Okay. Good morning everyone. Curt Nagle. I'm the Senior Business and Information Services Analyst here at BofA. This session's Iron Mountain. Very pleased to welcome CFO Barry Hytinen. We're gonna structure this as fireside. If there's time at the end, we'll certainly field questions. Barry and I go way back to when I used to cover TPX, and he was the CEO. From a personal perspective, it's a real pleasure to have you back on stage, Barry, and welcome and thank you.
Good to see you again, Curt.
Good to see you, too. Okay, starting from the top, Records Management business, building on that, adding Digital Solutions, Lifecycle Management, data centers. In terms of, I guess, just again, starting at the high- level, revenue growth, double-digit, visibility on that, confidence in that continuing, competitive edges, could you frame that? Yeah. First question.
Well, Curt, we've got a really interesting business in that, it's one that has been public for quite a while, but if people haven't kind of kept up with the story, it's a relatively, reasonably fast-growing company at this point. That is thanks to the fact that the team here started investing in areas some years ago that would more easily cross-sell off of our core. You know, we operate a company that has 240,000 client relationships around the world, and most of those have been doing business with us for literally decades. 95% of the Fortune 1000. They standardized with us and have continued to work with us for years and years because our retention rates are like very, very high.
Because we offer a very good value. We chain of custody, privacy, security is what they value in terms of our core service offering. Over the last few years, we have been expanding quite rapidly in three distinct growth areas that you just touched on. If I start with our core business, that today represents about 70% of the company's revenue. If you go back about five, six years ago, I joined six and a half years ago, at that time, the core business was round numbers like 90% of the business.
Now today, the business is much bigger. Like that core business is over $1 billion of revenue bigger and much better margins, and yet it's only 70% of the company, and that's because the growth areas have been growing and likely will continue to grow for a long period of time at very high rates. You know, data center, it wasn't that many years ago, we were a $200 million revenue data center business. Last year we did $800 million in revenue, low- 50s% EBITDA margin. That's up over 1,000 basis points over the last few years. It's still got more opportunity to expand because we are far from, if you will, stabilized.
We're operating a portfolio of about 490 MW, 98% leased, and we've got enough megawatts that are either under construction or held for development that we could more than double the business, if triple. Data center is a secularly growing part of the economy. We are seeing very good returns on the deals we write, and we've become a trusted vendor to several of the largest cloud hyperscalers, and that's where we've done more and more repeat business. That business is gonna go from strength to strength. In fact, just to give you a sense, we could build out to $1.35 billion of annual revenue without even leasing anything else at this point. That's what we've already contracted for.
We'll do over $1 billion of revenue this year with EBITDA margin that is up in that segment every quarter year-over-year. Our ALM business, it's a story that I think the investment community is only starting to appreciate because it's a newer story. You know, in 2021, I think we did $38 million of revenue in our ALM business, Curt, and last year we did $633 million.
Sorry, $40 million on?
$38 million.
$38 million.
Yeah, in 2021. We did $633 last year. We guided to $850 this year. I think in time it will be our biggest business on a revenue basis, and I'm sure we'll talk more about that business, but it's a huge TAM. I think the TAM is something like 3x what we address in records. It's much more revenue opportunity per client than what we have in records, and it's a very fragmented industry. It cross-sells quite well off of our core client base because all of our core clients have a need for the same service, and they can't have it serviced by a single partner like they do with us in records today because that offering doesn't exist. We're already the largest player in that market.
I think we've got a tremendous amount of growth coming in front of us on ALM. Our digital solutions business, it is also growing very, very well. Wasn't that many years ago, we were doing $150 million. At that time, it was principally scanning. Today, last year, we exited the year on a quarterly run rate of doing $600 million annualized, and it was growing high teens to 20%, continues to grow strength to strength. We just won a very large contract with the United States government. That adds nearly 10 points of growth to that business this year and another nearly 10 points of growth to next year on top of the secularly growing business we have in digital solutions.
Our team built a platform called what we call DXP. It really helps clients significantly lower cost and in terms of a whole variety of use cases. We see incremental traction. That business is becoming more recurring in nature. There's a lot of growth in our business and we, you know, it's all about execution.
You and defense are the only verticals growing in government at this point.
