Good afternoon, everyone. I'm Alex Hess, a Vice President on Andrew Steinerman's Business and Information Services Equity Research Team here at JPMorgan. We're really excited to have with us today Barry Hytinen from Iron Mountain. Many of you know Barry very well. He's been the EVP and CFO of Iron Mountain for about six years now. Barry, it's always great to have you, and we're really excited that you're back with us at Ultimate Services today.
Thank you, Alex. It is a pleasure to be here.
Great. I want to start with the growth portfolio or the growth businesses. Back in 2021, those were about 15% of revenues. Now they're tracking towards 28% of full year revenues. The exit rate might be 29%, 30%, what have you. It'll be a little higher. Walk investors through the history of how you found these growth factors. Those are digital solutions, data center, asset lifecycle management, as areas where you thought were the right places for Iron Mountain to be participating.
I appreciate that question, Alex, because our growth portfolio is, I think, a very underappreciated element of our story in that we have been growing it quickly, collectively north of 20% for quite a number of years. We see that trend continuing for a long time in light of the segments we are operating in. The team started investing in most of those areas over a decade ago. The concept was having really consolidated records management and developed a massive client list. We operate with over 245,000 clients around the world. The team was looking for natural adjacencies whereby we could serve our clients in a high-quality way in compelling areas where we could drive value for clients. Bill and the team, long before I joined the company, started looking at a variety of areas.
Digital solutions was one that kind of came quite naturally because it cross-sells off of our core very easily in light of the fact that much of what we initially started doing was scanning of content that we had on behalf of clients because perhaps they wanted to do a level of analysis on a piece of the inventory, that sort of thing. Where that has naturally now grown to is our digital business unit team members have created a platform initially built off of some Google technology. Some years ago, we were the AI/ML Partner of the Year with Google. Since that time, our internal team has built out the application to be a true software as a service platform, which we call DXP.
DXP is a platform that allows us to significantly save customers' cost, drive considerable efficiencies for clients, particularly in areas where there is physical and digital information that needs to go from an unstructured format to a structured format. I am sure we will talk about that more. That was a very logical adjacency and one which we feel has a tremendous amount of incremental growth as we are continuing to find increasingly new applications for DXP, whether it be in our government clients or in our corporate clients. In our data center business, that was really one that started in a very small respect years and years ago, Alex, in the fact that we had clients that were looking for very secure locations where they could have some level of co-lo. We started putting servers in our mines, for example.
In our Western Pennsylvania mine, we have a data center now, and we've had one there for some time in which it was specific to client needs where they were looking to Iron Mountain to help them solve problems that they had in their own infrastructure. Over time, as we continued to test and learn as it relates to enterprise co-lo, we found more and more of our clients, particularly on our data management business, were shifting load to the cloud. They had an on-site data center, for example, that was naturally becoming less efficient because it was getting less load. That was an opportunity for us to come in and help solve another client problem whereby they could co-lo with us in sites.
Over time, that's just been a growing and growing area of our business, which I'm sure we will spend more time on, so I won't belabor further. Lastly is the asset lifecycle management business. Bill and the team started investing in that area in 2017 again as another way of helping our clients. As you know, we go out to see clients and pick up boxes, do other work for them on their sites. What we found is many clients were looking for a solution to help with obsolete or end-of-life IT gear. We started helping them by picking that up and having third parties process it. What that has grown into is a really compelling market opportunity for us. In 2021, our ALM business was $38 million of revenue. This year, we'll do about $600 million of revenue.
We're continuing to consolidate what is a very large space. I would say, Alex, it comes down to the team has been looking for ways to serve our clients. We're very customer-centric, and we are continuing to explore additional adjacencies and vectors for growth that we'll talk to you about at future days.
Great. Yeah. I very fondly remember when the Iron Mountain ALM business was CITAD. We were shifting the name because you guys had just bought IT Renew, and you had to keep track of all these different names. Now it's just ALM, which is very easy. On this industry, on ALM, we'll start there. You sized this market at $30 billion. You're the clear number one here. I don't think it's close. I think it's three or four times larger than number two, perhaps. It's $600 million out of a $30 billion, excuse me, TAM. How do you size this market?
