She gets a lunch break.
Good morning, everyone, and thank you for joining us. My name is Brendan Lynch. I cover REITs here at Barclays. Very pleased to be here with Iron Mountain's Chief Financial Officer, Barry Hytinen. Barry, thank you for joining us.
Thanks, Brendan. It's great to be here.
Great. Maybe just to kick things off, you're coming up on the end of the Matterhorn strategy that has kind of contributed to the double-digit growth that Iron Mountain has been generating over the past couple of years. Maybe just give us a review of what that strategy accomplished and how we should think about your longer-term growth algorithm going forward.
Yeah. So Matterhorn was all about being very focused on sustainably driving double-digit growth for the company and exploiting what is a very large addressable market in several of our core growth engines and cross-selling off of our core business, and where we have 245,000 client relationships built on trust and decades of experience working with those clients. And over the years, we have invested in distinct other growth engines, which I know we're going to get into in terms of the growth businesses that can power double-digit growth. And so Matterhorn has been very, very successful for the company as we've been on a trajectory now for several years of growing the business, 11-plus %, something of that order. This year, we're going to do over 12% growth for the full year. We're exiting the year at about a 14% growth rate, if not faster.
We're generating considerable incremental EBITDA. Matterhorn for us was about growing the business, continuing to power a business that can grow at that level for a sustainable future. When we guide in early 2026, I fully expect we will be guiding to another record year of double-digit growth, Brendan.
Great. Matterhorn was the second of two restructuring programs. We're coming up on the end of it now. Should we anticipate any additional restructuring going forward?
No. This is the last quarter of Matterhorn restructuring, and we don't have any additional restructuring dollars, and just to kind of answer a question that might be out there, those were one-time in nature, so you should not be anticipating additional of that sort going forward, and in fact, what you should be expecting to see is significantly incremental free cash flow, which we will use to further grow our business and/or need relatively less debt for the same amount of growth.
Great. Maybe just one more on the restructuring charges. That had been kind of discussed in a recent report. Just want to give you the opportunity to kind of discuss that consideration and maybe any other accounting policies that have been topical recently.
Yeah. So, look, that report is highly inaccurate. I know there's been some questions, so let me just say it very clearly. There's no investigations of any sort going on. And of course, we shared the full report with our board, and they agreed with us that it lacks any merit to even publicly comment on. So that's why we haven't addressed it in a public forum prior to this. And I think about our performance, Brendan, as our team has been delivering outstanding results for several years now. And I'd rather let our performance, both our past and our future performance, speak for itself. And as it relates to accounting specifically, our accounting and our disclosures are all very consistent, accurate, in line with all reporting standards. And what they particularly focused on was our non-GAAP measures.
In that case, all of the things that we add back are fully disclosed and, frankly, are very much in line with all of our REIT and data center peers. So I consider that matter behind us, and we will let our performance speak for itself going forward.
Great. Very clear. Let's dig in on some of those growth drivers going forward. Maybe first on data centers. It's been an interesting year with. We see the emergence of NeoCloud players, still some ongoing power constraints, large training sites versus cloud and inferencing demand. Where does Iron Mountain's position in the data center realm sit, and how should we anticipate growth going forward?
Yeah. We are an operator of third-party data centers. We are operating today about 450 megawatts, which is 98% leased. We do not have much in the form of capacity to sell. Probably about a third of that is colo, and two-thirds of that is hyperscale. As we move forward over the next few years, we will generate incremental revenue and profit on the data center business just as we finish construction of the leases we have already pre-leased with major hyperscalers. We are generally in the cloud and inference area in tier-one markets. We have got a lot of capacity to energize and sell in Northern Virginia, in Richmond, in London, some more in Madrid, and Amsterdam, Mumbai. These are all very high-quality markets where we will be energizing over the next few years.
And if you look at what we're under construction on, Brendan, we're under construction on about 200 megawatts, of which about two-thirds of that is pre-leased at this point. And we would anticipate continuing to pre-lease before we build. We're not a speculative builder in that sense. And so what I'll say about this year is certainly the leasing window that hyperscalers were looking at between when they lease and when it energizes, I would say, has shrunk over the course of the last couple of years. We're seeing some indications now where that might be actually getting a little wider again, which would be positive. But having said that, we have 250 megawatts that will energize in the next 18 months that is available to sell. And then we have another 200 on top of that that will energize about 19 to 24 months from now.
