 Welcome to JSATV where we're covering the latest stories, trends and innovations from the leaders in global connectivity, digital infrastructure and the networks within. And we are coming to you live from Times Square and beautiful New York City at DCD Connect. And I'm here with who is soon to be my best friend here at DCD Connect, Mr. Bill Kleiman. Bill is the CEO of Apollo. Bill, welcome to JSATV. You are no stranger to JSATV. No, not at all. That was one of the most energetic introductions I think I've ever had. Thank you so much, Gene. It is a pleasure to be here. What's up everybody live on LinkedIn? Hey, drop a comment. I kind of know what this looks like right below. There's a like and a comment section. Yes, you can tell me I should have less caffeine. That's totally fine. But hey, Dean, it's really great to be here with you. There are like a million people walking around. So I'm so sorry if I wave at somebody. It could be you or it could be my friend Yuri over there. So thanks for joining us. Also, hi, Yuri. So, Bill, speaking of energetic introductions, why don't you introduce our audience to greener data? Oh my gosh, I'm actually glad this is the first question. Hang on a second. I'm facing the microphone in the right direction here. All right, take it down like a little bit. So greener data is a really amazing project that started off, I believe, last year. It was a journey to get thought leaders from our industry to talk about, I mean, I'm not going to be honest with you. All of the stuff that you see, the way you're even watching this depends on data centers. So what we want to do is talk about what does greener data look like? How do we create an architecture that's more sustainable and share ideas and thoughts from brilliant thought leaders on what they're actually doing to establish that kind of infrastructure? And drumroll, it was an Amazon bestseller, so... It absolutely was, as will volume 2. I can almost guarantee that. Exactly. So JSA said, hey, this is pretty good. Let's get some more brilliant thought leaders together and put together a second volume. I'm really excited about it. Our chapter, the one that I wrote over the past year, I've been working in the space of AI and data centers. And I'm not talking about the stuff from like 20 years ago, by the way. It's literally like the actual implementation deployment of the technology is happening right now. The most advanced technologies as well, Dean, like the generative architecture, retrieval augmented searches, all the really crazy acronyms that you and I sometimes don't know yet. And I talked about that a year of nonstop innovation, I had a chance to discuss everything that we learned, what we learned in the data center industry, where there are leaders in the space, where there are sometimes lagging behind a little bit and how you can catch up. It's been an extraordinary journey. I'm so excited for this book to come out to hear and read about the other thought leaders and what they put in there. But, oh my goodness, if this is an opportunity for you to learn of how data and greener data is going to impact you, especially on the AI front, you got to read it. Bill, that was the finest introduction to the greener data project that I have ever heard. Thank you very much and thank you absolutely for your contribution to the next edition of greener data. But aside from all of that, let's talk a little bit. You mentioned we don't let anybody leave the interview without discussing AI. You are no different. So let's talk about AI in the data center. What does AI in the data center even mean? Wow. I know we don't have enough time here to go into all of the details. But, you know, when we take a look at artificial intelligence, obviously, AI is like the big name right now. Everyone's talking about, but let's be perfectly honest. Dean and everybody listening, AI has been around for like 20 plus years. And the biggest difference right now is that, you know what, I'm going to use Pugs. Pugs is a way to explain the difference, right? In the past, traditional AI, now please understand this is a very high level explanation. In the past, I would have Dean, I would sit down with him and I would show him picture after picture after picture of really cute Pugs and French Bulldogs. I do love French Bulldogs. Alison likes French Bulldogs as well. And so eventually I'm going to take all those away and I'll ask, hey, Dean, what do you think the next picture is going to be? And Dean is going to say, well, based on what you showed me, I bet it's going to be a black Frenchie or maybe a fawn pug just based on the stuff that you've shown me. Now, generative AI is exactly the same thing except now I take all those pictures away and I go to Dean and I say, hey, draw me a pug. Generate original content. So that is the biggest shift that we've been seeing. So old school legacy pattern based recognition to something that we can generate. Now, before I go on AI in the data center, why is this important? What we're experiencing is not a technology shift. It's a shift in humanity. Please follow me on this for a second. Thank you, Bill. Yes. It's an important point to understand. You, me, Dean, we've been conditioned to interact with data in a very specific kind of a way. Dean goes to his favorite search engine, Alta Vista, Ask Jeeves, Yahoo! And if he's feeling feisty, maybe even Google. I can't believe I gave you the last Ask Jeeves. If any of you remember that, drop a comment. Yeah, yeah. And he'll ask a question. And for 20 plus years, it's been a blue link. A question and