Yeah. We have looked at it a variety of different ways, and we have also utilized third parties that have done primary-level research as well as secondary research on the sizing of it. What we find is that the total market is, as you say, $30 billion. Let's call three-quarters of that what I would consider is the enterprise ALM segment, Alex, and 25% is hyperscale. When I talk about enterprise, what I mean there is our classic corporate clients who have that IT gear that they need to deal with as they generally are kind of consistently having a flow of IT gear that obsoletes, whether that be laptops, on-site data center servers, screens, printers, what have you, all sorts of IT gear. That is a business, and that is a market, I should say.
That is a market that is growing kind of mid to high single digits on a consistent basis. One of the compelling needs there, if you talk to clients, is that they are naturally concerned about confidential information, PII, other things that have been written to devices, and making sure that they are very securely wiped and disposed of appropriately in a sustainable way. That is an area where there is a distinct need for more consolidation. The market in enterprise ALM is highly fragmented. As you note, we're already the largest player. That portion of our ALM business, about 60% of the revenue of the $600 million we'll do in ALM this year, so call it $360 million. I estimate that puts us at two or three times the next largest player.
We're continuing to grow on an organic and inorganic basis because there's a real compelling need. What we find is that medium to large businesses that have operations in multiple regions, multiple states, multiple countries, they have to string together a network of many small IT asset disposal companies to service themselves. That's really because there's nobody that has brought together the power of a consolidated go-to-market approach where the consistent chain of custody is very secure and very capable of dealing with privacy and other safeguards that need to be done in that area, as well as, of course, sustainably dealing with the material. That is what we aim to do, Alex, in the enterprise side. We are in the early innings of consolidating that market. If it sounds familiar, you know our story very well.
It's a strategy that we did really in the record side earlier on. If you go back 30 years or 40 years ago, records management was a very fragmented market without a single player that could provide the service around the world to clients. Over time, the team here consolidated that market and brought together what is a really compelling offering, especially for multinationals. We think we can do the same thing on the enterprise side. In the hyperscale section of the market, that's, call it, 25% of that $30 billion total TAM. That is a more concentrated space from a customer standpoint. Think about the major cloud hyperscalers because what we're doing there and other participants in that part of the market are doing is helping clients with retrofitting their data centers with gear.
As a major hyperscaler may have a fleet of data centers that they've opened and continue to open over the last many years, they tend to retrofit the servers in there about every five years on average. They're not taking the gear to obsolescence. They're renewing and refreshing the gear for reasons related to getting better compute, maybe better power efficiency, etc. That gear still has a considerable amount of value inherent in it, unlike on the enterprise side where gear tends to be taken more to obsolescence. On the enterprise side, it tends to be a service-oriented revenue model. In the hyperscale side, it tends to be more of a revenue share model where we are sharing what we can garner for the assets. What happens is we take the servers out of their sites. We wipe them.
In some cases, we destroy elements that have been written to. Then we will physically disassemble the servers and sell off the components, CPUs, DRAM, what have you, whereby we will take a 20% on average take of whatever we're selling that gear for. That part of the market is growing high single, low double-digit growth from a market standpoint, Alex, because as you would appreciate, there is so much more data center capacity each and every year. What we are really doing as an industry is retrofitting on that five-year kind of horizon that I mentioned. It just tends to be a market that has a tremendous amount of volume. It is subject to a little bit more component pricing volatility on that side of the business. It tends to be a little thinner margin than the enterprise side.
I will tell you that we love the ALM business. We think the market is going to grow for a long period of time. It's an underserved market, as I was describing earlier, and it's one that I think plays very well to our strengths.
Yeah. We'll have to check in with you in five years when we find out what all these Blackwell and Hopper chips resell for. I think that'll be a if you could answer that now, we could all go home today.
I will not do my Carnac impression today, but I will say that it does appear that it would generally bode well for the ALM market, what you are pointing to, because I think there is a fairly significant correlation in terms of the value of the gear going into the data center. It is fairly correlated with the value coming out, obviously, with a depreciation schedule. While none of us necessarily know what that is going to look like in the future, I think indications would be that since that gear is relatively more expensive on the way in, it will probably be at a better value on the way out. Therefore, with a revenue share, we are likely to see some, if you will, mixed benefit to headline revenue in the future from that. You are right, that is a few years off.
