Maybe we'll begin now. So, my name's Calvin Ling. I'm, I run the, Communication Services Banking Business at Morgan Stanley. I'm joined here on stage today by, Barry Hytinen, the CFO of Iron Mountain, and Gillian Tilson, the Head of, Investor Relations. So welcome, and thank you for coming back.
Thanks for having us, Calvin Ling. It's a pleasure to be here.
Great. So maybe before I begin, let me just read the research disclosure quickly. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com Research Disclosures. If you have any questions, please reach out to your Morgan Stanley sales rep. All right, so maybe we can start off, just by, framing the, the opportunity here. I think most, people here would be surprised to learn that, Iron Mountain's been one of the top five performing REITs over the last year. There's obviously also a lot of talk about, AI at this conference and what it means for various businesses. So maybe you could give an overview of Iron Mountain and kind of how the business model, benefits from some of these AI tailwinds across your businesses, and we can start there.
Okay, Calvin, sure. There's at least three areas that come to mind very quickly on AI. First off, would be in our data center business. As you know, our data center business is about 10% of revenues and been growing very fast over the last few years. We have a data center business where we are currently operating about 250+ MW, about 95% of that is leased. We're under construction on a similar amount, where we're about 95% pre-leased. Our total portfolio in data center, if we didn't buy any more land or power, would build out to 860+ MW, but that's not the plan. We're gonna continue to aggregate more land and power.
If I was talking to you a year ago, that number was about 740 MW we've been adding on to the business. We see the opportunity for data center to continue to grow at a high rate. I expect revenue growth this year of at least 20%. A couple of weeks ago, we had our earnings call, and we gave guidance for the year of 100 MW or more of new leasing in the year, which is about 25% more than we said at this time last year. You know, we continue to see the business building, and clearly, the pipeline is seeing more and more AI-oriented hyperscale opportunities. We are continuing to see our deep relationships with several of the major hyperscalers result in considerable incremental opportunity, and that's...
A large portion of that is AI. And so, that's a key factor. Secondly, is in our Asset Lifecycle Management business, where we do a lot of decommissioning of data centers for many of the hyperscalers. And in this case, what I think we will see is, over time, as that gear needs to be decommissioned and retrofitted, let's say, in four or five years, that gear that's going in place right now, that's the year we'll be decommissioning them.
Right.
And that will likely have a higher average selling price in the future, in light of the fact that the new gear has got a higher price. And so that yields a nice tailwind for us as we move forward. Secondly, the second order effect is, as there's just more investment in data center and potentially having some hyperscaler refresh sites to become AI-oriented faster, that gives us another opportunity to see incremental decommissioning over the next few years. So I think the price and volume side on ALM is a long-term tailwind in light of AI. And then lastly, but certainly there are other ways that AI positively affects our business. But the other one that I would highlight is within our own infrastructure.
We have a center of excellence focused on AI, where we are doing a variety of examples of case studies where it can drive significant cost efficiency for us, help us do things more productively. That's in the way of, say, one that comes to mind is in customer care, customer support. We have over 225,000 client relationships. We do still a lot of person-to-person contact in the form of emails, phone calls, chats, and even WhatsApps. And AI gives us an opportunity to be much more efficient in that area and be able to get the client what they need, or also get our customer service representative the information they need to be able to be very responsive to the client.
There's lots of other AI opportunities in the business, whether it be in our, finance and accounting areas, where we're seeing opportunities for transaction processing improvements, whether it be in IT, you name it. There's a lot of opportunity in AI for us. And then, of course, I guess I should also point out, we've got a large and growing digitization, digital services business, and we see AI requirements that the clients have, driving need for our digitization, offerings. So there's, there's quite a few places where AI shows up in the company.
It's really interesting. Maybe we can dive into each of those in a bit more detail. Maybe I'll start with data centers, just because that's the first one you mentioned. I think we're hearing a lot of the similar AI comments from some of the other data center folks here at our conference, and even just across the street, they're clearly seeing that demand too. How do you think or how does Iron Mountain think about funding that growth, fueling the capital, obviously, given the confines of your own balance sheet? I know you've done some JVs in the past in areas like Europe, but maybe you can just talk about your data center strategy, in particular, how funding it is planned.