a blue link. God forbid that Dean ends up on page two of Google. That's like Mordor. Nobody belongs there. And so now, if you use Google and being in the past a few months, congratulations, you're a user of generative AI because the first response you get is a generative AI response. In the data center industry, what we've learned over the past year is that we have been working in this sort of 5, 10 KW array. Now, if you're new to that, it's basically how much gear you can put into a rack. The challenge is that AI infrastructure is way, way more dense and requires you to rethink your environment inside of the data center. So all of a sudden, all of these data center leaders here thinking about AI are like, all right, hang on a second. I can't do business in traditional ways. I need to shift and shape in a different manner. The challenge for our industry and everybody. Now, this is from one of my mentors, Peter Gross. All right, everybody. The data center industry loves innovation as long as it's 10 years old. That's right. We don't have 10 years. We barely have 10 months. That's how quickly this pace is evolving. So the question becomes, how do I impact my customers? How do I impact my use cases and my workloads? How do I shift my business to going from supporting email servers and exchange databases and pizza box servers to actually working with artificial intelligence? And that has been one of the easily the biggest impacts that we've seen on the data center industry is the challenge for them to understand how they can impact the market and where they can shift and shape that industry based on their requirements and their customers' requirements. Once again, an absolutely amazing response to the question. With all of that in mind, how obviously monumental shifts in the infrastructure necessary to make everything that you just said happen. What does that look like? So I don't want to scare anybody. So we've taken on a few... We're already scared. We're already scared. AI is coming for you. It's not. AI is not coming for you. Oh man, it's his life. We can't... Now listen, if my dad sees this, he's going to be like that. Bill Kleinman, he knows what he's talking about. No conspiracy theories here. It's funny. A lot of the things that we've actually built has been more like assistant levels. But when we start talking about use cases and architectures around AI, what it means for the industries is a new way of thinking. It's a new approach to how we work with data. And the challenge becomes how do we better understand it? How do we better apply use cases and technology? It's an extraordinary time to be in this space. It really is. I was talking to my daughter. She's a sophomore in college. And I was talking about AI. My kids just kind of intuitively kind of know this stuff. And I'm just like, it took me a long time to figure this stuff out. But we were talking about it. And I said, AI is going to be the thing, the engine that saves your dad's life one day. Because maybe it's the augmented reality. It's the machine learning. It's all of those things that are guiding the surgeon's hands that are making the proper diagnosis, that are doing all of those things better, infinitely better than we have ever done it before. That's what the future feels like and looks like to me anyway. It's extraordinary. And you mentioned something really important. One of the things that we do at Apollo is that we, listen, we can totally create an AI architecture that takes a unicorn riding a unicycle. Is that going to make you any money? Probably not. So what we've been trying to approach is we'll go to Dean and we'll say, what are you trying to solve from a business problem? What data do you have? And then ultimately, what measurable outcome are you trying to get to? And then how do we use that data to use generative AI, all these new kinds of technologies to help you achieve something that you've never been able to do before? Yeah. In a big way, we want you to have that childhood sense of wonder. Remember how to dream again and broaden your imagination because that's what these technologies are effectively is they're eliminating your limit of what you can do with data. Now, you can do cool things that don't make you any money. I guess that's fun. But what you can really do is apply these technologies to extraordinary use cases, saving lives, predicting energy costs with levels of probability. Climate crisis. Being much more effectively able to manage these kinds of things or see more deeply into environments. That's where AI could. Obviously, there's challenges and there's faults. And we can talk about the actual impact, the physical infrastructure impact on that. But from my perspective, I think AI is extraordinary. You know what? I'm so glad that you were talking about some generational shifts in thinking about technology, where it comes from and all of those kinds of things. I just had Phil and Nabil on from NYI, the Nomad Futurists. Very, very fun conversation with those cats. But that's kind of where we're going, right? I mean, it is the future generations that will really reap the benefits of what's happening right now in the data center. So for those future generations, one of the things that I want to talk about. So everybody listening here, you are a user of generative AI, whether you like it or not. ChatGPT took five days to hit a million users. That's faster than any other application that we've ever seen in history. We don't have a precedent. Now, I'll caveat that with, I just lied to you because there was one app. One app