Ep11: What Drives Data Center Investment Strategy

(upbeat music) - This is uptime now, a North and Rose Fulbright podcast dedicated to conversations about data centers, from development and construction, to cutting-edge technology and operational breakthroughs, and the people driving it all. Joining us today is Julian Eden, managing director at Infrared Capital Partners, where he focuses on originating and executing infrastructure investments across America's, with the growing emphasis on digital infrastructure and data centers. - Julian, welcome to the show. - Thanks for having me. - How are you? - Very well, thank you. - Been a while since we last shared it. - I'm sure you've taken very busy. - The weather in New York has changed, plus 50 degrees, minus 50 degrees, but that's part of the fun. - Yeah. So Julian, for the benefit of the audience, can you tell us a little bit about yourself? How did you get to be here? Do you always want to be an MD focused on digital Infrared, or how did you land in the space? - Yeah, happy to. Might not be a surprise, but I didn't grow up thinking, one day I'd love to be an MD talking about digital Infrared, I can't wait to talk to Samir about it on a podcast. When I was a kid, I really, my goal was to be David Beckham. That was the dream. - My name is Michael Jordan, by the way. - Michael Jordan, okay. Well, yeah, we have appropriate choices, or, you know, more kid stuff. I remember having a phase, wanting to be an archaeologist. I had this weird obsession with artefacts and skeletons, so I'm sure there's a joke about background fiber in there somewhere, but to be honest, what really drew me to the space in the first place was infrastructure rather than being on the investing side. I'd say I ended up on the investing side almost by accident. So I've been interested in economics and infrastructure, my whole life as far as I can remember. My mom and dad are both in the architecture and urbanism world, so I've always had an appreciation for physical assets and how they can shape daily lives, how they can provide a public service, and how they also, even when you don't see them, they're quietly keep the economy moving. So a couple of years into my career, I realized that, you know, if you're thoughtful about where capital goes, and if you can influence those decisions, you can actually have a meaningful impact on what gets built, on what gets built on how it's run, on who it serves. And so I've now spent over a decade of infrared, almost all of it here in the US, and I guess moving to the US is as far as it gets in terms of being like David Beckham, and much more than that. And I've had the chance to invest across pretty much all of the major infrastructure sub-sectors, their energy, transport, utilities, digital, which I'm lucky to have had the opportunity to do. I think it's a great vantage point, because once you do that, you start to recognize the common patterns of what makes an asset durable, what makes it fragile. So I like to call myself an infrastructure specialist and an infrastructure journalist, and to tell you a bit more about my shop infrared. So we're a global infrastructure asset manager. We've been investing in infrastructure for close to three decades, so we're very early to the asset class. Within that, we invest in the full gambit, so across strategies from core, core plus, all the way through to value add and growth capital. That last part is where I spend most of my time. Infrareds fully owned by some life, but we manage capital for wide range of investors. And the nice thing about that is that I benefit from having brilliant colleagues across the broader group. They might cover real estate, they might cover fixed income, and so on. So on a topic like data centers, which really cuts across disciplines, which really has been of interest to the broader market, you get a much larger perspective, and I think that makes for a better debate and better decisions. So as you've highlighted, my recent focus has been data centers in North America. We recently, last year we set up Q data centers. That was off the back of a carve out. What's interesting there, it's a brand new company, but the operation has been around for a while. The operation consists of nine data centers across Canada. They offer location and private class services, and there's a strong emphasis there on the sovereign Canadian context. So very happy to talk info and talk data senses with you today. - Yeah, glad to have you on and excited about this topic. So let's get right into it. You've had a broad set of experiences in infrastructure. How do you view investing in data centers versus more traditional infrastructure like roads and airports? Is it more appealing, less appealing? What's different? - Yeah, so I'm gonna focus on different rather than more or less appealing because it always depends on the lying investment. I'll say again, I started an infrastructure out of love for the assets, whether the roads, whether the airports, whether the power networks, whether the digital, because they're so tangible, but whatever work you do in life, it's always important to remember who your customers and what your customer's priorities are. And so for me, that means really asking myself, what role does that asset play in a long-term portfolio that a pension fund might hold that underlying everyday investors effectively hold as well as what role that plays in the real economy? So we're always very focused on the definition of infrastructure and very disciplined around the key characteristics that that has to hit. So you're looking really for hard asset, proven technology, barriers to entry, cash for visibility, and some linkage to inflation over time. What investors want to see is economic cycle resilience for infrastructure. So the definition