Conference Transcript - EQIX 2026-09-14

Brendan Lynch, Analyst — Barclays

All right. Good morning, everyone. I'm Brendan Lynch. I cover REITs here at Barclays. Very happy to be with the Equinix team, Arkel Shaw and Ryan Burke. Arkel, you have been the president of the Americas for just over about a year now. Maybe you could tell us a little bit about your responsibilities and maybe about your prior responsibilities as SVP of sales in the Americas.

Speaker 1

Sure. Thank you. And it's a pleasure to be here today. And thank you for having us. So in my capacity as the president, I am responsible for our growth strategy in the Americas. And the Americas for us is everything from Canada all the way down to Chile and Argentina. And I have responsibility for corporate development, I have responsibility for our business development, and then also the growth that comes from those investments that we're making. And prior to that, I ran the sales for the Americas as SVP of sales. So I did that for about six years. So I think that brought a unique lens to where I see things from as a president now.

Brendan Lynch, Analyst — Barclays

Great. I think you kind of led the change in the go-to-market strategy in your prior role at the firm. Maybe you could talk a little bit about the changes that you wanted to see and how you executed on that.

Speaker 1

Yeah. So I came into the company, you know, we've been such an incredible story of growth over our 28 plus years. And when I came in about seven years ago, it'll actually be officially seven years at the end of this month. We were doing exceptionally well and we needed more. And so we had been approaching the enterprise space very effectively. We were starting to see, and there was a journey to the cloud. There was starting to be a conversation around repatriation potentially, and we also wanted to significantly increase revenue growth. And so what I went about doing when I came in is putting together probably a more structured approach against the segments that we had. Rather than treating every customer equally, we have small and medium-sized customers and large customers. We had verticalization that was important in terms of stronger. We played very well in the financial services, for example, so making sure that we had the right focus around that. But it was really bringing a consistent focus around how we managed the go-to-market strategy against the different customers that we had so we could really differentiate. One of the things that we saw was the fact that we had so many Fortune 500, Fortune 100, Fortune 1,000 customers, but we had a small bit of their business. And so really understanding the role. I think people, seven years ago, people thought of data centers as a box and didn't know very much about what happened in that box. And we have such a different value proposition in terms of an ecosystem that lives in that. So working with our customers around the digital transformation that they were going through and aligning our sales organization so they could effectively go after that opportunity in the market. And it proved to be quite good in terms of the growth rate that we saw coming off of that and bookings growth.

Brendan Lynch, Analyst — Barclays

I think the whole world is getting an education in data centers over the past couple of years.

Speaker 1

Yes, definitely, definitely.

Brendan Lynch, Analyst — Barclays

In terms of kind of the go-to-market strategy, certainly the data center market has evolved over the last two to three years. How should we think about your kind of positioning the sales force and your go-to-market strategy now? And is it continuing to evolve as we see kind of AI enterprise and AI inferencing demand start to ramp?

Speaker 1

Yeah, absolutely. So we have a new CRO. He's not new anymore. I think he can officially be called tenured. He's been here over a year, but Shane Paladin. And he has introduced, again, a go-to-market strategy that's evolved with what's happening in the marketplace, which is important for any sales organization to do. And so what we're seeing is the differentiation in terms of where the opportunity is and the different ways in terms of how customers are served. We have almost 11,000 customers, which is really different from any other data center operator. and those customers are not all served the same way and we have very small and we have very large customers and everything in between. And so being able to effectively serve those customers in a way that allows us to move as quickly as is necessary to capture the opportunity, but to do it in a way that is meaningful for that segment is important. So for example, in the small size business, we use many more partners and we have a robust channel organization whereas you go up market and you start to look at some of the very large enterprises that we have and they have very dedicated, bigger teams that support them on a global basis and then everything in between. So that's how we align.

Brendan Lynch, Analyst — Barclays

Maybe we talk about enterprise demand and how customers' capacity requirements are evolving as AI adoption accelerates and what you're seeing on the ground there.

