In this Risk Capital Insight episode of the On Aon podcast, Caroline St. Clair and Jon Chapman discuss how the next generation of digital infrastructure is changing how leaders think about risk, capital and long-term growth. As AI accelerates demand for data centers, organizations are pursuing larger, more capital-intensive projects with greater interdependencies across power, construction and operations.
The conversation explores why risk strategy must be embedded from the outset, how insurability influences investment decisions and what separates the organizations that can scale with confidence from those that face constraints later in the project lifecycle. Leaders will gain practical insights into aligning risk, capital and resilience to support sustainable growth and stay ahead in a rapidly evolving market.
Key Takeaways:
Experts in this episode:
Key Moments:
(01:15) The AI infrastructure boom and how data center investment, power density and project scale are creating new concentrations of risk unlike anything the insurance market has previously experienced.
(06:15) What's at stake when organizations fail to consider insurability early — including the impact on financing, campus design, site selection and long-term resilience.
(16:50) Why digital infrastructure leaders should adopt a lifecycle approach to risk, connecting construction and operations into a single strategy for risk transfer and resilience.
Soundbites:
Caroline St. Clair:
“Risk is no longer something you transfer after a project's been created. It's something that's determined very early in the process in the way you design, power and even configure your assets. And the people that get these choices right unlock capital, insurability and resilience. And if you get them wrong, it can be very difficult to fix down the line."
Jon Chapman:
“The organizations that win in the next decade won't just be the ones that build the biggest campuses. They'll be the ones that are credible under stress for insurers, lenders, customers, regulators and society.”
Caroline St. Clair
Risk is no longer something you transfer after a project's been created. It's something that's determined very early in the process in the way you design, power, and even configure your assets. And the people that get these choices right unlock capital, insurability, and resilience. And if you get them wrong, it can be very difficult to fix down the line.
Intro:
Hello, and welcome to another episode of On Aon, Aon's global podcast exploring the issues shaping business decisions around the world, with each week dedicated to either a risk capital, human capital, industry, or global topic.
This week it's our Risk Capital Insight episode, where we explore the trends, challenges, and opportunities shaping the world of risk and insurance.
And we're discussing the rapid growth of digital infrastructure and how AI is transforming the scale, complexity, and risk profile of data center investments around the world.
Joining us are Caroline St. Clair, Data Center Practice Leader for North America for Aon, and Jon Chapman, Practice Leader for Construction and Infrastructure for Aon. Together, they discuss why risk strategy is becoming a critical part of infrastructure planning and how insurability can influence investment decisions and what it takes to build resilience into digital infrastructure at scale. Let's get started.
Jon Chapman
So, hello, and welcome to this Risk Capital Insight episode of the On Aon podcast. My name is Jon Chapman, and I'm Aon's construction infrastructure practice leader for EMEA.
And today we'll be examining the recent explosive growth in data centers and digital infrastructure around the world. Data centers have quickly become a critical part of technology driving our global economy. They underpin AI, cloud and computing, financial systems, healthcare, government services, and energy networks.
So what has changed is not demand alone, it's the scale, speed, and concentration of investment now embedded in these projects. With me today to examine the issues surrounding digital infrastructure and why traditional approaches to risk management no longer apply is Caroline St. Clair our North America digital infrastructure practice leader at Aon. Welcome, Caroline.
Caroline St. Clair
Thanks, Jon. Thanks for having me.
Jon Chapman
So let's set the scene for our listeners. Caroline, can you paint a picture on how do we get here and what are some of the biggest issues surrounding digital infrastructure?
Caroline St. Clair
Sure. So I would say when you think about where we are today with data centers and digital infrastructure, we're seeing a mass AI infrastructure boom with power density and scale of investment surpassing anything we have seen in the past.
And a large data center, about even a couple of years ago was perhaps 750 million to maybe a billion. And we've seen a rapid escalation in data centers coming to the market and the surrounding infrastructure that is exceeding probably 3 billion on average, and then the upper side of the market getting much larger than that. And a lot of that's driven by the power density and the needs that the AI infrastructure requires. We're also seeing the speed to market demands being much more than we've ever seen, and the demand and supply misalignment, misbalance right now is in a position where there's so much demand and not enough supply.
