Ep1: The Data Center Boom: AI, Power Demand and What’s Next

Welcome to Uptime Now, a podcast dedicated to conversations about data centers from development and construction to cutting its technology and operational breakthroughs. And most importantly for me, the people driving it all. I'm Samir Kuznavi, partner with the Lafram Norton Rose Fulbright. I'm so excited to welcome my very first guest on the show. He's someone that, if you're in the data center space, if you've been to a conference, if you're listening to other podcasts in this space, you already know him. He's the Steve Irwin of the Data Center world, the CEO and co-founder of Apollo. Welcome to the show, Bill Clayman. Samir, it is an absolute pleasure to be here. Now, I know that people can't see us, right? Unfortunately, so we're going to have to really bring this energy. No pun intended. That's literally the higher gravity of what's happening and they should just shoot right now. What a great setup. And I'm just Samir. I'm just so personally honored to thank you for having me beyond. And not only that, just your first guest. Let's set the bar pretty high here. I appreciate you being on. Great to be talking to you today, my friend. How are you? What's new in your world? What's going on? What's new in our world? There's a really interesting saying, it's from a person on our board, right? So the data center industry loves innovation as long as it's 10 years old. Well, what's new in my world is that if you blink 10 months goes by, it's just an extraordinary moment. Samir. And I think we're going to talk about this just a little bit. I just got off the airplane from the Las Vegas event, the Schneider Electric Innovation Summit. And I don't want to just set the stage here. We are literally 90 seconds into this conversation. But our industry, and you're asking what's new in this world is slated to consume 200 gigawatts of power by 2030, 200. Now you can't see Samir's shocked face right now. But it's there. He's like, I can share those numbers right now. Yeah. Those are the latest numbers that we're seeing, right? And it's a generational investment opportunity. And you're asking me what's new, honestly, it's just everyone trying to figure out how to corral what's happening. The piece of it all is just kind of getting people a bit uncomfortable, but I'm sure we're going to talk about it. Everything is new. It's extraordinary. Yeah. Everything is new. We should change the podcast line too. Everything is new. So for people who don't know you as well, I'm sure they would find it really interesting. But can you tell us a little bit about your origin story? How did you land in this data center world? Tell us a little bit about your role. No. Well, Samir, thank you so much for giving me a chance to speak. By the way, everybody listening, this is, you know, this is me being really calm, usually remarkably. I'm even more excited, right? So I'm just trying to get that little radio voice for you so you don't have to be like, who is this energetic jet bug? I'm Bill. I was born originally in Kiev, Ukraine, so a real name is Vitaly. And the origin story, Samir, this is where like, you're either born a hero or a villain, right? Unfortunately, I'm the good guy, at least I hope I am at least. And in Ukraine, in Kiev, my brother used to compete in telegraph competition. So I legit, everybody listening, more as code, like I grew up with boops and beeps in the house. Now, I know it's Soviet Ukraine. We weren't completely backwards. We did have a phone on the wall, right? This was in how we communicated with people all the time, but it was the coolest thing in the world. The R.I. Amazon is an eight-year-old learning Morris code. And that was the spark, right? Coming in the States in the early 90s, undergraduate and network engineering and telecommunications management, an MBA and other matches and information security. Right off the bat in 2004, I worked with data center racks. I said, you know, let me back up. They didn't even call them data centers back, that back then, right? I'm going to age myself, everybody. There's a new term for my age group. This is called a geriatric millennial. I don't know why you have to call me that. That's completely unnecessary. Or you could just call me an older millennial. 2004, I started working in data closets and network rooms. Listen, before they even call them data centers, before that word was even cool, I've had a chance to rack, stack and work with these solutions, had a chance to spend some time at a place called MTM technologies. They were one of the nation's largest Citrix virtualization Microsoft partners, VMware as well. And then I got my, let's call them big boy data center pants, working for a really wonderful organization called Switch Data Centers in Las Vegas, joined them in 2018, spent about four years there. And I was very fortunate to be the executive vice president, reporting to the CEO and, oh, my goodness, this is where I really grew up, I feel, and I learned a ton, a ton. We took that company from four and a half billion publicly traded and helped them sell for 11 billion dollars in December of 2020, so it was an extraordinary moment. Now, Samir, you say, wow, you say, wow, right now, for a time, I was really proud of it, right? We were the second biggest acquisition in the data center market, 11 billion dollars just behind KKR, and when they purchased was, was a compass data centers for 15.1 billion dollars. Now, drum roll, if you heard what happened just a couple of weeks ago, aligned data centers was purchased for 40 billion dollars. Here I am, four times X, and we should have waited a little bit longer. So after Switch, after the exit in the sale, I had a chance to found this really wonderful organization called Apollo, you can check us out at Apollo.us, and we are the industry's