Equinix Aktie

All right. Good morning. Thanks, everybody, for being here. My name is Frank Louthan. I'm the senior analyst at Raymond James covering data centers and telecom, and other things. I'm very pleased to have Chief Financial Officer, Olivier Leonetti, here from Equinix. We're going to go through a few questions, and then we'll leave a little time at the end for some questions from the audience. But you guys have to start on with...

And just a quick disclosure statement. Some of what we will be talking about today contains forward-looking statements. Please read our SEC filings for more information about factors that could affect these statements.

All right. Great. I'm sure everyone will have that committed to memory.

Yes. Phillip and I will go through the motions together. Phillip and I are partners in crime.

Great. All right. Well, Olivier, maybe walk us through kind of your vision for the company as you step into the role here. What can we expect from you that may be different from what we've seen in the past from Equinix? And what do you see as the opportunities that really attract you to the position?

Yes. So again, thank you for having us. Thank you to everybody for being in the meeting room. I hope you're having a great conference. It's my first one, and I really enjoyed it. A lot of things going on in our company, so we love your interest. So I joined Equinix about a bit more than 2 months ago. And I joined this amazing institution for 4 reasons: one, attractive end markets being served. Two, amazing opportunity from the company to compete in those markets. We'll discuss a lot about that with Frank. Three, our culture; and four, this partnership is important, as is the vision from our CEO.

And after -- it looks like a marketing comment, but after a bit more than 2 months, I was wrong. Our end markets are better. Our ability to compete is better. Our culture is better and the vision from our management team is better. So what I'm going to do differently is not a lot. I'm going to be a member of the team to really augment, accelerate, facilitate the vision, but I'm going to be one of many shaping the future for this amazing company in those amazing markets. We'll talk about that, I'm sure.

Okay. Great. So maybe update us a little bit from -- you had the Analyst Day last year that you do every 2 years. Maybe update us kind of what's changed there? And what have we seen in the year since that, roughly a year.

Yes. So two things. I mean, if you go back, we have been in business for 3 decades. We were the pioneer to facilitate the Internet. That's first decade. The second decade was to facilitate the cloud revolution. This decade is going to be to facilitate AI. Our CEO has said many times, I think it's a great way to talk about how excited we are. This moment is our moment.

Why? The AI revolution, which is at the start, is creating the need to have a diverse marketplace where players are going to meet and exchange data. That's what AI is about. That is going to play to our advantage. Now to answer to your question, what is new since Investor Day, this vision is playing out in a way which is augmented relative to what we had expected. Those end markets are stronger and our ability to deliver what our customers need is stronger.

What do they need? They need a global network, they need a diverse network with different participants being on it, cloud, neo-cloud, AI lab, enterprise. They need neutrality. They need latency. AI will need to compute fast, close -- being close to where the enterprise is important. And we need also connectivity solutions, which are going to be fast, easy to use. All of that is at play. Equinix is delivering on this.

Equinix is the biggest marketplace on Earth, able to deliver this. As a result, you started to see the numbers being better than what we thought. I give you some numbers for this year. We're going to grow recurring revenue by 10% for the year. We did that in Q1. We did that in Q2. Double-digit growth is the start, and we have been able to clock on those numbers now for a period of time. Two, EBITDA margin, 51%. And three, an important statistic, one, which is one we're going to focus a lot is AFFO per share growing at 10%. So you see this is our moment, the value proposition of the company resonating and us being able to deliver attractive returns as a result.

So maybe take that a little further and your value proposition as you approach customers, what is that? And then how are you differentiated from the competition to be able to accomplish all these things?

Yes. And Phillip and I don't hesitate, Phillip. So if you see today, in the world of AI, you want -- go back to a rich marketplace, a marketplace where all the cloud providers are going to be participated. You want the Neo-cloud and new entrants to be part of this marketplace. You want the AI lab to be part of this marketplace. And of course, you need the enterprise. And you need this marketplace, which is global, neutral, and close to the action to manage latencies.

And what is different is that nobody can offer all of this. Could you have players offering a one-to-one connection, a cloud to a cloud? Some people have started to do that. Yes. Could you have some carriers, which are local doing some of that? Yes. But nobody can do everything we are doing. So that's different. And one of the key success factor of our company is to nurture a rich marketplace. So we are expanding. We're going to double our capacity. I'm sure we'll talk about that with Frank.

This new capacity is created to make sure that we have all the participants needed in those. So that's different. I'll give you some statistics. Neo-cloud, 4 of the top 5 are part of the network. They have deployed with us more than 110 nodes in our network, 8 of the top AI lab as part of the network. All the cloud providers are part of the network, right? So difficult to replicate. Then another part of the value proposition, and our CEO has mentioned that many times is we are not in the compute business.

We are in the connection business. And we want connections we're going to be easy to implement with no human being involved, managed at the press of a dashboard, and probably powered by AI to allow -- to identify which players do you want in your network. We do this. That's different. And we want to today keep increasing the barriers to entry or competitive advantage to make all of that even stronger than now.

And maybe just to add on to that and take a step back when you -- I think there's been a little bit of a homogenization of the data center industry. Now all these things that Olivier talked about around our differentiation. Keep in mind that we are -- we've got over 10,500 customers in our customer base. We did 3,800 transactions in Q1 with over 3,100 unique customers. We're not selling 100 megawatts to one customer and selling out a facility.

We're a very differentiated business model as it relates to the broader data center industry. I think it's really important to take a step back and make sure that folks understand that, that differentiation and all the statistics of how we're curating ecosystems. And because of that, we're driving returns and yields on our assets that are in the mid-20s. That's very different than a lot of others in the data center industry.

Let me build on this, if you don't mind, Frank. If you see the evolution of what we do, in the era of AI. I'm going to make an obvious statement. AI is to compute data at the edge from via sources, right? We are starting to be the place where this compute at the edge is happening, evidence of this. The average -- sorry to mention a technical term, rack at Equinix is 5 kilowatt. This is the average rack density.

We have in our network today racks, which are 100-kilowatt plus. Why would you need a rack like this in our network? They are liquid-cooled. We're enabling this. Why do you need this to compute at the edge. So you see why do you need to compute at the edge? Again, you need to send the fast action to an object, but also it's too expensive to compute in other places because of the traffic cost. So you start to see really the value proposition of what we do increasing step by step. The marketplace, the edge compute, and the ease of connection. So again, increasing the competitive advantage and what we offer to our customers.

That's great. So that actually touches on a topic that I wanted to bring up. I think I've been covering you guys since 2009, give or take a month, something like that. And from day 1, I get the question about obsolescence risk. And as you mentioned, averages 5 kW per rack, you've got customers wanting 100 kW. Walk us through the obsolescence bogeyman here that always seems to come up. It's had quite a resurgence based on inbound calls I've got in the last couple of months. So talk to us about how you -- why that is a threat and how you've managed that all these years? Because so far, it hasn't been an issue. But of course, it's coming now, of course, you never know. So talk to us about that.

It's a great question. So let me give you an example. If you go to headquarters in the Bay Area in San Francisco, we have one of our oldest site. It's in San Jose. You have 4 generations of data center, the first one and the last one. The last one, the biggest microchip in a rack from the #1 microchip manufacturer, 100 kilowatt to a few 4, 5 kilowatt. This full ecosystem is attractive to our customers. All of those cabinets with different level of power are today growing because, again, you need compute at the edge, use the 100 kilowatt. But you need also to use the cabinets to -- for communication. So those worlds need to be part of the ecosystem. So obsolescence today is not something, which is a concern to our company.

Yes. And we've always talked about putting right application in the right workload, in the right data center. And exactly to Olivier's point, I mean, when you look at some of our older facilities, they're the most network-dense, they're the most valuable. And when there is space, trust me, there are customers who want to go into that space. And we've constantly been maintaining these over the years.

And so they are absolutely fit for purpose for those specific workloads. And as we talked about, some of these higher-dense applications and workloads, they're going into our newer facilities. Plus we have 100 of our existing 280-plus data centers, which are fit for purpose for liquid cooling. So we always are thinking about the right mix of applications in the right locations for our customers.

All right. Great. So with that, maybe let's talk a little bit about competition. We've had a very, very strong market for data centers the last few years. It's been fun. But with that always, it attracts new competition. Talk to us a little bit about what do you see as some of the biggest competitive threats in the next couple of years and how you're positioned to address that?

Yes. So the competition is intense, and it's keeping us up at night. It's a good thing. It's keeping us on our edge. And the level of competition is actually the proof point that what we do is important in the era of AI. Carriers are on our network and competing with us, and also collaborating. You have other players as well doing the same. A lot of cooperation, competition happening on the Equinix network. We serve a different purpose.

Nobody today is able to do what I've mentioned. Could you do a 1:1 connection, a 1:2 connection? Yes. Could you do to many everywhere all at once, latency, neutrality, density, nobody can do this. But we're not sleeping. We want to increase the number of data centers we are launching. We want to make our connection business humanless, to be implemented.

And we want to create fast, the best marketplace in the market. So competition is good, validating what we do and good for our customers. And clearly, as a result of that, we -- you see the execution, the speed of execution, and the sense of urgency at our institution being stronger than ever before.

Yes. And I think the marketplace, obviously, we've all talked about the demand environment. There's room for all boats, so to speak. And I think that we feel like we occupy unique place in the data center space, but there is room for all these different workloads to live in different places and room for all of us.

All right. Let's talk about 2 of your products. You talked about interconnection and then also on the xScale side and the larger hyperscale facilities. You've got -- I constantly hear questions from investors, folks like Lumen that have looking at multi-cloud connectivity, worried that's a threat to some of your interconnection revenue and so forth. You've talked -- touched on this a little bit, but maybe address more specifically how some of these telecom carriers that are looking to do that business. How does that impact or not impact your business and what you're trying to do?

Again, so why do you have some carriers and other players part of the network. By the way, we know they exist. We know what they do. We know the footprint they have. It's a cooperation and some level of competition, too, right? We serve different purposes. One-to-one or mainly maybe a U.S. connection is what they will serve. But our customers today, they want a global footprint. A data center in Europe will have a lot of tracking, a lot of traffic coming from other parts of the world. So we want to be differentiated. We are. And we play in this dynamic place in complex connection businesses, and we can cooperate with the other players as well.

And then on the xScale side, you're building larger hyperscale facilities, it's a smaller part of your business, even some of your peers. Talk to us about how you're incorporating that with what you do and where you're seeing success there.

So xScale is not a retail, but it's not a full wholesale either. Why do I say this? They are not gigawatts centers, right? There are hundreds of megawatt centers. They are close to a metro without getting into too many details, you need to be about 30 miles, give or take, to metro to manage latency. Those xScale sites are smaller, close to a large metro.

We will share some of those sites. So we see xScale as being synergistic with what we do. They allow us to do 2 things in addition to have access to a site, they give us purchasing power with the power management company and more intimacy also with some of our large customers.

Okay. Great. All right. So we can't have a data center conversation these days without about the power question. Talk to us about the power. How do you see your ability to get that access for -- you have some pretty broad development needs. How do you see your ability to access power to reach your development goals?

