Then with obviously Capital Markets, I cover the communication infrastructure sector, and pleased to have Iron Mountain CEO and CFO with us. Bill Meaney and Barry Hytinen, welcome.
Thanks.
Thank you, Jon.
We are going to hit on data center ALM, some corporate topics, capital allocation. Maybe start with data centers. The energization pipeline declined from roughly 450 MW to 325 MW over the past year as you leased capacity. At the current leasing pace, the pipeline could substantially shrink within the next 12-18 months. How are you thinking about the cadence and scale of land and power replenishment needed to sustain 100+ MW of annual leasing beyond next year?
Okay. Well, thank you, Jon. Maybe I will start with that one. First, we are really happy with the recent momentum that the business has built since January this year. As you noted that when we got on the Q2 call at the beginning of August, we had actually leased already something in the third quarter, over 50 MW in Mumbai, which brings that land bank down to the 325 MW that gets energized over in the next 18-24 months that is not leased yet or is leasable. First of all, we feel really good about that. Over the next 18-24 months to have that much energized and permitted, fully permitted land available to talk to our customers about. The pipeline that we see across those assets is both wide and deep.
We feel really good about the momentum that the business is building over the next period. Your question is really what is beyond that? Beyond that is if we think about it as a company, we think about maintaining that kind of freeboard, if you will. The 300+ MW over an 18- to 24-month rolling basis. If we go the next period out, we have another 300+ MW on additional land that is permitted and is committed to be energized by the utilities after that. 200 MW of which, I should say, is additional capacity coming online in Manassas. If I look forward the next 36- 48 months, we feel really good. That is not to say that we are not continuing to build to our land bank.
In terms of regions or markets, what is your sense of utility delivering pipelines, lengthening, shrinking? Is it any unexpected developments that you are sensing?
I would say both in North America and in Europe, they are definitely lengthening. That being said, the land that we have talked about when we talked about energize, that is where we already have the power that is committed. That is why we are going out even further when we are actually acquiring land for the next 24 months after that. I think both places have become constrained in terms of the amount of capacity in the U.S. Quite frankly, for decades, we relied on efficiency to take care of economic growth. We did not really build a lot of transmission or generating capacity. Now we see that we are making up for kind of lost time or kind of past sins. Europe has a similar dynamic. It is also a little bit more complicated in Europe because of trying to balance the renewables.
We have a data center campus, obviously, in Madrid. I am happy to say that functioned perfectly during the brownout and blackout last summer. That was really trying to balance the increase of renewables into the grid. Then we go to India where we, as I just leased over 50 MW, I cannot have a meeting with a chief minister or one of their deputies in one of the states in India where they say, "How much power do you need? How much land do you need? We want our state to be the largest data center market in India." In India, we get a lot of support at every level to grow the business.
And then just the kind of the pie chart of demand. That can fluctuate between enterprise, hyperscale, social networking, AI startups. What are you sort of seeing and expecting over the next several quarters? What is your calculus around underwriting deals with AI startups essentially?
Maybe I will start with the first bit, and I will let Barry talk about how we think about credit risk in terms of people that we lease to. I think in terms of where we see the demand is that, and I think we talked about this last year when we were here, Iron Mountain, at least to date, has not played in the large language model campuses. Our capacity historically continues to be focused on what I would call the now inference, where they are actually going to run the models. We still see that the top hyperscalers are by far our very customers for our leasing. Over 90% of our leasing activity is to the usual suspects who you think are the largest and most secure credit risk of the cloud providers and AI providers into the infrastructure across the globe.
Those are typically 10-15 year leases. Barry, you might want to talk about some of the Neoclouds and some of the other customers.
Yeah. So Jon, and thanks again for having us here. I would say that the pipeline is very robust, as Bill is mentioning. We have done a lot of repeat business, as you know, with the major cloud hyperscalers. I expect that the vast majority of our business going forward will be continuing with those very high investment grade type clients that we have become a clear partner to over many years now. As it relates to smaller clients or new upstarts, et cetera. We like all of our customers, of course, but I would say that it comes down to economics. We really like the long duration leases with the major high profile cloud hyperscalers, the 10-15 years or longer that Bill was just alluding to.
And we have been writing deals for the last few years at cash-on-cash unlevered returns of like 10%, 11%, 12%, something of that nature. And we certainly get inquiries from other would-be customers, like Neoclouds and others. And to date, our business with those, that portfolio of clients has been quite small. It is 5% or so of our portfolio. And that is partly because, if you look at what we have had available, we have generally been leasing it to the largest players in the industry, Jon.
Look, a few years ago, five, six years ago, Bill and I had a plan with Mark, our head of data center, around, okay, this asset is going to be a colo site for enterprise, that one as well, that one as well, and none of those three assets I was just referring to ended up that way because we had the opportunity to fully lease them to single tenants for a much longer duration. So that is the continued plan.