Speaker 1

Well, it's definitely evolving. And what I would say is we've seen this over the last few years just in terms of the increased size of capacity that they need. So the workloads that they're placing with us in our data centers and across the industry are drawing more power. They require greater density. And so we're seeing that shift. So whereas a 250 kVA deal was a large deal for us, a megawatt is kind of the going thing now. And that's a big shift. It's occurred over a number of years. But what we're also seeing is AI has really taken off. Enterprise is a little bit slower in that, and I think there's reasons for that. The enterprises are the businesses that have to make the investment in infrastructure and all of the changes that AI might be driving. And so I think that up until last year, I was just sharing this in one of the meetings we were in, you know, customers were, you know, if you talk to an AI company or you talk to a hyperscaler, they were all in on what was happening with AI. When you talk to enterprise customers, we ask the question of, what's your AI strategy? And the majority in the room were contemplating it, but had not embarked on developing a strategy. They were dealing with what we call shadow AI shops, where there were people using AI, but they still hadn't stepped into actually driving a strategy. This year, when we met with that same group of customers, they all were heavily invested in building that strategy and implementing a strategy. And that's reflected in what we see customers doing with us and the work that they're doing trying to understand the growth patterns that they have.

Brendan Lynch, Analyst — Barclays

Customers commonly think of Equinix as the go-to location for latency-sensitive workloads. Maybe you could elaborate on what element of demand is coming from those latency-sensitive workloads versus maybe some other types of AI workloads that might not be as latency-sensitive.

Speaker 1

Sure. So I think, and Ryan, you want to jump in here at any point. So I think what we're seeing, first of all, is the latency sensitivity has always been a factor in workloads. And AI is potentially driving a greater need for latency sensitivity. We're still looking at that because there are some, you know, what we're seeing from what customers are telling us is not all AI workloads are the same, just like any other workload is not, they're all different as well. And so what we're finding is that the customers, when we build, it's probably important to do this, when we build a data center, it's typically within what we call a metro. It's a campus of data centers. And we are very thoughtful about where we place the workloads in that data center, depending on the latency that's required. And so in non-traditional, before AI, you think about financial services, high-frequency traders, where the latency is so incredibly important. What you start to see with AI is the response time on more and more applications is really important. So if you're ordering an Uber, you can tell, right? within seconds if you haven't gotten a response on your phone. And AI is driving more and more of that need for instant response rates on that, and that's driving latency sensitivity associated with that. We also build our data centers in the sense that, again, as I said, it's almost like concentric circles where the high sensitivity around latency gets placed in a certain part, certain IBXs, and where it's not necessary, will work with the customer to place those in different areas.

Brendan Lynch, Analyst — Barclays

I think that's an interesting example of the Uber because for a lot of people it's hard to envision what the latency sensitive, why it's important when you can in some ways get low latency from across the country but I believe every time we order an Uber there are 300 workloads that are involved with making sure that that journey goes successfully and getting those all timed right and with super low latency is part of the reason that it works the way that it does.

Speaker 1

Yeah, that's an incredibly good point because the complexity of what takes place to make that happen is something that most people don't know about. Sure, sure.

Brendan Lynch, Analyst — Barclays

You recently introduced some new products, Fabric One and Inference Exchange. How do these enhance your current product suite?

Speaker 1

Yeah, I think that they are two products that actually enhance the product suite by helping us move with the needs of AI. So I think of it as fabric as the connection and the exchange as what's operating. When you mention that it takes 300 different activities to occur, that's what inference exchange helps us do. It helps us to actually make it simpler for the customer to make the data, use the data that they need in order to drive the business so they don't have to think about that. And the connection, the fabric, it makes it easier for customers to connect. So we know where they need to connect to, how they need to connect, whether it needs to stay in one location because of sovereignty or the data can move to another location, all of that. But they don't need to think about it. We do the work behind the scenes for them.

Brendan Lynch, Analyst — Barclays

Great. On that point of sovereignty, I think that's a growing consideration in kind of workload placements. How are you guys thinking about that internally at Equinix?

