Beyond the Build: AI Infrastructure, Exit Readiness, and What's Next - Part B

Speaker A: Welcome to Private Equity Perspectives, a UH podcast by BDO USA's Private Equity practice.

Speaker B: Each episode, BDO connects with leaders in the private equity space to discuss the latest trends driving deal activity, fund strategies and portfolio company optimization.

Speaker C: Hello and welcome back to BDO's Private Equity Perspectives Pipeline podcast where we explore the trends impacting private equity today. I'm um, Todd Kinney, national relationship director in BDO's private equity practice. Today we're continuing the conversation with Jackson Pay, Managing Director and the co Head of PE at UH Closed Loop Partners Michael Ryder, Partner and Co Head of North America at Igneo Infrastructure Partners and Blair Barlow, Managing Director at Lime Rock New Energy. If you haven't already, I'd recommend listening to Part A of our conversation which went live a couple weeks ago. It'll give you a good idea of what kind of work the guests do as well as their point of view on investing in assets that are evolving super rapidly and diligence risks in their verticals. This episode we'll take a closer look at the materials required to build and maintain these assets. What an exit ready data center looks like today and we'll close it out with a look toward the second half of 2026. Maybe get some predictions from the guests. Just a quick reminder M to our listeners that the remarks and opinions of our guests do not necessarily represent BDO's view. So Jackson, I'm going to have you kick off part B of the episode. PE sponsors are pouring capital into data center assets and we're hearing a lot of talk about what goes in including power cooling, capex. I wanted to ask you about the critical metals and materials used to build these assets. How are you thinking about solutions to ensure stable supply of these resources? I know Closed Loop also focuses on what comes out the other end. So what should sponsors be thinking about on the hardware lifecycle and sustainability side? And is that even on most sponsors radar right now?

Speaker B: Yeah, Ah, no. Great, great question and thanks for the intro. Glad to be back. On our last episode I think we covered a lot on the macro topics related to energy and power and I think Michael and Blair did a great job on that. The demand and the bottlenecks related the build out of data centers I think as we pivot a little bit here and hone in on the hardware lifecycle and sustainability at the sponsor level, I think it is still a critical part to the underwriting process given the geopolitical landscape and life cycle management and sustainability acts as ah, the tip of the spear for many of the blue chip customers that own and operate data cent. So as we think about the sustainability angle, there's an opportunity to actually acquire customers throughout portfolio companies, at least by using a reporting system, by being able to track where materials are actually going at the end of life and naturally giving value back to those customers themselves. I think there are a lot of funds that are actually focused on energy empowerment. We have guests today that do that. But alongside that energy question, there's always a materials question that you alluded to. Do we have enough materials or physical minerals required to build and power AI infrastructure? I think 80% of the critical minerals is being manufactured by China and that causes a huge bottleneck in the U.S. but that's because we don't have strong supply chains that recapture materials coming out of the end of life. Materials from data centers, from enterprise IT equipment from PCs, just your traditional electronic equipment. I think the growth of data centers has resulted in such a large demand surge for copper, aluminum alongside many other valuable materials like gold, silver, palladium, tungsten steel and even water. So you're seeing the current market drivers really creating a case to invest in platform solutions that enable domestic material circularity, supporting the transition to a circular economy in the US and beyond. Today I think shifts in regulation and m global supply chains are uh, further increasing domestic demand for these materials. Meaning we need to think about other ways we can access materials. Traditionally you think about just traditional ground mining and now we want to shift that conversation to above ground mining which is a way to capture or recapture material that is already in circulation. Altimet Closed Loop Partners has invested in companies that are making above ground mining possible through the recovery of critical materials from electronics and technology assets and metals used for major industries. Most recently we acquired a majority interest in Sudden Metals, uh, Metals and Critical Minerals Processor focused on collecting and sorting high value metals from aerospace, from automotives, from electronics and a significant part of the industrial manufacturing supply chain that's resulted in really strong prices on the back end and bottom line margin for this asset which uh, we're really excited about. I think I've mentioned the company Sage Sustainable Electronics throughout our conversation in the past and we made a couple of add on opportunities to Sage to create a larger, that's a national footprint to provide to these blue chip customers. So I think being able to support a large data center infrastructure provider gives us a lot of value on the back end. Being able to support the repair services, the data clearing services, the reverse logistics services. There are a lot of additional ancillary requirements related to handling this data center. Equipment that I think people are just not thinking about, at least not being top of mind. So these portfolio companies are really creating value for customers by deriving value from their waste streams, whether it be data center equipment, critical minerals from industrial manufacturing. It helps support the sustainability and financial goals for stakeholders, which naturally drive a really good sustainability story for the end customer.

