Ep15: Insuring the Future of Digital Infrastructure

(upbeat music) This is UptimeNow, a Norden Roseville Wright podcast dedicated 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 Caroline St. Clair, data center practice leader of North America at Aon where she advises developers, operators, investors and digital infrastructure stakeholders on risk, insurance and capital strategy across the data center lifecycle. She has extensive experience building specialized expertise in data centers, digital infrastructure and large scale real estate developments. Caroline, welcome to the show, how are you? - Good, how are you, thanks for having me. - Of course, I'm sure folks who are listening would be interested in learning a bit by your story. How did you end up at Aon and leading the digital infrastructure practice? - Sure, so I joined Aon over 15 years ago right after college. I spent most of my career doing account management focused on real estate and then about 10 years ago started focusing a little bit more on data centers during the cloud buildout and then kind of through that learned a lot about the industry supporting clients and through my own curiosity. And then when the AI buildout came around, I was one of the people at Aon who knew the most about the risk in this industry and naturally grew into this role as our practice leader. And today, as the practice leader, I oversee strategy across our North American book of business for hyperscalers, developers, asset managers across the ecosystem and also very involved in our prospective client base as well. - Nice, I'm sure that practice has grown a lot over the past several years. - Yes. - For people who don't know and or don't know exactly what you do Aon, can you help us understand what you and your team actually do? Like when a hyperscaler or a colloid developer calls you, what problem are they trying to solve and how do you guys help? - Yeah, definitely. I say when a hyperscaler or collocation provider calls us, they're rarely calling because they simply need an insurance policy. It's usually trying to solve a bigger business issue. So that can be developing a multi-billion dollar campus, trying to secure financing, negotiating complex contractual risk transfer or just managing their growing operational risks. Our role is really to help them understand and manage risks so they can kind of deploy capital and scale faster and just inherently operate more resiliently. So we spend a lot of time helping clients align risk strategy and capital strategy. And really insurance is just part of that solution. Ultimately, we're helping organizations make better decisions around risk resilience and growth. - Do you find that a lot of your calls come when there's already a project that's an operation and now they're suddenly realizing maybe they're risked that they weren't thinking about or as a people who are more, I guess, an earlier development stage or thinking about where they're gonna get capital from and starting to deal with, you know, lenders and their concerns? - It's really all of those things. I wouldn't say it's like one or the other. I would say we can be the most effective when we're involved at a very early stage. There's a lot of work that goes into, you know, building insurable assets that ultimately are financeable. And if there's early consideration around a lot of things that can be an easier process for everyone, but we get involved at all stages of that lifecycle and different stakeholders as well. - And risk can mean different things depending on who you talk to in the industry, could be contractors, the developers, tenants, equipment manufacturers, lenders kind of sit in the middle of it all. And what are the risks that lenders are most concerned about right now? You know, it goes without saying, but the AI build out is, you know, using a significant amount of debt of all types or, you know, coming into play to finance this build out. And so, you know, that takes the form of construction loans, perhaps like take out financing once stabilized so that could be CNBS, ABS, even 144A bond offerings. But I think, you know, there's a lot of different, I guess, forms that the financing can take. And so across that, we kind of see a spectrum of concerns from lenders rating agencies and investors. And I would say lenders investors generally are more conservative than hyper scalers who are kind of driving all this demand. And so, you know, there's been a lot of articles recently in the press about different projects that maybe were quote unquote, you know, underinsured or, you know, the insurance market wasn't able to get the lenders what they needed to be comfortable. And I think what's interesting is like the hyper scalers and developers to an extent really think about risk and insurance in different ways than lenders and investors. And even insurance companies to an extent view risk more on, you know, EML or MFL basis, which is probable maximum loss or maximum foreseeable loss, which is an engineering calculation of basically what a loss would look like if a bad day happened. Verse investors and lenders really think about things and the frame of mind of, you know, total loan amount percentage of investment, percentage of loan. And so I think where a lot of the divergences happening right now is, you know, one community looking at things from an engineering insurance perspective and the other looking at things from a more conservative winter perspective. So, you know, I think just in general, the narrative that exists today in the media that there's not enough insurance on these large projects might not necessarily be true. That being said, lenders sit at the center of this ecosystem because they have, you know, exposure to nearly every risk on a project and they don't have a lot of control over how it's executed. So at a high level, lenders are kind of focused on, will this asset be completed, operate as intended, and generate the expected cash flow, right? And then I guess ultimately repay the debt. So when you think about during the construction, they're really concerned about delays, cost overruns, perhaps labor shortages, you know, all those types of things that could impact the project timeline. And then once the facility is operational, they kind of focus more on resilience and uptime. And, you know, a major outage or something that could affect ultimately revenue generation and asset value. So I think we fully appreciate where that side comes from and I think there's a lot of work going on with the insurance industry to kind of build education around what is standard and attainable in the insurance market to cover off risk and backstopping with kind of what the lenders are expecting and looking