Yes. So I'm kind of embarrassed by what I'm about to say because it's so different than what you have heard, right? We read an article this morning on the Wall Street Journal about how difficult it is to access to everything, right? We build something different, 60-megawatt data centers. We build those in metros. We have been doing that for 30 years. We have great relationship with the utilities companies. We have great relationship with our general contractor, with the power management company.

So today, we have been able to manage these constraints pretty well. We have not experienced any delay. If anything, our team has been able to accelerate the launch of data centers. So we have been largely immune from those constraints. So again, we have access for the next 5 years to more land, power, water, power management equipment than we need. But again, different use case. Let me speak about the community for a second, and we are proud of this approach at our company. We have a team across the world only doing community engagement for the world.

We have a playbook. We are part of the community. Our playbook is as follows: We have been in the community for 30 years. We are green. We manage water, we recycle it. We invest in your schools. We invest in apprenticeship. We're different. We're going to stay. We've been with you for 3 decades. We're going to stay. That resonate, and that allows us to navigate. It's tougher than before. No question, but no delay. Our team is doing an amazing job.

And what about the self-provisioning power? How have you thought about that process? And is that the solution that's right for Equinix? How do you think about that?

You cannot be in this business without thinking about alternative sources of power and local provisioning of the power, particularly to manage peaks in power load. So we do all of this. I mean we're even looking at data center space, I think, is something which is fashionable nowadays. But all of that is being part of the playbook of our company. We have local power provisioning using various sources, gas and green energy as well. So we have to do that, and we do.

Yes. We were one of the first users of fuel cells, for example, in our Silicon Valley. We use gas turbines in some areas in Europe. And as Olivier said, we've kind of taken an all-power approach as we look forward. But I want to make one other comments just about the power situation. We've got 3 gigawatts of land under control. And we -- we're not in the business of speculatively buying land and then figuring out the power situation later. We are doing all this in parallel. And so when we buy a parcel of land, we either have fully contracted power agreements or we are in the very late stages that have high, high confidence about our ability to get the power. So it's a key distinction that we're -- again, we're not out speculatively buying land and then figuring these things out down the road.

Four conditions for us to sign a contract for a piece of property, land, power, water, community at the same time, yes.

All right. Great. And with that, so where are you on the build -- the ambitious build plans? You've got the resources you need, the plan and power. Where are you on the build plans you laid out last year?

Well on track. We have solved the capacity constraints. We have solved the financing constraint. Now we want to offer great network created and great full stack interconnection solutions. That's a work in progress. But the other constraints have been solved.

Okay. And then I'll have one more question, and then we'll take -- see if there's some questions from the audience. If not, I've got a few more. But -- so new CFO, I have to ask a capital allocation question to you. Talk to us about your view of capital allocation for -- as it relates to investors.

Yes. So there are a few things, which are going to be foundational regarding the way we manage our company. And I'm going to state some obvious facts, but they are always good when repeated. One, investment grade is table stakes. We think we can leverage. We have a 3.8x leverage. We believe we can increase the leverage, but investment grade is going to be critical. We want to maintain this. And we believe we can finance our expansion while staying an investment-grade company. So that's one constraint.

The other constraint is we do not want to sacrifice the short term for the long term. 1/3 of our investors want us to do both, and we believe that we're going to be able to do both. So what does that mean? FFO per share is part of the algorithm. Our capital allocation has 4 tenets: one, investment grade; two, top line; three, EBITDA expansion; four, an important our AFFO growth per share, which is attractive to you. So that's what we're going to focus on. Largely, I mentioned it, Phillip mentioned it as well. We have the land, the power, the water we need. At times, we'll be opportunistic to do M&A. AtNorth was one of them, but they are not needed for us to achieve our goals. So that's the framework.

All right. Great. So folks somebody has a question there, there's a microphone or you can just raise your hand, we can grab the questions there. Yes, go ahead.

How fast do you see that power to rack growing. You mentioned kind of 5 kW now and a few that are 100 kW. Is this growing at 10% a year? Or are we going to be at 100 kW for a good bit of?

So this is -- I mean, if you go to the site, those 100 kW rack exist. They are already deployed. It would take time. It will be a diverse set of use cases. So if I give you a statistic, the average kilowatt per rack today at Equinix is 5 kW. In Q1, the increase in density went up by 36%, right? So we're still going to increase in this increase, you're going to have Phillip mentioned it earlier, it was a good set of statistics, low power, high power co-existing. And again, the high power, something very new, which is the need to do compute at the edge.

But just to put it in context, you talked about the 36% growth, that's still single-digit kilowatt per cab, right? So it's -- there's a big installed base for sure, and there are needs for those for all different types of workloads. So I think it's going to be a slow climb, but we are building our new facilities are building towards these increased densities. So we feel like we're kind of going to where the market is and where it's going. We feel very confident about the ability to meet the diversity of the demand around those requirements.

REIT of change is always slower. Any other question? Yes, go ahead.

SpaceX and others are data center space, they still need the interconnection [indiscernible].

Yes. So I mean, everybody has to look at this, right? Elon has been pushing the idea first. People thought it was crazy originally, right? And some people are saying, okay, maybe we need to pay attention. We're looking at this. The physics are going to be challenging. It's over my pay grade. Would you use -- I mean, we're speculating here. Do you need that for large language model? Yes.

Can that be the solution for what we do, retail network proximity? The answer is no. The physics do not allow this footprint to serve what we serve. If you are in the wholesale business somewhere in Arizona or Texas, maybe. If you do what we do, we do, no.

Okay. Great. Any last questions? All right. Well, with that, Olivier, Phil, really appreciate the time here. Thank you very much. Thanks, everybody, for joining...

Good day.

Good afternoon, and welcome to the Equinix First Quarter Earnings Conference Call. [Operator Instructions] Also, today's conference is being recorded. If anyone has any objections, please disconnect at this time. I would now like to turn the call over to Mr. Ryan Burke, Vice President of Investor Relations. You may begin, sir.

Good afternoon, and welcome to our first quarter conference call. Before we get started, I want to remind you that some of the statements that we make today are forward-looking in nature and involve certain risks and uncertainties. Actual results may vary significantly from those statements. and may be affected by the risks we identified in today's press release and in our filings with the SEC, including our most recent Forms 10-K and 10-Q.

Equinix assumes no obligation and does not intend to update or comment on forward-looking statements made on this call. In addition, in light of Regulation Fair Disclosure, it is our policy to not comment on our financial guidance during the quarter unless it is done through an explicit public disclosure.

On today's conference call, we will provide non-GAAP measures. We provide a reconciliation of those measures to the most directly comparable GAAP measures in today's press release on the Equinix Investor Relations page at www.equinix.com.

We have made available on the IR page of our website a presentation to accompany this discussion, along with certain supplemental financial information and other data. With us here today are Adaire Fox-Martin, CEO and President; Olivier Leonetti, CFO; and Phillip Konieczny, VP of Finance. At this time, I'll turn the call over to Adaire.

Thank you, Ryan. Hello, everyone, and a warm welcome to our Q1 2026 earnings call. This quarter's results reflect continued strength across the business as we capitalize on a large and growing set of opportunities. Demand is broad-based and durable. Execution is driving efficiency and AI continues to fuel infrastructure investments that play to our strengths.

Before I get into our results, I'd like to start with some important market context. Over the course of the past year, my conversations with customers have changed. A year ago, they were about piloting AI. Now our conversations are focused on enterprise-wide adoption at scale. New forces are driving this shift.

Inference has grown from experimental workloads to an engine of real-time business decision-making. An AgenticAI is moving from demos into distributed deployments with agents acting autonomously to achieve business outcomes. The reality is that most enterprise architectures are not optimized for these workflows.

Agents need private low-latency paths to data wherever it lives. They perform best at the edge. Closest to where the decisions get made. And they must be able to move freely across models and clouds whilst staying within jurisdictional boundaries. Performance, cost and compliance all suffer when today's agents run on yesterday's network.

Simply put, this deployment gap is an architecture problem. Enterprise is a need infrastructure that's purpose built for the way AI operates. Distributed, interconnected, sovereign by design and in close proximity to the date of that matters most. This is a market that we are built to serve.

Equinix is not simply the world's largest digital infrastructure company. We are the world's most deliberately curated digital ecosystem. And our Q1 results demonstrate the progress we are making to capture the market opportunity.

In Q1, our recurring revenue grew 10% on a normalized and constant currency basis, coming in at the high end of our expectations. This is our second straight quarter of double-digit MRR growth. At the same time, we are driving continued margin improvement. Q1 was also the largest quarter of total sales activity in our history, inclusive of annualized gross bookings and preselling activity.

Total sales activity was up more than 35% year-over-year. We drove significant interconnection and cape billing growth, whilst reducing churn. Reflecting ecosystem strength across our key operating metrics, and we are expanding our capacity whilst bringing new products to market that extend our runway for growth.

Our progress stems from the extraordinary efforts of our team, and I'm proud of the way our employees are stepping up to meet the moment. Let me now provide some color on our overall results and what's driving our performance. As you saw in our press release, our Q1 results do not include the xScale Hampton lease.

We are nearing execution on expanded mutual beneficial terms with our customers. Olivier will provide additional details on how you should model Hampton. Adjusting for the timing of Hampton, our Q1 revenue, AFFO and AFFO per share results were all ahead of our expectations. Overall, our xScale pipeline is robust given that our remaining capacity is in major metros.

Our momentum reinforces our confidence in the trajectory for the year. As such, we have raised our guidance across several key metrics. I am especially pleased with the strength of the position we are building across the AI inferencing ecosystem. The expansion of our relationships with the world's leading hyperscalers, Neo Cloud, AI security vendors and model providers serves as a magnet for agentic AI workloads.

8 of the top 10 AI model providers and 4 of the top 5 Neo clouds are actively expanding with Equinix. They have placed more than 110 separate network nodes with us to support mission-critical and latency sensitive elements of their architectures. Consistent with the prior quarter, approximately 60% of our largest deals in Q1 were AI-related.

Additionally, large capacity fabric connections have tripled from just a year ago. We believe there is meaningful upside to come, given we are still in the early days of the agentic AI wave and inferencing adoption. This momentum is part of a broader uptick in customer demand, spanning a wide range of AI cloud and networking workloads.

Now let me highlight some recent wins and associated use cases. Cupid Pharmaceuticals are quantum AI-driven drug discovery company relies on Equinix for the high-performance, low-latency infrastructure required to run millions of GPU-intensive molecular simulations.

By deploying a dedicated GPU cluster in Equinix data centers with direct cloud interconnection, Cupid has reduced experimental cycles by 20x whilst lowering cost by a factor of 5. Most importantly, our solutions are accelerating the path from discovery to potential therapies that can save lives.

Gammon construction, a leading construction and engineering services company in Asia chose Equinix because of our neutral platform, presence across major metros and connectivity solutions to enable their multi-cloud AI platform.

They are using our fabric interconnection portfolio to power their network infrastructure. which is the base for inhibitive solutions such as AI-powered robotics and drones for on-site risk assessments and smarter decision-making. During the quarter, we expanded our partnership with Options IT, the #1 provider of infrastructure to global financial services firms.

They selected Equinix because of our presence in the locations that matter most to their operations and ecosystems, including London, New York, Singapore and Tokyo. We are enabling options IT to deliver private cloud and AI managed infrastructure solutions to grow their business whilst meeting the date of sovereignty requirements of their customers.