Speaker C: All right, Jackson, well, I certainly appreciate you detailing that for us. Next, I want to pivot over to exit readiness. All the buyer expectations around power, capacity, cooling, energy sourcing and sustainability credentials are all raising the bar for exit readiness in this asset class. So Michael, I want to come to you first. What in your eyes does a truly exit ready data center asset look like today and how early should sponsors be building toward that?

Speaker D: Yeah, Todd, great to be back with you again and appreciate the opportunity to continue the discussion. To answer the last part of your question first, simply said, any sponsor that knows what they're doing should be building towards an exit from the day they close on the acquisition. In fact, I would say they should be planning towards an exit while in due diligence and approving the investment at the investment committee stage. So, you know, you really gotta be thinking about what do we need to do to drive value in this business with these data centers during our ownership so that we are best positioned to be able to exit. And that means thinking about who are the likely buyers, what the market conditions need to look like or would be best suited for a successful exit, and obviously what levers do you need to pull, what, uh, capital do you need to invest, et cetera, during your ownership to set that up. So for data centers in particular, I think the key things for exit readiness in my mind for an investment, first and foremost that your data center or data centers have multiple use cases. Optionality around the use of the data center is really important to future proofing, which as we talked about during the last discussion, it's critical to position the data center portfolio to be strong and resilient for future demand, both from a hardware perspective and importantly as well from a location perspective. We didn't talk about that before, but I think being closer to the customer, and I'll get to that in one second, really does drive optionality in how you use the data center in the future. Secondly, I'd focus on the customer base, ideally a uh, diversified portfolio of investment grade customers. Ideally those customers have been in the data center for a while, so you can see resiliency and duration in their relationship, that the data center is an important location for them in their business. And that they have long term contracts in place. I think the biggest challenge that some of these very large hyperscale developments face, where the hyperscale development's being funded by third party capital, is what happens at the end of the initial contract. You've got a single tenant. A number of these hyperscale facilities are located where power is cheapest, which means that they're a long way away from the consumer, the market, and most likely the releasing contracts are going to come up at a single point in time, at some point in the next 10 or 15 years. So although you can underwrite to get a reasonable return perhaps through that first contract, exactly what happens at the end of the contract, how much capital needs to be spent in order to upgrade the facility to be used again and to sign another contract, I think are, uh, unknowns that we just don't have a good view on today. And so we've tended to stay away from that type of investment and focused more on data centers that have multiple tenants or portfolio of data centers that are in urban locations where there's many opportunities and use cases. Thirdly, having a distributed portfolio of assets is helpful. So multiple locations in a lot of end markets, I think that's ideal because it then means you have good diversity across often end market industries and local regional markets, et cetera. As an example, at US Signal, our uh, retail colocation business based up In Grand Rapids, Michigan, we have 17 data centers today and hundreds of customers. At Altum Digital, we have seven data centers spread across the country. We have customers in the finance space, in healthcare, in enterprise technology. And we just believe this diversity really sets us up for a more successful exit, particularly if you have good long term customer contracts, as I mentioned before. And then finally, ideally you've got a data center portfolio that still has future growth potential for the next buyer. That can come from a demonstrated ability by the existing management that is running the portfolio to be able to manage and drive growth through acquisition of existing facilities or development of new facilities. And it can come through additional opportunities for capital investment and upgrading in the existing portfolio as well. So I think if you can put all of those things together, uh, a diversified portfolio of assets with good customers, multiple use cases and growth potential, you should set yourself up for a good exit after the end of your whole period.