Speaker 1

Yeah, it's a good one, because I feel like we were kind of built for sovereignty, because we've always worked with that in terms of understanding that we're in countries, we're throughout the world as the largest data center, and we work in countries where sovereignty has always been important. And so we have the ability, because we operate within the countries, we understand that there is data that needs to stay within the country, and then there's data that can leave the country. And so we are able to help facilitate that for our customers and acknowledge that there's some data that's going to stay and we can afford that. We provide that to them. And then the data that needs to traverse and can leave, we help facilitate that for them as well with our interconnection. Is there anything you want to add to that one?

Ryan Burke, Head of Investor Relations

Just that today, given the rise of sovereignty needs, having sovereignty built into the network layer for us is a huge, huge draw to customers, right? that's very different, very much more automated and much less complex than having to deal with it on an app-by-app basis or a software-bound basis. So we're sort of naturally sovereign, and that's been one of the many selling points to our customers today.

Brendan Lynch, Analyst — Barclays

Great. You've had a relationship with NVIDIA for years. Maybe you could talk a bit about how some of these new products that you have rolled out, specifically Inference Exchange, is different, which is also a partnership with NVIDIA, is different from kind of past iterations of the relationship that you've had.

Speaker 1

Yeah, I'll take that one. Yeah, so we've had a longstanding relationship with NVIDIA, and I think the way I would think about it is when we first started working with them, it was really about the infrastructure that customers needed, and so facilitating access to that infrastructure. As it has evolved and our customers are now looking at how do they use inference and how do you get inference to the customer, our relationship with NVIDIA has changed as well. And introducing this product was an important part of that journey in terms of working with NVIDIA and also Together AI to build out how we can actually facilitate the customer using the data. Not just you've got the infrastructure, but what do I actually do to get the data so I can now use it? um coming off of the different places that that data comes from and if we facilitate that and that's that's how that has evolved for the for the with the relationship with nvidia and i assume it's going to continue to evolve that way as we think about how customers are using ai good partner to have that's for sure um maybe you can talk a little bit about your pricing strategy and if you anticipate that that would continue to evolve uh with different partnerships Yeah, that's a big question, a broad question. Is there a particular part of that just in terms of the particular product or in relation to everything?

Brendan Lynch, Analyst — Barclays

I think with Febber 1, if I'm not mistaken, there is a suggestion that it's going to be more outcome-based pricing. Maybe just if you could clarify what that means.

Speaker 1

Yeah, well, I think part of that ties to simplicity for our customers. So when you simplify, instead of having to price out a multitude of components and navigate that from a pricing perspective that can provide one price based on the outcome that the customer is looking for, we simplify it for the customer. And then for us, it also makes it more manageable for us in terms of expectations on what we're building. We anticipate that what's going to happen with Interconnection is that it's going to continue to grow. It's one of our fastest growing products at 9% this year. And so we feel that as customers, it becomes essential for customers using AI to retrieve their data from many different places, whether it's a model or it's compute, in order to get it to where the workload needs to be, in order to utilize it, that interconnection becomes greater in importance to the overall platform that they're using. And so simplifying that for our customers and making it easier to deliver becomes the most important part.

Brendan Lynch, Analyst — Barclays

To touch on that, the interconnection component, how is AI adoption driving increased demand for interconnection? I think there's a thesis out there that historically your customers would have had an interconnection with maybe some of the fiber providers and one or two of the hyperscalers, but now with OpenAI, Anthropic, et cetera, all in the mix, is that driving, I guess, increased interconnection density?

Speaker 1

Yes, very much so. So in the past, when data lived kind of in one place, maybe two places, and your interconnections you would see were typically point-to-point, with AI, it's driving a greater need to retrieve data and move data to many different places. So data doesn't live in one place anymore, right? It can live in a cloud. The model might be in one place. Your compute might be in another place. And then you need to ultimately get the data over to an end user. and that's far more complicated than it used to be and requires more interconnections than were required in the past and they need to operate more flexibly as well. And so we're seeing that this is driving two things. One, a significant increase, which I think is why we're seeing that growth in our interconnection product, but increase in interconnection and the flexibility of that interconnection. But it's also creating a lens on networking that maybe hadn't been in the industry for quite a long time and two years ago three years ago you'd be in a room talking to people about networking and you don't even go to university to learn about networking anymore and now with AI it has had a kind of a rebirth in terms of the critical element of running a platform is that that networking that you create so yeah we see it as being a absolutely important element of driving AI or utilizing AI and driving a platform that helps a customer to use that AI.