only true white label, multi-tenant MLOPS and AI platform, basically a facility will deploy us, they have their own GPUs and infrastructure, and BAM, Apollo allows them to become a little neo-cloud in a box, a mini Amazon, up source rights, doing AI with the data lives, doing all that dirty plumbing in the back, so networking, ticketing, configuration control, all the stuff you can actually use those GPUs. Now, outside of all of this, I do, I promise I do sleep, I do hang out with my children, and I do get a chance to actually enjoy life. I'm really proud to be the executive chair of Data Center Programs for Informa, and I oversee and help put on and keynote and work with brilliant individuals to support the world's longest running and largest data center conference, and that is Data Center World, and Data Center World Power, and then sort of in parallel, I am really fortunate to be the author of the ATHCOM State of the Data Center Report, and writing my ninth one right now, and contributing editor to Data Center Knowledge, Data Center Frontier, I've got a couple of books that I've co-authored, I have ADHD, I can't write it entire book, I can do a chapter at best. And those are called Greener Data Volume One, Volume Two, and now we're just finishing up Volume Three, Volume One and Two, we're drumroll Amazon bestseller, so you're curious about some really cool things that are happening in the Data Center space, you can check out Greener Data, and then finally, finally the one of the ways that I give back is I actively participate in some philanthropic organizations called Infrastructure Masons, and Nomad Futurists where I work with everyone from K through 12 to graduate and undergraduate students, I'm a regular lecturer at USC School of Engineering, and really just selfishly trying to get more young people, more brilliant minds in this industry, so that was a very long-winded reduction, I do apologize in advance, but that's me. No, I love it, it's great to hear, and it's great to hear the de-7 enthusiasm, and I want to talk about the comparison to Steve Irwin, because I love that comparison, he was such a passionate person, just loved what he did, and anyone that has met you even for a few seconds knows that you share that same kind of passion, so where does that passion come from? Was there a turning point where you're like, "Man, I really love this stuff." In case you can't tell, I'm a gregarious individual, I do love this stuff, and I feel similar, I feel like we're doing a disservice for our audience here, they're only getting 50% bill, yeah, there's a lot of energy, yeah, there's a lot of excitement and passion, but imagine I'm waving my arms right now, you can't see that, but that's all right, I'm going to try and express that energy, radio voice, right? It is certainly happening, I'm really fortunate, a part of this is personality. For my entire life, I loved, loved, loved, explaining really complex topics to people. I mean, heck, when I was in college, I was a stats tutor, because I had a really great professor, she was an actuarial scientist, and she was the first one in my life to bring what was a complex topic, really advanced statistics, to something that I could practically understand, right? I love that, and when it comes to technology, I'm really proud to write pieces, speak in front of massive audience, do master classes, one-on-ones, that are meaningful, right? They're not just like this whole 200 gigawatts that I just talked about, for some people like, "Well, that's a great number," well, putting that in context, I'm going to put my teacher hat on for a second, one gigawatt is enough for a million people, 10 gigawatts is the size of Los Angeles, New York City, and 100 gigawatts is about 10% of all the lights that are currently on in the entire world. So when we say you need 200 gigawatts of power by 2030, holy cow, that's like 20 LAs or 20 New York cities that it's so, so much. But now, we've explained it, at least you have a context to what 200 gigawatts actually looks like. So, Samir, I don't want people to be bewildered, I don't want people to sort of be afraid of new technologies, I really believe in fearless curiosity, because that's really the way the only way we're going to break out from any of these paradigms. And right now, what's really fueling this enthusiasm is I've always loved explaining data centers, the thing that makes not just everything run, but everything runs on data centers, you believe in all the connections, all the connectivity runs through some kind of physical infrastructure, and more than ever before, I'm getting approached by so many individuals, because unless you've been living under a rock, the data center industry is having a spotlight, we are having a moment, and it's making some people a little bit uncomfortable, where the first questions that I usually get right now are bill. How much water are you taking, bill wires, so many electricity, so much electricity going away, right? Why, you know, what are you putting in our ground? And I'm, I'm challenged with the fact that the first impression many of the folks get in our industries is, is slightly negative. So yeah, yeah, I'm going to funnel that inner steward would cry. Keep looking at that server. Let me go press the button just to see what happens. The working on my Australian accent. I'm just going to embody the energy of Mr. Irwin instead of just trying to mimic him. That's where it came from. You know, I love getting in front of big crowds and stages. I really do enjoy. And it's not just for the fanfare and then they'll pause and all that stuff. I really do love explaining what we do, just how important it is, just how vastly connected our world is and how much more it's going to become connected in these new emerging markets. So that's why I do it. That's why I