We also grew our relationship with Maersk, a global leader in integrated logistics as it digitizes critical supply chain infrastructure. Maersk recently selected Equinix as its primary data center partner to support high-performance and AI workloads, including its first liquid-cooled AI deployments in Frankfurt.

Our global footprint, secure and resilient operations and industry-leading interconnection capabilities are supporting Mark's ongoing network transformation and long-term growth strategy. I'm exceptionally grateful to all our customers and partners for trusting Equinix to help move their business forward.

The outcomes we are enabling for them reflect rigorous execution against our strategic pillars. Starting with Star better, we delivered annualized gross bookings of $378 million in Q1. We up 9% year-over-year with approximately $140 million of pre-selling activity on top of that.

As I mentioned earlier, that 35% growth in total sales activity in the quarter, resulting in a record backlog. Transaction volumes continue to demonstrate a broad base of workload requirements with over 3,800 transactions spanning more than 3,100 unique customers in the quarter.

Importantly, we also saw increased customer adoption of our South service portal. Our portal is a key area of focus as we work to create a better customer experience. It also drives efficiencies within Equinix compared to a traditional cost-based ordering. This is one example of our broader focus on digitizing processes and workflows across the company.

Customers placed 20,000 orders through our portal in Q1, up 12% year-over-year. and we intend to continue driving enhancements to this solution. Turning to Solve Smarter, our customers consistently raised 2 key challenges to us. The first is AI infrastructure fragmentation. Enterprises are spending too much time and budget. Navigating dozens of disconnected AI model providers, GPU cloud, data platforms and security services.

The Equinix distributed AI hub, which we introduced at NVIDIA GTC solves this by getting enterprises a single private low latency connection to the entire AI ecosystem. Unlike AI marketplaces built by providers with their own services to sell our distributed AI hub is completely neutral, providing access to all models and cloud so customers can select what's best for them.

The second challenge facing customers is network complexity. Most enterprise networks are not designed to handle distributed AI workloads and it's resulting in degraded AI performance, inflated costs and compliance risks.

Equinix Fabric Intelligence solved these problems by monitoring network performance in real time automatically adjusting configurations and flagging anomalies before they become outages, all without human intervention. Unlike other network management tools that sit on top of the network fabric intelligence is built directly into our fabric interconnection platform.

This is a structural competitive advantage given the more than 500,000 live interconnections across our ecosystem. Our innovation is extending our market leadership and driving growth. Total interconnection revenue was up 9% year-over-year in Q1 and boosted by fabric revenue growth of 26% year-over-year.

Fabric bookings were up 70% year-over-year as our attach rate continues to increase. These growth rates are all on a normalized and constant currency basis. On Bill bolder, we continue to expand our capacity to meet demand. We have 46 major projects underway across 32 markets, including 6 excel projects.

More than 70% of this retail expansion CapEx within our major metros with the remainder focused on critical expansion markets, particularly in our Asia region. Given the strength of our presales motion, approximately 25% of our 2026 retail capacity expansion has already been sold.

We continue to meaningfully grow our pipeline for new power land and capacity expansion opportunities that can enhance our long-term growth prospects in key metros and deliver attractive returns. And we're not just growing, we're doing it responsibly.

Last week, we released our annual sustainability report. It shows how we are building essential infrastructure the world needs in ways that are affordable for our communities, sustainable for our planet and reliable for our customers.

These have long been core Equinix values, and they will continue to guide our future investment decisions. In Q1, we announced an important investment in 1 of the world's most sustainability-focused market as we signed a joint agreement with Canada Pension Plan Investment Board to Paratus at North.

This deal will further enhance our position in the Nordics by giving us access to an installed and active development pipeline of approximately 800 megawatts expected to come online over the next 5 years. At Nord's footprint in key markets such as Copenhagen is complementary to our existing EMEA operations and is well positioned to serve enterprise, cloud and AI growth.

The transaction is subject to closing conditions and is expected to be immediately accretive to AFFO per share upon the closing. Overall, Q1 demonstrated continued momentum across the business, and we see significant opportunities to accelerate growth as we deliver on our strategy.

I'm now going to turn the call over to our new CFO, Olivier Leonetti to go into more detail on our financials. Olivier joined us in March. He has already proven to be an excellent addition to our leadership team. Previously, Olivier was CFO of Eaton and Johnson Controls to large suppliers to the data center industry.

He has a strong track record of delivering profitable growth and creating shareholder value, and we look forward to his contributions to our success as we work to deliver healthy revenue growth margin expansion and superior returns. Olivier, over to you.

Thank you for the kind words, Adaire. I'm delighted to be here, nearly 2 months in, I'm excited about the strength of the markets we serve and very impressed by Equinix company culture vision and unique positioning to serve accelerating customer demand.

I look forward to helping enable our vision by prudently allocating capital and thoughtfully utilizing our balance sheet to drive durable, profitable growth. As Adaire summarized, we are executing well across our business. This was the largest quarter of total sales activity and record up 35% year-over-year.

Reflecting broad demand and strong execution. Customer activity increased across all of our verticals, products and channels. Turning to Q1 results on Slide 7 and with all figures discussed on a normalized constant currency basis. Recurring revenues were $2.3 billion, up 10% year-over-year as our bookings performance from the second half of last year is converting into revenue.

Total revenues were $2.4 billion, up 8% year-over-year. Adjusted EBITDA was $1.2 billion, up 13% year-over-year, resulting in a 51% adjusted EBITDA margin which is up 190 basis points quarter-over-quarter and 300 basis points year-over-year. This is a result of our continued cost discipline power cost benefits and scaling our operating leverage.

As we have discussed, driving additional efficiency would be a focus moving forward. Quarterly AFFO surpassed the $1 billion mark for the first time, increasing 11% year-over-year and AFFO per share was $10.79 a up 10% year-over-year. Please note that adjusted for the [indiscernible] excel lease signing, which I will provide details on in a moment. we came in above the midpoint of our Q1 revenue and adjusted EBITDA guidance ranges.

As Adaire mentioned, we are near execution on the Hampton xScale leads. These types of negotiations are fruit and we have adjusted the expected timing while discussing expanded mutually beneficial terms with our customers. Here are the moving pieces as they relate to guidance over the past couple of quarters.

Our guidance for Q4 2025 assumed $54 million of nonrecurring revenue from the deal based on the original terms being considered. Our guidance for Q1 2026 included the expanded terms with an expected contribution of approximately $80 million of revenue, $65 million of AFFO and $0.65 of AFFO per share. The expanded economics are now included in our guidance for Q2.

This timing shift does not impact our full year outlook because the economics were already incorporated. Now to our nonfinancial metrics, which also demonstrate strong momentum. We increased physical and virtual net interconnections by 5,800 with particular strength in fabric additions.

We added 4,100 net cabinets billing and our backlog of cabinets sold but not yet installed is at a record level. Churn came at 1.7%, primarily due to the benefit of some delayed churn and to focus and execution doing our renewal process. For the full year, we are tracking towards the low end of our 2% to 2.5% guidance range.

And MRR cabinet increased to $2,524 up 7% year-over-year. Reflecting the firm pricing environment and continued increase in density. On Slide 12, our capital investments continue to deliver very strong returns. Consistent with prior years, this quarter, we completed the annual refresh of our stabilized pool, which increased by 5 IBX data centers.

Our 192 stabilized assets increased recurring revenue by 6% year-over-year are collectively 82% utilized and generated a 26% cash-on-cash return on gross PP&E. Turning to our capital structure on Slide 10. At quarter end, we approximately had $3.1 billion of cash and short-term investments on the balance sheet, and our net leverage was 3.8x annualized adjusted EBITDA.

During the quarter, we issued $1.5 billion of senior notes at the blended effective rate of 3.1% and reflecting positive execution in the market and our ability to take advantage of lower cost debt around the world. Our balance sheet and diversified capital program are competitive advantages in all macro environments, particularly so in the kind we see today.

In combination with significant retained cash flow, we continue to access lower cost sources of capital to fund our robust growth opportunity. Now looking at capital expenditures on Slide 11. Total capital expenditures for the quarter were about $1.3 billion, approximately 90% of which was growth and value-accretive capacity expansion. We continue to expect mid-20% unlevered cash-on-cash returns on investment.

Since the last earnings call, we opened 6 projects, adding critical capacity to meet demand across 6 metros. Before we get into guidance, I'll briefly address the energy environment given developments in the Middle East. We systematically hedge energy cost to provide predictability to our customers and broader stakeholders, particularly in volatile periods.

Globally, we are more than 90% hedged for 2026. And as usual, we are progressively hedging into the future. As a result, we expect minimum impact for 2026 even if energy prices were to remain elevated. Finally, please refer to Slides 17 for an update of 2026 guidance with all growth rates discussed on a normalized and constant currency basis.

Based on the robust environment and the team's execution, we are raising guidance across key financial metrics. For the second quarter, we anticipate continuing strength across the business, including MRR growth of 10% to 11% year-over-year. For total revenue, the largest piece to consider is that it includes the expanded economics from the Hampton xScale signing that I provided a moment ago.

Again, please note that these economics were already included in our guidance for the full year. They simply shifted from Q1 into Q2. For the full year, we are raising total revenue guidance by $21 million based on our Q1 outperformance, improving expected total revenue growth range by 100 basis points to 10% to 11%.

We are raising adjusted EBITDA guidance by $24 million, resulting in adjusted EBITDA margins of approximately 51%, a 200 basis point improvement over last year. Additionally, we are raising AFFO guidance by approximately $40 million improving our expected AFFO growth range by 100 basis points to 10% to 12%.

This corresponds to a similar 100 basis point improvement in our expected AFFO per share growth range to 9% to 11%. We -- we continue to execute on our capacity expansion to meet robust customer demand. Excluding xScale and land acquisitions, we now expect total capital expenditures to approximate the top end of our prior range at $4.1 billion, including $280 million to $300 million of recurring spend and approximately $3.8 billion of nonrecurring spend.

Given our confidence in the growth opportunity in front of us, the team continues to evaluate opportunities to accelerate our capacity to deliver growth and value to our shareholders. Overall, we are pleased with our progress and confident in our plan.

We will continue executing with discipline to deliver on our goals and create shareholder value. I now turn the call back over to Adaire.

Thank you, Olivier. Our Q1 results demonstrate strong performance and our outlook reflects underlying strength across the business. We see immense opportunity ahead to drive revenue, enhance margins and deliver attractive AFFO per share growth. But we take nothing for granted.

Our continued success demands focused execution against our strategic priorities and disciplined investment to unlock structurally higher returns. Above all, it calls on every member of our Equinix team to deliver exceptional value for our customers each and every day. This is the mindset guiding us forward, and I'm confident in our direction. We are well positioned across our markets.

We are building momentum in key growth areas, and we remain focused on delivering against the goals we have set. With that, let's open the line for questions.

[Operator Instructions] Our first caller is Ari Klein with BMO Capital Markets. Ari Klein with BMO Capital Markets. We'll go to the next caller, Michael Rollins with Citi.

I had a question about some of the comments you made earlier in the call. So I think if I got this right, you mentioned that 8 of the top 10 -- I think it was maybe hyperscalers and 4 of the top 5 -- no clouds are actively expanding with Equinix for AI, 110 separate network nodes.