And so the data center developers are being asked by the hyperscalers to build larger and larger data centers, power infrastructure, and all the surrounding infrastructure on campus at a speed that's never been done before. And we're starting to see this shift from single facilities to more multi-building campuses and global portfolios, especially when you think about AI, the LLMs (large language models) and the AI training models have really driven one side of the market to have substantially very large projects. And now we're starting to see AI inference at scale.
So that's where we're starting to see the average $3 billion projects cropping up all over the place. And AI really brings a concentration of insurable values to single addresses like the insurance market has never seen before.
So when you think about what can happen if you have a gigawatt scale campus, for example, you can have a concentration of a third-party developer coming onto that site with anywhere from 20 billion to 30 billion in value just for the infrastructure. Then you have a hyperscaler coming in bringing GPUs (graphics processing units) or TPUs (tensor processing units), and the insurable value replacement cost of those is oftentimes 200% of the value of the infrastructure.
And then you could have, and most times do have, on-site power generation as well. So that presents the market with an incredibly dense and highly concentrated risk, all put at a single campus or an address. And so that is what is challenging our industry the most right now from a concentration of a durable value perspective, and where our industry has the opportunity to change and bring more capital to the market.
This also presents to the market as incredibly complex risk from an interdependency standpoint and contractual risk standpoint. There's very high uptime requirements once these are completed assets. And so that infiltrates all the way downchain to the developer and the power requirements. And there's a ton of risk transfer that happens in the contracts and then associated penalties if things aren't delivered in the way that they've been agreed to.
That can create more risk than the market has ever seen, and that many stakeholders in this environment have ever seen. This means that risk transfer needs to happen earlier in the process than usual. Historically, insurance has been considered much later in the process, and that's not really relevant at the scale that things are happening today, because that can lead to friction in financing and delivery, insurance renewals, and all sorts of things.
So that's where we're seeing a shift in insurance being much more of an upfront conversation versus something that's happening midstream.
Jon Chapman
I couldn't agree more. I think coming in early, the insurance needs to align exactly to what is available in the market. So looking at contracts, making sure that they make sense from what is insurable and what it isn't and available in the market. And lenders also coming in need a bit of education also around what can be done and what can't be done, right? So looking at all of that and getting it right will have an effect across the whole of the life cycle of the project around what can be done from an insurance perspective and a risk transfer perspective.
So looking at those risks within the contract, understanding also the scale versus speed of deployment and complexity versus predictability and commitments versus what can be controlled.
So all of this means that the industry needs to approach digital infrastructure, as you said earlier looking at the financing and insurance and resilience of a project in a much earlier stage. What do you think, Caroline? What is at stake? What happens if companies can get this wrong?
Caroline St. Clair
Sure. I think if companies get this wrong, I think on the front end, the easiest thing to see that is an issue is financing and insurability, right? So if decisions are being made around campus design or ultimately like delivery of power without the lens of insurability or financing that's really where people run into trouble.
The decisions are made well before shovels go into the ground around campus design, building placement, the separation of buildings, and just the overall location of the campus and whether you can source a power supply not at risk from natural catastrophe? So we're doing a lot of consulting work today, helping people understand different addresses that they're looking to develop and the insurability factors of each one.
And our climate modeling team is doing a ton of work for our client base, helping people evaluate addresses and then figure out the factor that they would have to apply to build the same data center on multiple campuses. Compared to other asset classes, this is new and at the forefront of how we're seeing people develop is really thinking about insurance and insurability in the long run out of the gate.
Because I guess what could go wrong if you don't do that is you spend all this money, you build an asset in a place where ultimately the insurance market will not be able to transfer risk in the long run. I think that's what's at stake and what could go wrong in a nutshell.
Jon Chapman
I think that that will totally take an effect though around the bankability, the resiliency of the asset, and whether it's insurable or not, but it might be insurable, but at what price as well, and how does that affect your financial model? So, like you say, I think bringing in at the land acquisition phase, climate modeling can be a key decision maker on whether the asset, the land to be acquired is worth it or not, right?