Brendan Lynch, Analyst — Barclays

You could talk specifically about Equinix's competitive advantages and interconnection relative to some of the other options that might be available on the market.

Speaker 1

Yeah, so besides the products that we have and the reliability that we have, we have over 522,000 interconnections globally. We also have the largest number of data centers globally as well. Is it 282 at this point I think we're at. It changes, grows. But that gives us a level of density that is unparalleled. And there are other companies that absolutely have interconnections, but we have this density of interconnections that is unparalleled. When you combine that, so that's good, but what really differentiates us is the fact that they're connected to ecosystems. And we've been building ecosystems for almost 30 years, and it takes a long time to build a really viable, rich, dense ecosystem. And so that's what we've done, and that's where we have focused in our major metros in terms of making sure that we've attracted the right customers, whether it's clouds or network service providers, the customers who actually live there, the right neoclouds. Making those connections is critically important, and that's what creates a magnetism and customers come to us and say, I want to be in that ecosystem because half of my supply chain lives there, or I've got an important customer base that's there. And so we've shown and proven that it creates this magnetism that draws more business in that's very, very important. And so I think the two are important, not one or the other alone. That interconnectivity to that ecosystem is what's most critical.

Brendan Lynch, Analyst — Barclays

One of the elements that's kind of newer to your ecosystem is the X-Scale offering. How should we think about the value proposition of your X-Scale assets relative to maybe some of the single-tenant assets that might be available through other developers?

Speaker 1

Yeah, so we don't build gigawatt data centers to perform training. That's not what we do, and that's not what our X-Scale product is. But what we do know is that the clouds want to be close to where those ecosystems live. And those ecosystems sit in large, densely populated communities. So you're not going to build a gigawatt data center in a large, densely populated community, but you want to get closer to it. And that's what XScale's purpose is, is to bring those clouds closer in so that they can connect into those ecosystems and the customers that are there, and so that our customers have a way to connect into the clouds.

Brendan Lynch, Analyst — Barclays

I believe you have about half of the Hamptons campus still available. You've also got the Manuka site. Maybe you could talk about those two assets or two locations and perhaps where else in the Americas you're considering expanding XScale next.

Speaker 1

Well, those are the two right now that we're looking at in terms of, and they're on track. We're excited about them. And I don't know if there's a whole lot else that we've shared publicly other than that they're on track and progressing well on plan.

Ryan Burke, Head of Investor Relations

And that's good, good XScale growth for 28, 29 for those investment-grade hyperscale customers.

Brendan Lynch, Analyst — Barclays

Maybe I'll open it up to the room if there's any questions here. Happy to relay them. I can keep going. Maybe on the topic of development, how is Equinix adapting facilities and your design to meet the operating standards of some of these higher power density workloads that you're seeing through AI?