Well, the thing about it is, you know, government efficiency. That's a tailwind to our business. Because if you look at our business, we, you know, we don't do a tremendous amount with the government because a lot of processes and a re internalized at the government. Having the opportunity to outsource and do things in a more efficient, you know, kind of, you know, done through process and software is a means to significantly lower costs for the government. The IRS example that I was just speaking to is just one where that opportunity saves the government a tremendous amount of money. Think hundreds of millions of dollars. There's more use cases for that. Another example, like, you know, the government stores a lot of physical records. We don't store much for them because they have it all internalized.
We're working with the U.S. government as well as governments around the world to help them become more efficient. I think that efficiency effort is a tailwind to us for the next several years.
Maybe just a quick follow-up on that, just interesting points on the government. IRS is tip of the spear, maybe?
It's a big opportunity. It's obviously a big win. It's probably one of the largest opportunities. We've got with the government, and obviously we already won it. It's a multi-year contract and with renewal options, and we feel super good about where we are. It also is sort of a means to get into more opportunities because as a result of that deal, we've recently announced we're now FedRAMP High, which is one of the key standards for being able to do more work with the government. Yeah, there's more opportunities. Our team is in Washington and in dealing with other governmental agencies routinely as we pitch more opportunities. I mean, the IRS one, just to be fair, is a very big one.
It's a large one. Okay.
Yeah, it's a large one.
Understood. Switching gears a little bit and then we'll go back into some of your business segments. Big theme of this conference is AI, huge topic within info and business services. How are you capitalizing on AI risk exposure to the business on the other side?
AI is a huge tailwind to the company. I mean, you may not hear that from every participant here today, but I'll tell you, in our business, it is a tailwind.
All will tell you it's a tailwind.
Is that right? Well, let me tell you how it's already a tailwind for us, both on the revenue side and on the profit side. AI intersects with our business across all of it. On our digital solutions business, AI is enabling us to help unlock dark data for clients, where they can now start doing analysis on information that they didn't previously have the ability to, and our DXP platform allows us to do that. It starts with digitization. We can combine both physical and digital data, which we do for clients. That is a very growing portion of our business. What we're doing is we're then meta tagging the information.
Our DXP platform has agentic technology built into it, and it enables us to save clients a tremendous amount of cost reduce labor in what are fairly manual i ntensive applications. Think like mortgage processing. The IRS example, that's all about manually processed paper tax returns, which, you know, people might.
It's a nightmare. Yeah.
It's a nightmare. There's a huge savings opportunity there. There's more and more digital projects of that sort, and then AI is driving our own ability to serve our clients and save them money. In data center, look, we're principally playing in all Tier 1 cloud hyperscale locations. While major hyperscalers are building out very large gigawatt campuses to do training of their models, where are they gonna monetize it? They're gonna do the inference in cloud locations where latency matters, and that is strength to strength for us. It is another significant use case. In our ALM business, AI is making us more efficient in terms of how we work with the gear that we're bringing in, monetizing it. It's also making us much more opportunity on the volume side because AI new gear is creating, in some cases, refresh cycles speeding up.
Which helps us because we're on the other end of that. We help them with decommissioning the gear. The gear that we'll be decommissioning over the next few years is higher priced generally than what we've been decommissioning. It's more volume and more price. AI is a big opportunity for us, and we're monetizing it. On the cost side, you know, we have a lot of opportunity that we're executing against. We're in a very early stage of harvesting the benefits. For example, you know, because we have a large client base in our core business, we have something on the order of 10 million customer contacts a year. Most of those are voice.
There's still email and chat as well. Like, you know, most every company that has that kind of customer interaction, we did the labor arbitrage on call centers a long time ago. What AI allows us to do is be much more effective for the client. Like in any call center, we have attrition. You get 15%-20%. AI makes us get the new representatives up to speed much faster because it can listen to the call, it can read the email in advance, and suggest to the human what the response is. We're starting to test things like responding more directly to the client vis-à-vis AI. It makes us more effective in ways like on our procurement side, we're starting to run more and more of our commodity through RFPs that are AI driven.
With 245,000 client relationships, our commercial team responds to a lot of RFPs. It used to be a terrible manual process with many humans involved. AI is enabling us to respond to more RFPs faster and more accurately. In our HR area, it helps us do a lot of training and recruitment much more efficiently, and when we're bringing on the sorts of growth rates that we have, we're hiring a lot of people. You know, AI is a big help to us.
Okay. Very clear. Go back to the core business, records, strong recurring revenue base, you know, cash hub of the business, right? You know, in terms of, you know, I think one debate is pushing back on just it's kind of a legacy business, right? You know, very simply, less use of paper. You know, in terms of, Barry, how you think about the sustainable bit of volume and then pricing, right, in that context, digitization, you know, how would you answer that?