is expanded, but even though you've seen this expansion, the asset classes become extremely popular and investors are now looking through it, not only for resilience, but also for return enhancement driven by these big mega trends. You have to remain true to the underlying characteristics. And so where I'm going with this is that, for data sensors specifically, I think it's important to say that data centers are not necessarily an AI play. Data growth and digitization have been going on for a long time. And what's happened recently is that the focus on AI has accelerated attention, accelerated capital flows, kind of like Jet Fuel on an existing trend, which is why the hype has been so intense. So, you know, coming back to your question, comparing that to say a tool road or an airport, I think the core difference is that data centers and the growth and data centers are tied to digital demand growth. And that means you can see why the range of growth pathways or why the range of outcomes, depending on exactly which subset you play and which location you play in. Whereas the tool road, that's much more directly tied to local economic growth, local demographics. But I would say in both cases, you have to be disciplined and you have to study what is truly durable demand versus what's cyclical, what's speculative, what's concentrated or not mitigated, which might not hit those infrastructure characteristics that I look for. - Yeah, one thing that we think about often is the use case matters and also not every customer is the same or built equally. Okay, so help, I guess, bring this to life for me. Help me break down how a private institutional investor like yourselves originate attractive deals in this space. - Yeah, very happy to give you the general framework as how we think about origination and I can explain how that played out in data centers specifically. So we like to think about origination in two prongs. One I call thematic and the other one I call systematic. So the first thematic, I would say that's almost an academic approach. So you identify trends, you identify subsectors, you identify business models that you believe meet infrastructure tests, that you believe offer attractive reward dynamics and importantly that you as an investor are well suited to sponsoring through execute. The second is systematic. That's really no big secret there. It's all about pounding the pavement, not just with advisors, but with the whole gambit of industry players, management team, et cetera, the holy co-system so that when something surfaces in a high conviction area, you're not starting from zero and you can move more decisively. And so for data centers, we really started looking at this at the beginning of the decade and we did that semantic work. And I would say through this, there were two areas, two subsectors that stood out as particularly compelling for our strategy, for meeting that infrastructure test and for our ability to deliver the best outcomes and the most compelling risk reward story. So on one side, co-location is a subsector that we honed in on. It's a well-established model. It's got a long history. You can study the chair and you can study the stickiness and so you can underwrite it with some confidence. You know, the simplicity of the service makes it more accessible, more infrastructure-like. The second is more of a characteristic of the type of assets of the type of platforms we look for. It's what we call or gets called edge characteristics, which really at the end of the day is a pretty simple sort of real estate, principle, location, location location. You know, location matters, proximity to the end user matters and to the customer matters. And on top of that, by being in metros, by being in more dense areas, some sites are generally hard to replicate. So you've got a natural barrier to entry that forms there. And, you know, being perfectly honest here, this wasn't off the back of some grand vision, not that long ago of AI shifting from training to inference, it is much more basic, right? Where's the digital demand actually anchored? And how defensible is the asset? So from that point, you know, we had a good roadmap. We targeted opportunities globally. It's the benefit of having a global team is that you can compare cost geographies with what's most attractive. And we invested in platforms with different remits. So some have focused on delivering builds. Others have focused on optimizing and existing portfolio, sometimes simply by using a variable power more effectively. But in every case, the common threads are infrastructure style resiliency and not to be underestimated. Partnering was excellent management teams who deeply understand and value their own customers. - For sure, very important. Can you talk a bit about why infrared has prioritized co-location versus other types of assets that are available in the space? Obviously, they're, you know, certain risks involved with co-location and in terms of, certainty of revenue. What is it that makes it a tractor for you guys? Yeah, it's a very good question. We get a number of times, particularly because a lot of the hype in the market is around hyperscale, and that isn't part because hyperscale is, as the name indicates, very big and commands a lot of a lot of capex. So just to be clear, at a macro level, I don't think the answer is either all co-location or hyperscale. I think they both matter. They're both essential to the digital ecosystem. And so the future is not about choosing one. It's about leveraging both. And I actually co-authored a piece with our head of research, General Caminella on this topic exactly where we as infrared have leaned into co-location today is because it tends to offer a particular set of characteristics that we like, flexibility, proximity to