to risk transfer. - Yeah, I was curious because you mentioned the size of these projects are, are the projects becoming too large for any one single insurance company to insure over? - When you think about like the very, very large end of this spectrum on gigawatt plus scale campuses, there is no one single insurance company that will ever cover that risk. It's multi, multi billion dollar risk very, very concentrated in a single location, which inherently is challenging for an insurance company, right? It's just too much risk concentration. And so there is a substantial amount of work that goes on the broker side, bringing together different sources of capital and different insurance companies to kind of come together and insure a project of that scale. And the market today probably could bear, I don't know, anywhere between 10 to 15 billion of property damage cover, whether that's under construction or operational. And we're continuing to expand that market and help our insurance company partners bring more capital into the space, but it takes time and the insurance industry does not move as quickly as the AI buildout has moved. And so it will continue to grow capacity, it will continue to come into the space, but it will definitely take time to get there. - Yeah, what do you think is the single hardest thing to insure in a data center today? - The easy answer is back to that gigawatt scale campus. If you put a gigawatt scale campus where you have 10s, 20s, sometimes even 50 billion dollars of risk kind of sitting in the same campus, and you put that on the campus, that might be exposed to natural catastrophe. That is incredibly difficult to insure, especially if it's not built in a way that the insurance company's deemed to be highly resilient or a best in class from a risk perspective. So it's increasingly important today on campuses of that scale for developers, hyperscalers, investors to think about how you can build insurable and financeable assets before you put shovels on the ground. Because if that consideration isn't made early, some of those assets are going to have an incredibly difficult time getting insurance and the long run. - And how do you build an insurable asset? What are the things that you should be thinking about? - Insurance companies are focused on concentration, right? So to the extent possible on campuses of that scale, like being really thoughtful about spacing things out as much as possible to perhaps present to the insurance market separate and distinct risk versus one concentrated risk. And then just being thoughtful around construction type that's resilient roof type if you're in a wind zone power redundancy, all those types of things that most developers are very focused on anyway, but making sure that there's consideration being made to kind of best in class from the insurance company. perspective is definitely important. You mentioned something earlier that, like when I think about insurance and risk products, I, it doesn't, I guess it's not a topic that really comes to mind for me, but you mentioned workforce shortages. How, I guess, how do you ensure against that? Because we have clients who are working on, you know, in the same area, there is a, you know, gigawatt data center project plus a gigantic, you know, power project that's being built next door plus there's a transmission project that's running through the area and you imagine like, you know, in the middle of nowhere, United States, there's just, isn't the amount of labor necessarily that everyone's going to need to do this and so that's going to, that's going to become a problem. It's already becoming a problem and a lot of these areas, especially where there's just fewer construction workers, how do you, how do you deal with that risk? Our human capital team has done a decent amount of work on that side. I think it just comes into, you know, being thoughtful before you decide to build a project and doing some consulting and thinking around, you know, can your contractors deliver what they say they can deliver in that jurisdiction? Do they have the workforce available and as companies scale really quickly as well? Like, do you have the, are you hiring the right people internally that can kind of, you know, scale the company the way you want? Yup, that makes sense. We've spoken with guests about power, capital, talent, and community engagement. When you look across industry, what risks are becoming most critical to successful project delivery and operational resilience? When I look across the industry, I think the biggest challenge isn't really a single risk. It's kind of managing the growing interdependencies between power, capital, talent, supply, coaching, community support, you know, all those things. And I think the industry is building at a pace and scale we've never seen. And it means that execution risk has become really important. And you know, securing land and capitalism enough, projects have to secure power and navigate regulatory requirements and attract talent as we were speaking about, you know, and ultimately operate with, you know, exceptional uptime and reliability. So I think the organizations that are most successful or will be the most successful over the long term are the ones that view this holistically and not as a, you know, insurance purchase or a single risk, so to say. And you mentioned power, power seems to be a part of every single discussion that we're having when it comes to data centers and developing new data centers. There are all sorts of delays and good constraints and interconnection cues. Is there anything that the risk world is doing to respond to this risk and just how significant power is in the context of just developing and ensuring that this data center is going to be here and generating revenue early in the life cycle. We've gotten very involved around interconnection though, you know, financial security is becoming kind of a key caveat as there's longer weights and cues to get, you know, power. And so with those, with those cues for power, there's, you know, pretty significant security requirements that come along with that, which can strain developers' balance sheets. And so our insurance and risk world continue to push, you know, surety and bank-fronted instruments to satisfy those security requirements where possible without then sacrificing, sacrificing developer liquidity. I think there's a lot of work to be done around advocacy with the utilities and ISOs to make sure they will accept surety and, you know, different forms of security are viable. But, you know, that's, that is an area that we're getting pretty involved on a daily basis. So I think, again, is the market keeping up? I think there's more work to be