Speaker C: That's awesome. Thanks Michael Jackson. A little bit different question for you. What about the technologies and solutions that support material supply and recovery for data centers?

Speaker B: I think it's a really good question and I put into three specific buckets that I'm looking at as, ah, it touches data centers. I think one is inventory management, two is the data destruction software and then three is the robotics that are touching data centers and its existing kind of recovery of, of that supply chain. So on inventory management, the way I think about it is reverse logistics. I think a lot of people are focused on new builds, new configurations, new deployments. But a lot of the reverse logistics and decommissioning supply chain is extremely antiquated and they don't talk to the front of the house from a fulfillment standpoint. So being able to provide a uniform and consistent tracking service to a lot of these global companies allows that company to be more informed in its own data tracking, which is important from a data security standpoint and asset recovery as they're looking at refreshing and redeploying a lot of new assets into the field because of the need for additional memory processing power, what have you. The second is data is a key security issue for anyone operating a uh, data center, anyone touching cloud. And given all the cybersecurity breaches in our day and age, we need to be able to protect the data that's on those assets and be able to wipe and reassure the customer that all that information is off your data center equipment, your server, your networking equipment, et cetera, before it goes into a redeployment process. Otherwise the end customers really focus on shredding and landfilling, which is wasted profit and allies. Finally, I think from a robotic standpoint, we're seeing in the third party maintenance space, in the decommissioning space, more and more robotic solutions that create a cost efficient way to recover or maintain your existing data center infrastructure that I think is naturally critical to maintain the useful life of this data center equipment. So at closed Loop, we've actually invested in a couple companies that support to end to end solutions of these data centers through their decommissioning and recovery services. I think ones that I haven't touched on include Mold, which actually a venture portfolio company of ours that operates robotic microfactories that dissemble complex electronic products to recover valuable components for reuse in the data center space. And the other is valis, which is another venture portfolio that's used to actually maximize the data analytics coming out of the data centers and the metals that uh, are required to support this material. And how do you actually maximize the resale value of these outputs?

Speaker C: All right, Jackson, certainly a unique perspective there, so thank you very much Blair. To close out the topic, I'm going to pose A slightly different question to you. What does a data center need from a power delivery perspective?

Speaker A: Yeah, sure. And thanks for having me back on again, Todd and Jackson and Michael, fascinating perspectives. So thanks for sharing those. I think it's probably three things that make a data center more attractive from a power delivery perspective, I think. One, I believe that even if a project starts as behind the meter solution, you're increasingly seeing momentum behind islanded data centers because of the interconnection issues and getting capacity on the grid. But I think even if a project starts out like that, uh, and is not grid connected at the outset, I think there is significant value in having a clear and credible path when you are, you know, developing a data center and looking to exit it. A clear and credible path to getting grid interconnection. I think if you accept that, then two things really matter. One is how quickly can you get interconnected and then how much power can you access once you are? I think speed and access to power are increasingly two of the most valuable attributes in this market. And one way that the data center operators can achieve both is really by positioning their data centers as a resource to the grid rather than just a load center. And I think specifically, if you can operate as a flexible or even a controllable load, providing that predictable curtailment during peak periods, I think you meaningfully change how the grid operators view your projects. I think this is not a new concept. You know, we're seeing hyperscalers and developers actively exploring it, but I think the opportunity is still underappreciated because the reality is that the grid constraints are often driven by a small number of peak days a year. And if a data center can reliably respond during those periods, whether that's at the chip level or at the facility level, can unlock additional capacity that otherwise would not be available to it. And so I think in that scenario, you're not just another interconnection request in a long queue. You're helping to solve a real system constraint. And I think that could translate into a faster path into interconnection and potentially down the road, greater access to power once you're connected. I think the second point I'll make, and Michael touched on this in the last episode, but I think when you are constructing a, uh, data center, designing to maximize fungibility for future chip design, rack density, cooling technologies, and really all facets of the power infrastructure, I think ensuring you have flexibility at the outset to adapt to whatever the future state is, and importantly, without a ton of capital investment required in the Future, I think that's going to be increasingly important and I think every developer of an asset is thinking about that. But just given the, the rapidly moving technologies and, and the racks getting more power intensive, I think you really need to be thinking, uh, about an asset that isn't just there to meet today's requirements, but it's engineered in a way that can accommodate these higher power densities in the evolving technologies, such that you can have a greater overall power draw without requiring a full redesign. So I think being thoughtful at the outset about system architecture and redundancy is really going to matter. And then, last one, this is maybe me talking my own book here, but I think integration of renewables and storage into the energy mix is important. I think there are advantages of not going 100% gas and having a more diversified mix of power that includes renewables and storage. For a bunch of different reasons I probably won't get into on this podcast, but, uh, there are a lot of positive benefits, including on the cost side, speed to power and resiliency and redundancy is which, which I think between renewables and stores, it can provide a lot of that.