Yeah.
In light of the pipeline, that is what I expect it to continue to look like, Jon.
So we had a hyperscale on one of the earlier panels talk about their topology and kind of the rigid AZ architecture is morphing into something that can be a bit less stringent. Still need to be relatively close to GDP centers, but does not have to be quite where it used to be. And then there is remote locations. So as you think about your growth into new markets, whether it is domestically or internationally, greenfield, JVs, like with Wisetek or M&A, what the landscape look like in your appetite to allocate capital there versus, say, ALM or other segments?
Maybe I'll start kind of where we're looking, then Barry can talk about how we think about the capital allocation across the portfolio, which might be a good segue to talk about ALM a little bit. But I think that to your point is, either lucky or we're clever is that you think about we were the first to move to Manassas, where people were saying, "Well, that's not Ashburn." And now, Manassas is very much, or Prince William County, is very much considered. People thought it was in the U.K. when we went to Prince William County, but now they realize it's just down the road from Ashburn. Then Richmond, almost the same thing. We built conviction around Richmond about three or four years ago.
We started looking at maybe five years ago, and that's turned out to be a very good asset where people are starting to embrace some of those areas. And I think, given the scarcity of powered and permitted land, I think people are opening the aperture. That being said, is the key markets are the key markets. So we still look when we kind of, I would say, go out of one of those areas where we have line of sight that it can be kind of a tier 1 or close to tier 1 data center market, because that's where our customers first want to go, especially for cloud and inference build-up, which is where our focus is. Then more broadly internationally is that we like India a lot, right? We like a number of market, the key, I would say the FLAP markets plus Madrid.
In Europe a lot. I think those will continue to be there, and there is strong demand both from corporates, from cloud providers, as well as the government in the E.U. on trying to build more data center capacity to catch up on what they feel is they're behind on AI. So I think across the market, in terms of joint ventures, as you said, we started off in India with a joint venture. But that was really because we wanted a partner that we felt comfortable with that could help us navigate the Indian market, especially when it was acquiring land. And we actually had a path to take that to majority, which we did quite quickly by just putting in all the capital for the growth. Then eventually we bought them out.
Then in the Middle East, we had a different approach, is that, the Middle East is more than one country, so we looked for an operator that could really work with us across the region, and it's someone that kind of understood foreign capital like ourselves. And so, Ooredoo, which is the local telecom operator, which is controlled by the QIA. I mean, QIA is a very sophisticated global investor. They understand how people like us think, and they understand how to have a partnership. And I was speaking to the CEO of Ooredoo just last week, and that partnership is working really well. He's getting what he value that we're building a minority, but a footprint across the Middle East, which made a lot of sense for us. Yeah.
Jon, from a capital allocation standpoint, the first thing to note is other than data center, the vast majority of our business grows with very limited CapEx. Our core physical storage business, which continues to grow both on an organic volume basis as well as dollar value. We have never stored more physical volume than we are storing right now, Jon, for clients, it just generates a tremendous amount of cash flow. We are utilizing that cash flow together with what I think is a pretty conservative or prudent level of leverage, which is just under five turns currently. To put that in perspective, seven years ago, we were cresting six times, and we are now at about 4.8.
With five turns of leverage on the incremental EBITDA we have been generating each year, together with the cash generation from the business, because we have actual retained cash flow from our core business, we can fully fund our build-out. Generally speaking, we have got a digital business that is growing teens to 20%, and it is not particularly capital intense. We have got our data center business, which obviously is capital intense. It is somewhere between, let us say, $9 million-$13 million a megawatt to build out, but with very good returns and very good cash generation with excellent clients.
Then we have our ALM business that you mentioned. In the ALM business, the Asset Lifecycle Management, for those people that are not as familiar with that part of our business, this is where we are helping clients with IT gear as it either reaches obsolescence or it is time to refresh or renew it. In that business, we are addressing a massive total addressable market. It is a $35 billion annual TAM. Of that, 75% of it is in what we call the enterprise side. So think corporate clients, Fortune 100, Fortune 1000, that sort of thing, whereby they have gear that is consistently going obsolete or time to refresh every year, every quarter, year in, year out.
What we do in that case is it is largely a fee per service, whereby we are building a worldwide capability to serve clients in a way that they have not historically been served. Today, that market is extraordinarily fragmented, with very significant large number of small, mom-and-pop, kind of founder-led IT asset disposition vendors that service very large corporates and on a regional basis. That is just how the market looks. Today, what we are doing is we are building a worldwide capability to service a client. When you talk to really large corporates, they are very focused on wanting to have this service provided based on chain of custody, consistent process, trust, privacy, security. They recognize, and increasingly so, that anything that has been written to can turn into a liability, because you do not want a hard drive with confidential information or personally identifiable information just getting out there.