Caroline St. Clair
Yeah, 100%. And what I think is unique right now, too, is like we're in a market right now, especially from property. So first party property damage, builder's risk, where the market has a lot of capacity. So it's a buyer's market. We are in what we call in the insurance world a softer market. And so there's a lot of capacity. There's not a lot of rate increases. And it's been interesting to overlay that market dynamic with this digital infrastructure evolution because right now we can put together really meaningful size programs and there's a lot of capacity in the market and a lot of appetite.
We all know that the insurance market is cyclical. And so when the market inevitably makes a turn the other direction, there will be insureds that have not paid attention to this insurability points on the front end that will have an incredibly difficult time finding risk transfer to some of these assets if they haven't been thoughtful around insurability early on.
Jon Chapman
So Caroline, as we're talking about bankability, insurability and resilience of the assets and talking about being in early, it's going to be key to do some risk engineering and what we call possible maximum loss studies or equivalents to try and set the right limit of indemnities on the policies, right? How do you see that?
Caroline St. Clair
We're spending a lot of time with our client base on this very topic. What's interesting to kind of think about is the way that investors and lenders tend to consider risk can deviate a bit from how insurance companies see risk.
And so starting with the insurance side of that, we're very used to evaluating risk from an engineering perspective, especially on risk such as digital infrastructure and data center, where they're highly engineered, highly resilient.
We can send engineers and insurance companies often send engineers out to site and gather information on how the site is engineered and what a bad day looks like from an engineer's perspective. And so at its core, the engineers can come up with a number called a MFL or maximum foreseeable loss. That really is worst case scenario on a campus. Like that is what the bad day looks like. And the engineers assign a number to that.
That is then how they underwrite. What is the likelihood of the MFL scenario happening? I think what where we deviate from how investors and lenders think about things is the amount of debt being taken out or the full replacement cost of an asset or the investment in an asset is how some of that other side of the fence wants to view risk and wants risk to be transferred. And so we have a little bit of an inflection point where everybody is talking about this MFL or PML, as you referred to, John, which is probable maximum loss, a little bit different number, but a similar type calculation. PML is just usually a little bit less severe. It's maybe not the worst case scenario, but the most likely scenario that could happen from a maximum loss perspective.
And that's where we have this divergence. And so you have lenders and investors really seeking full replacement cost or looking for full cover of their investment.
And insurance companies thinking about things from an engineering perspective. Because we have seen people get caught off guard from a developer perspective, being asked to place insurance to full replacement costs on an asset that maybe $10 to $15 billion. And while we certainly can work to achieve that, it's a conversation on what should be insured versus what can be insured. And I think that's where there's a lot of conversation going on in the industry right now.
Jon Chapman
If we think about all of this and move on around the bankability or the resilience, what does this mean around the downtime for clients, right? On the contracts, is there any insurance available for it? What does it mean contractually for developers and what does it mean for the hyperscalers?
Caroline St. Clair
As of 2025, there's a new product in the insurance market to respond to operational SLA risks, a service level agreement risk. It's really a product meant to be purchased by a developer for their operational SLA exposure. And it covers off if there's cash flow concerns from an investor based on these SLA penalties. There's more of a protection and it can create that operational SLA backstop for power availability, temperature in the data hall, humidity in the data hall.
What it doesn't cover today necessarily is construction-related SLAs, which is where there's a lot of crossover and candidly a lot of interest from our client base. And that is around can you deliver the project on time? Can you deliver power on time as you've agreed to in a lease? And is that insurable? And we're having a lot of conversations today on what that looks like from completion guarantee perspective and the different ways you can ensure ultimate delivery of a project.
Our team that works around contracts and contractual risk transfer and consulting can help our clients understand the different dynamics and the different ways to protect themselves.
But it is a very high-stakes environment right now with people taking on more risk and agreeing to incredibly tight timelines to deliver speed to market as they're being asked. And so I think it's just important that people pay attention to - it's a very fast moving environment. And we just help our clients make sure that risk and risk transfer is not one of those.
Jon Chapman
I think if we can maybe go into the amount of risk that is accumulating across the project life cycles, we've said about making decisions at site selection stage and then leading on to the design, to site layout, power, compartmentation, procurement, supply chain. Supply chain is a very big one, actually. Long lead times, transformers are taking longer and longer to get to sites. If you lose one, what does that mean for your project?