love doing this stuff. You know, that's why I'm honored to have this conversation with you here. But really, it's just to help people understand my goodness. This is quite literally the lifeblood of everything that we do today. And especially with this little thing called chat GPT. My gosh, that's only going to accelerate the adoption of technology. I definitely want to get into what is making people a little bit uncomfortable with the massive amounts of growth that we're seeing. But before we go there, can you tell us a little bit more about your company, GPU as a service? I find stuff really fascinating and I am not very technologically savvy. And I love that you said, you know, you're great explaining complex concepts. So if you could like help us understand, help me understand what it is that you guys do. One of the easiest ways is obviously you're welcome to order our website Apollo.us. But let me back up a little bit, right? Without getting too technical, okay? Follow me on this everybody. We live in a world that's just discovered oil, but hasn't invented the internal combustion engine. Oh, this raw material. And no one really knows what to do with it. But using that oil analogy, it's in everything, right? It's in my clothes. It's in it's in M&M's. The red one specifically. It's in it's in medicine. It's in shoelaces, right? Where we are in the world right now is how do you get from this barrel of oil to a shoelace? Not just any shoelace. One that doesn't cost as much as the entire dang shoestore and one that can drumroll actually tire shoe, right? Something that's functional. So Apollo was built on on this premise, right? To be able to democratize, to give access to AI ecosystems above and beyond what the hyperskillers are doing. So think of Apollo as the factory floor. This fast, wonderful, powerful ecosystem with all of the tools that you possibly need to build something. Now, that's how you get to the barrel of oil. Well, cool, Bill. How do you get to the dang shoelace? On top of Apollo, we have this little thing called launch pad. And launch pad is, well, think about like a car frame, right? It's pretty much done. But Samir comes to me and says, Bill, I need this really cool specification, Agente AI thing built. The cool thing about that is that in Apollo, it's kind of already done. The car frames done, the shocks are there, the tires are there. You might just need a little bit of a tweak of an engine, maybe some leather seats, some tinted windows, and boom, you're often running. So what we've effectively done is simplify the adoption of artificial intelligence, Agente AI, traditional AI and ML to actually something that's practical. What the heck does that mean? We refuse to build an elephant on a unicycle. Looks great. But it's not going to do very much of your business. And in that sense, we have all sorts of different kinds of organizations, manufacturing, financial services, healthcare, banking, government coming to us and saying, Bill, we don't want to do AI in the hyperskills. We can't, right? We can't even use coal pilot. We're afraid of the API. We just don't want to put the stuff in the cloud. We want to do AI where our data already lives. And that's why they would deploy on top of Apollo in a completely isolated, sovereign AI type of architecture. But it's not just the platform. It's the practical application. But so my team does the research. We do all the homework. A company comes to us and says, Bill, our engineers spend 100 hours a month reviewing these specifications. And we can't put them in the cloud because they're highly confidential or I know there's some compliance bound things that are wrapped around it. And in that situation, we will validate proof of concepts, build something for them and say, you know what? Your theory is spot on. This will cost you $1,000 to build. Well, just estimating, right? But it's going to make you $100,000 in six months. This is a really good use case. You should build and listen. Sometimes it's the opposite. Sometimes a theory is just way too outlandish and just a little bit ridiculous where it's just not going to work. And before a company even spends a dime, we don't want these wasted resources or a GPU even spins up. We go back and say, listen, we need to tweak something. Change your data. Change what the shoelace looks like because right now it's not going to tie your shoe. Now that's what Apollo does. We provide the factory floor, but also bits and pieces of the actual solution upon which you can build. So my clients are data centers and telcos that want to go after a very, very fascinating market. All sorts of different kinds of individuals that are trying to leverage artificial intelligence. And the theory there, some here, everybody listening, right now, the market's really broken down into some really big segments, right? 80% of our efforts are used to train the next big foundational model. That's next like OpenAI or Bard or Deepseek or Clawed. Insert your favorite large language model here. And 20% of it is used for inference, which is the practical application of artificial intelligence. Well, I'm going to be willing to bet that in the next 24, the 36 months, that's going to flip right on its head because eventually the investors are going to be like, cool, cool, cool. Can we actually see how this thing makes so money for us? Like how, how does this actually work? So we're going to spend those resources that 20 to 25% on still building models, but you better believe that 80 to 75% of those folks are going to start to apply that. We're starting it early. We're getting some lessons learned. We're actually deploying some real use cases. That's what we're trying to accomplish, especially with the the world that we live in. So I guess to put it in a very simple pardon the crude analogy, see what I did there? I