the edge, carrying neutrality, and multi-tenant dynamics, which is that diversification is appealing in itself because you don't have all your eggs in one basket effect, but it's simply a different demand driver set than hyperscale. So again, from a risk perspective, that can mean more diversification, it can mean multiple growth levers, but it can also mean a bit more operational complexity. Now, that's not to say that hyperscale doesn't have a, it's a place. Again, hyperscale is attractive because of strong credit tenants, high efficiency, long-term bespoke sites can be very sticky, but you have to be honest about the trade-offs, and you have to equip yourself to manage them. You have to accept concentration risk, capex intensity, and execution risk, assuming you're doing built-to-suit or greenfield, and that's where the broader group experience can help, and is interesting, hyperscale, built-to-suit can be attractive when it's de-rest, for instance, we're releasing and clear power delivery, and a focus on the execution length is how some of our colleagues in the wider group have approached large power land and power shell strategies. You've heard me talk quite a bit about how to address this risk, how do you manage this risk? I would say that that's one of the key lessons is the driver returns isn't just the label isn't just the type of risk. It's how those risks are structured, how those risks are managed, and whether you're best equipped to do that, and to date, collocation is where we have found the best opportunities and where we have had the highest conviction of being able to deliver the best risk reward proposition to our underlying investors. Yeah, it sounds like much of the world is about finding opportunities. That's where the risk reward is attractive, and within that, the opportunity and platform that offers the best risk reward proposition, has that equation evolved in the data center space? Yeah, exactly. It's that simple, and you've summarized my job in three words, which is a bit scary, but to answer your second question, absolutely, it's definitely evolved a lot, which weren't surprised anyone given the pace at which the sector has grown, has shifted. Data centers, they've moved from being seen as high-growth alternatives to real estate, to mainstream infrastructure. That means that competition has increased, and that means that you have less room for error and more emphasis on the writing discipline. So the conversation really shifted from is this demand real, is this demand sustainable, to, well, where exactly is this demand? How do we best address it? And what could disrupt it down the line? I can focus on two, the two biggest areas that, that always flag up as risk, power, and AI, on power. Obviously, it's no longer a background assumption and an afterthought, it's absolutely central. You will spend a lot of time on the writing certainty of access, timing, reliability of connections, backup solutions, and the regulatory framework that underpin it. You have to be prepared for delays. That's one of the golden rules of infrastructure in the real world. There are frictions, things take time, but we're very lucky in my shop that power is actually our largest subsector. We were very early in the renewables game, for instance, so we have a deep understanding on the dynamics and on individual power markets. In fact, before my previous two deals, we're actually a wind farm and an electric transmission network in Texas and Nevada, so power runs through the DNA of infrared. And then on AI, you know, from the outside in, it can be framed as a pure tailwind. And just for the record, you know, I absolutely agree, it's hit to stay and that it has run to grow. But I would say that that's not enough to underpin an investment thesis by itself, not an infrastructure anyway. AI, as I see it, you know, should be a source of upside, but it shouldn't underpin you on the writing case. Your base case is still rooted in a broader mega trend of structural data growth. And somehow maybe paradoxically, where AI has changed things is it's pushed us towards more conservative underwriting in the sense of stronger stress testing. So it's not just of the asset itself and its readiness, but also of the customers of the asset and how they could be disrupted. So I just talked about, you know, two major risks. And I feel like I've been talking about risk a lot because, you know, you ask me about risk, but you also ask me about reward. So, you know, I want to be clear that the strategy I primarily focus on is called value ad and it's called value ad for a reason. It's a, it's a really exciting space to be in because we get to grow these businesses, you know, invest in and be alongside on the journey, growing these businesses and these assets. We get to contribute to local and national economies and doing so. And so day to day that value ad can also take a variety of very different shapes. You know, in my remit, it might be investing growth capital to upgrade and densify sites or to expand them to deliver more capacity. It could be improving utilization, customer mix, customer tenure, customer satisfaction over time, or similarly simply providing the best platform possible for critical strategic services. So healthcare, government, financial services, including hosting their own AI applications and development in a secure environment and a high compliance environment in a sovereign environment, which is a big theme, particularly outside the US. And, you know, a bit like electric transmission or freight, as you, as you get deeper and deeper in the infrastructure space and you get nerdy or nerdy about it, you tend to like more and more those things that are that I like to think of as the dark side of the economy. And I think data centers are absolutely within that