Yeah. Just staying maybe in the data center world, right? We are going to talk about all the assets you have to energize. A point that you have made repeatedly over the last few years is that Iron Mountain do not build on spec. You do have right now a notably high percentage of construction in progress that is not pre-leased. Maybe not notably high relative to the industry, but notably high relative to your history. How should investors think about that unleased construction in progress? Are you building to indications of demand based on the stated needs of your recurring tenants? How do you decide, "Hey, look, we do not have a person yet leasing this facility, but this is not a spec project"?
Yeah. It is a good question because our strategy generally, Alex, as you know, is to not build on spec. One of the things that we really like about our hyperscale data center business is we are dealing with very large AAA kind of credit counterparties in which we are signing leases of 10 year-15 year duration with multiple renewal options and at a very good return, think 10%-11% cash on cash on levered returns. In light of the build schedule and the pre-leasing elements of that, frequently, we can sign a lease before we put the shovel in the ground to start the construction. In the case that you are pointing out right now, where we have a slightly elevated level of not yet leased construction-oriented development, I would say there are a couple of very specific things going on there. As you know, we signed our Chicago asset.
We totally leased that building post the last quarter end. It is not in our current schedule as leased. That is with a shift from London, as we talked about on the call. It is a net additional 11 MW for the year. However, we had been working with the client on that potential move for some time. As you saw, we have been doing construction in that asset, and there is a considerable amount of spend against it as we are prepping because a portion of that will go live for the client even next year. There is that element. There are a couple of other assets in there that are under development that are clearly very, very highly sought-after locations.
When you look at the assets that we have energizing over, say, the next 12 months - 18 months, Alex, we've got 28 MW in Northern Virginia. We've got megawatts in Madrid and Amsterdam. All three of those would clearly be significant hyperscale locations. Then we have 25 MW in London. I would tell you that if you look at the energization schedule there, those should be very, very good leases for us as we get closer to energy.
I do want to just step back on this for one second, which is you clearly do not build to spec. You clearly build with some sense of tenants in mind, but you are also selective. You work with a very select number of tenants, and you build and you develop conservatively. When investors understand those two facts, the stylistic third fact that falls out is, "Hey, when we decide to go and start building something before it is pre-leased, we have a reasonable indication of a small handful of tenants that will probably be interested." Is that a fair characterization?
I think that's generally fair, Alex. I mean, the market for the type of leasing that we're doing on the hyperscale side is you're talking about the top 10 hyperscalers. Really, there's probably four or five for us that are really meaningful as it relates to our client base in that sector of our business. When we talk about Northern Virginia, London, Amsterdam, we have a very active pipeline of discussion there. While the deals tend to be a little bit lumpy in the sense of it's hard to predict precisely which week or which month a deal is going to sign because there's always elements of technical due diligence, and the design can change slightly, which naturally affects the underwriting and the contract terms.
As you would appreciate, there's a lot that goes into those sorts of deals because you're talking about a 15-year initial lease with perhaps the likelihood of renewal options to go 30 years-40 years. The details matter. When we get into detailed discussion with clients and something changes, we need a little bit more time to get the contract done. I will tell you that when I look at the tier one assets that we have, I feel and I know our team feels extraordinarily good about the ability to lease over the next 18+ months as we look at the assets we have coming for energization.
Perfect. Thank you so much. I want to turn to the Treasury contract that you guys just signed. This isn't global RIM, but it is part of the growth portfolio as well. It's a digital service that you're providing them. As we understand it, you're going to help the Treasury reduce their reliance on paper-based tax filings. There's sort of a long-running digitization revenue stream. It appears based on what's sort of in the public domain to be a revenue stream that coincides with tax season. Can you elaborate a bit more on sort of the scope of the contract? Specifically, what are you doing and what's going to govern volumes in that business and what the revenue mechanism is?