mean, basically, the oil thing. I've got plenty of dead jokes. I got a seven-year-old. That's a professional eye roller at me at this point. To put it into that oil analogy, right? We're trying to find how do you get from a barrel of oil? Because there's no going back. There is no going back, right? And as far as like a corner called bubble is concerned, we can talk about that if you're curious, but you need to understand Samir and everybody listening. I promise I'll be quiet here in just a second. Since 1998, you, everybody listening, you've been trained to interact with data in a certain way. Samir might go to his favorite search engine. It could be altivista, maybe even ask Jeaves. You know, actually tried doing that bit in front of some high school students, ask Jeaves, and they're like, "Is this what it was?" He's asking his Butler for information. Who is Jeaves? That in the world, rolodex can't be used anymore in front of high school students unless you want to be thoroughly embarrassed. But right now, the crazy part about that, right? Let's use some practical use cases, right? So, September of 1998, Samir would go to his favorite search engine. In this case, he would be Google, because that's the first time you could ask Google a question. And you and I have been trained to interact with data through a very certain means of Lulink. That's it. God forbid, you find yourself on page two of Google. Forget it. At this point, you're lost. Just throw the ring in the fire in the mortar and run, Frodo, because you shouldn't be there. But that's the wild part of it. We've gone from blue links to wholly original contextual answers. And I'm phrasing a little person listening to this podcast right now. All of you are users of Generative AI by simply going into Google or Bing and asking a question. And you see now that first half page, it's not a Blink anymore, is it? It's a Generative AI response. That's how fast all of this has happened. This transformation for all of you listening, including myself, including Samir, all the wonderful folks at NRF, that's the future that we're going to be holding onto. And right now, we're just trying to figure out, what does that, what does that car ultimately look like? Yeah, I guess that takes me to my next question, which is what for you guys as a company, what's the main or one of the biggest challenges that you guys are currently facing? It's fear and certainty and doubt, right? And I think anything that goes along with artificial intelligence, especially something that's net new like this, there's still a lot of uncertainty. And you hear about it from the mainstream media, whether it's a bubble or whatever the case might be. And we can talk about physical infrastructure. A lot of it revolves around security. A lot of it revolves around practical utilization of artificial intelligence. Like, how does this actually make me money or improve processes? And there's still a vast amount of unknown, but we've learned in terms of the agentic or simple AI apps that we've created. For now, they haven't really replaced anyone. What they've done is they've given back something truly finite. And it's time, right? All these engineers who are still there, they're just doing more valuable things. The biggest challenge for us, honestly, and we're lucky, we're patient. It's the sales cycle. It's just, it just takes time for people to understand just how crazy impactful this this technology is. I'll give you an example. We're working with a fairly major oil and gas organization out in Texas. And they receive specifications like 60, 70 specifications to build one or two components that are confidential, heavily compliant, can't put the stuff in the cloud. And it's a learning process. Everybody, it's a learning process. And we built a little spec AI agent. You can actually check out the agent on our website. You can go to Apollo.us, click on launchpad and you'll see the spec AI agent. You could read about it right there. That situation was extraordinary. We take an agent, an agent, a AI that is able to see all of these specifications and whatever language that you want and translates them into something this engineer can actually build. Now, this is a learning process. There was nothing out there. They could ever do this before. But now, they're like holy cow. This saves us hundreds of hours a month by simplifying this process, but also it's a gen A.I. engine that learns our business and builds upon the capabilities of what this AI can do. Again, that's the big difference between quote unquote traditional AI and this gen A.I. stuff that we keep hearing about. But again, some year, the use cases are extraordinary eye opening. I mean, we can call it whatever fun, fancy, happy word that you want to call it, but because this is so new, because we're still exploring our imagination of what we can do with these technologies, what is possible and what's practical. I think those are two really important things to consider. That's a challenge. This isn't something like you could just go buy off the shelf. This isn't Microsoft Excel or PowerPoint, right? This is a technology that could potentially revolutionize the way an organization does business. And it was something interesting. So Kevin O'Leary from Shark Tank was one of my keynote speakers at Data Center World this year. And I had a chance to spend about an hour with him backstage. First of all, he is a very nice dude. Super, super interesting. We talked about guitars and wine and his own custom Gibson that he's getting built that says Mr. Wonderful on the neck because of course it does, right? And he told me Bill, unless you're investing in two specific areas, you're completely missing the boat. And that is data centers in real estate. And then he said something very profound before he went up on stage and he's