And I'm curious if you could provide more color, is that 110 in addition to whatever cloud nodes they typically would have? And can you characterize the types of interconnectivity demand that you're already seeing for those AI nodes and how that's informing you maybe early in this environment of the type of growth that's out there from AI for your business model?

Okay. Mike, thanks so much for the question. And maybe let me just clarify a couple of points. So I mentioned that it was 8 of the 10 AI model providers, the LLMs and 4 of the 5 Neo clouds have deployed between the 110 or so separate network nodes to Equinix. And that is in addition to all of the nodes that we see that are being deployed by the hyperscalers in order to manage their connectivity journey.

When we look at the role of the neos here, we can see that for many of them, their journey is evolving a little -- their value proposition was always based on pricing and based on GPU access and largely facilitating large-term training footprint.

Mostly focus with the SaaS and the hyperscaler. As we can see, they're transforming into AI offerings workloads and looking to pursue enterprise customers and medium-sized SaaS companies. we see them as potential inference magnets for our ecosystem going forward. And we see many of them converging, as I've mentioned already, and engaging at Equinix.

It's about a couple of things in terms of the use cases -- it's about network outs that provide connectivity to the CSPs and the NSPs for the NEOs and the L&M. It's about AI inference notes for densely populated metro so a little bit of a different picture. And it's about fabric access to the enterprise customer base of Equinix. So that sums up the 3 things that we're seeing for the OUs of our environment.

Our next caller is Cameron McVeigh with Morgan Stanley.

I wanted to ask about the $140 million in pre-leasing activity. Just curious how tenant appetite is changing and if tenants are willing to commit further in advance for longer terms and really how that's translating to the terms for Equinix, whether through pricing, terms or deposits? Any color there would be helpful.

So pricing remains firm, whether we're looking at presales or bookings within the quarter. And I think the presales booking really provides our customers with security, security in terms of the infrastructure that they're defining and the opportunity to ensure that they are solving for their own compute and energy future. .

So this is something that I think we've done only in the recent past, but we're seeing a great benefit from that in terms of the conversations with our customer and our long-term ability to serve them.

Matt Niknam with Truist.

My question is more big picture on macro. Have you seen any macro dynamics, particularly around rising memory or fuel and energy costs and the prospects for higher IT costs later on in the year, affecting customer behavior at all, whether it's pulled forward demand or pushed out deals if customers are running into supply shortages?

I think as it relates to concerns about energy costs, Olivier mentioned our hedging program which means that we're in a position to be able to continue to support our customers at the price points that we're operating today. .

I would say based on the demand environment that we see that it is a very durable and broad-based demand environment. It is very diverse. And we're not certainly seeing any pullback from customers as it relates to increasing costs, et cetera, at this point in time.

I think you can see that reflected just in the sheer scale of the numbers of transactions and that those transactions occurred across all of our customer segments and also actually equally enough across all industries that were all growing at roughly the same percentage in Q1.

Our next caller is Frank Louthan with Raymond James.

As you see the rising demand for AI inferencing, is there any difference in the incremental capital required that you're seeing to fulfill those new workloads versus what you've traditionally seen? And can you quantify that if there is?

No, we don't see any difference in the capital that will be required notwithstanding the fact that our strategy has been to be very metro focused.

We are located in 77 metros across the world, and we will continue to build on that footprint but that's already embedded into how we've managed our capital because that's part of our 27-year history, and therefore, we don't anticipate any capital differences. I'm going to ask Phillip to add an additional comment here.

Yes. The only thing that I would add on to that, Frank, is that as we are always kind of skating to where the puck is going as they say. And thinking about the types of requirements that are needed for the deployments. And so when you look at some of our facilities that we're going to be bringing online in the next few years, the densities that we are building towards are much higher and much more suited for a lot of the requirements that we're hearing from our customers.

So we're always thinking about where we need to go and what the requirements are of our customers, we're building towards that.

Is that increasing or decreasing the returns that you're looking at going forward with that higher density requirement?

Now the density -- the returns we're underwriting against even those higher densities are still in that mid-20s percent that we've been talking about for a long time.

Our next caller is Vikram Malhotra with Mizuho.

I just want to clarify 2 things. One, just the bookings dipping sequentially. How much of that is seasonal? And maybe you can just some composition of traditional enterprise versus maybe chunky bits? And then just secondly, the interconnection business, given kind of the rapid tripling almost of the fabric business.

How is that playing into interconnection revenue growth overall? You mentioned sort of network enhancements needed there. So I'm just wondering like how does that flow through? Does that mean in the future, we see a greater pickup in the interconnection side?

Yes. So just to comment first on the sequential nature of our annualized gross bookings. First of all, where we've concluded Q1, and Q1 is seasonally a quarter that has traditionally been lower.

But I have to say that I am especially pleased with the performance that we had in given that we came off the back of such a large Q4. And so I think that the team worked really hard to deliver what was our largest Q1 ever and driving our largest backlog ever.

So I look forward to moving that into revenue in the future. And I'm proud that the delivery of our bookings in Q1 isn't just related to top line, but we did it at margins that are growing and profitability that is growing too.

Across the Q1 booking profile, we saw strength, as I mentioned already, across various different industries, but we also saw some very broad-based strength in our under 1 megawatt deal cohort.

As it relates to the second question around interconnection revenue and interconnection revenue growth. And we're obviously very pleased by the performance that we've seen here our Internet connection revenue growth was at 9% on a normalized and constant currency basis.

Fabric revenue growth was at 26%, and our fabric bookings grew 74% year-over-year. And this kind of growth, the value proposition that we're delivering to customers is really behind our investment strategy around our distributed hub and our fabric intelligence, which is in pre-preview with a number of customers and partners who are very positive about the outcomes that we're driving with this solution set.

Our next caller is Jonathan Atkin with RBC.

Yes, I wanted to just follow up on that last response and maybe ask you more directly. Is there a scenario over the next several years, their interconnection growth would exceed the growth that you're seeing and would represent a meaningfully increased percentage of your overall revenue composition?

I guess in some way, Jonathan, we're probably seeing that in our stabilized assets where our stabilized assets are growing at a 6% and interconnection within that asset group is growing at 9%. .

I do believe that there is opportunity for us to continue to grow our footprint and the range of services that we are offering to our customers here because we feel a very specific niche in the market in terms of providing that neutral environment where the ecosystem around AI converges.

And so there is potential for upside here, but that is not yet factored into our plans.

Our next caller is Irvin Liu with Evercore.

And welcome, Olivier. Appreciate the color on energy hedging -- just given your exposure to the Middle East, I wanted to understand whether recent geopolitical cross currency in the region have -- or have had any impact on your operations. specifically related to your ability to sell and/or add IBX capacity?

Yes. Thank you very much for the question. First of all, I think the most important thing for us is the safety of our employees, our customers and our partners, and that was our most important priority as we navigated recent events in the Middle East.

Thankfully, all of our people have remained safe, and our facilities are fully operational. We do have a limited footprint across the region.

We have a total of 6 data centers across the Middle East region, and they're comprising about 1% of total revenues. We have one project underway in Dubai at our DX3 facility. Construction project. And we have seen the RF state of that project be impacted due to the conflict. So limited operational impact, we were able to keep our facilities up and running. But we're watching the situation very carefully.

Our long-term view is that the region will continue to see growth in investment in digital infrastructure as the Middle East itself looks to position itself as a global AI hub.

Our next caller is Nick Del Deo with MoffetNathanson.

First, I want to congratulate Olivier on his appointment. And my question is also for him. I was wondering if you could elaborate kind of share with us your high-level capital allocation and operating philosophies. And whether your previous vantage point as a supplier to the data center industry provides any initial insights into areas where you think Equinix might improve the business or things you'll be focused on? .

Thank you for your question, Nick. First, regarding capital allocation, we're going to keep the course that has worked pretty well for the organization. First, what we want to do to fund our ambition -- ambitious CapEx program, growth program. We want first to use debt as a way to finance our growth.

We can do that based upon the leverage we have today, 3.8x. I mentioned that in my remarks. We will use equity on an opportunistic basis, but the key is going to use debt. Relative to impressions, I guess, that's the question you had as a supplier of Equinix in the formalized I was -- we were all of us very impressed by what we have seen.

We use Equinix always as pioneer in this market. And after 2 months, I've been -- it looks like a marketing comment, but it's true. Impressed by the quality of the team, the culture and also and mainly the rigor with which we run the operation. And what we have said before, you see the play -- we have high-quality data centers in top-tier markets.

We're connecting the world and we are ready to power the AI agent workload. So very happy. We are very differentiated and looking forward to help their and the team to grow this business even more.

Okay. Any particular areas where you're looking to drill down more or too soon to say? .

No. We want to enable the strategy that out there as a lineup, billboard, Solvesmarter, serve better. I'm going to be too many among many others, to enable this strategy, but no change today, not that there was a need to?

Thank you. Our next call is Richard Choe with JPMorgan.

I just wanted to follow up on the churn, 1.7% super low, but I think you mentioned some of it's delayed. And should we go back into the range. I mean, should be above range and -- or the rest of the year at the higher end?

Or could we be seeing a kind of maybe low end or towards the lower end for the full year?

Thanks very much for the question. As you saw at 1.7, we were below the low end of our range. And I think there were probably -- 2 elements as to why that was so. One was the timing of some churn, including our metal business moving forward into this quarter.

And others really is just the continued focus that we've had on the renewal process from our teams and so we're very pleased with the performance that we've seen in Q1. Notwithstanding that we don't want to call victory too early.

And therefore, I think to keep our churn in the range of 2% to 2.5% for the rest of the year, it's the right thing to do. We do believe that our focus on our available to renew contracts. We've been doing that much earlier in the cycle is actually starting to have an impact.

But we will watch those trends closely over the next several or so quarters. And obviously, our aim is to bring churn down consistently over time. But for now, we're holding into the 2 to 2.5 range for the year.

Our next caller is David Guarino with Green Street.

As we think about modeling in these large onetime fees related to xScale leases, I was wondering if there's any framework you can provide us to estimate and forecast how large they might be.

And then kind of tied in with that, we heard some rumors that the Monocle campus might have been pre-leased, but you guys didn't comment on that at all. So could you give an update on what's happening with that project? And how soon we could maybe expect another large xScale leasing fee after the Hampton one?

Look, these transactions are always very complex and multifaceted and particularly as we have very high demand assets locations that are energized within the right time frame in great locations. So I think as we look forward into the second half of next year in terms of Manuka, it is not timing that we have put into the short term is something that we are still working on.

We have a very robust pipeline of interested parties. And obviously, we want to ensure that we're maximizing the outcome for our customers, for our shareholders, for the company. As we look forward into the second half of the year and into 2026, the guide assumes a total NRR of approximately 5.8% for the full year, and a portion of that is associated with xScale leasing.

An additional comment, if I may, David. If you look at the balance of the year, with the exception of the xScale deal we have mentioned many times now, the rest of though the excel deals are relatively small in nature, and we believe that the risk is balanced for the rest of the year. Thanks .

Thank you. Michael Ng with Goldman Sachs.

Adaire, you talked about agents performing best when closer towards the edge -- have you seen some customer workload repatriation or a shift in investment away from public cloud as a result?