Yeah. The business is a phenomenal business, okay. Number one, we've never stored more physical volume than we're storing today. Sometimes people have a hard time imagining that. As you said, most everybody thinks, "Well, there's less paper in the economy." Yeah, so let's talk about that some. The stuff we're storing really matters to clients, and they've been sending us boxes for literally decades on average. We're not looking for the volume to grow substantially, Curt. You know, I've been saying since I've been here, like, our expectation is for the volume to be flattish to slightly up. Think, like, 30-40 basis points, and that's what we've been doing for the last several years. We've been growing organically year to year, quarter to quarter. We know that our inventory that we store for clients, it has a natural life cycle.
The average is about 15 years, and it's been at 15 years for a long time. We start every year knowing we gotta destroy about 5% of the volume, and our commercial team goes out there and wins new business. We consolidate business with existing clients. There's some markets that are still in the early stages of outsourcing, and then we get new volume incoming from clients. That's kinda how it works. You know, let me point out one thing about less paper in the economy. You know, Curt, when you and I were doing conferences like these, like, 15+, 20 years ago, there were then stacks of analyst notes all around and you know, conference books, tons of paper. Let's look out there. There's none of it. None of that was coming to us.
Clients didn't send us that kind of paper because they didn't need it in the future. No offense to the sell-side community.
It's okay.
You know, today it's all electronic. It's all on the iPad. We don't get that paper, but we never got that paper. The stuff we get from clients are things that they absolutely need to store because they might need in the future, and those use cases, they aren't changing. You know, there's been levels of digital transformation in the economy for years. We absorbed that, like, years ago, and we continue to absorb it, and we continue to grow. The pricing capability we have on the businesses has been proven to be very strong because we offer clients a very, very high value. The offering that we have, which they chose to standardize on years and years ago, is a high one. If you're, you know, Bank of America, you're probably sending us boxes from Charlotte and Boston.
New York.
You're all over the place. If you not that they do this at this stage, but if they wanted to kind of create that sort of relationship with someone else, you'd be talking about stringing together literally dozens of vendors. 'Cause we're the only global player at all. We're not looking to, you know, certainly overprice. We price to value. The value we offer is super compelling, and we have a lot of additional products and services we can sell to c lients like a BofA. You know, do we see some elasticity? I always say we absolutely see elasticity. It skews generally to our smallest clients, and generally speaking, that's where we would add the least value.
Because if you're a two-person firm here in Manhattan, you might be sending us one box a month. You know, you don't care that we can service you in Germany or in Florida. You don't do business there. For the vast majority of our larger clients, they can't replicate what we have, and they don't want to. Because they standardized with us for a reason. We give them great service.
Yep. Yeah. Well,
By the way, The Wall Street Journal had a customer satisfaction survey in the last year or so, which they looked at B2B companies and customer sat, and they ranked us number one.
Okay. There you go. Maybe just going back to that point before we pick up on another on the dark data, how that fits into record storage and maybe a new vertical.
Yeah. Dark data is an interesting one. It's part of the AI tailwind I mentioned that we have, which is clients now through advanced digitization, meta tagging, our DXP platform enables us to help them by creating data information that they can then analyze that historically they didn't have the opportunity to. While clients are, it's still, you know, it's still expensive to do that, relatively speaking, there are many use cases and increasingly so as costs come down, where we're seeing more and more of that sort of information flow. We see a fair number of digital projects every year, a growing number, which is on information that clients are asking us to convert to meta tagged data that we weren't even storing.
A lot of our digital businesses is essentially either new paper coming into us that may or may not be stored in the future as opposed to significant amounts of digitization of the inventory we have on the shelf. We do some deployments of specific project-oriented work of digitization that's on the shelf, and the vast majority of the time, that goes right back on the shelf, by the way.
Okay. Fair enough. Pricing, I wanna go back to that, as you talked about an important driver, I think across a number of your segments. What are you assuming for this year? You know, what's the sustainable rate of growth? Then again, just you know, generally elasticity. You know, sounds like it's fairly concentrated.