like, Bill, I'm building this eight gigawatt facility. I'll know Bird to Canada and Wonder Valley, right? And he's like, what's remarkable about that is people think that my biggest clients are going to be like the Lockheed Martin, the Coca-Cola, the Boeing, the big organizations. When really the power of this technology is going to lie within the small and medium enterprises, because they're going to be able to invest not tens of millions of dollars, but a little bit of money to build a genetic AI application solutions, which Drumroll are going to allow them to compete against the Microsoft's, the Lockheed Martin's, the Boeing's, the Coca-Cola's, the world. So what we're going to see is a massive uptick from those mid-sized organizations in using these technologies to carve out entire new segments of market customer bases, solution selling and everything in between. But that stuff takes time. That stuff takes understanding when the Model T4 came out, you know, you better believe people then unfortunately crashed into curbs and trees a whole bunch before they figured out how to drive that thing straight. And I'm not saying that's where we are, but you know, there's going to be a few bumps in the road. And that's a learning process. And that's not something that we mind doing because ultimately, whether we win the business or not, we're creating more educated people. We're creating people that are less afraid and more curious. Very cool. I like that, Mr. Wonderful was building a data center in Alberta. That's where I grew up. Get out of here. Yeah, no, you should check it out. It's called Wonder Valley. And of course, it's called Wonder Valley because Mr. Wonderful was building it out in Alberta, Canada. But listen, it's not without its challenges. Eight gigawatts of power, right? Again, that's basically going to Canada and be like, cool, give me the power for Los Angeles. Give me the power and the Canadian governments like we've got the connectivity, we've got the lab, we got the water, we got everything you need. Where do you think we're going to get a gigawatts from? So obviously, Mr. Wonderful is looking at obviously things like behind the meter, kind of off-grid power solutions, straddle natural gas. He's going to put up some really cool natural gas turbines. And, you know, drum roll, he's going to be able to build it and those clients, you're going to come. But there you go. And it's in this, we're something we're going to see happening all over the world. This is kind of critical infrastructure being built. Sovereign AI, that's going to be a big one, as more countries and more institutions really want to grab a hold of their data and their models, how it's being trained and how it's being used. And a lot of that's going to result in more of these sort of private type of AI deployments, which is great. I think it's wonderful. So everything you're saying makes a lot of sense to me. And I think at our firm and in my practice, we're seeing it every day, just different ways that people are incorporating AI. And for many of our clients, they're demanding that we find ways to utilize AI to be efficient because they're doing it internally and they want their counsel and their other advisors and parties that they work with to do the same thing. But on the same time, we're seeing just recent criticism of this potential overbuilt and this bubble. There's a lot of talk about this bubble. It's hard to feel like we're in a bubble because everywhere, everywhere you look, this great demand is there. What are your thoughts? What are you sharing? What do you think we're going? I love this question about, are we experiencing a bubble? And we try to do often enough, and everybody listening, this is your moment to really sort of try and grasp or wrap your hands around this. So if you're on your Amazon and shopping, let's come back here for a second, because we're going to talk about literally the number one question you can see on every single financial and new station. When we start to take a look at this market specifically, I feel that obviously we need to learn from history. I'm not saying exclude that, but we're trying to compare what we're seeing right now to previous experiences, whether it's the.com bubble, the subprime mortgage situation that we had, et cetera. And I was recently in Las Vegas for the Schneider Electric Innovation Summit and I had a chance to tour my alma mater, switch data centers and some of the really cool stuff that they're building. And we were doing a Q&A session with their chief strategy officer, Jason Hoffman. And the question was asked, what do you think about this bubble? He brought some really interesting points up. So when, first of all, yes, you always need to be concerned. You should always err in the side of caution, question everything, trust nothing, that's the zero trust security coming out of me. So you should always validate things. But this is the important part. When we take a look at bubbles in general, what I want everyone to sort of do your own homework on is look for the catalyst, look for the catalyst of why that bubble happened, right? And we started to take a look at the subprime mortgage situation, right? We see that people were taking out mortgages that were directly not aligned with how much income they were making. And so as a result, you know, eventually that thing flopped over because these people couldn't unfortunately afford their homes. And in the.com bubble days, that was a result of, you know, overbuilding potentially, right? And not really understanding what the scale of that architecture was. But even ultimately, it's not like the internet one away just changed. So when we start to take a look at this, this market right now, and I know this is going to be released soon. And just just a week or so ago, we were all very worried