And then when enterprises decide to do more in the edge, could you talk a little bit about the customer decision tree between colocated data centers versus on-prem today?

Sure. And so I think the reality of the environment that our customers operate in is the environment that we've been describing on many of these calls, and that is a hybrid multi-cloud environment. where data sits across the plethora of all of those platforms.

And that creates the opportunity for a neutral platform like Equinix to serve customers who want to run agentic workflows across those environments, but need to access the information that fits in more than one location.

So I would certainly say that customers have a multi-cloud environment that they are, of course, looking at the cost associated with their environments as well as important considerations, particularly in locations like Europe around sovereignty and the compliance to the sovereignty legislation, which may mean that certain parts of their data set need to move into a private environment or be repatriated from cloud.

But I wouldn't say that this is a broad-based conversation that we have across our customer base. I think as we talk to CIOs, it's a conversation that is less about on-prem and cloud and more about the journey from token management, token cost all the way through to those kind of sovereign data controls that ensure that the organization is compliant to whatever set of data governance rules that they have in place for their own business.

And that's certainly a conversation that's an important 1 because we can help customers navigate that by providing through the distributed AI hub access to all of the players as well as to private SLL models, which companies have for smaller, less intense AA type activity. So I think the conversation is really about how you navigate from the token and the training all the way through to that compliance conversation often driven by sovereignty in some locations.

Our next caller is Madison Rezaei with Bernstein.

You've talked about potentially building multiple incremental gigawatts with the Build bolder program. With the full year CapEx plan now around $4.1 billion, is this a kind of annual spend we should anticipate for the next couple of years? Is it more front loaded? Do you think the intensity will ramp as you are sort of moving into more large campuses and a short follow-up to that, are you anticipating maintaining the cash-on-cash return level throughout that build process?

Okay. Thank you for the questions. Maybe we'll take that between us given that there's portions for each of us in here. First of all, as I think we mentioned in our materials, we have 3 gigawatts currently either in land under control or in development today at Equinix.

So that's the broad base of the portfolio that we are working with. But as Olivier mentioned in his prepared remarks, we are at the top end of the range that we mentioned for CapEx earlier at Analyst Day last year.

We are continuing to meaningfully grow our pipeline for new power land and capacity expansion opportunities to enhance what we see as the long-term growth prospects in key metros, which, of course, we know delivers very attractive returns.

So we're very pleased and excited about what we see in the business. We're very excited to position ourselves for growth. But you can see that we are at the top end of our range as it relates to from the Analyst Day event when we provided that guide last year. But perhaps I'll flick to Olivier and allow you to comment a little on the returns and so on.

So the diligence we have before to do deals, deploy new CapEx is very strong. The 25% is a target. That's not an aspiration. We are seeing that quarter after quarter. And we feel very comfortable with achieving that return target as we are in a market where demand is oversupply.

So we can be very selective about the deals we take. We are, and we are very differentiated today. And interconnection is more and more an important part of the value proposition of the company. So we feel very confident about this mid-25% target.

Our next caller is Erik Rasmussen with Stifel.

And Olivier, good luck and look forward to working with you. I wanted -- you talked about Maersk, one of your customer highlights. I know you had a liquid cooling deployment in Frankfurt. But maybe just overall, can you give us a sense of where customer demand is for liquid cooling activity today and how many active or signed deployments are using direct to chip or even immersion cooling? And how quickly is that moving from pilots and maybe scale production?

Yes. Thank you. Thank you so much for the question. So we had quite a significant quarter in Q1 as it relates to liquid cooling orders generally, of which Maersk was one I believe it was a 50% growth in terms of our liquid cooling deployments. And today, we have 36 deployments across our footprint of customers using liquid cooling to facilitate the workload and density of the systems that they have put in place. .

It's active across all our regions, and it's something that we continue to evaluate and work closely on with our customers. So that's, I guess, the landscape that we see as far as liquid cooling is concerned. I said, 36 deployment, 7 orders within Q1 across all of our regions, up 50% Q-on-Q.

Our last question comes from Joseph Osha with Guggenheim Partners.

Wow, I made it Kind of a follow-up from the previous question. As you think about these fairly power dense genetic workloads out at the edge of the network, are you encountering situations where either from a physical space of power or just a thermal standpoint, you're running into constraints?

I'm just trying to understand how much of a challenge that is.

I think probably the availability of power would be the largest constraint in our environment. So as densification increases, quite often, we would need to put some space on hold around that particular implementation in order to ensure that IBX, we're meeting not only the obligations of the workload that is the highly dense workload, but also the service level agreements and the obligations that we have with the other customers who are sharing that space and power. .

And that's, I think, one of the reasons why you see the yield on our MRR per cab grow so effectively up to our 2.5% to 4% of 7% year-on-year. partly due to the increase in densification and of course, the association of a value-added products like interconnect with every installation as one of the measures of that.

Thank you. I'll turn the call back over to you for any closing comments.

I just want to thank you all for joining us for our Q1 call. Have a great rest of your day.

Thank you. This concludes today's conference call. You may go ahead and disconnect at this time.

Okay. Might as well get started. Thank you all for coming. My name is Cameron McVeigh. I cover communications infrastructure here at Morgan Stanley. And pleased to welcome Jon Lin, Chief Business Officer at Equinix.

Thank you, Cameron.

Welcome. And before we get started, I have to read this. For important disclosures, please see the Morgan Stanley research disclosure website. If you have any questions, please reach out to your Morgan Stanley sales representative.

Okay. Done. Jon, let's get started. Equinix reported fourth quarter and full year earnings in February. CFO, Keith Taylor, called it the best quarter ever. Jon, what in your view was the main driver of this momentum?

Yes. I think a couple of different areas there. I'd say, first and foremost, when you look at the performance of the business in terms of our bookings, which is how we treat our customer sales essentially best quarter ever. And I think in a recurring revenue business, we're always under pressure to do that. But to be able to deliver that in a momentum that we're seeing around that, I'd say, I think for the full year of '25, we end up delivering 27% growth in bookings year-over-year. I think for Q4 end up being closer to a 40-plus percent number compared to the previous Q4 of the prior year. So it really just shows, hey, you're starting to see continued momentum in the business in terms of customer demand manifesting across a number of different dimensions.

I'd also say in terms of managing the business, I'd say we're continuing to be -- see continued improvements on driving operating margins. Also, the treasury team has done a fantastic job of managing the balance sheet, continuing to do work on our capital raises, et cetera. So when you put all of that together, I think it just put us in an incredible position to win. And that bookings and that customer growth was also across a really distributed base. I think we had 4,500 deals across 3,400 customers. And so it's not one particular segment that ends up translating into across both service providers and enterprises and across different segments inside the enterprise base, really sustained demand.

Great. And as you just laid out, solid 2026 guidance was well ahead of market expectations in the previous guide given at the Analyst Day in June. Maybe just talk about your key priorities and key focus for the year.

Yes. I think 3 different areas there. First, continuing to capitalize on the customer demand that we're seeing out there, satisfying the requirements. It is definitely an area where, one, we've got to make sure we've got the capacity to be able to satisfy the customer requirements. We've got to make sure we've got the coverage and kind of mechanisms from a sales and marketing standpoint to get in front of customers, hear the requirements earlier to be able to understand and shape the demand as they're thinking about these AI requirements.

Over the course of the last 3 years, I think we've been looking at AI at large and our early read around the industry trend was going to be -- there's going to be a lot of activity on the AI training side, which I think we've all seen. There's going to be a lot of investment into that, again, which we've all seen. Our belief was always the durable value creator for AI is going to be around inference. And when that materializes in a real way, that's when we'll want to go ahead and continue to step up in terms of our activity. And I would say that's what we're seeing. I think some of the announcements that we had at Analyst Day around our CapEx intensity and the increase there was a real realization that we're starting to see the enterprise takeoff here around durable use cases, around value creation for them.

And it's not just on the service provider side where, of course, you'd expect service providers to invest early to be able to see like build in front of where the revenue capture was going to be. And so now we're seeing that kind of long-term trend across multiple different segments around various use cases, both ones that are going to be agentic in nature as we're all hearing about now around how that's going to change things around generative AI, around also though core machine learning systems across different modalities that are going to need different intersections with different data sources. And we believe our -- again, Equinix' fundamental value is when data is from different locations, whether that's geographic, whether that's in different clouds, whether that's from different counterparties, we're the logical place to interconnect all of that and really bring that to life. And so we're starting to see that demand.

So the priority #1 is keep building as fast as we can around that, keep delivering that capacity to support these requirements, keep staying in front of customers to understand not just where are those requirements to land immediately, but also where are those requirements into the future so that we can -- again, you've heard us talk about we're preselling more than we ever have in the past. We want to be able to capture those requirements to inform our development pipeline as well.

The second area is just we're continuing to drive improved operating margins for the business. I would say, again, using a lot of the tools that we're looking at for digital transformation that our customers are talking about, we're also seeing the opportunity to transform our lead-to-cash systems, our kind of back-office systems to be able to drive increased efficiency there, and really flatten the curve in terms of the growth of those costs.

And then the third area is managing the balance sheet. Again, capital-intensive business, making sure we've got kind of best-in-class cost of capital access to be able to deliver services and build as we're -- again, a lot of our construction can be multiyear in front of when the revenue is going to show up. We want to make sure we have the best balance sheet and structure possible to drive that.

Great. That's helpful. Jon, I wanted to ask on latency. How often does the issue of latency come up when you're speaking to customers? And are there any particular workloads that are latency sensitive? Maybe how do you expect Equinix to serve a greater share of those workloads going forward?

Yes. Certainly, all the time because we bring it up. But you think about it from 2 lenses on for interconnection requirements for our customers, there's both the latency and also the throughput required for these workloads for AI in particular, but across the board. And the way Equinix has decided where to locate ourselves geographically, operating in 77 metros in 37 countries now around the globe, it's to be closer to where the eyeballs are, right? We're not developing our data center campuses where there's cheap power. We're not like looking for empty farmland. We're actually located where the nexus of population and like fiber connectivity exists so that we can serve those workloads.

And so whenever we're out there, we like to say, I think we're within 90% of the world's population within 10 milliseconds, right? And so when you think about 10 milliseconds, it's faster than the blink of an eye, it really lets you serve almost any kind of workload that is latency sensitive, whether that's real-time interactions, whether that's gaming, whether that's financial trading, whether that's media distribution most effectively, there's a lot of different areas that we can drive around that. And so those conversations come up a lot.

I'd say for AI, in particular, it ends up being twofold. One, when you're -- when all of us are interacting with our chatbot of choice, like that's not particularly latency sensitive, right? We're all waiting for the dots there and like the token generation takes longer than the latency of the network. Where it does become valuable, though, is when you start actually dealing with multimodal systems, so different data pools that are coming in there being ingested by your AI, well, every time that machine needs to go out, establish a new connection, find a new data source, get that and process that, well, that's like essentially hair pinning that traffic back and forth. And so when you start seeing those effects, that's when you really start to see latency sensitivity come into play.