And so all of that is capacity that will be very attractive to hyperscalers. And we gave guidance earlier this year to do between 30 and 80 megawatts. I can tell you, as we sit here today, as we discussed on our last call, we had leased an additional asset early in the fourth quarter, and that was a net of an additional 11 megawatts for us. That was leasing our entire Chicago asset offset against London. And as I look out today, when we report the fourth quarter, I expect that we will be at 60-plus megawatts for the year. And I expect next year, in light of our pipeline strengthening, we feel really good about the conversations we're having with hyperscalers.
You look at the blocks of inventory we have to energize and sell over the next 12 to 18 months. I expect 2026 will be another year of over 100 megawatts of leasing for us. Frankly, it could be a lot more than that because we've got a lot of megawatts energizing.
Great. Terrific. I think it's underappreciated by a lot of investors just how volatile leasing can be quarter to quarter. But if you look back over Iron Mountain's history over the past couple of years, clearly has had a very significant ramp in bringing megawatts online and getting them leased.
Yeah. The three years prior to this one, we leased over 100 megawatts each year and kind of averaged about 120 megawatts. And the thing about it, Brendan, is we're leasing in kind of like 25-30-megawatt blocks. And so it can be kind of lumpy in that way in terms of when exactly a signature comes in. So some people have asked me, "Why 30-80?" That felt a little wide to folks. Well, if we get two deals done, we'll be at the higher end of that, if not even beyond. If we get one deal done, we'd be kind of at the midpoint, that sort of thing. So it can be a little bit of timing one quarter to another, but we feel really good about where we are here in the fourth quarter, and we feel great about next year.
Great. How should we think about delays in accessing power? Is that getting any better? Is it getting any worse?
I can't speak for the broader industry, but I'll just say in our energization, all of our energization schedules are pretty much on track. The vast majority of them are, as I mentioned, energizing over the next 18 months and then some more the six months after that. I would say certainly energy is the long lead time item, and I'm pleased that our team has very good relationships with the energy companies in the various markets we do business in, principally because we've been operating in those markets now for years. We've been dealing with the same power companies and GCs that help us construct and electricians, and also, we know the local municipalities and the markets that we're operating in, so we feel really good about our ability to deliver over the next year or two years.
Great. How should we think about your exposure to hyperscalers, which you've had contracts with a number of the largest operators in the world, versus maybe some of the NeoClouds that have come up, maybe a little bit lower credit quality? How should we think about your mix and your approach to that type of contract?
Yeah. When I talk about hyperscale, that is disproportionately in the largest cloud hyperscalers out there. I can't name them, as you know, for confidentiality reasons, but they're all the household names that we know. And those are the customers that we've been consistently doing repeat business with and that we expect to continue to do repeat business with. As it relates to NeoCloud, Brendan, at this point, we don't have any exposure to NeoClouds.
Okay. Great. Let's pivot to the ALM business, Asset Lifecycle Management, for those who aren't as familiar with the company. This is a business that you've gotten into over the past couple of years. It's ramping very significantly. Why don't you just walk through your rationale for getting into this business and the growth that you see?
Yeah. So I've been with the company six years now, but back in 2017, 2018, the company really started investing in a small way, test and learn kind of format on the enterprise IT asset disposition side of the equation, Brendan, where it was a cross-sell offering where we were already going to client sites to get physical records and/or bring them back. And we could pick up and help those clients dispose of obsolete gear, like, think laptops, computers, printers, things of that nature. And over the next few years, it grew, but by 2021, it was only a $38 million revenue business. Now, earlier this quarter, we gave guidance for our ALM business, and we've been kind of notching it up quarter by quarter through the year because it's been strength to strength.
We indicated we thought the ALM business would do $600 million of revenue this year. That is a business that is organically growing the last couple of quarters in the 30% year on year. The total growth has been much higher than that because we've been tucking in some acquisitions. We like the market for ALM, Brendan, because it is a very large market. We're going to call it a $30 billion TAM. About three-quarters of that TAM is on the enterprise side, which are large. Most of that is medium to larger size clients who we do business with already on the enterprise side of our physical business. The other 25% is very fast-growing hyperscale decommissioning. Now, we have businesses that serve both ends of the market there.