about this bubble. And Nvidia came out with their earlier reports, there were 60% up, right? And so that well, the conversation for a while, right? For a while. And then it comes, comes bubbling up again. Let's look at an example. Let's try to find this catalyst together really quick. So open AI, open AI is a $20 billion revenue organization, right? That's approximately plus or minus. You think that just going to go away overnight? Like it's, it's not 20 billion in revenue. It just doesn't just disappear. There just has to be something absolutely catastrophic. Catastrophing happens. And that can happen to any industry, right? So you can't really call that a bubble. It can happen to anything. But something absolutely catastrophic would happen to happen for that 20 billion to go away. Realistically speaking, if a company like AI were to start quote unquote winding down, that would take years, years to happen. 6789 years for that 20 billion to ultimately go away. That's not so much a bubble. That's like letting the air out of a balloon. Really, really, really, really, really slowly. And you're going to see that balloon shrinking before you. So you can make some decisions around it. The reality is you have to take a look at the broader aspect of it. There's been nothing in our history of humanity that has ever seen this type of adoption. I might be talking about chat GPT, everybody. I'm talking about just generative AI, this new type of artificial intelligence in general. And the other thing that we have to remember is that generative pre-trained transformers, GPT and LLMs, here's a secret, they've been around for a while. They just weren't very effective. Why? Because we didn't have the hardware systems that could fully utilize this type of architecture. Well, drumroll, now we do, and another drumroll, they're getting even better. So while we're building vast architecture, these new up-and-coming solutions from GP200's, GP300's are more efficient. They're smaller. You will need less space, obviously, maybe more power. But we're already working around optimizations. It's difficult for me to define what a bubble would look like. I will comparably say there will be ebbs and flows, ups and downs, just like with any kind of market. But from a reality of the situation, is we're still seeing demand. We're seeing easily the largest capital outlay of funds that we've ever seen in the entire industry. I think overall, this was a really interesting report for McKinsey in the next five to six years, global infrastructure investment. That's not just data centers, but everything else that supported is going to be $106 trillion. This is what's known as a generational investment opportunity for so many different industries and verticals and people in different businesses. I don't think there's going to be a quote-unquote bubble in any traditional sense that you all might be thinking of. I think there might be obviously market expansions and contractions, but nothing, nothing that would be defined as like a.com bubble bust or even like a subprime mortgage bust. If we see ebbs and flows, it'll be a natural progression of the market. But please understand, the foundation of how we interact with data has changed. That's a fact, right? That's it. We're not going back. We require physical infrastructure and systems to support this change. So that's not going to change either. The way we design facilities and and data centers, all these crazy pieces of information. infrastructure that we need every day to connect, that's not going to go either. And if anything, we're seeing new use case like AI at the edge, inference at the edge, net new markets emerging. And most of all, here's where I really don't think this is a bubble. Our industry, data centers specifically, aren't just power consumers anymore. We're power producers now as well, right? So we're actively connecting into the grid. And we're becoming a much, much bigger part of it. I know that was a very long answer. Some of you are in everybody listening. I don't think there's going to be a bubble, certainly not in the traditional sense that you and I might be trying to define it. I do believe there are going to be ebbs and flows in the market, certainly just like anything that's natural. But my goodness, if this is a time for you to look at some kind of investment, I recommend it. I got to say, I completely agree with you. When I think about just the difference in the way that people can do whatever it is, whatever the thing it is that they do every day. And how much easier AI can make their lives is already making their lives in so many different ways. The analogy that comes to my mind is like, if we were living, say at this point, 100 years ago, and somebody was proposing the car bubble or auto manufacturing was just a bubble. And we're going to go back to carts and horses as soon as that would be incredibly exciting about. Because here's this thing that has the potential to make your life so much easier, better, more efficient. Does it take up additional resources? Absolutely. But is there any going back from that? I don't think so. I love your car analogy, right? And even back when the Model T4 and all that came out, right, there was pushback. But ultimately, what the saying is, the car saved the horse. I like that. Yeah, I like that. So I guess on the topic of more resources, we've talked a lot about the amount of power that's necessary to power these data centers. And it really is a shocking amount of generation that's needed. We already have an infrastructure that's old assets that need to be retired, even without these data centers. We would need a lot of power plants. But with these data centers, we need multiples more than what we would have needed before. There are potential issues around water and water consumption. How are you seeing