The second is any kind of machine-to-machine communication. Again, when you start interacting with humans, we're all slow, that's okay, especially on a Monday morning. But when you start seeing the machines actually trying to converse with each other, if you're not able to have the lowest latency possible, well, that's just costing you either like time to token or time to value, whatever your transaction might be. And so in the case of financial services, that might look like, hey, for risk management calculations for end of month runs, that's not particularly latency sensitive. When you're trying to do trading and you're trying to make sure that like you understand what your next decision is going to be on a purchase around that or you're trying to go ahead and inform what your next set of currencies that you might want to move during a day period might look like around that or commodities today, right? Like any of that, the latency sensitivity can be dramatically higher. And so like that's an example of a workload.

But I also mentioned like in addition to latency, the throughput has actually become a more and more important element of the conversation. And like for a long time, we've been dealing with data across the network is increasing. But I would say post-COVID, it had kind of plateaued a little bit, right? I mean HD content is HD content is HD content. 3D-s didn't come about. 4K like streaming is there, but it's not massively relevant in terms of passing this. So the part that we had, as an industry, been thinking about was like what's going to be the next big driver of traffic across the network.

And I would say AI is an example of that, right? The ability to process AI is entirely reliant on the ability to process data and actually like process those workflows quickly. Every time that, that data is hung up on the network instead of hitting your GPU, you're costing yourself money, right? The GPU is the most expensive thing on the planet right now in terms of bringing that to life. And you'll hear Jensen talk about that tomorrow, like the unit of value that is most important is making sure that these GPUs are maximized. And that's why we're focused so much around like, well, in order to maximize that GPU, you've got to make sure that data is flowing into that thing as quickly as possible. So that's increased the amount of throughput required.

So we have -- for Equinix, a software-defined network that we call Fabric. It connects to both the cloud and all of our locations together. When we started that, those were 1 gig circuits, growing to 10-gig circuits with like 10 gig ports, growing to 100 gig ports. And now it's like we're upgrading that network fully across the network to 400 gig ports with 100-gig DCs, right? And so a 10x multiple in terms of the throughput and capacity required. And that can be anywhere from our service providers that are looking to drive AI traffic and workloads into enterprise deployments in private clouds to, again, enterprises trying to move data and workloads from different clouds into a central data repository so that they can have their private AI work against that.

Great. Related to that point, during your presentation at the Analyst Day, you mentioned partnerships with NVIDIA, Dell, Groq. How is Equinix leveraging these relationships? How has the partnership evolved? How do you expect these partnerships to help capture more of these AI workloads?

Yes. We're in a unique position in the marketplace as Equinix, where like we've talked about the neutrality of our business, and that's been core to our thesis since we were founded 27-some years ago now. And neutrality means a couple of things. One is we're carrier neutral. So we invite all of the different carriers to come to our facilities and exchange traffic, but it also means we're technology neutral, right? So we want to make sure that we understand and can bring to life all of the most important technologies in the world to our customers. And that's both from a sales perspective, but more importantly, from how can we accommodate that and how can we bring that to life for these deployments.

Our partnership with NVIDIA started maybe 5 years ago, right, before AI was a real thing and before enterprise AI became a real thing. But what we saw was, hey, they're making incredible progress around machine learning, and that's going to continue to be more and more relevant. Well, when that infrastructure comes to life, how are we going to be able to support that for our customers in our data centers. And so a lot of our conversations start there. So they've ended up evolving and obviously becoming an incredible partner from a go-to-market perspective.

But first, we want to understand how can we help them bring this technology to the rest of the world's customers. Again, our customers are the world's digital leaders at 10,500 customers now, roughly divided, 50% of them are service providers. So we've got all of the major cloud providers, all the major networks and all of the emerging ones, the neo clouds that are developing and all of the enterprises that then need to consume those services.

So it's really important for us to then not just understand, but then also be able to take an opinion to these customers like when they're asking how can I support these liquid cooled GB200s? I have no idea. I don't have liquid cooling in my enterprise data center. We can tell them, oh, we've already thought this through. We've worked with NVIDIA for 18 months on designing solutions around this. We can operate this for you with the same reliability of any production facility tomorrow for you, right? And you can go ahead and start deploying. And so that comfort level that can come from that, but also the value then that they can get from that technology becomes very powerful.

That's helpful. And on the topic of liquid cooling, curious if you could provide an update on the rollout of these liquid cooling solutions in the new and existing capacity. And then just how important is liquid cooling for AI inference?

Yes. So Equinix has about 280 data centers around the world. Around 100 -- slightly more than 100 of them at this point across 45 metros, again, globally, are liquid cooling like able to support, right? We've done pre-engineering work, and these include facilities of ours that are 10, 15 years old, where we understand what the central plant looks like. We understand how to bring water in cooling systems directly to a customer infrastructure, and we're actually comfortable operating that, right? I'd say 3 or 4 years ago, if you asked a data center operator, oh, we're going to have water coming to customer equipment that's like heavily electrically loaded. Is that going to be okay? Everyone is like, that might be pretty challenging for us.

And so we spent a lot of time figuring out, well, how do you actually operationalize that, right? The plumbing is one piece, but knowing what to do in the event of a leak, knowing what to do in the event of a fire, knowing what to do in the event of an operational situation where if you lose cooling capacity into that liquid, these machines are so hot that you will thermally overrun essentially within seconds, like that's obviously a more complicated scenario, right? So to be able to do this in production and be globally consistent around that.

So we did that work about 2 years ago, right, in terms of understanding how to productize the capability for liquid cooling, bring that to life for customers and have been solving for that in what we would consider like a standardized, productized way. We had been doing this as one-offs though for customers for over a decade, right? And so like customers would ask us and we try and figure that out. So that is going great.

I'd say we continue to see customer demand scale now, which is super exciting to see. I think we built honestly, in advance of where those requirements were going to be. And now we actually understand that pretty well. We're doing this in a repeatable motion. It's still, though, from a like total percentage of deployments or like customer request basis, it's still the minority, right, to be clear. It's -- when we see these requirements, again, in the 3,600 transactions that we talked about there in the like low double-digit percentages of total deployment. So call it, maybe several dozen to maybe 100, we'll have inquiries around that and figure out how to support that.

So it's still relatively low volume. But these workloads are generally -- when they are liquid cooled, it's almost always going to be an AI deployment. That's probably the most expensive gear that's on the planet that's going into those deployments. And so we want to make sure that we're satisfying the customer on that delivery.

That's great. Jon, I wanted to hit on cabinet density. Maybe just an update on your power density across your existing portfolio versus some of the planned data centers coming online, and how we should think about the average power density currently and at what rate that might start to tick up over time?

Yes. It's an interesting one. Obviously, you look at like Grace Blackwell, and that's at 120 kilowatts of cabinet for liquid cooled deployment supporting the latest NVIDIA, and that's going to go up, right? We all know like NVIDIA is talking about 300 kilowatts of cabinet for future architectures around that. And that's true. Like, again, the deployments that we have supporting some of that are at 120 kilowatts of cabinet. But when you think about the cabinets that are also supporting that, right? So it might be 4 or 8 cabinets of 120 kilowatts a cabinet GPU capacity, it will need 20 cabinets of storage. It will need another like 5 cabinets of networking gear and infrastructure to support that. And so that average can actually come down pretty rapidly.

Across our new deals, I think new deals booked were at 6.6 kilowatts a cabinet across the entire Equinix fleet. Like our existing base, though, across all of the things that are deployed and live is somewhere just north of 4 kilowatts a cabinet. When we think about data center design for the future, we're certainly upticking that though, right? And like the latest design that we're deploying in Dallas, I want to say, is around 18 kilowatts a cabinet. We generally think somewhere between the 15 to 20 depending on the mechanical infrastructure that we have in place. That feels like the right average.

And the reason I say that is we'll have liquid cooling to support as dense as you want. I think the latest designs where the team was telling me it's more like 500 kilowatts of cabinet or maybe even a megawatt. It's just you need a bigger pipe fitting for that, which blows my mind. But the important part is it gives you the flexibility, though, right? Again, if we used up, let's say, it ends -- let's say we're wrong on density that ends up being overly dense. We can support that with the liquid cooling infrastructure that we've deployed there. All it means is we strand shell, right? And honestly, the shell of a data center is the cheapest part about the construction. We want to maximize the utility feed that we're getting out of the utility provider. We want to maximize our electrical and mechanical systems. So we're delivering that.

If we're wrong the other way, if you build too dense and you actually can't sell that dense, you've stranded all of your capital investment on all of your electrical and mechanical plants. So that's a really bad outcome. And what we've seen over the years is it's really hard to tell where the puck is going to land. Our data center investments when we're building them, this is not a 3-year investment, right? This is a 2-year construction project, and we're expecting that to be able to deliver revenue for us for the next 20-plus years.

Like we want to make sure we're thinking about flexibility of that infrastructure then and the long-term value that, that thing can drive for us because we're not just worried about the customer for today, worried about the customer on the renewal. We're worried about the computers and workloads that will be supporting there. And what we've seen over time is that can move, right? Again, everybody has been focused on GPU density and like the training in that density. It is certainly very dense.

When we look in -- and we've been talking with silica start-ups for the last 5 years, when you think about inference and even like NVIDIA's acquisition of Groq, the AI inference chips are actually much less dense, right? The compute required to drive that, the complexity of the silica is significantly lower. And when you think about inference, a lot of that can be air cooled, that can be lower density. And 5 years ago, everyone was saying, hey, Arm is going to be the -- replace all of x86 chips in every data center, densities are going to go down to 4 kilowatts a cabinet. And maybe that will happen, right? So we want to make sure we're designing though so that we can go ahead and drive revenue and value for the long term.

Got it. So I'm curious how you think about price elasticity of your customer base. Monthly recurring revenue per cabinet has been going up. And would you expect this to continue over 2026?

I think there's definitely some amount of price elasticity there. I think it's certainly something that we're always trying to identify and figure out, hey, where are the opportunities for us. And it comes from a couple of directions. One is our pricing per cabinet will go up like inflation is a real thing around the world, right? And like our input cost in terms of data center construction, development, all of that, it has been going up over time and probably will continue to. Utility costs are highly variable. That can go up over time. The density of the cabinet itself, we're delivering, again, more kilowatts into a cabinet than in the past. So that can drive pricing and yield up on a per cabinet basis. Our interconnection offerings continue to drive more and more value for our customers. And again, one of the reasons Q4 was so great for us, like when Keith says it's firing on all cylinders, like interconnection actually started to reinflect in terms of growth, and we're starting to actually see that go up again, that drives more dollars into that yield.

So I'd say, overall, it's a complex algorithm to be able to get to exactly how far you can push that. But I will say the part that's very like interesting over the course of the last like 2 years, especially around AI and the importance of the workloads that are being driven there, the value creation being done out inside of that infrastructure now and also the pricing of like the GPUs and that technology stack is significantly higher than your standard like web servers of the past, right? And so when more and more value gets created out of that deployment, then there's a lot more elasticity around like, well, it's going to cost more to make sure it relies -- runs reliably is always on and can be connected to all of the data sources that matter like people are like, okay, I'll pay up to be able to go ahead and make sure that I'm maximizing my dollar value out of this massive capital investment that we're putting in.

Great. Jon, I wanted to be sure we mentioned power and talk about it. On the last earnings call, Adaire had mentioned the 3 gigawatts of developable capacity is powered or close to being powered. Can you talk about maybe just the power procurement environment and potential challenges Equinix is seeing at the moment?