Yeah. I mean, our view on the record side, on the core business, is you look back about six years ago, we said, "Okay, we expect the core business growing at about a 5% rate." We all said that's probably pretty conservative. We've been growing about double that for the last, you know, five years. You know, last year we did about 6% round numbers on average, pricing, and our core business grew a little bit faster and we, you know, we've been saying for several years, like, "Hey, mid-single-p lus is probably the right place to be." That's kinda what we guided to this year. I think, you know, that's a healthy place for us to be because that business requires very little capital to grow.
As a result, as you described, it just generates a tremendous amount of cash and we use that to go deploy into other parts of the business, principally data center. Pricing in our other areas of our business, on digital I'd say, you know, we do compete with a variety of competitors in that space, so that's, you know, the pricing is slightly up, but not as strong as in the core business. In the data center side of the equation, and we have a colo business and we have a hyperscale business. Our colo business is sort of, you know, in place and we renew, you know, call it 1,500 leases every year. We have low churn, think like 2%, 3%, 4%. Mark-to-market's been like double-digit for years now, for several years, so it's very good.
Double-digit growth.
Double digit. Then on the hyperscale side, pricing has been rising some over the last few years. Frankly, costs have too. The cash on cash unlevered return is how we kinda think about that business on hyperscale deployments. That's been running the 10%, 11%, sometimes you see a return that's a little bit higher than that, but we love that business because it's the hyperscale business, you know, you're talking about investment-grade clients. You're talking about leases that are 10-15 years in duration. It's a pre-lease business for us. We don't generally speculatively build. We generally are looking at signing a contract with a hyperscaler, you know, for let's say 10-15 years, before we principally put the shovel in the ground. That is a really effective capital- deployment strategy for us. In the ALM business, it's you know, on the commodity side, it's you know.
As the market goes.
Yeah, it's as the market goes.
You're not wildcatting for data centers.
No. No.
Fair enough. Okay. Yeah, I guess sticking on the records business, digital solutions, maybe just talk a little bit more, you know, again, about the synergies between the two and then, you know, again, just, you know, how big that business is and kinda where you expect rates of growth.
In digital?
Yeah.
Yeah. The digital business is last year, you know, as I mentioned earlier, it exited the year at a $600 million annualized rate in the fourth quarter. That business has been growing high- teens to 20s% for quite a while now. The cross-sell is quite good because nearly all of our clients o f size have some level of digitization opportunity, and frankly, there's lots of other use cases beyond digitization that we're doing, and that's a smaller portion of our business now. We're doing automated processes where we're saving clients tremendous amounts by taking what is largely a manual process and doing it in a much more efficient means and with high accuracy and big scale, and we're seeing that continue to be a larger and larger piece of our recurring business. We have a software as a service model inside our digital business.
That's our DXP deployments are growing. We did a record number of DXP new wins and logos in the fourth quarter. We think that business is gonna continue to grow at a high rate for a long period of time, and that's even before talking about that IRS deal I mentioned. This year we guided to what I think is a conservative posture of $45 million of revenue from the IRS deal in 2026, and we only did $6 million or $7 million on that business last year, $6 million in the quarter. We expect to do over $100 million in that business line next year. Yeah, there's a lot of growth in our digital business, Curt. It's very synergistic with the core.
Okay. Fair enough. The ALM, the lifecycle management business, touched on it a little bit. Again, very robust outlook. I think you said, you know, largest business or could be. Again, just, you know, maybe reiterate kind of why you think it's gonna be bigger and, you know, larger than your core records business. Thinking about, you know, just the allocation between logo expansion, you know, growing wallet with existing customers, M&A, if that's a factor.
Yeah. We and others scope the ALM market as being a $35 billion TAM, so it's a big market, and it's also growing. Why is it growing? There's more IT gear. Frankly, you know, if you look at the market, it segments two ways. One is you got enterprise clients like a BofA or, you know, a whole host of major. All the Fortune 1000 would be very good targets for us. Then you've got the hyperscale business. You know, directionally it's like two-thirds to three-quarters enterprise and the balance is hyperscale.
Now, the hyperscale side of the business is more of a revenue share model where we're helping hyperscale clients, the same clients we sell to on the data center side, I might add, that we're helping them with the decommissioning of servers that are inside their physical data centers. The quantum of those servers keeps expanding because the refresh rate is about five years. As you know, the amount of capital that's been going into data centers has been increasing for quite a while. Now that part of the business is a little thinner margin. It's kinda like teens to 20% because it's a revenue share model.