this issue being tackled? And how can folks who are concerned about this get comfortable with how we're going to resolve these issues? I don't think anybody should get comfortable. I don't think anybody should be, oh, I got this. I've figured out I'm in a good place. I'm comfortable right now. For the first time in humanity's history by 2028, demand will be greater than peak supply. And you should think about that, right? So the Devons Paradox, right? Where efficiency is going to pace consumption is out the window. And as a result of that, by 2033, we may experience a shortfall equivalent of enough power for 100 million homes. And 25 years from now, in 2050, electricity demand, just from our industry alone, is going to more than double. Now, that's not to scare anybody. I don't want to put like this whole fear and certainty and doubt in anybody, people like listen, oh, you're going to consume all of our power and all that stuff. It's not easy. We certainly need to understand the ramifications of this. And I often get-- that's the first question that I usually get, Bill, is that you've taken up too much water, our power systems, and so on. Our rates are going to go up. That's not entirely true. Over the next, I think, 24 to 36 months, we're going to see a revolution. In-- no, forget that word. Scratch that. Don't delete it, because it's about a renaissance. We're going to see a renaissance in our industry, right? And I really do mean that, right? Because there was an interesting statement from a Rob Roy, the CEO of Switch Data Center. He's like, I don't like how we measure rack density in this kilowatts per rack. And you know what? I think he's right. We're going to be reimagining what these things that hold these AI servers, what they're actually going to look like. But I digress away from your questions, Samir. I think I truly believe that we're going to be seeing-- because back up, we don't have to corner gigawatts of power. We don't. That's a silly number, right? We certainly don't, right? And our grid here in the United States really hasn't experienced a meaningful update since 1960s, 1950s or so, right? So we're seeing these isolated grid architectures, but you better believe that very actively, our infrastructure, these data centers are becoming an active part of the grid environment, right? So these energy producers. So going back to my Switch example, we learned that Switch Data Centers, you ready for this, consumes a third of all of the power for the entire state of Nevada. I do want to do that. You heard that, right? And if Switch were to turn off and produce generation into the grid, they could power the entire strip, right? That's how much power they consume and can also generate. But here's the kicker. As a result of their efforts and their direct alignment with utility economics, ratepayers are actually paying less for energy because of the relationships they have within Nevada utility commissions. I firmly believe that we're going to be seeing those types of relationships in the near future spread across many, many new and emerging markets, North Dakota, South Dakota, West Virginia, Indiana, Ohio. I'm literally named Utah. These are emerging markets that are already double digits growth in the amount of energy they're going to be consuming. So I think the really very visible silver lining here is that these facilities are going to actively try and go after natural gas, clean natural gas. They're going to go after clean energy, like nuclear systems, right? We're going to see things like micro grids that are capable of aggregating different power sources to give a base load of power to these facilities. And here's the other reality, a lot of these facilities, not all of them, please understand. I'm not trying to speak for every single data center out there, but many, many, many, including my friends at Switch at Compass Data Centers, for example, they are actively going out of their way to become more sustainable, to become greener, just imagine for a second out in Texas, Compass Data Centers is building a 400 megawatt facility that there's 10, 40 megawatt halls. And each of those halls, these massive facilities, is 70% prefabricated, right? So it's already done. Just need to put it together on site, which results in less trucks being on the road, right? They do cement and concrete batch mixing directly on site, which means they don't have to outsources done directly there, which all these little things are wonderful. Now drumroll, they also use this thing called carbon cure. It's an AI platform, ready? That's a co-wester's carbon from the atmosphere and injects it into the concrete mix. How cool is that? Yeah, you didn't think you'd be on a podcast here with Samir learning about concrete and cement mixes and how neat it is, but yeah, here we are. But that's how creative and innovative these facilities have become. But again, going back to your point, look, we're trying to create systems that are what known as closed loop water systems, so use the water once and keep recycling it as much as you can. We're building these facilities that are at 60 decibels or less, guys, that's this conversation. That's really this kind of a back and forth. It's nothing louder. 