Yes. It's -- on a global basis, it's interesting and like for the first time to be able to see in the mass media like talk about data centers all the time. So infrastructure is sexy again, that's great, except when it's negative press. It's fascinating, right? I'd say, one, you have to be long-term thinkers in this space to be able to have the relationship with the utilities that matter, and they have to trust you to know that you're going to be there. And candidly, there's a lot of money that's come into the data center space. They aren't long-term operators. They're putting a lot of weird requests in. And as regulated utilities are, they have to respond to those, right? They're not allowed to just say like, I have no idea what you're talking about, and I've never seen you before, I'm not going to give you power. They have to respond to that.

And so that creates a lot of challenges for the entire market of like, well, how do you sort between all of these different requests that may actually be at the end of it with one customer underneath that load, right? And so you end up seeing gigawatts worth of power requests coming into one particular market or jurisdiction. And again, many of us in the industry are like, oh, we know exactly what hyperscalers are, that customer wants that requirement. It's not actually like 2 gigawatts, it's maybe 500 megawatts, and there's this many people that are all trying to sell to that one customer, trying to get power certainty on land.

So we've been working with the utilities, and I'm really, really happy with the work that we've done there and just say like, hey, like raise the bar, right? Like actually force people to make it a little bit harder to get access to the utilities actually put processes in place that require like payments for these load studies. So it can't just be random land developers doing that. We want to make sure that there's real long-term providers that are credible making these requests so that we can get some of the noise out of the system. Again, do it fairly. That's obviously very important for us, and we're very serious about our neutrality. Like we want to make the rules for everybody the same, but like just raise the bar a little bit so that you can get some of the noise out of the system. And that's been a great help.

And then the fact that we're willing to do that, and we're willing to invest in all of the development, and we've been very like loud and open with the communities at large, along with the utility providers like the data center industry needs to make sure that we're paying the fair share of what we're doing, whether that's net new generation creation or transmission costs that we're creating so that we're protecting consumer rate payers for paying for what we're doing as an industry, right? Because -- and one of those is just from an ethos basis, we take that very seriously.

Again, we're going to be long-term owners and operators of these facilities. Our people live there. I live in Northern Virginia. like I see the data center development. I want to make sure like my kids are having the benefit of the tax revenue, et cetera, around that. And so we're going to be in there for a long time, and we're really deeply partnered with the communities.

But the second part is like if we look like more selfishly around that or more commercially, if we're doing bad things and consumer ratepayers are like ending up being impacted, you're going to get regulated. And like that's not a good outcome and you're starting to see the pushback now, right? Because again, I would say some players that were working in this industry weren't thinking as long term around that and you start seeing people say like, we should put a moratorium, like these data center things are going to like impact customer housing.

So I would just say it's an opportunity for the industry as a whole to kind of grow up, take this seriously and work with government, right, work in close partnership to say like we need to do this. Like one of the good things that I'm seeing out of the administration right now is like seeing, hey, there seems to be that lean of -- we need to make sure we're protecting the consumers from the impact of this net new piece.

We've been working with whether it's ComEd or PG&E, you might see some announcements that went out over the past couple of months where, again, because we're working with the utility around this and consuming load of generation that was already there, it actually reduces the consumer payer's price, right? And again, we're paying for substation development. We're paying for actually upgrades to the transmission system to be able to take that. If you have unused generation that's out there in the system, well, the consumer is paying for that as well, right, in terms of a higher unit cost on their power. And so the ability to have that load is useful.

You're starting to see more and more happen in the space, though, right? And I think there's a lot of talk right now about, hey, if you're going to bring large load on to portions of the grid, you've got to go ahead and make sure that you're actually kind of bringing your own electrons, so to speak, creating net new generation around that. And like total believers in that, right? Huge believers that we need to make sure that this industry as a whole is paying our share of the way around that. And whether that's ourselves or our customers around that, like the total economic benefit that we're driving around this is high enough that we should be able to take that.

And I think that, that's a really important message to communities at large. When they hear about data center development, and I'm sure all of you guys have been covering something like that and watching the space, there's a lot of push like what are these things actually doing? And it's like, yes, the data centers are not producing like cat pictures and stuff like some are. But really, like it is the fundamental like driver of economic activity now, right? Like digital infrastructure is the factory of today. It's the economic -- it's the mall of today. It's the economic center that's happening because so much is happening inside of these facilities.

And so it's really important for us as an industry and for our investors, et cetera, to understand that and be able to tell that to your stakeholders around like, hey, like this is actually how like GDP is created now, right? It's a meaningful portion of that, that's going to show up digitally. That's what's happening in the data center, right? It's actually coming online, converting power into money for somebody. That money for power is actually not for the data center developer generally. It's for the customers inside that data center.

Great. And just a follow-up on that point. How does Equinix internally, how are you thinking about using grid connections versus some of these behind-the-meter power solutions, you think about fuel cells or turbines? How does that fit into the strategy?

Yes. I mean we're operating, like I said, in 77 markets across 37 countries. So it's really all of the above strategy. In some jurisdictions and again, in the vast majority of the time, we want to work with the local utility, right? Like grid power is the best solution because you end up creating more shared resource across a broader set of constituents to be able to do that. In cases where the grid is not able to accommodate our power requirements, so we have done and actually created like net new power generation.

And we're looking into the future around that as well as like, hey, how can we scale that up? How can we do that? It's -- on a global basis, post kind of World War II and kind of the industrialization of the global economy, like we've actually generally seen power utilization and draw go down, right, for -- I want to say like a 40-year period. And only recently is it starting to inflect back up because of electrification of the grid, electrification of vehicles, more efficiency around heat pumps, like all of that's driving electric utilization up and then data centers as well. And that was like before AI, like you were actually, for the first time, actually seeing like draw increase across the globe. And then now you layer on, hey, all of a sudden, there's also all of this data center activity associated with AI and the electricity consumed by that. So that's net new like load that's coming on.

And for a long time, again, because utility was going down, well, those power plants were still there, right? You hear about all these like shuttered power plants. It was like, well, they were created because we were smelting aluminum. And I will say, from an environmental impact to data centers far, far better than like having aluminum smelting in the community you're operating in. And so like that's kind of the replacement that we're talking about now, just like, okay, it's fundamentally converting power into financial benefits.

That's helpful. We have a few minutes left. I wanted to open it up to any audience Q&A, if there are any questions. Now you can think about them. I'll ask one.

Go ahead in front -- there's a mic running here.

Just want to -- if you could just talk a bit more about the 20-year life of the assets and how you see that playing out? What do you -- obviously, the CapEx you've spent, a 2-year build cycle and then you're talking about the 20 years. In terms of the upgrade cycle you expect to have to do to make that 20-year life cycle work for the assets. Can you just talk a bit more about that?

Sure. Yes. And I would say it's 20-year plus, right? Again, most of our assets we actually think about as essentially perpetual. Like we've constructed the data center. We're putting in the maintenance CapEx to keep it going there, but it's going to be generating revenue for the entire -- like outside of our planning horizon, I guess, is the way I would say it. Some of that does require maintenance CapEx, right? Obviously, we're putting that in on an annual basis. There's also requirements like for some of the longer lead elements of that, like we started categorizing like redevelopment CapEx to say like, hey, if we have to replace the 15-year plant, like some of the core data center infrastructure, whether that's the generators that are running or whether that's like some of the electrical switching gear or some of the central chiller plant, that could be like a 15-year asset, and we might be able to say, okay, we can put enough maintenance in there to get a little bit longer out of that. But at some point, you're going to go ahead and want to refresh that.

The plus side of all of that work is there's been so much additional work that's happened in terms of efficiency gains and like just upgrading and understanding of how to do that mechanical and electrical work more efficiently that we can upsize that, right? And so in a lot of cases, we've ended up taking like -- take our DC2 facility right now that's going through some of this redevelopment, we can add incremental power availability into that site from both efficiency and from upsizing the infrastructure, which means we can actually generate more revenue into the same facility, right? And so that's how we think about that from an underwriting perspective. But yes, I mean, we're essentially continuing to model our ability to generate revenue out of these things for a very, very long time.

Could you briefly touch on xScale? If you think about it 5 years down the line, has it become a bigger part of your overall business compared to how you thought about it a year or 2 ago?

Yes, it's a great question. I'd say yes is the answer. And certainly, that's the reason that we -- when we started xScale, we were thinking, okay, it's -- our customers have been asking for it for a long time. For folks who don't know what about xScale, basically, our core business for Equinix on the data center side, they're multi-tenant facilities that we call retail, right? In any given facility, we'll have anywhere from dozens to multi-hundreds of customers inside one data center. A large part of the market, like the hyperscalers and some of the other large customers may want an entire facility for themselves, right? And so when we look at that, our average expectation on returns across where we're developing capital for our core business, it's in the -- I think our target that we've said publicly is 25%. It's in the 20% like mid-20s and up range, right? And it's like, again, it's a beautiful business.

I'd say when you're developing a single-tenant facility for a customer, like your costs are pretty well known, right? It's pretty apparent to the market at large around that. And yields around those facilities are generally, call it, in the high single digits to low teens. You can take some leverage on that, work that up into the mid-teens. So for 2 reasons then, Equinix had not wanted to pursue that as one of our growth vectors: one, it can be highly capital intensive with a much lower return; and then two, from an actual like the structuring of that because you want to apply so much leverage to that, we thought we wouldn't be able to maintain or at the time when we were thinking about it, get to investment grade or maintain investment grade, if we were doing a bunch of hyperscale development around that. And that's largely been true, right? When you look at outside of the data center landscape, I think there's 2 maybe like depending on how you squint, maybe 4 public players right now. And that's the reason why, right?

And so when we looked at it, though, our customers were asking us loudly enough for this product that we said, okay, let's see if we can figure out how to do this. We ended up creating a development joint vehicle structure. Our first one was with GIC, the sovereign wealth fund in Singapore, then we did one with PGIM. Now we announced one recently in the U.S. with both GIC and CPPIB, the Canadian Pension Investment Board together for $15 billion in the U.S. So that's where we're doing kind of solving for the hyperscalers around that.

I'd say -- so yes, in the sense that our initial investment for xScale was we had gotten $7.5 billion of equity commitment there. We've tripled that in total size now with the equity commitment from CPP and GIC, and we're executing against that. And I'd say it's delivered what we were hoping out of that, right? One, it gives us capital efficiency, but still gives us the customer -- being able to solve for the customer demand in the ways that we wanted to. And it's also given us quite a bit more scale in terms of our procurement capability on utility as well as equipment. So it's been good, and we'll continue rolling forward with it.

Excellent. We are out of time. Jon Lin, thank you so much.

All right. Thanks all.

[Technical Difficulty] with us, Equinix and CEO, Adaire Fox-Martin. This session is for Citi clients only and disclosures have been made available at the corporate access desk. [Operator Instructions]. Adaire, we'll turn it over to you to introduce your company and team, provide any opening remarks, tell the audience the top reasons an investor should buy your stock today, and then we'll get into Q&A.

Okay. Good morning, everyone. Thank you so much for joining us so early this morning, and thank you to Citi for giving us the opportunity. I'm joined by Phillip Konieczny here, who is our SVP for Finance, our treasury function and our IR function reporting to Phillip. And together, we look forward to addressing your questions today. And then I will have Phillip read that disclosure. Thank you.