Today, on that $600 million of revenue for this year, we indicated it would probably mix to about 60% enterprise and 40% hyperscale. Enterprise as a market is growing kind of like mid-single digit. Hyperscale is growing faster than that because of the nature of what happens. Because on the hyperscale side, you've got embedded fleets of deployments that hyperscalers have made, and they refresh the gear about every five years. There's underlying value in that gear, which we help them by wiping the gear. In some cases, we destroy something that's been written to. In all the other cases, we end up disassembling those servers and then selling the components into the secondary markets in which we get a revenue share, think like 20% of whatever we sell it for.
On the enterprise side, it is a fee-for-service, whereby we are helping corporates with what is an ever-growing concern around privacy, chain of custody, security of assets that have had private confidential information written to, and we are stitching together. Our ambition is to stitch together a global offering whereby we can service a client's needs on enterprise IT gear anywhere in the world, and historically, up until we started doing this, there had not been a supplier that could do that for corporates. In fact, if you look at most large corporates, they're dealing with dozens of really small IT asset disposition vendors because there's no large player. We're already the largest player, so we've been growing that business organically around the world, as well as tucking in small acquisitions that are very accretive and improving our margins in that business.
We think there's a tremendous amount of growth ahead of us in our Asset Lifecycle Management business, and we are really pleased about how the business is trending.
In terms of the acquisition opportunities going forward, you've already organically grown and combined that with some tuck-in acquisitions to become one of the largest players, or the largest player in the world. Are there still a lot of opportunities out there to acquire either on a geographic basis or for additional capabilities in this realm?
The short answer is yes. And we have a target list that my corporate development team is looking at and talking to potential acquisition targets all the time. I mean, it's hundreds of companies long, spread out across the world. I always tell people, "You need a willing buyer and a willing seller." So when the next deal shows up is always sort of, "We'll let you know when we have a deal." But if you look at the last couple of years, we've tucked in two or three deals kind of per year. And these deals are pretty small, just to put it in perspective. I think our average, if you look across the last few deals we've done, is about a $50 million revenue kind of business.
But we're paying somewhere in the five to seven and a half times EBITDA on a trailing basis, and we end up synergizing that down quite rapidly. And importantly, this is a part of the business where, as we continue to do that, it gives us more presence. It gives us more capability, as you were describing, to be able to bring what we think will be a very sticky offering to our clients, whereby they want this. Our clients are telling us that they would like to have a consistent process, a consistent chain of custody, and a certification that things have been appropriately wiped, if it's something that has data, as well as appropriately and sustainably recycled or renewed. And so there's a lot of secular megatrends in asset lifecycle management that makes it a great market for us.
Maybe one of those secular megatrends would just be the absolute growth in data centers. All of those data centers have servers in them that need to be recycled. Where do you sit today in terms of recycling some of the GPUs that are getting so much attention versus maybe that opportunity over the longer term?
Yeah. That is certainly the fastest-growing portion of the general market. And as you would imagine, it's more concentrated because there's, call it, five to 10 large cloud players that have the vast majority of those servers. But those fleets of data centers have been growing very rapidly over the last five or 10 years. And the refresh model, as I said earlier, is about, on average, once every five years. And so each year, the volume that needs to be decommissioned is considerably larger than the prior year. It's kind of building on itself. And we have good relationships with quite a few of the large cloud players. We're really doing business with essentially all of them. Some of it can be on a project-basis scenario. In other cases, it's a kind of continuous flow as they're decommissioning sites. You asked about the components specifically.
We don't see a lot of GPUs at this stage because, as you would imagine, in light of what we're refreshing for them, it's stuff that was put in service, say, five years ago, and so there weren't that many GPUs. We do a few GPUs, and the pricing is very good on those, but it's a small portion of the volume today. It's more of an opportunity as we go forward, as the business continues to grow, but generally speaking, the way that market works for us, Brendan, and what we're decommissioning, if you think about the servers that are coming out today, we're wiping them, as I mentioned, then we will physically disassemble them. We'll sell off CPUs. We'll sell off drives. We'll sell off memory.