Okay. As we said, for the last few years, we're gonna continue to ramp our development dollars for data center, and it's completely driven by the fact that we pre-leased so much activity. When you look at what we have under construction, around 250 or so MW, the vast majority of that, like 90+, 95%, is already pre-leased. And so that's to the major hyperscalers. And what they're expecting us to do is deliver them a quality product with power on budget, on time. And so to do that, you know, we are very systematic and process-oriented about our construction and our design, and of course, before that, our site selection.
And so, we have the situation where we're not building to spec really at this point, Calvin. We're really building-
Right
... to client demand. And as it relates to funding, a couple things there. One, you know, we're obviously growing EBITDA a lot over the last several years and expect that to continue to grow. When you look at our Global Records Management segment, that is a very cash generative business, and that is the bulk of the company at this point. It requires very limited CapEx to grow and has very high incremental margins. And as a result, that business throws off a lot of cash, which we can then go to reinvest in the company, in our data center development, which is the primary call on the capital for the company.
Together with that, we have a leverage target range of 4.5-5.5 times, where at the end of last year and for quite a while now, we've been right at the midpoint of that, like 5, 5.1 times. And we're at the lowest level of leverage the company's been at for about a decade. And so you should expect us to continue to operate inside that leverage target range, which gives us a fair amount of incremental borrowing capacity in light of the growth of EBITDA. Together with the cash-generative nature of the business, it more than funds the company's growth as well as our dividend.
Got it. So with that kind of liquidity and access to capital, how do you think about inorganic opportunities in this, in the data center sector?
Yeah. So, what you should expect is we are going for the development returns, and so, you know, we are going to continue to acquire land to build and power to build out our platform, as you've seen us do over the last several years. We've expanded the platform from what could have been, say, 500 MW a few years ago, to now over 860 MW. It'll continue to grow. It is possible that we could do some small acquisitions in the form of, like, a very small operating site that has power, because we really look at that more as almost like a land deal.
We did that in Madrid not that long ago, where we bought a very small operating data center, but it's because we saw the opportunity in light of how much land and power the site had, and we felt, and we believe very strongly, it's going to be very interesting to a large, large number of hyperscalers. And so you might see us do that, but what you should not see, and this goes to my earlier point about the development returns, is you should not expect us to be buying any sort of stabilized assets or anything of that sort. Because we see the opportunity with our network of client relationships, with being on a shortlist, so to speak, of their vendor list of the hyperscalers, that that's how we kind of create value, is by aggregating land and power and then building it out.
Got it. And then what do you see as the longer end goal or strategy for your data center business? I mean, there's some folks in the past have seen separations of businesses like that where, you know, they're valued higher by the market than your core business. And I think you guys have a pretty interesting value proposition as it relates to your data center business with your other two businesses that you mentioned.
Yeah. So one of the things that, you look at when, when you see our business, one of the things that we think is a differentiator is, with our client relationships, there's a large opportunity for cross-selling across all of our solutions. So the vast majority of our clients, nearly all of them, are records management clients, but still relatively few use our digitization services or use our ALM, asset level, Asset Lifecycle Management solutions or use our data center solutions. And there's a lot of cross-sell opportunity between those. We have a and a stated intention of growing the business. Said this at our Investor Day in 2022, that we'd grow 10% revenue CAGR between 2021 and 2026. We're actually ahead of that, over the first two years.
With that, we expected a considerable amount of growth in Data Center. To drive that growth across the total company, including Data Center, cross-selling is a big initiative. In fact, we've changed management compensation plans to incentivize having cross-selling as a major feature, as well as our comp plans of our sales team to be focused on cross-selling as well. So when I look at the cross-selling that we've done recently, I've been talking about on the last few calls, we're seeing the vast majority of our Data Center megawatts are the direct result of cross-selling off of other parts of the business.
