites sort of shovel ready, utility ready, sound and ready for a customer. We've been transitioning our business for raise cap building our execution team to be a vertical developer. So, you know, let's see, we're the exciting year for us and our first projects are coming out of the ground. So, finally, but what's really exciting for us is to be able to kind of build in, arguably, the top market, and for all we're able to build a portfolio of sites that are in demand today. So, it's really about our, we're focused on execution, and that's really, it's really our, our make a break for our companies, can we execute these properties, and become a trust developer and provider with some of our bigger recruiters. But by nature, you have to be thinking five to 10 years old at any given time. Where do you think they'll five to 10 years old from today? Good question. The, you know, knowing, knowing that our industry will look different in two years. It might have five years, it's it's challenging if our thesis around geographic expansion of the major markets and, and powered land, you know, scarcity, certainly has played out. That's going to continue to play out, but then, you know, because of the demand, we've, we've seen in the change of compute to this, you know, the city, I demand, there's, there is a, there's a lot of people currently speculating on power land, and they're a lot of energy companies entering the space with their, with land adjacent to their, there is a shutter. So, I do think, I do think it'll, when we continue to five years, there's no question it'll be continued to your graphic change in the way where data centers are built, and that's just going to be because supply constraints in the places that, the data centers want to be, you know, the fact that, and the fact that, the US and, you know, like places like Canada are, are stable governments that have, they know that are, that are trying to break down corn, warm red tape, and, and support the industry, I think North America data centers can also get, this is a tremendous opportunity to come and support. I just, it's, it's, it's, it's, it's, it's, it's, it's really, it's really challenging for me to think about, um, power land in, uh, in, uh, in an unestablished market. But, I guess in five years, Edmonton could be an established market, it's, it's really arts. It's, it's, it's, it's, it's changing to that fast. And for a dull day, when you're in an industry that changes that dramatically, at a fast, I'm always something you would think, what I, I do know that is, if you've got, you know, and I would invest in, if you've got, you know, a developable site in a tier one or tier two market that's, that's established data center, but then you have a utility and you can deliver a data center within 24 to 36 months, you got something that's, that's really exciting, you know, so that's, that's the, that's what we're all looking to try. Okay. Well, it's a good advice for people that they're, any other questions, uh, having to ask you that, uh, you know, you know, that's what we're talking about. should have brain topic something that you get asked that I that I haven't asked. She's recovered a lot. We did. We hit her a lot. It's off excerpt. Chevy is called the heartbeat of America for a reason with SUVs made to move with your rhythm. The spacious traverse fits your crew and your whole weekend and tracks bring style with value you can count them all infused the tech that has your back. So your drive always hits the right cord. Chevrolet together let's drive. Find quality you can trust at a value you will love with a wide assortment of products from our exclusive brands. Fire up the grill with cookout classics like burgers and brats and don't forget delicious produce like fresh juicy melons or beat the heat with frozen treats while chilling pool side. However, you're planning to send off summer. Croger makes it easy to enjoy high quality fresh food at affordable prices. Croger fresh for everyone. Yeah, I think I think maybe the bigger the biggest thing I think in my mind anyway is that I think people lost count. You know, when we're talking about a gigawatt, even we have a gigawatt side, but that's 10 4,000 foot buildings. Right? And that like to take the problem to Alberta is using something like 14 gigawatt. That's what I call right there. So he's here. People have sort of lost scale. And I think it's going to be extremely even though there's a ton of excitement in AI and a ton of demand. I think we've kind of lost scales to not if I were to make a lot project, which is multiple data centers, multiple buildings. So it is extremely, it's extremely capital intensive, extremely expensive. The customer talked about how much the customers have to invest in their their compute and network and cyber data centers. Yes, they all have announced 50 billion and 60 billion of CapEx every year in data centers, but that's a global. Right? And so I do think there's a there's a land and power rush now. I just think that they kind of, you know, I was talking about a gigawatt. And I think the people should be talking about is not necessarily those grand slams, but thinking about singles and doubles and then hand in the your megawatt range and in the stock market. So I think it's is where the where things you couldn't really is. And I think that the three mile island example that we were kind of queer kind of joking about and in some of these behind the behind the meter deals with diverse values are doing. I think that's I think those are serve you the climate's handles are one off customer back deals like so from real estate to the moment perspective, I think people just lost scale a little bit in terms of thinking about ever using the gigawatt word. It's just it's just it's just an extreme scale that sits hard to start to think through. That's a very good tip. I know one last question for you because I put out a post on Twitter or X now who also was a big user of data now with it's pretty. And I asked. Yeah. I asked for questions and one came in that you know, with this one from the curve ball favorite pizza toppings and ice cream. That said, just favorite pizza topping. Okay, I hear how's this when we're really syruped. Twitter beat pineapple. That's sweet. That's good. Do you see non-traditional? That's called charge that. That's what it has said pineapple. If it's not like you know, the pizza and apple we you know candidate the New York or tell you pizza. I like pineapple and pizza as well. And ice cream. I'm a vanilla guy. It's hard to be good. There's a lot of, there's a lot of, you know, weird stuff out there, but I haven't been straight down a little bit of a guy. It's cool. It's called and thanks so much for joining me. So this is me. Yeah, I appreciate it. Covered a ton of stuff and I really appreciate it. Oh, last last thing. People wanted to get in touch with you. What's the best way? Hey, if you want to learn about coin one, our business or our website is.1 development pair. Put my corporate column, I should say. You can find me a LinkedIn and put it out of that to put it in. But the best way to find is probably on LinkedIn. We need to get you on Twitter because there is a vibrant community right there and you would add a ton of value to it. So we'll try to get you on Twitter and I feel like I already typed it off. I guess I had to respond. But yeah, you're not wrong. Well, in the meantime, we'll get the website and LinkedIn and the show notes and encourage people to reach out and touch based with you. And we have once again, thank you and thanks for everybody to win. It was awesome. Thank you very much. Relax and let Kroger delivery handle your grocery shopping this week. We start with only the freshest items. Then carefully pack your order in our refrigerated trucks. And our cool trucks keep your food fresh right to your door. Fresh groceries your way with Kroger delivery. And right now, enjoy free delivery on orders over $50. Kroger, fresh for everyone. I didn't like what you said. Odin pauses about the future. This is the love story of real hinged couple Odin and Edward written and read by me Curtis Garner. Listen to the free audio book now.