Having a company of the wherewithal of Iron Mountain, I think, is extremely compelling proposition. That is a business, just to give you a frame for the growth. in 2021, we did about $30 million of business in the enterprise Asset Lifecycle Management. This year, we will do $600 million. That business is growing both organically and inorganically at a very high rate, and we think we are just getting started there. This cross-sells very well off of our core, where we have 245,000 clients, most of them standardized with us on our records business decades ago, literally. We are working to cross-sell ALM services to them. The other part of our Asset Lifecycle Management business is in hyperscale data center decommissioning.
Here, what we are doing is helping the largest cloud hyperscalers, many of whom are tenants of ours in our data center leasing business, with the decommissioning of servers inside their and third-party data centers. On average, we find that cloud hyperscalers are refreshing that gear about every five years, plus or minus a little bit. They are refreshing for different reasons than on the corporate side. They are doing it because they can get better compute or better power efficiency with new generations of technology that have come along since they installed those older servers. So those servers that come out, they have residual value left in there.
What we and a few other players do for them is we will take that gear in, we will wipe it and give them a clear certification that we have sanitized anything that has been written to, and then we will physically disassemble the server and sell the gear off in a revenue share model, which we can talk about further. This part of the business is also growing quite rapidly because you think about what are we doing? We are working with clients that are refreshing data center infrastructure. As we all know, data centers have been growing quite rapidly over the last, let us say, decade. As that continues to refresh, there is more and more gear. To give you a sense, the addressable market for hyperscale data center, we estimate, is about $3.5 billion, based on last year's number.
Just based on the growth of data centers and what we will refresh over the next few years, we estimate that the TAM for that piece of the business will double to $7 billion in the next four years. So it is a huge growth area. While we have been acquiring on the enterprise side, we are generally allocating a relatively small amount of capital to that part of the business, Jon, because it is not particularly capital intense. As we find new deals, we will acquire, but we have been generally buying in between 5x and 7.5x trailing EBITDA, synergizing those down very rapidly. We have done, I think, seven deals in the last three years, and all of them have been very significant successes. As we find more targets to acquire, we are happy to do that. In every case, it has been founder-led businesses that we have acquired.
We've convinced the founders to come over and work for us as part of an earn-out process, and all of them are still working for us. It's just a super interesting part of our business that is growing rapidly alongside a rapidly growing data center and digital business.
I got one more question, maybe kind of putting it all in the mix here. You got ALM, data centers, digital solutions, obviously the core business. As the mix maybe approaches more 50/50 for your growth businesses, I don't want to say the core business isn't growing. But as that profile emerges, they'll have different margin profiles. So how would you kind of help us think about the next several years and what are the puts and takes around margins, given that each of your segments has a different profile?
First, if you take a step back is, Barry and I have been consistent both five years ago and today that when we embarked on this, the growth strategy, we said that we could actually deliver
Roughly 10%+ AFFO per share growth, and we've done that, and we said that, and my contacts are pretty good, so I can see pretty far ahead, is that we continue to see 10%+ AFFO per share growth. Of course, last quarter, so far this year, we've done a lot better than that. But from an underwriting standpoint, we see that capability. So we start there, right? Because we think with a total TAM for all our products and services of $175 billion, shame on us if we don't continue with approaching $8 billion in sales today, but we don't continue to drive growth with that in mind. So when we look at capital allocation across the businesses, and you're right to point out they have different margins.
They also have different investment profiles in terms of how much capital it takes or fuel that you need to put in it. Some things you can grow much faster with very little fuel and other things to drive the growth, you have to put more fuel into it. We look at it through the capital allocation because we start with what we have underwritten to the community and what we are going to deliver as a team. The 10%, so people should take away is that as we do that is kind of in the back of our mind. As we go, it gets easier. What do I mean by that? Usually when you have a story, as you are compounding on a bigger base, it gets more difficult.
In our case, when we started this, we had 15% of our sales in those growth portfolio, ALM, data center, and digital. Today it is 35%, and as you point out, it would be 50%. At 35%, we are delivering 700 basis points of consolidated growth, and that will just build so the tailwinds of the business become stronger. The last thing I would just leave you with is that if you think about the story that we are underwriting, if I said to you, there is a company out there that has consistently driven 10% AFFO per share growth, that drives 10% annual dividend growth, and maintains leverage flat to slightly down, what would you say the dividend yield of that stock should be today? I think it is probably more like 1% or 2%, and today we are 3%.
Barry and I feel that we have a financial model that allows us to do that without using equity, and we are generating the cash, and the track record will speak to itself, and that we will continue to drive those kind of shareholder returns.
Thanks very much for your time.
Thank you.
Thanks for having us, Jon.