And then, as you know, we've recently been expanding our capability and the size of our business in Asset Lifecycle Management, where we're frequently doing a level of decommissioning for hyperscalers of their data centers, together with enterprise, IT asset disposition, recycle, reuse. There's a tremendous amount of synergy between those businesses, between the ALM business and the data center. Frankly, we're talking to clients, major hyperscalers, about scenarios where they lease an entire building from us, and then we do the retrofitting of it through our ALM business multiple times over the time that they're in that site. Similarly, we have situations where our ALM business is bringing us in on the data center side.
Right.
Because we have while we have very significant relationships with many of the largest hyperscalers, we have some relationships that are unique on the ALM side and not in the data center today, and the same the other way around. So there's a great degree, I think, of synergy between those two businesses, as well as, frankly, the rest of our company.
Got it. Maybe we could actually just pivot to the ALM business, on, on that front. I think people would be surprised that you're the largest player in that segment. You've been pretty acquisitive, or you've made some acquisitions, ITRenew and Regency, most recently, to help you cement that status. Maybe you could just talk about strategy in that business is, how those two acquisitions help set you up for what you have today, and we can go from there.
Okay. So, just a little bit of context. Asset Lifecycle Management, we think, is a very large market. It's something on the order of a TAM of $30 billion.
Right.
Probably half of that is served, meaning clients also do a lot of ALM on their own.
Right.
They haven't yet used a vendor. So when we looked at that business and that market some years ago, we started making entries on an organic basis in 2017, and we found that it was very good from a standpoint of cross-selling off our records clients to get them to use our ALM solutions on the enterprise side. So think about a large records client. Why did they standardize with us? They standardized with us on the record side, usually because of the chain of custody, privacy, wanting consistent process, and being able to follow them in any market around the world that they do business in. And we're really the one of the only players that can meet all those requirements for a large client on the record side.
They have similar needs on the Asset Lifecycle Management side, and frankly, this is becoming a larger and larger concern for most clients because of the privacy nature and things that can be written to in their IT environment. Frankly, there's just a lot more gear and there's more refreshment. So that market, we think, is going to secularly continue to grow. Going back to when we started investing in it, over the next several years, from 2017 on, we were growing the business at a very high rate-
Right
... say, like 30+% CAGR, and but it's still relatively small. When we looked at the opportunity to continue to cross-sell off our records business, we think it's just immense. And so we wanted to get into the space in a much larger way because it's a very fragmented market, as you say.
Right.
We're already the largest player, and we think markets of this size, with this sort of profit pool, they don't generally stay fragmented for long periods of time. So there is likely going to be, and there is market leadership drives economic benefit.
Right.
And so while I think there's a lot of space for lots of competitors in that market, we think we can offer something that's truly compelling, especially to our large and medium-sized clients, who have needs across the world or across various regions of a market. So it is a space that you're gonna continue to see us play in. You noted that we just recently acquired Regency Technologies. They are a U.S.--primarily a U.S. player. They have eight sites where they do the processing of enterprise gear, and it's really kind of a, from start to finish. So that team wins client relationships and then goes and gets gear, processes it, in some cases, like, mines the gear for even, like, copper out of wires and things of that nature, and then sells it on or recycles or reuses it.
And it's a business that's been in development for over 20 years now and has some very long-term relationships. Importantly, they got a lot of incremental capacity as well in their processing sites. And the thing that I wanna help people understand about what Iron Mountain has historically done, in many cases, we would go get the gear from our clients, and then we'd sub out some of the processing. With Regency, we can internalize that-
Right
... and we can thereby capture more of the value chain and be that much more compelling of a solution for our clients. It gives us more reach, and frankly, the team at Regency really knows what they're doing. So they're taking on a larger level of leadership responsibility in our operations of Asset Lifecycle Management as we speak, actually.
Got it. So and obviously, these businesses are growing nicely on the top line, like you mentioned, you know, well into the 20%-30%. What does it do for you on the bottom line and from a cash flow perspective as well?