done. But I think in general that we have a lot of capital and we have access to a lot of capital from different sources. And I think if we can continue to educate, we can continue to deliver meaningful report on that side. Yeah, really, really helpful to know. We have a lot of clients who, you know, as you have projects kind of spread across the country and interconnection requirements, deposits increasing, definitely something that, I guess a lot of people are struggling with. We're hearing more and more about behind the meter power data centers a year ago, I think it was a very small percentage of data centers that would consider behind the meter power. This year, been seeing studies come out say, you know, 40% or maybe even more than that are considering behind the meter power or some sort of, like, bridge solution until they have connectivity to the grid. Is this creating any new challenges or new insurance products that data center developers or power developers or vendors are asking for? So I think when you think about behind the meter power generation, I think it's important to remember that it compounds the, you know, aggregation risk that I spoke about earlier where you have a data center and a hyper-skiller with GPUs already on site. This brings sometimes a third insurance buyer to that same address and presents the market with an even more concentrated risk. So I think there's many things being done in the industry. And I think one of which is our ANS data center life cycle program that kind of helps present risks like this to the market as a life cycle risk and a single, you know, asset or so to speak. And you can create a situation where insurance can be purchased for everything together versus kind of bifurcated. So I think other than that though, I think it's a pretty fascinating period and it's kind of, we're seeing a ton of developers hire in, you know, their own power resources. And it's just becoming a very interconnected situation. We're seeing a strong interest in solutions that address construction, commissioning and kind of operational performance all together. And that's exactly the type of challenge that benefits from the life cycle approach like ANS data center life cycle program because they aren't isolated risks. So they're connected across development, financing power and long term operation. Does a life cycle program include cyber or is that something that you all think about separately? It can include cyber. It has the option to include cyber for the asset for construction. And I think cyber is, you know, fundamentally a risk that many people are paying close attention to on these projects and so we're seeing a pretty large interest in kind of adding on that cyber under construction element for many projects. I'm curious too. I mean, you guys are global. The industry is building like all over the world, places that kind of, in my mind at least, as a non-risk insurance person would seem to present different and unique risks based on their particular regulatory makeup, the workforce, even things just like transportation and roads and all that. When you move into a new market, whether it's a new country or a remote location, does the risk profile change or is it pretty consistent? When we kind of consult around, you know, site selection, I would call it, the variable that we often consult the most on is climate and how does climate ultimately impact the insurability of building an asset on that site. So if you build the same data center on 10 different addresses, we can run models for clients that show what the insurability number looks like for those sites and give you kind of a variable that people can use in their underwriting model for cost of insurance. So that's been kind of the most impactful thing we've been doing on site selection. And you know, I think the biggest thing that comes into play there would be just like, you know, especially in Texas and Midwest, like severe convective storm, tornadoes. We're seeing some of these crop up and named when hurricane zones, earthquake, you know, you name it. But I think having that consulting view kind of before you put shovel in the ground of like, what how the insurance, I guess the insurance market will view the risk based on those underwriting models and we can help clients kind of have that as a variable in their decision-making process. And we fully understand like, insurance is not going to be the full decision maker on site selection, but definitely gives clients the ability to make more informed decisions. I'm curious too. There's been an enormous amount of new capital coming into the space in the data center world. You know, you have private equity, you've got sovereign wealth funds, you have institutional money. Does that change your job at all? Are the new owners bringing new problems? I wouldn't say they're bringing new problems as much as just high expectations and, you know, these investors are typically deploying enormous amounts of capital and they want confidence at their projects and secure power and deliver on schedule operate reliably and ultimately generate long term returns. And so as a result, I would say our role has expanded beyond traditional insurance where we're helping investors and lenders and developers, you know, evaluate and manage risk across that entire life cycle of the asset. The common theme is that everyone is looking for greater certainty because when you're investing billions of dollars into, you know, AI and digital infrastructure, resilience and risk transfer strategy become pretty key drivers of value ultimately. So I want to be respectful of your time. But before I let you go, what's one, I guess, final piece of advice that you would have for data center developers, lenders, other other other developers that are out there in space to understand. have some people who would be reaching out to you who are considering building a project right now. Looking at like five to 10 years from now, you know, the winners won't just be companies that build the most capacity. I think the winners will be the ones that can reliably secure power, execute insurable and financeable projects, maintain uptime and adapt to change at scale. So companies that succeed will be the ones that create confidence for customers, investors, and insurers alike and, you know, resilience is critical on that. Awesome. Well, thank you so much, Caroline. I've really enjoyed having this show. Yeah. Thanks for your time. Thank you for listening to Uptime Now. If this episode sparked an idea, pass it along to someone in New Network. You can leave a comment. I'd love to hear your thoughts, ideas, and recommendations for guests. Till next time, thank you and remember to subscribe.