category, even though lots of people talk about them these days. Those are the things that no one normally thinks about, but are so important to enabling others. And, you know, that message transpires at the macro level and at the micro level. At Q data centers, actually, the branding strapline is infrastructure for the inspired. The original version of that before the marketing folks improved it was the boring behind the bold. And really, that's the key message, right? Where you're, you know, slightly boring, but reliable partner, and we can enable you, the other line user, the other line customer, whatever you may be, we can enable you to go and do the cool things that you want to do. Very interesting. You mentioned themes outside of the US. Most of the conversation whenever I turn on the TV or check the news or one of the many articles I get on data centers every day in my email is about conversations around the US market. But you guys have invested in Canada and Liberia, Germany, are those more attracted than the US in your eyes? And if so, what are the differences? Yeah. So just to be to be clear, I think the US is clearly a leader clearly at the forefront of the momentum in the conversation. And we're very active in this market in the digital space and in the infrastructure space generally, including in in data centers. But I will say at the same time, the opportunity set is global. And it's generally helpful to have a global team because it lets you compare markets side by side, right? It lets you sort of arbitrage or as the best risk reward opportunity here. So in terms of some of the other jurisdictions that we've invested in, starting with Canada, I think Canada is very interesting because it's not an either all with the US. It's a complimentary market, which has its own demand drivers on top of that, including sovereignty, including domestic hosting requirements in certain sectors and strong fundamentals and key metros. On top of that, you have a favorable climate for cooling. You know, remember, there's no greener energy than the energy that's not been actually consumed in the first place. And so in the case of Q, Q data centers, platform was built around a portfolio of nine high quality facilities with a mix of co-location, private cloud managed services and with capacity available for both traditional enterprise customers as well as for AI and high performance compute requirements. All of that in a Canadian context for sovereignty matters. And obviously right now, the accent is quite strong on that particular point. In Europe, we have similar dynamics. So what I would add is you can also frame it as a bit of a catch-up in the digitalization and the cloudification when compared to the US. And there is also a real sovereignty compensation and a perceived overreliance on US hyperscalers. I bear you see that continued catch-up. You got good land availability. You actually got subsequent connectivity to North America and relative energy advantages. In Germany, what's interesting is that you have a decentralized economy, right? So suppliers concentrated in Frankfurt, but enterprise and public workloads actually create openings in other cities. Regulation is also pushing energy efficient facilities. So you know, all in all, I wouldn't say necessarily more attractive than the US. I'd say it's different. And that with the right strategy. And again, the right execution team can be very compelling. And Julian, I always like to end the show by thinking about what's next in the area for my guests and the respective areas. As you look ahead to the next two to five years, what do you think is going to shape the next phase of data center investment? Great question. I'm going to take the macro investor lens here. And I'm going to say a couple of things. It's a fast moving space, but I think what we're seeing already is continued increased institutionalization. That's going to mean more scrutiny. That's going to mean fewer easy wins. That's going to mean more differentiation between teams who can truly execute those who are jumping on the ban wagon, including the ability to reckon with some of the slower deployments compared to the initial ambitions that we saw. Secondly, power and policy are increasingly going to determine where the growth happens, not just at demand. And I would say that's another cardinal role of infrastructure. Government, regulatory bodies, they might not always be in the driving seat, but they're often around the corner. And the corollary of that is, I think, thirdly, you'll start to see a clearer separation between generally infrastructure grade platforms, professionalized businesses exposed to long-term data growth, and the more speculative sort of tech adjacent developments that are more directly exposed to the AI arms race dynamic. So all in all, I would say, discipline, selectivity, patience, matter, again, not just on the economics of the asset, but also the selection and the working relationship with the management teams. For all the talk of artificial intelligence and hard assets, we have to remember that much of this is still a people business. And after all in life, there's no more important decision than choosing who you marry and how you nurture that marriage. So good to apply some of those same principles, whatever area of work you're in. Yeah, I agree. Or who you have on your podcast or who you partner with. Do I need an managing director at Infrared Capital Partners? Thanks so much for being on the show. Appreciate your perspective. And I look forward to continuing the discussion with you. Thank you very much to appreciate the opportunity. Thank you for listening to our time now. If this episode sparked an idea, pass it along to someone you knew 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 subscribe.