When we decommission servers for the hyperscalers, the first thing we do is wipe anything that's been written to. That's incredibly important to the hyperscale clients because t hat's all about their brand. That's a key element of the security and trust part of that relationship. We're physically disassembling them and selling off the components into the secondary markets, and it's kind of on average like an 80/20 split, where the hyperscaler gets 80% of whatever we're selling and we get 20%. We gross up the revenue for the whole sale, so the actual margin is relatively low here, but it's very incremental business for us.
We're doing it increasingly in a more and more profitable way because we're getting cost efficiencies. That business, I think, has got a very clear, significant amount of volumetric growth over the next few years as clients continue to decommission. You kind of know the volume's coming because i t's all about supply chain, right? They're not gonna decommission something where they don't have the new gear to write in. It's a very much a planful kind of area of our business with the clients, where it's more concentrated also, to be clear. On the enterprise side, which is the bigger TAM, we're cross-selling increasingly so from our existing client base. Now, for us, of the $630 million we did in revenue last year in ALM, 60% of it was enterprise and 40% of it was hyperscale.
We punch a little bit higher on the hyperscale than the market just because of the more concentrated nature of those clients. On the enterprise side, though, Curt, we are in, I think, probably like the first inning, maybe just getting up to bat on the opportunity to cross-sell. I'll give it to you like this: We have 95% of the Fortune 1000 as a customer. When we started last year, we had about 270 of them using us for some bit of ALM. We exited the year at, like, 360, so we picked up 90 new logos. We're not penetrated anywhere close to full penetration with any of them.
With any of them.
I mean, we're very early stage, and it's very much a land and expand type of situation because we're offering them something that is a replacement of a whole string of vendors across their network because it's all, as I said earlier, it's a really fragmented market. You know, you've got major corporates and even financial institutions that are sending IT gear that has PII, confidential information, et cetera, to lots of small little mom-and-pop vendors around there. Why do they do that today? It's 'cause that's the only thing they've had to be able to send it to. Our view is we're gonna stitch together a global model like we did on records, and we're gonna offer major corporate something that is super valuable to them, which is all about chain of custody, security, privacy, compliance.
This is a secularly growing part of the economy, and it is a secularly more important part of the economy to, like, CIOs and CSOs, et cetera, in the future because, you know, as we all know, PII confidential information, you probably want a compliance certificate from somebody like an Iron Mountain a s opposed to a $30 million or $40 million vendor. In addition, we augment our organic growth through tucking in some small acquisitions. They're small because I said all the p layers that we're competing with, for the most part, are very tiny players. It tucks in really well. We're buying those businesses at a 5x-8x t railing EBITDA, synergies down below 4%. It means it's a means to both extend our capability but also get in the door with certain clients to start doing that expansion. It's a great business.
The margins are like 20s%-30s%, so it's not as strong a margin as our records business, but the quantum of the opportunity is much larger. The TAM for this business is much, much bigger.
Yeah. You know, corollary or similar playbook to your point, right?
Definitely. It's very similar to the way the team here rolled up the records business over the years. You know, if you go back 30 years ago, our records business, or 40 years ago, was like a $3 million-$5 million revenue business annually. Over time, the team grew it and grew it and grew it and created an offering that was super compelling to corporate clients, and that's what we're aiming to do in ALM, Curt.
Okay, fair enough. Maybe just going back to the hyperscaler business exposure in terms of component pricing memory. I guess what is the exposure, market dynamics? How to think about for this year and then, you know, again, just think about the longer term.
I'm gonna give a little bit of context and then come right to your question. When we are decommissioning for the hyperscalers, as I said, the first thing we generally do is wipe anything that's been written to. In some cases, we wipe it twice. Then we're disassembling. We're selling CPUs, drives, memory, et cetera. Memory and again, this is 40% of our ALM business. Memory, quarter to quarter, it can move around a little bit, but it's about 40%-50% of that revenue that we do. It's the lion's share of what we're doing. After that, it's like the drives and CPUs, et cetera. Of course, this is all used gear, just to be clear. We have varying generations of what we're decommissioning based on what the client is sending us.
Memory pricing, as it feels like everybody knows, has been pretty inflationary. Part of that is driven by there's a supply and demand imbalance for new. There's also less capacity for certain generations of memory that's still very much needed like DDR4. As I said in December at a financial conference, and then as we reported on our most recent call, look, memory prices moved up pretty appreciably on the used side in the later part of the year. It was tight on new throughout the year, but really kind of got into used more later in the year. Secondary prices are continuing to be high. You know, we share in that through the revenue share model that I mentioned to you earlier.