Sure thing. It may be counterintuitive, but to most folks who might be using electronic means to file their tax returns, the U.S. government is still processing a tremendous amount of inbound paper tax returns and various other tax correspondence, Alex. Historically, up until present day, the government has generally been processing that largely through a labor-based solution of personnel at the Internal Revenue Service. What we are offering the U.S. Department of the Treasury is a digital transformation effort whereby we showed them a proof of concept of where we had built large language models to process the various tax forms and correspondences in a high-quality, consistent, and rapid way such that we would meet the naturally exacting standards they have. Naturally, as you would imagine, the government does not want a lot of errors in processing of tax reform and tax forms.
When we pitched this opportunity, and it's been in the public ether now for several quarters, what has manifested itself here recently in this win is a five-year contract in which the U.S. Department of the Treasury has awarded us a $714 million contract for the estimated value over the next five years. They did award to three other, what I think is fair to say, much smaller players than us in this area. While we are not exactly certain yet, there are mechanisms by which the government will use to disperse the volume. We're the only one processing forms at this time. It's not a huge amount of revenue. As I mentioned on the most recent call, I expect we'll do probably $4 million of revenue in the fourth quarter with the IRS. You are right. We expect it to ramp into tax season.
In light of our ability to process, our ability to be ready to process the inbound rapidly and across a variety of forms, and the fact that we've already become FedRAMP-ed and we have a tremendous number of people that are now cleared, as you can imagine, there are quite a lot of hurdles to get over to doing this kind of processing. I think we're in a very, very advantaged position vis-à-vis that. If you want to think about it from a standpoint of if we got 100% of the volume, and I'm not saying we would, but just to kind of frame the opportunity, we estimate it would be about $140 million a year, Alex, based on the volume that the government is currently seeing.
The way that was brought together was the government put out a bid form that said this number of ten fortys, for example, this number of other forms and various correspondences, and then had anyone that wanted to bid that was a qualified bidder come in and say what they would process each form on a per-form basis. Naturally, more complex is more expensive, less complex, less expensive. When you mix effect that under our bidding, it came out to about $140 million a year. The opportunity here is, I think, quite immense because if you kind of factor that against the size of our existing digital business, our digital business unit is now at a run rate of about $550 million of revenue on an annualized basis.
Even if we got the majority of this, this is a big move in terms of our digital team. Furthermore, I think it becomes a huge case study for us to go and talk to both corporate clients as well as other government agencies as we show the massive value creation we are going to have for the U.S. government because published reports would suggest, in fact, some from the government suggest that the United States spends, I think, in the vicinity of $600 million plus a year doing paper processing of tax returns. When you think about it, our cost to do that to the government of $140 million is a huge savings. I think it speaks to the opportunity for digital transformation and what our team has built in DXP. This is not the first large contract we have won using DXP, Alex.
We've been doing mortgage processing for a couple of major financial institutions. We're doing a variety of other similar exercises, but this is a very meaningful one and I think can be a great case study for demonstrating value.
Excellent. That is the growth portfolio. Now, we are 24 minutes in. We still have not talked about box volumes. It might be a record.
I thought you'd get there.
Might be a record for you guys. So just again, let's just go through box volumes. And specifically, you've had an uptick, and I'm not talking sea change, but a modest uptick in retention rates for the core global RIM storage rental business. And that's been despite continued mid-single-digit price increases. What's driving that increased retention and what might it suggest about your customer's elasticity of demand going forward?
Yeah. I will say this, Alex. It's really a wobble. It's an improvement. I'm trying to be balanced. Even when the retention rate was coming wobbling down slightly coming out of COVID and through a couple of projects that we were doing for very specific small pieces of inventory, but it can move the retention a tenth or so. I told people a year or so ago, I was like, "Hey, in about a year, it's going to start moving back up because we'll get through that project." Fundamentally, the underlying health of our client base is really good. Our customer retention rates are extremely high. Most of the clients who we do business with in the record side, they've been doing business with us literally for decades all around the world.
In most cases, they've nearly standardized with us if they haven't totally standardized with us for this as we deliver tremendous value to those clients that need a partner in this area across their enterprise. Look, the retention rate will kind of continue to be in this vicinity, I believe. It kind of also factors into the concept that the average box stays with us for about 14.5 y ears, and that's been kind of steady for a long time. There are some boxes that end of life much sooner than that. There are some that stay much longer, but the underlying health of our records business is strong.