Speaker 1

Yeah, I would be glad to do that. So we do this, and we've got probably our real estate guy in the room as well, who does this every day as part of his job and livelihood. So we're always thinking about this. And I think, let me back up for a moment and talk a little bit about the metros that we exist and kind of these campuses that we build. We have assets that have been there for many years. And then we have new assets. We're continually growing. And so as we think about what we're building and we think about what's needed in those metros, we're constantly assessing existing workloads, the new workloads that are coming in. We have a degree of churn that happens in those metros as well and where we're placing customers to best meet their needs. And so when we think about what we're building to in with AI workloads, it's absolutely creating a need for us to build at a larger scale with higher density. And so when we think about, for example, I think it's DA12 that we've announced that we're building, we're going to be delivering, I think, about 67 megawatts, and it's going to have a density of about 18 kVA per cabinet. That's outside of, not outside, but that's denser than we've ever gone, and we're very comfortable with that, but that's what's required. And so we'll continue to think about that as we think about our designs, how we're evolving that to address the needs of AI and how we can do that. I think what's also really important, this has come up in lots of other conversations we've had today, how do we do that in a responsible way as well? Because we have environmental impacts that we have to be conscious of in terms of how we cool things, the amount of water that we use, et cetera. And so keeping a lens on that while also making sure that what we're building accommodates the needs of what AI is driving. I don't think that we, sometimes I get the question of, well, what is that going to look like? And do you have a lens on that? I don't think that we are building to say this is what AI is going to demand. I don't think anybody can say that right now. I think it's too unknown. But I think we're comfortable with the fact that we have a history of evolving how we build, how we're thoughtful, and we're responsibly looking at what's happening and changing and how we change our builds to accommodate that for our customers. And so I think we feel well positioned for what AI is going to bring forward and what we're going to have to build to support the AI in the future over the next, say, five years.

Brendan Lynch, Analyst — Barclays

I think it's an interesting dynamic for Equinix in that you're basically planning on doubling the size of the company over the next five years, which would be all the most modern assets. So even as AI requirements evolve, your starting point of where you might need to alter a facility is much more progressive than having to retrofit an asset from 20 plus years ago. and even some of your assets from 20-plus years ago are the most valuable ones in the portfolio because of the ecosystem density that you discussed earlier.

Speaker 1

Exactly, exactly. Yeah, and I think it's an important differentiator for us, and it's also important in terms of how that ties into how we think about those metros and how we continue to build them. And also, the conversations that we have with our customers are probably different. They're not coming to us saying, I just need space and power. They're coming to us, and they're talking about, this is what we're doing in our business. This is the next initiative that we might be rolling out, or we're having much more business-led conversations with our customers in terms of what they're trying to accomplish, and then working with them on how we design where they end up being placed and what they're going to utilize.

Brendan Lynch, Analyst — Barclays

How should we think about either the requirement or the option to retrofit older assets for liquid cooling and any other next-generation AI-type deployments?

Speaker 1

In terms of retrofitting them, I think, so to your point, we have existing assets that we utilize with the cooling that we've had in place. When we think about new builds, we think about liquid cooling. So we're less prone to retrofitting a building for liquid cooling, and we're anticipating where we need liquid cooling in the portfolio and building to that. So I would say that's how we're handling it. We're not going back and retrofitting a bunch of buildings for liquid cooling.

Ryan Burke, Head of Investor Relations

We've got liquid cooling in 100-plus properties right now, and Adair, our CEO, likens it to a NASCAR pit crew in the sense that a customer wants liquid cooling. They'll pay for it. We can come in, put it into what they need in terms of the cage. They'll use it. They'll pay for it. They no longer need it. We can pull it out of that cage and deploy it elsewhere. So it's surprisingly nimble relative to what some people think the retrofit need might be. But it all goes back to the idea that, you know, when you take the new properties that we're building and the old properties that we have in place, we think what we have is a portfolio that's well set to serve enterprise need for the future. There aren't many products or real estate types where what you built 25 years ago is still very relevant today, but that's very much the case for us. So we're building to what we believe is core enterprise demand, just regular way, but also AI oriented.

Speaker 1

Yeah. And I think what's also really important is building in that optionality in our new builds is really important because what we saw when we started looking and the industry was looking at liquid cooling is that there was I think there was an anticipation that customers, they're going to buy AI and they're going to need liquid cooling. we did not see that uptick on AI in the enterprise space. And so you had AI companies that were very excited and growth was incredible. But over the last year, you didn't see, or I would say three years, you saw a lot of customers looking at AI, the enterprise customer looking at AI from afar, but not actually leaning into it. And so what we've seen over the last year is this acceleration in enterprises is stepping into AI and needing to really grapple with what's our AI strategy and how are we using AI. And so when we originally, you would see a build happen with liquid cooling and it would sit there with nobody, you know, the customer and say, oh, that's really expensive too, right, because it is more expensive. And they say, okay, maybe I don't need liquid cooling. Now it's a requirement. And so I think we've been really thoughtful in creating that optionality and flexibility in our builds so that we're not stuck with, you know, liquid cooling that we're not utilizing and we've paid for the infrastructure to build it. Instead, it's there now and available as needed.