And to contextualize this for you, as I mentioned on our last call, our total ALM business this year, at that time, I was projecting to be about $600 million. 40% of that was the hyperscale business that we're talking about right now. Within that piece, about 40%-50%, give or take, quarter to quarter, is from memory. And a lot of folks have been asking and noticing that memory prices have been rising. And I'm happy to tell you that it has been. That is the case. And it continues to be strength to strength. In fact, I would say that memory prices, on average, are probably in the vicinity of 50% higher than when I gave the guidance.
If I play it through, that would, if you just work through that math I was giving you in terms of the addressable piece of the revenue, that's probably, call it $15 million, maybe more additional revenue in the quarter that we would expect to generate. So I think the one I think we guided to just under $160 million. I expect we'll probably do north of, right in the vicinity of $175 million or more in ALM revenue in the quarter. And frankly, I'll tell you, the larger cloud folks, they know that the prices have been rising. So some of them are even holding back a little bit because the expectation and the trend on forward pricing is for those to continue to move higher. And that's just in light of the supply-demand situation for memory. So it's a good market. It's a good trend.
It can be volatile, and I think if pricing were to stay where it is at this point, when we look into next year, that's naturally going to be an incremental tailwind to our ALM growth and our total business.
Great. Maybe we could address one more question on ALM, thinking about the margin profile of the enterprise business versus the hyperscale and how that fits in with the greater Iron Mountain margin profile.
Yeah. And that's one of the reasons why we often talk about the two businesses, because on the enterprise side, which is more fee-for-service, that's margins that are being in the 20%-30%. And we're absolutely seeing them rising. And then on the hyperscale side, it is a revenue share model, as I mentioned. So we're taking, depending upon the contract, let's say, teens-20% of whatever we sell. And so therefore, the margins on that are more like low double-digit upwards of as much as 20%, that sort of thing. So it mixes to a margin depending upon whatever % of the business you expect to come out of hyperscale versus enterprise. Importantly, both of those businesses have the ability to continue to improve their margins with their own business.
Frankly, that's the same for our digital business, which is seeing improving margin trends, and of course, our physical business. Data center has been, as you know, very strong. We're probably on track to do low 50s% EBITDA margin this year and again next year. Our total blended EBITDA margin will be basically a function of how fast we grow the various elements of the business. We've got very good momentum in our data center business and in our ALM business. Of course, our digital business, which we haven't talked about yet, has got the big opportunity with the IRS contract.
Great. Let's go there next. In terms of the global RIM business, there is a perception that volume is in decline. Maybe you could just address that.
Yeah, so it is not in decline, so as you can see from our disclosures, our organic volume, and this is on an organic basis, and in total, has been slightly up for the last several years, and that's kind of like quarter by quarter, and the way that is, sometimes people are confounded by that concept. It's a function of the fact that we are the world's leader in this segment. We have been doing business with clients for literally decades. We service them very, very well, and each year, our team is working to aggregate more volume from those clients. Because even clients that we've been doing business with for as long as we've been a company, there's still pockets of inventory out there that they might be managing themselves or may have with a competitor, and we win a little bit more volume.
We are also exposed to some markets. An example would be India, which is still in the early stages of outsourcing physical document storage. Whereas in larger markets like the U.S., U.K., Australia, they outsourced largely 30, 40 years ago, India, the most noteworthy, largest business for us, that is still in that early stage. We get a little bit of incremental volume, but sometimes people say, "Well, it must be all coming from India." India is not even 1% of our revenue. While I talk to you about that it's growing fast, it's just a nice little adder. In total, our physical volume that we're storing has never been larger than it is today. That business is what I would call very healthy.
Great. On the pricing front, you've been able to push kind of mid-single-digit price increases over the last several years. How sustainable should we think about that going forward?
Yeah. We always approach it from a standpoint of adding value for our clients, Brendan, so when we look at the offering that we're giving to clients, certainly they save money using us versus, say, trying to do it themselves or using some sort of network of a whole bunch of other smaller suppliers, and for medium to larger size clients that might be doing business in multiple geographies or many countries, we're really the best option by far because we're the only player that has broad reach in the market, but it's not just that. It's also things like we've got a long track record of consistent process and chain of custody and safety and security. In addition, we continue to add new offerings like our Smart Sort or our various other offerings, such as our DXP platform and Asset Lifecycle Management.