Building that into what you contract and then into the insurance solutions. And then thinking about interfacing and testing and commissioning phases where we have construction and operations ongoing on the same building, where there's some data holes which have been put into operation, or even with campuses where you have buildings happening next door next to each other.
So what does that mean? How do you think leaders should think about this at the pace of planning and executing a digital infrastructure project?
Caroline St. Clair
It's a really good question. I think for us, it's helpful to view things from a life cycle perspective and then think about risk transfer in that same mentality. And really that is most critical when you think about when you have a project that's a campus that's multiple buildings that you're building as a greenfield, there's a very complex construction schedule and delivery schedule. And to your point, there's a lot of overlapping construction operational transition phase assets.
When we take that type of risk and present it to the insurance market as one risk instead of two. So what do I mean when I say that? We're not presenting that to the insurance market as construction risk and then separately operational risk, but rather as one asset that is what we call life cycle based. So you can cover your operational concerns and alleviate the scenario where you have a claim on one side or the other, one policy and one lifecycle policy covering all of that and creating contract certainty.
So that's probably the biggest takeaway and the lesson had is this industry, especially on some of these really large data center campuses, is bringing in that lifecycle mentality, presenting to the market as a single risk and alleviating that adjacency risk of construction operations damaging operational data centers and kind of complexities around that.
Jon Chapman
Yeah, I agree, Caroline. I mean, that exposure around business interruption and delay and start up on a project where you have the construction operations happening at the same time, a construction project triggering a business interruption situation, or vice versa, the operations triggering a delay and startup exposure is key to wrapping around the coverage to the projects.
Especially with the size of these projects and the campus solutions, it's really critical for making sure that the actual risk is properly transferred to the market.
Caroline St. Clair
And I think obviously we both know a lot about this, but that's why our data center lifecycle facility has been such a valuable part of our value proposition and been incredibly popular with our client base, it solves for this problem. And then gives our clients up to $3.5 billion of capacity predetermined and ready to go for projects of that scale.
So developers can put up very substantial limits on a project very quickly and respond to that speed to market demand that they're facing from their tenant base and also have a well thought out risk transfer strategy.
Jon Chapman
Totally, Caroline. I think the facility wraps around all of those issues and concerns. But also thinking about as we mentioned earlier, the supply chain, the marine cargo, marine delay, and startup market is playing a critical role, around covering that delay from shipment of key equipment to a project.
And also we need to think about the cyber exposures which these projects are now facing as well. I think it's one of the key risks which is coming up. And how also the property market is excluding any physical damage arising from cyber events through the cyber exclusions that they're putting onto policies around how the market, cyber market is responding to those exposures.
Caroline St. Clair
Yes, I mean, this risk is just so interconnected across all of these different solutions, right? And it's just back to our earlier point. It's so important to pay attention to that, the contractual risk transfer and then how does that emanate downstream to all of these policies and that's really what we've been keying in on the most is that interconnected.
Jon Chapman
Okay, Caroline, as we're coming to the end of our podcast, what's the one thing listeners should remember? We've covered a lot of things here and there's a lot of thought to go away with, but what would your 30 second key piece of advice be?
Caroline St. Clair
I'd say the biggest takeaway for me is risk is no longer something you transfer after a project's been created. It's something that's determined very early in the process in the way you design, power and even configure your assets. And the people that get these choices right unlock capital, insurability, and resilience. And if you get them wrong, it can be very difficult to fix down the line. John, how about you?
Jon Chapman
The organizations that win in the next decade won't just be the ones that build the biggest campuses. They'll be the ones that are credible under stress for insurers, lenders, customers, regulators, and society.
So that's our show for today. Thank you all for listening. Don't forget, you can find out more about how Aon can help in planning and protecting digital infrastructure and talk to an Aon colleague about our solutions by heading over to Aon.com.
In the next months, we'll have more episodes on risk capital topics, including claims, trends, and supply chain disruptions. Until next time.
Outro:
Thanks for tuning in to the latest episode of On Aon. If you enjoyed this episode, don't forget to subscribe wherever you get your podcast and visit Aon.com to learn more about how Aon helps organizations make better decisions.
We'll be back next week with another episode of On Aon, bringing you insights from Aon leaders and industry experts on the issues shaping businesses today and into the future.
Thanks again for listening.