Yeah. So I'll take Data Center on that one as well, since we didn't talk about it. So on Data Center, you know, we're driving about a low- to mid-40s% EBITDA margin, and we expect that to kind of be at that range or even slightly move up as we get more efficiency. But I'll say, right now, we've actually been somewhat investing ahead of the growth, and that's one of the reasons why the margin hasn't gone up even more than it has, because, you know, we're seeing such significant growth. It isn't that many years ago on the Data Center side where we were doing, like, 10 MW a year. Last year, we did 124 MW. The year before that, we did over 100 MW. We've set our guidance for 100 MW this year.
We're adding to our site selection team, we're adding to our construction-
Right
...organization, we're adding to our sales team because we just see the pipeline continuing to build and at record levels. On the, So I expect that's a very good margin incremental to the company. On the Asset Lifecycle Management side, I'll give the margin on the two pieces of the business I mentioned. On the hyperscale decommissioning, that's generally a revenue share model, whereby if we decommission a server and start selling the components, we get a percent of whatever we sell. So think something like, depending on the client relationship and the gear, maybe 20 points is what we'd get. So the margins on that business are relatively lower, so think teens to 20, as pricing continues to come back. As we've talked about publicly many times before, component pricing has been very depressed and still is at pretty low levels-
Right
... albeit it sort of came off the bottom in the fourth quarter some and is expected to continue to rise. That'll help us very significantly because it'll both drive up the revenue as well as the profitability, because of course, the cost to decommission-
Right
... a server is similar for us, whether or not we're selling it for X or X divided by two. And then on the enterprise side, that's more of a service offering. It's not generally a revenue share, and the margins there would be more like 20%-30%. And both businesses, importantly, have decent operating leverage. So as we scale them, and I think we can grow those businesses quite appreciably, our targets are out there for that total business, ALM, for us to be a nearly a billion-dollar business in the next few years; that is big. You know, there's a good amount of operating leverage, but it will have a little bit of a mix shift on the company if we grow it really fast, because of course the company's total margin is ahead of those.
Got it. And actually, while we're on the financial topic, then maybe you could also talk about the CapEx differences between the two businesses. Like, obviously, data center is more capital-intensive. What kind of returns are you targeting there? And then contrast that with the CapEx requirements for the ALM business.
Yeah, thank you for that, Calvin, because I should have mentioned, one of the other reasons we like the ALM space is it is not particularly capital-intensive at all. The cost to set up an incremental processing site is relatively low, and as I mentioned, we just acquired a bunch that have incremental capacity. And similarly, on the IT decommissioning side of the hyperscale side, we've got a lot of potential capacity, and we just need to bring on more labor as volume comes in. So the returns there are good, very good. And then on the data center side, it is, as you know, a very capital-intensive business. But with that come those long-term leases we've signed with the hyperscalers, exceptionally strong tenants in the Big Tech , 15-year leases with multiple options to renew.
What we're seeing is returns have been rising, as we've mentioned on the most recent couple of calls. If I look at sort of hyperscale contracts that we're writing, they'd be in the nines plus on a cash-on-cash, unlevered return at this point, which has been rising pretty steadily over the last, you know, year and a half or so. But of course, it should, because construction costs have moved up some-
Right
During that period of time, coming down a little bit now. But also, you know, cost of borrowing has moved up. So I think that the returns on data center are quite attractive. We like the space.
Okay, great. Then, the last question on the ALM side, like, good cash flow profile, high growth. What's the growth strategy there, aside from organic stuff? Can we expect more acquisitions or inorganic stuff, like more Regencys, or how are you thinking about that?
Well, couple things about that. One, as you kind of highlighted in one of your earlier questions, I think we're probably the largest player in the space now. And so the next largest player might be $200 million, and there's only, you know, I think two or three-
Right
that I'm aware of that are that size. And then after that, it's a very long tail. I think we just bought. I know I'm biased, but I think we just acquired the best operator in Regency. They're a really strong team, and we've known that team for quite a while. We've been kind of talking to the team there and looking at that deal for a couple of years, nearly. And so we're very pleased to get that one. I think we can learn a lot, and there's synergies in both directions. In terms of one of the benefits of being the largest player, is we certainly see, you know, all the deals that are out there.