Brendan Lynch, Analyst — Barclays

It's a journey for you guys.

Speaker 1

It is a journey, yes.

Brendan Lynch, Analyst — Barclays

Arkel, one of the things that you mentioned was kind of building responsibly. Maybe you could talk a little bit about the impact of growing community opposition to development and how that is affecting your pipeline.

Speaker 1

Yeah, absolutely. So community sentiment and backlash And backlash is a real thing. And so let me back up a little bit. We've actually, as I mentioned, probably 10 times up here, sorry, but we've been in the industry almost 30 years. And where we're located, we've been in those communities for 15, 20-plus years, 30 years in some cases. And so we have a reputation within the industry and within the communities that people can point out and say, oh, you've been a good operator. We understand who you are. That said, the issues, the growth in the data center industry is real and has happened. There have been bad actors in the industry. And what the communities are pushing back on are, in some cases, very real issues. Now, some of what they're pushing back on are things that they're hearing in social media or on the news. And not all of it is fact and not all of it is real, but it certainly has picked up steam. Our position on it is that, first of all, we always listen to the community. And we get that they're concerned about rates, you know, rate increases. They're concerned about, you know, what it's going to do to their community. Does it really bring jobs in? And so we're actually able to point at the fact that we've always paid for our usage. We've always invested, and we invest heavily in partnership with the power companies to build out the grid that's needed for us. We've done a lot with workplace development, workforce development, as well as community development. before we even acquire a piece of land or think about expanding, we'll sit with the community leaders and talk to them about what we're thinking about doing because we need to have an understanding, agreement, right, that we can serve what their needs are and we can build what we need to build. Our sustainability projects are always like how we build a data center directly correlates to regulations that the communities have or what they're asking for. So I think for us, we need to get that story out more. I just did a, probably four or five months ago now, but issued a document around our community principles, which we've been doing for almost 30 years, but we haven't really had a need to talk about it. It's just been the values of how we are as a company. It's part of our principles as a company. And so we've done some advertising on that and putting it out there and are trying to get that message out. I think it is from a, is it holding up our development? I would say that our data centers are smaller than the big, we're not building big gigawatt data centers. And we have worked extensively with communities that we're in right now where we do have specific builds happening. And so we see less, we see an ask of work with us and help us to get that message out so that, you know, the elected officials that actually said yes and the community leaders will say we want you to come out and meet with the community and do town halls with us which we're happy to do we just did a extensive one in Manuka that was really well received talking about what the project is and and working with the community um but we do a lot of that and I think that's that's very important and that and so therefore we have not seen some of the the delays or or issues that have come up um but I think it's something that we're very sensitive. It is a real thing going on, and I think continuing to get out what's fact and what's fiction is going to be important. And I think also using fact about what specifically data centers have done is an important message for people to hear.

Brendan Lynch, Analyst — Barclays

Sure.

Speaker 1

It's interesting just in the industry how we've gone from ribbon-cutting ceremonies at data centers a few years ago to pretty strong opposition, but to your point, a lot of it is based on misinformation and just having the right communication apparatus to kind of spread the spread the word yeah one of the things that a lot of people don't realize is what actually lives in a data center in an equinex data center and so when we start to talk to them about the fact that 9-1-1 runs through the data center that the local hospital uh you know and all of the things that a doctor might be looking at to transmit your x-rays to another doctor you know across town. The emergency services, the information that's going from an ambulance who's got a patient who's coding to the hospital so that when they arrive at the hospital, that patient, the vital statistics are there. That all runs through a data center. So when you start to talk about that, it brings to life what happens in an Equinix data center.

Brendan Lynch, Analyst — Barclays

Great. Maybe when we think about development, you guys are clearly undertaking a very large development program now. How should we think about the outlook for development yields given the rising costs and constraints on supply?