Speaker C: All right, Blair, appreciate the perspectives and closing us out there. So unfortunately, guys, we're reaching the end of this episode, but I can't let you go without asking my classic closer the crystal ball question. So, given everything we've discussed around data centers and the role PE is playing in this ongoing evolution, I'm curious how you're thinking about what's coming next. Looking ahead through the rest of 2026, what do you think PE sponsors in this space should have on their radar and where are each of your firms focused as that plays out? Blair, I'm going to let you take this one first.

Speaker A: Okay. Yeah, I mean, I think from our perspective and uh, mentioned on the last podcast, we really are viewing this as a generational build out of infrastructure. And what's unique about this one is not just the scale, but it's also the pace at which things are happening. And that's just showing up in a lot constraints and bottlenecks across the value chain. And so I think when we think about what's happening, it's not a demand problem. The demand is there accelerating. It's more of an execution problem. And so when we think about what we have on our radar and what other sponsors should have on their radar, it's really where are those bottlenecks in the broader ecosystem emerging and then how persistent are those bottlenecks going to be? And so I think today it's very obvious that the biggest bottleneck is power, and that is availability and the time to access the power. But also we think it extends beyond things like interconnection queues, permitting supply chains and skilled labor, which I've already talked about. But just getting into that mode where anywhere you can accelerate the timeline is going to be really beneficial and unlock a lot of value. So I think our focus is on alleviating any and all constraints and accelerating the time timeline. You know, where we are spending time is thinking about which platforms can unlock the power bottlenecks. So that's distributed generation, uh, grid supporting assets, technology services, grid enhancing technologies, anything that can improve the utilization of the existing infrastructure so we don't have to build new and don't have to get new infrastructure permitted. And we're also spending some time on solutions that enhance flexibility, uh, and that comes on both the supply and the demand side. And that's around, you know, storage, storage, load management, uh, demand response, and even sort of the more integrated side of the energy equation. Because ultimately out of all of this, I think everyone recognizes that the grid is going to have to become a lot more flexible, digitized and dynamic if we're going to try to support this level of growth. Maybe the last thing I will touch on is, and I think it was Michael who hit on this, that uh, these sort of locations of where these things are getting developed are evolving as well. And they're sort of leaning towards maybe not the sort of population centers, but actually the places where power is more available. Um, and I think when we think about sort of what that implies is like what is the second order challenges around that development? Right. So what in these regions needs to happen in order for those to become more valuable assets and thinking about, you know, bottlenecks and constraints in those more localized regions. So I think, you know, taking a step back, the maybe theme in the message is that we're not super focused on the data center themselves. We're focused on the power and grid infrastructure that is going to allow them to scale. And we're just looking for where we think the most durable and defensible opportunities are. Not just for the near term, but as we look out over the next five to 10 years and think about where we should be deploying capital to accelerate these opportunities and debottleneck the broader value chain.

Speaker C: Yeah. All right, Blair makes a lot of sense. Thanks. Michael. How do you feel? What are your thoughts on what's to come?