That's great. I appreciate that you're flagging a positive change as a wobble. It's good to be balanced on that side. I want to just touch briefly on capital allocation from here. You've got all these businesses, some of which have, I'd say most of which actually have very impressive incremental margin profiles. You're not throwing losses after this revenue growth. It all comes with incremental EBITDA dollars, which is lovely. You have taken on more capital as you've grown your top line. You've grown your capital base. In a juncture where, let's say, we go into a recession or there's a slowdown or one of your businesses wobbles more significantly for a year, which of these metrics are you going to allow to downshift first? Would you take down your CapEx? Would you moderate the pace of dividend growth?
Just trying to think how it's a balancing act. There's a lot of work that goes into this, right?
Yeah, sure. There is. Our team gets up every day and creating value for our clients, and it's hard work. I would tell you this, Alex. If you look at the increase in capital that we've been deploying, it's been disproportionately to data center over the last five years. I think my first year with the company, we might have spent $300 million on data center. This year, we'll do closer to $2 billion. Why is that? Coming back to your earlier point, it's to support the pre-leases we've already signed. If you look at our data center business, this year, we're going to do just under $800 million of revenue, call it a low 50% EBITDA margin, super strong business, writing very strong returns on deals.
If you look out to next year, I've kind of said a few times that in light of the backlog of deals that we are constructing and that will commence between now and earlier this year and anniversary next year and through next year, we will be over $1 billion of revenue with no additional leasing. Obviously, we're on a lease. On top of that, we have another $250 million of revenue that will come out of backlog over the next few years after next year. That's all booked business with clients that are AAA-type credit and with very long lease terms. In your hypothetical, I will say this. We're a very cost-conscious company. We are very focused on driving margins in each one of our businesses.
You did not ask, but I will just tell you, I think that there is opportunity for incremental margin in each of our businesses. Now, the total enterprise margin will depend upon where you grow your revenue, but there is self-help, so to speak, and opportunity. Our team is working hard on driving profitability in every single one of our businesses. We would be very focused on driving capital returns in such a scenario. I will tell you this. You mentioned the dividend. For us, the dividend is almost formulaic. We have a payout ratio target of low 60% as a percentage of AFFO. Some years back, we were well higher than that. We said, "Look, we are going to pause. We are not going to raise the dividend until we get down into the payout ratio." At the same time, our leverage was a little bit higher than we wanted.
It was around six. We said we wanted to get it down into our target range of four and a half to five and a half times. What has happened during the next five-plus years since then? We got the leverage down to five times, right at the center of our target range. We have been there for quite a while now. We got the dividend payout ratio into, we got the dividend into the right payout ratio range. We just raised the dividend 10% for the third time in a row, third year in a row. Why 10%? That is in line with our AFFO growth. What you should anticipate is that the dividend will continue to rise in line with AFFO because we expect to continue to maintain that payout ratio. I will say this, Alex.
If the macro, on your hypothetical, if the macro is a little more challenged, there's always opportunity for us to kind of belt tighten, but it won't be on pre-leased data center because we got clients waiting for us to open up those, and those will then convert to revenue.
No, that makes a lot of sense. That makes a lot of sense. We've got about 10 seconds left. Maybe a little word association with something that's not on investors' radar that you think they should be paying attention to with respect to the Iron Mountain stock.
Yeah. I guess I'd point out two things. It's not that they're not on the radar, but I'd just underscore. Our core business is really durable, Alex. I mean, we've been doing a level of revenue management year in and year out. Our volume on an organic basis has been flat to slightly up, think 30 basis points. We've driven a lot of margin in there. That business is very capital-light. It can continue to grow in a way that requires relatively very little capital to grow. That is one of the secrets of our success because unlike a lot of other data centers, we've got inherent free cash flow coming out of that business, which we can then drive into growth in data center.
The other one I would point out is, look, I mentioned it earlier, our ALM business is growing strength to strength, right? 36% organic growth last quarter. That was 70%, I believe, total growth. Margins are expanding in that business. It is a secularly growing portion of the economy and one that we think is ripe for consolidation. Frankly, we're going to ultimately build a business that serves clients in a way that they want to be served, which is that we offer the same consistent process, chain of custody, privacy, and value all the way around the world, which is something clients are looking for.
Thank you so much for your time today.