Right.
You know, we and people know that we've been acquisitive in this space, and it is an area that we have been willing to allocate more capital to. You should expect the vast majority of the growth to be organic, because that's certainly our plan, but we will certainly be facts and circumstances driven. There could be scenarios in a market where we get faster time to market to serve a client. If we were to make an acquisition, that's possible. I will tell you, the thing about deals in the ALM space is, for the most part, I think they're very attractive because the cost is relatively low compared to some other spaces. So, for example, on Regency, we paid a 7.5x multiple.
I think most of the ALM deals that I see out there are sort of like high single digit multiple-
you know, at the highest. So, you know, that's a space where I think if, if we see the right opportunity, we, we will go after it, but that isn't like plan A, Calvin.
Got it. Got it. Okay. Well, maybe I'd be remiss not to ask about your core business. You mentioned that was the most cash flow generative segment of the business. But maybe you could touch on the records management business from two aspects. Just one, just how is that steady cash flow progressing over time? And then, two, where do you see areas of growth for that businesses or frankly, challenges, too?
Yeah. So in our Global RIM segment, which is the vast majority of it is storage, and then there's a fair bit of service as well, it's been growing quite nicely over the last several years. In fact, last year, that business grew about 9%. And on the storage side, the algorithm that I suggest people use is that we expect volume on the physical storage side to be flat to slightly up.
Right.
That's what I've been saying for, well, as long as I've been with the company. I've been with the company now a little over four years, and we've been-- last year, we were slightly up on volume. We expect to be slightly up again this year. And then, we expect services that go along with the storage to kind of be riding along with that. But the, the thing that's been driving a lot of our service growth over the last few quarters and the last couple of years is our Digital Solutions business, which is inside that, records management segment. And that's a business that we think can go from strength to strength. We've mentioned that for several years now, it's been growing double digit. We expect it to continue to grow at those levels because there's just secularly more demand.
Coming back to your earlier question about AI, frankly, clients looking to get, the ability to, look at data that we're storing for them that they've never been able to before. You know, they're just taking essentially hard copy paper and converting it into information that they can actually use-
Right
in their business processes, which creates a lot of value. And we're continuing to invest quite considerably in our Digital Solutions to derive more, more opportunities. So whether that be in the form of recently, we did Digital Mailroom , we have Smart Sort. There's a whole bunch of services we offer clients, and we think we can continue to grow. And then, of course, we've been driving a lot of revenue growth from Revenue Management, and that's where we try to make sure that we are delivering considerable amount of value for the client and pricing appropriately for that value. And so when I look at the sorts of service offerings that we offer, that I think really we are unique in the space and only we can truly offer to clients, whether it be on Smart Sort or on-demand digitization, same-day digitization, et cetera.
These are value-add services that clients, you know, as I said, value. And as a result of that, we've seen Revenue Management being running in the high single digit, mid- to high-single-digit level for the last couple of years. And as we've said for a long time, we think something on the order of like mid-single-digit is a reasonable place to plan year to year, plus or minus a little bit. And I will just round that all out by saying, on our last earnings call, when we looked at our guidance, somebody asked me, "You know, how do you think about the midpoint?" So, the total company's growth is in the midpoint of our guidance for this year is 11%. And when you work through it, that assumed data center would be growing in excess of 20%.
As I've said before, we have very high visibility on that. And then the ALM business, there's a component, of course, that comes in with the acquisition, and then there's an organic piece. That information is available on the last call, so I won't reiterate it. But then that works into our global RIM segment, needs to grow a little over 6%, let's call it 6.3%. Last year, it grew 9%. So you know, we feel quite good about where we're positioned and, you know, continue to see similar trends in volume.
Just on any challenges or what keeps you up at night with respect to the global, sort of the records management business?