So we can continue to drive more value for clients. When we look at elasticity, as I've said for many years, it does tend to skew to our smallest clients. That's where we would see a little bit of pressure. But our customer retention rates are in the very upper 90s, like 99%. And we expect to be able to continue to, by enhancing the value proposition of what we offer the clients, continue to see revenue management adding something like the mid-single digit for the foreseeable future, as we've been doing, which is what we're getting again this year.
Great. Maybe let's talk a little bit about the digital business that kind of falls within RIM. Maybe just at a high level, what are you doing for customers? I think there's a scanning component. There's a software component. Maybe just kind of walk through some of those nuances.
Yeah. So years ago, our digital business kind of started as scanning. But at this point, and there's still a level of scanning involved, but really what it is, is we're helping clients structure unstructured data. Something on the vicinity of 80%-90% of all new data is still unstructured. So clients have a distinct need as they want to do analysis. When they are looking at data, they want to get it into a point where they can be meta-tagged and then use AI to do a variety of analysis and machine learning. We have a proprietary platform, which we call DXP, which helps clients in a variety of ways. And it's very modular in that it can be used in quite a few different applications, and we continue to expand those applications.
For example, we have cut out a tremendous amount of cost for major mortgage processors, whereby we use our DXP solution to streamline the process, physical and digital data going into our DXP, and automating processes such that they save a considerable amount on labor. Going forward, another example of that is with the United States government. As we talked about publicly before, we're doing a multi-year engagement with the Department of Veterans Affairs, where we are helping them do a meta-tag and do analysis on historic medical records of veterans. Then we've recently announced that we were awarded under the Department of the Treasury's outsourcing of paper tax returns.
That will be another example of where our team has built the large language models to ingest the various forms that come in in paper form, structure that data in an at-scale, rapid way with very limited issues around quality. Naturally, there's high expectations for the service-level agreements on processing things of that nature. We've got the track record and ability to do it. So our digital business, Brendan, a few years ago, it was like a $100 million business. Today, this year, probably be closer to $550 million. That's before really any of the big IRS deal comes in.
Frankly, we think the IRS deal is going to be a great case study for us to go around to other clients as well as government agencies to demonstrate for them the massive value proposition we can offer them in return on investment in terms of outsourcing that sort of processing to us. We have felt like the whole government efficiency effort is a great thing, both for the country, but also in a way that Iron Mountain can help our country get more efficient.
Great. Maybe just one last question in regards to capital allocation. How should we think about capital investments, leverage, and the dividend going forward?
Yeah. So I'll do the leverage first. So if you go back five or six years ago, our leverage was cresting close to six times. The last couple of years, we said at that time we were going to bring it down, that our target range was four and a half to five and a half times. A couple of years ago, we got it down to 5.0 times, and we've been at that level steady quarter after quarter ever since. We expect to kind of operate right in that level. We have a dividend payout ratio target of low 60s% of AFFO. Similarly, some years ago, we were higher than that target, and we didn't raise the dividend for several years as we were saying we were going to intentionally get into that range. And we've done that, obviously.
For the last three or four years now, we've raised the dividend. I think we've done it three times in a row at 10%. That's just kind of, you should anticipate that's going to continue to happen, Brendan, because we're not going to fall out of our payout ratio range, and we're growing AFFO at kind of high single to low double kind of consistently. As it relates to capital deployment, the vast majority of our growth capital is spent building out those pre-lease data centers. The nice thing about that is we're essentially writing 10- to 15-year leases with multiple renewal options with some of the best credit quality tenants you could have. We write that lease before we basically put the shovel in the ground.
So we're deploying capital at what I think are very good returns, think like 10%-11% cash on cash on levered returns. And that's one of the reasons why you've been seeing our margin lift in data center as we commence more and more of those projects that have really high returns. So going forward, you should probably anticipate that our capital deployment into data center will be pretty consistent, maybe slightly up from where we've been here this year, because we've got all those pre-leases that we're building, which will generate considerable revenue. Just to put that in perspective, this year, our data center business will deliver just under $800 million of revenue at a low 50s EBITDA. We've said in light of the backlog we have as we commence and complete construction, we'll be over $1 billion of revenue next year with no additional leasing.
Then over the ensuing next couple of years, we have another $250 million of already pre-booked revenue that will commence as we finish construction. Incremental leasing on top of that will generate that much more growth. We feel really good about the way we're trending.
Great. We're just about out of time. Barry, thank you very much for doing this with us.
Thank you, Brendan. Appreciate it.