Speaker D: Yeah, firstly, I'd say, Blair, I think it's really interesting how you're approaching this. And there's just such a huge, huge demand for incremental power, not only generation, but a lot of the grid balancing and other, uh, grid build out that you're talking about in order to support the data center demand growth here. So very glad to hear that that's an area of focus. What do we see over the coming couple of years and therefore where are we focused? I think mostly it's actually a transition from what we see as being a focus on large hyperscale facility development, moving towards more distributed networks of facilities to drive future data center demand. We believe as AI moves more into an inference phase, and particularly as that inference phase becomes increasingly driven by consumer demand, we believe that the key use cases will move to more of a network type structure and that data centers closer to the consumer will benefit from that shift. They have lower latency, tend to actually be in areas where you'll have 102 megawatt or 105 megawatt data centers instead of a 50 megawatt or 100 megawatt data center. We think the other benefit of that is actually that those data centers in many cases exist today and where they don't, they can be developed and built uh, reasonably easily. It can be a little bit difficult to site in population centers, but there's generally power to be able to power a 5 megawatt facility. The other key trend that I think we're pretty focused on is. Jackson mentioned it just a minute or two ago. Data security. We completely agree that data security and integrity is going to become increasingly important as you see a lot of these, uh, data security breaches that have occurred. We expect enterprises to focus increasingly on storing their data and manipulating their data in data centers that are in private networks and on um, private servers to ensure that they aren't exposed to these security breaches. And I think that's going to drive a large increase in private network development. The need for more private cloud environments that often sit in data centers like we have at US Signal, where companies can run their own copies of large language models referencing their own proprietary data on their own service. So that level of security I think is where we will start to see a lot of large enterprises and corporates go so that they can be comfortable that they aren't going to suffer from a data breach if their data is in the cloud. So we think we're really well placed to take advantage of these with US Signal and with Altum and are uh, looking for other opportunities out there where we can continue to position ourselves to support this growth over the next five years.

Speaker C: All right, very interesting. Thanks for those perspectives, Michael. So, Jackson, I'm going to let you close us out here. So what should PE sponsors keep top of mind throughout the rest of the year and where is your firm focused?

Speaker B: Yeah, happy to close this out. Obviously, Michael and Blair touched on a lot and it's interesting to see how each of us are using our own strategy to apply into a sector that has obviously attracted a lot of capital. And as I look at key themes, even as my guests here also are touching on, I'm hearing circular principles that are very core to what we believe in, which is resource efficiency, local productivity, and then durable profitability. I think many sponsors are looking much, uh, harder at the execution and ability to find pockets of durable profitability. And you know, from closed loop standpoint, I think naturally investing in businesses or doubling down on businesses that have circular principles that promote resilient and materials security is core to our next few years. I think with the build out and the rapid build out of AI and data center infrastructure, I think we'll also see a rapid decommissioning process and uh, end of life assets come uh, back into the market over the next five years. And being able to create the right reverse logistics and end of life processing infrastructure will help us capture the valuable materials back in circulation. And then on a kind of adjacent note, we're looking at broader infrastructure services that provide a maintenance repair angle that naturally elongates the useful life of data centers. I think that's going to be very important as uh, we look in the near future. And then the last thing I'll touch on is alignment. I think investors and stakeholders will need to recognize the financial environmental opportunity of reducing waste and its associated costs to create value for both parties involved. And at closed Loop, we're going to remain dedicated to building the circular economy and advancing our platform for the forefront of this transition.

Speaker C: All right, Jackson, thanks so much. Appreciate all those closing thoughts. Well, unfortunately that's all we have for our episode today. If you want to hear more from Jackson, Blair and Michael, please go back and give part A a listen. Trust me, it's a really great conversation to our guests. I realize you're all super busy, so thanks for carving out some time to share your valuable perspectives. BDO certainly values our relationship with all of your firms and I think you shared some really great content today, so I appreciate you guys joining me.

Speaker D: Yeah, thanks for having us, Todd.

Speaker B: Yeah, thanks so much, Todd.

Speaker A: Uh, yeah, appreciate it to our listeners.

Speaker C: Thanks so much for tuning in. If you haven't already, we'd love for you to subscribe, rate and leave a review of the show wherever you get your podcasts. Until next time, this has been BDO's private equity perspectives. The views presented by our guests do not necessarily reflect the views of their respective firms.