Well, I think, you know, we try to focus on things that are within our control. And so for our business, that's keeping up with growth, that's keeping up with customer demand, that's creating solutions that clients want to have, investing in projects that are going to yield return. You don't ever get everything right, but I think we've been doing quite well in terms of how we're servicing our clients. We're pleased with our customer satisfaction scores. They can always be better. Client's the most important thing. And we continue to make sure that we're kind of investing at appropriate levels in our key growth areas, and that we are performance managing the organization day in and day out, because ours is a business that there's a lot of people in the organization.
Right.
We have 30,000 people. We need to make sure everybody kind of knows what the goals of the company are and what their portion of how to deliver that is, and that we're executing to serve our clients consistently all the time, everywhere.
Got it. And maybe keying off of that, the corporate comment, and zooming out a bit, you announced, kind of Project Matterhorn a few analyst days ago to help try to drive efficiency and streamline some of the company. Any updates on that front as it relates to your cost structure or anything else we can expect there?
Yeah. So for everybody's benefit, that maybe doesn't know our Matterhorn goals, we said in 2022 that our CAGR for 2021-2026 would be 10% on revenue, 10% on EBITDA, and 8% on AFFO. And as I mentioned on the last earnings call, we're running ahead of that on all those metrics, like multiple, you know, 100 basis points, 200 basis points ahead on each one of those metrics, and in some cases, well ahead. And even on a reported basis, because the dollar's been strong-
Right
... during this period of time, we're at or better than those CAGRs. And if you look at our guidance for this year, it was 11% and 12% on a reported basis, revenue and EBITDA, and AFFO was 9%. So we're expecting this year to be at the midpoint ahead of those long-term CAGRs as well. So you asked about productivity. There's a lot of opportunity in our business. I mentioned some of those that are coming from AI in response to one of your earlier questions, but there's significant opportunity in our operations and in our SG&A. You know, the team at this company has done a great job driving productivity over many years and improving the margin. It's not that long ago that we were kind of like a 20s% EBITDA margin business-
Right
... and now we're in the 30s. But I think there's considerable more. We are constantly evaluating which projects to execute against. And the nice thing about our business is, since we have consistent processes around the world in terms of how we serve, we can test things in one market and then proliferate them across the company after we see benefit.
Right.
So there's benefits in terms of continuing to improve routing, improving the way we service, improving the way we're dealing within the warehouses. Then, of course, we're getting more scale and productivity on how we construct data centers, how we service our ALM clients as we're ramping that business up. There's a lot in our gross margin. Then on the SG&A side, you know, frankly, there's a lot of leverage because we do not need the SG&As to grow at the level that the total company is growing. We see considerable opportunities for driving productivity across the infrastructure of that company. So there's a lot going on with Matterhorn. The important takeaway, I think, is that we are, we have a fully funded plan-
Yeah
... to deliver that kind of growth rate, and, we're very much on track, if not ahead. And I think there's a tremendous amount of opportunity. One of the key initiatives, as I mentioned earlier, is cross-selling, to drive that kind of revenue growth. And we have seen our cross-selling metrics increasing quarter by quarter as our sales team has been able to navigate within clients to get more big share of wallet from them. Importantly, we are still very low penetrated across the vast majority of our solutions-
Right
... within the client stack. So there's a big cross-selling opportunity over many years here.
... Maybe on that point, you talked about a fully funded business plan. We can talk a little bit about the financial policies and leverage targets that you've set out. You mentioned 4.5-5.5x EBITDA. Is that still an appropriate range given this new interest rate environment? How do you think about leverage as it relates to all the growth that you've just talked about?
Yeah. So we've had that target range for quite a few years now, and we do feel it is the appropriate level. It gives us the opportunity to operate, and, you know, with, given the size of our EBITDA, with a fair amount of room inside that. We're operating right at the midpoint at this point, and we expect to to operate inside it. Never say never, that we couldn't someday go outside, but I don't foresee that. And that gives us a great degree of financial flexibility. The capital markets have been very open to us from a, from a debt standpoint, and I'd say from a commercial banking standpoint, we're already rated better than we are from the agencies there, and we issue at a very competitive rate. And so I, you know, I feel very good about our access to capital.