Speaker 1

Sure. Well, we think that we are a very responsible, thoughtful builder with our strategy. And so while you see cost going up, first of all, we think that the way that we build and how we build always has a lens to doing it so that we maintain strong returns on that investment. We also are seeing continued growth in customer demand and what's happening there. And so we feel that our build is keeping pace with as demand accelerates and our builds accelerate. We feel that we're managing that well and still have no problem in the returns that we expect from what we're building.

Ryan Burke, Head of Investor Relations

We think we'll continue to achieve the mid-20% cash on cash that we have historically, that's driven by competitive advantages sort of across everything involved with delivering and operating properties. But at the core of it is just the individual mosaic at each data center, right? So, like, if you just pretended you had a bird's-eye view of each data center, you're looking down on it and think about a bunch of colored tiles. And those colored tiles represent different workloads and customer types. It's really curating that ecosystem, curing that mosaic that dries our ability to achieve the outsized yield. So we'll continue to be very focused on that. And therein lies the lease-up period that we do have. It takes two to three years to lease our properties up to Equinix level of stabilization. But Equinix level of stabilization is at that higher, mid-20% return.

Speaker 1

Yeah. I'll take a minute also to elaborate on the mosaic. So when we build a data center, before we build, when we anticipate the size and scope of it, we think about who are the customers that are going to be in there. And we think about it in terms of retail customers, so small and medium-sized deals, customers and large customers. We think about large footprint that goes in there. And that all plays out in terms of the returns on that asset. And so we build accordingly, and then we sell accordingly. So you'll never have an asset where we say we should have, you know, 60% should be retail customers who have a higher yield. And all of a sudden we've got two big customers come in and they want to buy out the data center. We'll never do that because that would so dramatically change the returns on that asset. And so that's a really, really important aspect of how we drive those returns is how we fill that asset. And that's all correlated then to market opportunity.

Brendan Lynch, Analyst — Barclays

Maybe we'll do one more topic here on guidance. When we think about your guidance through 2029, which is now for 9% to 12% annual AFFO per share growth, how should we think about the contribution from the various components, be it cabinet volume growth, cabinet pricing, interconnection momentum, or anything else that you want to throw in the mix? How should we think about that kind of growth algorithm?

Speaker 1

Why don't I take the first part, and then you can add all the little details that you want to around that one. So, I mean, the growth is, I mean, we're very confident about what we're seeing. I think what we also are seeing is when we think about the growth that we've seen over even the last three to four years, which has been significant for our company, and then you see what's happened in the last year just in terms of kind of when I spoke about the enterprise, maybe now coming into that AI period, and we start to see even greater growth happening. I think we're quite bullish on what we're seeing in terms of the future, and that's how we're thinking about what we're building and how do we keep pace with that in a responsible way. Do you want to add in on your...

Ryan Burke, Head of Investor Relations

Yeah, and just, you know, hopefully what you're all hearing from us is just that there's broad strength across our business, whether it's customer type, whether it's workload type, whether it's, you know, each individual product and service. So that's number one. Number two is, yeah, we updated our long-term outlook, which is through 2029 you know after having just given one about a year and a half ago and what was our old high-end is now our new low-end and that's because of a mix of demand strength which frankly came on faster than the company anticipated but also the team's execution as well but as if you think about it just from a simplistic perspective you can see our same store pool sort of migrating from mid single into high high-digit revenue growth and And so you can't expect that capacity expansion will do the rest. We're mindful that we're a very differentiated company within the data center space. We're focused on core workloads, core metros, co-location. We have a good feel of supply and demand conditions in our core addressable market. So we feel good about the execution over the next few years.

Speaker 1

I think the last piece is we're very diversified across the customer base. So there's no single customer that houses it over about 2.5% of the portfolio. And that's significant. So when you see shifts happening, and we've seen this over the years in the history of our company, when you see shifts happening, it does not have the impact on us that it might have on another company that's heavily invested in one customer type or one particular product.

Brendan Lynch, Analyst — Barclays

Diversification is always important. Let's leave it there. Thank you very much, Arkell and Ryan. We appreciate being here. Thank you, everyone.