In the fourth quarter, look, you know, as compared to when we gave guidance at the very beginning of November to the end of the quarter, we picked up about $15-$16 million of additional business because of memory pricing. You know, the thing to be aware of is pricing has continued to be high. You know, I don't tend to try to prognosticate pricing because, you know, this is it can move around. The folks that do suggest that there's gonna be this supply-demand imbalance for at least, you know, this year, if not throughout all of next year. I think pricing is probably gonna continue to be pretty good. We were, you know, and I hope will prove to be conservative with respect to how we guided because we just basically rolled the current price at that time that we saw.
You know, the other thing to watch out, and I mentioned this in December too, so I'll reiterate it, is, you know, the hyperscaler is only gonna decommission if they have the new gear already here. Right? They have line of sight for any other gear that they need. You know, as I said in December, we saw some clients that were delaying a little bit on decom because they wanted to make sure they had line of sight for the new supply chain. We've seen them do that before in the supply chain crisis. You know, I feel very good about that business, just to be clear. I think it's biased for upside, if you're gonna, you know, ask me one way or the other, and that's because of both the pricing as well as our organic volume.
Got it. Okay. Maybe just touching on data centers again. You know, touch on this a little bit, but in terms of you know, what you're guiding in terms of you know, megawatt capacity or you know, usage, I suppose. Just how you're feeling about the overall pipeline, and you touched on this a little bit ago. Market moving or shifting to inference, what does that mean for Iron Mountain?
Yeah. We generally are making our data center platform in all tier one kind of markets for cloud, and therefore inference in the future and now and in the future. If you look at our business, Curt, you know, we've signed about on average 100 MW of new leasing each of the last four years. It's averaged that. Last year was a little bit lower. It was in the low- 60s MW. We did 40-something MW in the fourth quarter. Part of the reason we were lower last year is we had sold so much in the prior three years that we were constructing that we didn't have near-term power coming to market that we could sell.
The reason why we guided to, you know, at least 100 MW of new leases this year, and that is bolstered based on the fact that our pipeline activity has meaningfully increased. As compared to this time last year, it's up a lot. Now, it kind of should be because if you look at it, we're lucky enough to have a portfolio of energizing land that's coming and, you know, in the next 18 months or so, we will energize 200 MW that is currently unsold. Then in the next 24 months, we'll do like 400 MW. Beyond that, we have another nearly 300 MW to energize out beyond that timeframe. We've got a lot of very, very good available capacity coming to sell. Generally speaking, we see the hyperscalers kind of leasing like, you know, 12-18 months before the energization process.
Which works really well with our model because we can get the pre-lease from one of these investment-grade clients before we really have to do any significant construction.
Got line of sight.
Line of sight and at good returns.
Okay. Terrific. Running tight on time here, so maybe just capital allocation, a favorite of yours, I think. In terms of just outlook for this year and cash flow priorities, how should we think about leverage and then just deployment, dividend growth, stuff like that?
Yeah. There's a bunch in there, so I'll try to go fast. Our operating cash flow, this is another big tailwind to our business, is increasing at quantums that are substantial, this year and going forward. You know, like last year, we had nearly $200 million of cash restructuring charges, and they're gone. You know, that's over. That right there is a big step up. Plus, obviously, the EBITDA of the business is growing really fast and much of that's coming out of businesses that are very capital light. We'll probably do somewhere between like $1.5 billion and $2 billion of operating cash flow this year, Curt. We still have a lot of working capital opportunity going forward together with growth.
To give you a sense, at the midpoint of that would be up like $400+ million year-over-year on the operating cash flow side. Unlike a lot of data centers, we have retained cash flow from our core business. That we will plow into covering incremental data center development. Couple other parts of the capital allocation story. We target a low- 60s% of AFFO for our dividend. The last four years, we've grown the dividend about 10% each of those years. We have guided to AFFO being growing another double digits. In light of the fact that we have no intention to move off that payout range.
You could expect the dividend to continue to grow. We're using what I think is a prudent, probably, if not low leverage level of 5x as a REIT. Principally, you know, a few years ago, we were closer to 6x. We brought it down. Last year, we ended at 4.9x. We haven't been that low in well over a decade. We aim to kind of be right there. With that, with the retained cash flow and the growth of EBITDA, we have plenty of capital to deploy to build out those data centers.
Great. Good way to end.
Good to see you, Curt. Thank you.
Take care.
Appreciate it.