You know, as I said on the last call, we had nearly $3 billion of liquidity-
Yeah
... at year-end. So we've got, we've got ample capacity to continue to build out our infrastructure. And so I think that is the right leverage range. And then from a standpoint of dividend policy, we've had a long-standing target ratio there, where we target to pay low- to mid-60s% as a percentage of AFFO per share in our dividend. Years ago, we were much higher than that, and we had-- we, we did not raise the dividend for several years as we kind of gradually came down into that target range. At the same time, we were lowering our leverage, I might add. And so we are now inside that range.
Right.
I think on a trailing basis, we were 62% at year-end. We raised the dividend in the second half of last year when we got into that range, and we've reiterated that that's the payout ratio range. So you should be anticipating, as AFFO per share rate grows, that it's a bit of a forcing function-
Got it
...on the dividend because we don't aim to be outside that range.
What about on the return of capital, just policy? Obviously, the dividend's there. How do you think about buying back your shares potentially, or even frankly, issuing shares to fund some of this growth?
Yeah. So, we've got a fully funded plan, as I've said many times, and that is thanks to the growth in EBITDA, capacity on debt and our target range. And so, you know, you mentioned earlier that occasionally, one or two times we've done some JVs on the data center side.
Right.
Those are opportunistic as another source. Then we've also incidentally done a little bit of asset-level financing in terms of construction debt on specific assets, whereby essentially we're using the client's balance sheet-
Right
... to get an interest rate that's even lower than what we could issue at, at times. And so we got ample capacity. We don't need any, you know, our business model is not predicated on equity. And as you've known, as you know, we've said many times that we don't, we don't see that. In terms of buying back stock, what I would say is, similarly, we've got a lot of growth ahead of us. And as you've mentioned earlier, data center is a fairly capital-intensive business, and we think the value we create for shareholders through continuing to build out our platform and become a larger and larger, partner to the hyperscalers and have data center be a larger and larger piece of our business is, is the right model for us in light of the returns there.
We'll always be very returns oriented, of course, but I would say that at this point, the marginal dollars are going into data center.
Got it. Got it. Maybe I'll see if there's any questions in the audience.
Can you talk a little bit about the sort of the contribution of data business, data center business today, and how you see that kind of percentage grow in next 3-5 years?
Oh.
Also talk a little bit about the competition in the data center business.
Okay, thank you for that question. Ours is a business that's been growing quite rapidly in data center over the last several years. We've actually been in the data center business for quite a few years, but if you go back to, say, 2018, we would have been leasing something in the vicinity of 10 MW. As I mentioned earlier, last year, we did well over 100 MW of new leasing. If you look at our total data center opportunity, as I said, we have- we could build out to 861 MW. That's the land and the power we already have assured. Over time, you should expect us to continue to be growing that piece, but if we, if we...
Just as a planning point here today, if we only built out to what we have, that'd be 861 MW. We're operating about 255 MW, and we are under construction on a similar amount, where we're nearly completely pre-leased. So just by constructing and finishing the construction of what we have under construction, we're gonna nearly double the MW we're operating, and those are on pretty good returns. As I mentioned, the returns have been rising on hyperscale. And so you should continue to expect us to be growing at a fairly high clip. In fact, in the Investor Day model that I shared and I was speaking about earlier in response to Calvin's question, that had data center growing at a CAGR of about 23%.
If you look at our last quarter, I think we grew storage about 35%. So there's a high degree of visibility on commencements. We had a nice commencement increase in the third quarter and the fourth quarter. You can see from my supplemental tables that we have a lot of construction that will commence over the next year, two years and three years, and we're continuing to see a considerable amount of pipeline. From a competition standpoint, you know, there is one player that's kind of got the waterfront property, so to speak, and they get paid a relatively high price for what they offer clients. We don't generally compete in that space.