that hit, I think was one million users, and I think it was five hours. Threads. Interesting. Right. And that is the appropriate response because I don't really use that that much. But ChatGPT, after just a few months, had over 100 million weakly unique users accessing the application. Now, I want to talk a little bit about what that actually means from an environment perspective, right? Greener data, making sure that we support this. So follow me on this for a second. A single Google search can power a 100 watt light bulb for 11 seconds. A single GPT-like instance. I say GPT-like because our company, we don't use ChatGPT, we develop our own kind of architecture. But a single GPT-like instance is 600 to 800 times more powerful than a single Google search. Now, that Google stat was from 2011. And when we do GPT-like searches and services, it's not just one query. They're done in processes and batches. Six to eight, six to eight of them per process. One to two kilobots of consumption per process. It is an extraordinary amount. Listen, every single time you enter in ChatGPT, you can charge this phone 40 to 60 times. One time, you ask it to make a really cute French Bulldog Dalí image. That's like, I have done this before and it's really cute. That will allow you to charge your phone at least one to two times. There is a massive impact that we have to understand, which is why here at this conference, while we're talking about our greener data, it's actually in the book in my chapter, the journey. We've taken many of these people here. This is something we say at Apollo, from what? To token in understanding how to restructure your environment. Now, Dean, before I let you talk, you said something, it's not scary to enter this. A lot of our data center partners have gone on a journey. A hybrid environment is the way to go. Rear door heat exchangers, better containment, better visibility into your environment. You don't have to dunk your servers into immersion in liquid cooling. It's totally cool. We are going to talk about that later. We are. But your method of entering into the AI space does not have to be scary. We've seen extraordinary success in taking data centers, traditional co-location retail data centers from 10, 15 KW into the 60 KW range. And they're like, cool, my building's still here. My customers are way happier. And I've got a new approach to this technology. Yes. I'm breaking all kinds of rules, but I was on a panel only because we're going long. Break them. Break those rules. And there's nobody here to beat me up or anything like that. I don't think right now. I don't see the virtual hope yet. Allison, I'm going to keep going. Is that okay? Cool. Thank you. We're having fun. Okay. So I was on a panel yesterday, and basically we were talking about colos and enterprise expectation on the colo with regard to AI. And I know it was deep, heavy, heavy stuff. But from your seat, do enterprises really know what they're asking of the data center with regard to AI? Wow. It's fascinating. It fascinates me. What a question. So Apollo, we were recently launched. Yay, go team. We're a VC backed, I would say startup, but originally we've been around since 2019. We had to move from Eastern Europe into the United States. So we're a restart up effectively. And what we've learned is that we are largely being driven by these enterprise customers. They want data locality, proximity. They want to be able to control their costs. There's organizations like the government, healthcare, FinServe that sometimes can't put some of these AI and trading models into the cloud. They're absolutely driving the conversation. So right now the challenge is that enterprise goes to a traditional co-location retail data center and says, you have my workloads. You have my SQL server, exchange boxes, VMware, Citrix, all this stuff. Now I want you to take my AI work. The typical response is, I can't. Because A, I don't have the software tool, which is what we do. And B, I don't have the density level. Usually if you're at 10 to 12 kW per rack, and one NVIDIA DGX node, it takes about six units in a rack, is about 10 to 12 kW. So what are you going to do? Put one in there and a whole bunch of panels. That's not going to look very good. So these enterprises, they're hungry for capacity. They're hungry to work with trusted partners in the data center space outside of the hyperscalers. Now, Bill Klayman is not here saying that Amazon and Google and Meta and all of them are going away. Please, I'm not. There's plenty of market, but there is an extraordinary opportunity for funded startups, small media businesses, even large enterprises to leverage amazing relationships with co-locations and data centers to support those AI use cases and those use cases, Dean, and everybody listening. First of all, we don't have time to go over them all. There really is no limit. We've actually had to go back to customers and say, this is amazing. I don't know how you're going to make money on this, but this is an extraordinary use case, and we've gone back and said, this is amazing. Not only can you create this for your business, put your logo on it, productize it, sell it to your competitors, and they're like, that's the kind of stuff we can create. Bill, we were out of time five minutes ago, but I tell you what, we're going to have to do this again. Absolutely, Dean. Absolutely. Thanks so much for being with us. Thank you so much. Hey, thanks, everybody. Yeah, and thank you viewers for watching JSA TV. Stay curious, stay connected, and we'll see you soon. Bye, everybody.