60 feet in height, we're putting shrubbery to make sure that we're not ugly, for example. We're trying to look at ways to reuse heat and put that back as a usable heat source for buildings, homes, even potentially. So we want to be good neighbors. We are actively trying. I was every single data center doing this, unfortunately, no. But I do believe as a proliferation of this industry continues, we're going to have to become better neighbors. We're going to have to look at better ways to be more economical and sustainable because ultimately that does help the bottom line and it does help an organization become more profitable because simply put, they're more efficient. Now, it's not perfect. We're trying to get there. But remarkably, I challenge everybody listening to do your own research and understand that we're not just energy hogs. We're not just out there trying to consume millions of gallons of water. We're trying to be good neighbors. It's just that we're not very good at PR. We're getting the message out there. And that's probably why I'm grateful to be here with you, severe to spread that, that goodness. Yeah, I know. I'm really glad that you showed that because we were earlier this year at some meetings with one of the largest owners and operators of data centers out there. And the entire conversation was about sustainability. And they want to do it in a way that makes their data centers that are not neutral or not negative. And there are so many companies out there right now. We're running into a bunch of different founders who are focusing on ways of making this entire sector much more sustainable, everything from the way that equipment is fabricated and manufactured to the delivery of that around the country and from around the world. Cooling technology is all sorts of cool things that are happening out there in this space. So I've one last question for you. And that is for folks out there that could be college students. They could be folks deciding to transition in the careers from wherever they're at now to something similar to what you do. What advice do you have for someone coming into the data center sector? What a great question, Samir. First of all, I'm hoping that at the end of this, My LinkedIn inbox is is blowing up with new connections. and friends, please do find me. I'm an active advocate of this industry. I do a lot of videos and podcasts. The whole, the whole steaver-win thing. It's as funny as it is that's true. You'll see me do videos. I've done everything from loading the data centers. I'm not kidding. They converted a barge into a floating data center. You gotta check it out to power systems and everything else. You can take a look at those videos and just educate yourself a little bit as well. Outside of that, I recommend a couple of really fun organizations that I work with philanthropic, not-for-profit organizations. One is called Infrastructure Masons. A lot of really good resources there. And the other one is the Nomad Futurists. Both of them are wonderful. I'm an active participant in both of those. I'm a board advisor for Nomad Futurist and I'm the co-chair of the People Committee over at Infrastructure Masons. Both of those are really good valuable resources if you wanna learn a little bit more. Obviously, you can read the trades. But I think the one thing that I'm gonna mention everybody, was a great quote from Ted Lasso. It was a great quote from from that TV show if you've ever seen it be curious, not judgmental. Okay, and if you hear something about our industry that's negative or just really bad, ask why, or what's being done around it, or do your own research to really ascertain your own opinion and your own knowledge about what's happening in this space, what's wild about the data center industry is that we've become so incredibly mainstream. I mean, this in South Park is doing episodes around AI and data centers, which is kind of like. - Rare. - That's when you know. - Rare, exactly when I'm gonna reference South Park and something you can use as a cultural reference. But that's my big takeaway here is, there's always gonna be fear and uncertainty around any new piece of technology that you could possibly put forward. But please understand that your users of this stuff, there's really no going back. And the best way that you can feel comfortable is by asking question, approaching everything with fearless curiosity, right? And if you're a business, just like Simeer talked about, if you're a business and you're looking into this, and you're wanna do something, I'm gonna leave you with one final thought. Vision with that execution is just hallucination. And that's not something that we want our AI models doing and that's certainly not what we want in your business doing as well, right? So if you hear something today or in the industry and you think it can make an impactful impact on your business, your life, whatever it is you do, see what is possible, see what the journey of AI where it can take you because I'll tell you, you can do a lot with oil, right? You can do a whole lot with AI. Hopefully obviously we're developing these systems that are much cleaner, much better and much more impactful to do something good for the human race that's certainly our prerogative. And also work with partners that truly mean that, right? So we had a pilot, we were built on three core tendencies, AI ethics, AI transparency and AI sustainability. Are we perfect? No, I'm not going to openly admit that. But we actively try to go after those core tendencies and the pillars of what we were built on, use partners and leverage partners that believe in doing something good. But most of all, you know, ask your own questions, do your own research and you'll quickly see that curiosity really brings so much value in terms of your own knowledge and understanding these systems that all of a sudden that fear just kind of goes away. - Love it, appreciate the advice. I really appreciate you being on, Bill. Thanks for the insight and I hope I see you soon. - I hope I see you soon as well, Samir. Thank you for letting me be your first guest and everyone, thanks for listening. (upbeat music) - Thank you for listening to UpTime now. If this episode sparked an idea, pass it along to someone in your network and leave a comment. I'd love to hear your thoughts, ideas and recommendations for guests. Till next time, thank you and remember to like and subscribe. (gentle music)