We're competing with a host of other data centers, but I would say it's very market by market driven in terms of, you know, the folks that we compete with in Phoenix are a different set than generally we compete within in Northern Virginia or that we compete with in Frankfurt. And frankly, it can even be down to the specific network or region of Frankfurt, or the network within Phoenix in terms of east side, west side, south, et cetera. But what I would say is, you know, there's healthy competition, but what we're increasingly seeing is, particularly on the hyperscale side, the hyperscale clients seem to be, you know, kind of, shortening their vendor list.
They don't want to deal with lots of partners, where I think some of them have gotten burned by folks that didn't maybe have the capacity to construct on time and to deliver the power consistently. We've demonstrated that, as have several of our peers, and I think the opportunity is to continue to grow our share with hyperscalers, bring on incremental hyperscale clients, which are in our pipeline, and, you know, continue to grow our colo business, but that will naturally be slower growth than the hyperscale side. I feel very good about where we are. Incidentally, pricing continues to be quite good on both the colo side as well as the hyperscale side. Thanks for that question. I think there's another question over here.
Just when you think through the power demands, it seems to be like a pretty big bottleneck regarding data center construction, particularly in certain regions as demand has, you know, gone up with AI. How do you solve for that? Or are you thinking about going into new regions, new areas with, you know, power might be more plentiful? Just, can you talk about that? I know it's a long lead time to build a data center, but the demand seems to be, you know, multi years out, and just trying to solve for that equation seems to be like a pretty, you know, topical point. So I'd love to-
Yeah.
As an operator, I'd love to understand how you view the power constraints and the opportunity.
Well, in some markets, it's very real. Thankfully, we haven't had those sorts of issues. I mean, when I talk about the portfolio we have, that's the land and the power assured. So, you know, Northern Virginia is a market that people have talked about a lot of power constraints, and we're over in Manassas, where there really hasn't been any issue, and frankly, we're building our own substation there. We're building a substation in Phoenix as well, where we'll be able to, you know, basically have our own supply, you know, with these very large subs that can feed our campus and give us an incremental efficiency. But it is a factor that in some markets, they're power constrained.
We haven't experienced that ourselves, and I don't see that in the sites that we've got under development or scheduled to develop in the future. We are pretty choosy, though, to your point about where we're going. In fact, one of the reasons we acquired that business in Madrid that I mentioned, is because we see that as... We and hyperscalers, I think, see that as a compelling market from a power standpoint, as well as just general demand dynamics, because as you probably know, it's off the North African grid, it's not off the main European grid. And also they're doing a lot of things with green hydrogen or testing things with green hydrogen in Spain and Iberia. So I think it's a, it's a key market.
You will see us continue to be very focused on aggregating land and power in key markets where we see big opportunity. And I'll tell you that, that includes places like India, Europe, and North America for sure.
I was just thinking about the 250 MW that you have under construction now, what is the time frame for when they go from construction to operation? Assume you start recognizing revenue when those facilities are fully operational. And then how does the pricing work up or down? How often is it reevaluated contractually with your hyperscale customers?
Okay, and I'll be quick because I know we're out of time. But what I will say about that is, if you look at the supplemental report that we publish quarterly, you can see the expected time to when the construction will complete, as well as when it'll stabilize as an asset. And generally speaking, there are assets that are finishing construction this year, as well as over the next 2 or 3 years in there. To the earlier question, it is a business that tends to be planful, in that you can sort of see the when these things are going to need to be available for the client.
In terms of pricing, on the multi-year relationships we have with hyperscalers, say, a 10- or 15-year lease, they, those are very bespoke contracts, as you'd imagine, because they're oftentimes for entire buildings, and they would have discrete escalators in them. Traditionally, that's on an annual basis. The nice thing about that is, it's a fairly open book business in terms of with the hyperscalers. So they generally know what it costs to construct within reason. We all know what the cost of capital is, and we've seen the returns rising, as I mentioned in response to Calvin's question earlier. I appreciate everybody's question, and those of you in person as well as on the web, thank you for having us, Kalvin.
Yeah, thank you for coming.