Am I on here?

yes.

Okay. Some of what we will talk about today contains forward-looking statements. Please read our SEC filings for information about factors that could affect these statements, and I'll turn it back to Adaire.

So the top reasons for buying Equinix stock. I mean maybe let me just do this. I'll just lay out 4 reasons, and then I'll come back and drill into each one and see how that lands. First, the demand for infrastructure is absolutely surging and Equinix is at the center of that demand. Second, Equinix provides an essential layer of connectivity. Thirdly, we're building as the backbone of AI inference economy. And fourthly, we're executing exceptionally well against the opportunity.

So maybe let me just go back and drill a little on each of those. First of all, the demand for infrastructure is absolutely surging. Customers are asking us the same questions that probably many of your companies are asking around how they can differentiate their business models, their business process, their business outcome with AI as an enabler of those process changes. And that question then leads to considerations about the infrastructure that will be necessary to support that. And there's no doubt that, that infrastructure is much more complex and much more distributed than it has ever been before, and this plays directly to a core strength of Equinix. So our infrastructure demand element is a key factor in driving the growth scenario that we're experiencing today.

Second, Equinix is the essential layer of connectivity. We are very much the neutral ground where AI, cloud and networking requirements converge. So we enable enterprises to connect and to extract real business value. And there are reasons why that is so. First of all, there are advantages that Equinix has built over decades. We have the market-leading number of cloud on-ramps. We are a global organization with presence in many critical metros. We have the broadest enterprise customer base in the segment and more than 0.5 million business-to-business interconnections that are active today. So we are that essential layer of connectivity. These are our core strengths, and we will continue to build and evolve on them.

In the Q4 that just passed, we saw the role that AI has in driving our business momentum. 60% of our largest deals in our Q4 portfolio could be attributed to AI workloads. Now interestingly, of that 60%, 50% of those were led by enterprises. So companies in the retail, financial services, health care and media sectors making use of Equinix to deploy their AI workloads. We also saw an uptick in our liquid cooling sales during Q4. We had 11 sales of liquid cooling during Q4 and 5 of those 11 interestingly went into New York campus to support financial services. So we're certainly supporting the ability to land these workloads in an Agentic world where we're supporting it with network diversity, with cloud proximity, with AI-ready interconnection and with the latency that these workloads require. Location and density do matter, and this is something that is important to us, and our platform continues to excel as every connection adds to the platform.

And then finally, strong top and bottom line growth. We're very much focused on delivering shareholder value. We do that through very disciplined execution. We focus on doing the important things well, on serving our customers so that we generate the highest return and that we deliver the AFFO per share that our shareholders would like to see and expect from us.

So 4 elements that we think are reasons underpinning why our stock is a good buy today, demand for the infrastructure, Equinix as this essential layer of connectivity, the backbone of the inference economy and very strong top and bottom line deliverables underpinned by excellent execution from the team over the last quarter.

Adaire, thank you for that. Maybe jumping in on some of what you were describing. Just from a high level, when you look at the business, what are the metrics that give you and the management team confidence in the durability of this retail-centric strategy given rising power density needs for IT infrastructure and as the hyperscalers are expanding their offerings to go after the same customers.

So I think that there are a number of signals around our KPIs that indicate the durable nature of the demand and also, I guess, the role that Equinix uniquely plays in fulfilling that demand for our customers. First of all, I think you can see it in the bookings performance in Q4. And I would say actually throughout the second half of last year, but particularly in Q4, we had a very, very strong bookings performance, delivering $474 million of annualized gross bookings. But I think it's less about the actual dollar numbers, although they are a significant growth that represents a significant growth on a year-on-year basis and on a quarter-on-quarter basis, but more about the sheer number of transactions that underpinned that number.

We did more than 3,400 transactions with over 4,600 unique customers. So in our model, we are very diversified. We're very diversified across segments. We're very diversified across industries. And we're very diversified across the workloads that we support from our customer base. So I would say one of the first leading indicators of the durability of the model is the bookings because bookings subsequently turn into revenue within that 90-day period.

The second indicator is our pipeline execution and our pipeline discipline. How we are managing to come into every quarter with a highly qualified pipeline of opportunities, which means that our conversion rate in situ in quarter is particularly high. This is nothing other than relentless discipline. And I think that the teams have been doing an amazing job looking at how we navigate our pipeline, but also how we navigate the opportunities that are there to the best location for that opportunity. Today, we're in a supply-demand environment that is very favorable to Equinix. And quite often, we find ourselves in a situation where we may have 8 or 9 customers who want the same piece of capacity in a particular location, in a particular metro. We work now to ensure that through a rubric mechanism, we identify the customer that we will work with to actually implement into that capacity, but that doesn't mean we tell the other 8 to go away. We work with the other 8 to demand shape the workloads into capacity that is available in our footprint. And that has been something that has been very beneficial both in terms of long-term customer relationship, but also in maximizing the use of our capacity.

So bookings, pipeline certainly a second measure. And then I would say the third from a top line perspective is how we are managing churn. When we look at our churn, we're at the lower end of the range that we've guided to. We guide to 2% to 2.5% churn. In Q4, we are at the lower end of that range. And a number of reasons for that internally. One, some new predictive capability that we built ourselves with AI to have a look at our customer base and define our customers into segments that we define as ATR, available to renew. So we know well in advance when a customer is up for renewal. Reconfiguring our customer success motion in order to have them focused on upsell and cross-sell opportunities and making those phone calls early. And we've started to see that have an impact on churn because obviously, keeping a customer means that the revenue continues rather than there being a gap in revenue when you implement a new customer into those environments.

So I think those are some of the top line measures that we're looking at in terms of growing our business. Obviously, we're very focused on our operational efficiency and effectiveness. How we utilize our power is a key component of that. And that's something that we have done very well. Also, we have had a tailwind because power prices have come down, surprisingly enough, given the narrative around energy over the past period of time, and that is something that obviously falls straight to our bottom line.

Very helpful. If we could also double-click on your comments around the bookings. So we think of Equinix as a very well-established company, been operating now for 27 years and you delivered 42% bookings growth in the fourth quarter after building momentum through 2025. So can you discuss as an organization, what has to come together to deliver that type of bookings result? And maybe give us a little preview of what that means also for the sales capability of your organization as you look forward?

Sure. So I think there are a few factors. I mean, first of all, we have a significant presence in all major metros as it relates to sales execution. So we have almost 700 quota carriers across our business who operate across the 10,000-plus customers that today are our customer base. And so when we open capacity in a new metro, we're not necessarily increasing our sales costs in order to prosecute that new capacity because it's already embedded in our metros and embedded in our team.

So first of all, we have a very focused sales force who have targets that they are obviously set to meet, processes that support them in doing that, systems that support them in doing that and looking as always for the opportunity to make that whole process as effective and as efficient as we can so that we convert as many opportunities as possible. I think another important measure of our success last year was our ability to accelerate capacity into the 2026 window, the 2025, 2026 window. Our design and construction team at any one time will have 50-plus projects, 50-plus major projects underway around the world. And our team were tasked with releasing more capacity into the sales force for the sales force to be able to execute against that. They accelerated 30% of the retail capacity last year. And they did that through a number of different techniques, but essentially, it resulted in an acceleration of 30% of capacity.

Now there's still some window to do some of that as we move into this year, but that will become a lever that you can only pull so many times once you've looked at your project processes. But I will say that our design and our construction team are generally accelerating builds and delivering capacity faster. So that was also a factor that enabled us to have the bookings that we enjoyed in Q4. And I think since we opened the visibility of bookings to our investors, we did that in June. You can see that bookings are ticking up and moving to the left -- beg your pardon, moving to the right. And we hope that we're in a position with our pipeline to at least be able to continue that trajectory because certainly, the demand in the market is there. But I would say that bookings can also be variable from a seasonality point of view and so on. One of the things that we've done is try to iron out seasonality across our year. And to have the teams really focus on Q1 and Q2 as 2 quarters that build in current year MRR for the current year, which is obviously how we're judged. And then 3 and 4, 2 quarters that then set us up beautifully to exit into the next year. So my team don't call Q1, Q1, they call it Q5, so that we continue to build on the momentum that we had last quarter.

Maybe just to add one other thing. I think another motion that's relatively new for us is because of the capacity that we've been able to pull forward is what we call kind of a preselling motion. And so we've opened up the window with which we're -- our sales team is allowed to sell into new IBXs that are going to be coming online that used to be -- if you were to ask me 5 years ago, that would have been about a 3-month window, 3 months before our site is going to go live, sales will be able to sell into that. Now we have extended that out to about 12 months. And so we've talked about a presales balance that we have of $170 million, which sits on top of the bookings. And the way to think about that is that will turn into bookings in the future periods. So we're able to derisk some of this capacity that's coming online for us in the future quarters as well.

For example, Mike, in Q1 at our earnings call, we indicated that at the time of earnings call, we'd already closed 45% of our Q1 pipeline, and we had $100 million in presale at the time of earnings.

And that's on top of the $170 million that you started with?

Just as we think about efficiencies internally for the company, how are you thinking about the opportunity from AI deployment internally? How are you thinking about build versus partner versus buy and those solutions and opportunities?

Probably like every other company, looking at it through the lens of how does this help us be accretive to our core measure. There are certainly some things that we would buy, but there are areas of our business where we have, we believe, a unique approach and that process differentiates us and enables us to execute in an extrapolated way around certain topics. I'll give you, for example. We have built AI capabilities into our capacity visualization, which is something that is quite challenging to manage in a retail colocation facility. Remember, we have hundreds of customers sitting across our various different data centers. And in many cases, you're managing your capacity a little bit like a Tetris block. So one customer may take this small piece here and another customer a small piece here and then you have to look at the capacity that's available in that location and see how you best configure that in order to make that available for our customer. So making that visible to our sales team through the lens of AI and being able to look at in real-time potential configurations of that available capacity is something that's been highly valuable.

And just providing a single source, enabling us to work with customers across regions, across geographies, just one example of where we self-build. We've also self-build an AI capability around our cage configuration. We collect a terabyte of data a day. It's incredible to think of the data that we collect and have available every day. And when we look at our customers, they're often in more than one location with us. And so we know already from engaging with them pieces of how they like to make use of Equinix and what kind of configuration they like to have in Equinix. So rather than start with a blank sheet of paper, we start with a twin. Based on all the data of their previous implementations, we can now say things where we visually build this cage, we can say to the customer, "Hey, you always have the camera in the top left corner. Is that where you wanted in this cage." And then we just place the virtual digital camera in that environment, regenerate the cage specs again. And so by the time that we get to closing that transaction, we're very close to understanding what actually has to be implemented. It just passes across, which means that we reduce the time frame between booking and billing, which means then that we can get faster to MRR. So lots of examples of where we're using this also on our construction side, too.

Thanks. And for our last 15 minutes, I'm going to give a little preview of the topics that we're hoping to hit? And then if there's any other questions or topics that our group here wants to hit, please raise your hand or use the system. So real quick, we'll talk about AI inference, the financial guidance. xScale enterprise a little more and atNorth, which was recently announced. So maybe first, moving over to -- or staying on this whole theme of AI. Can you describe a little bit more from a customer perspective, what you're seeing in terms of the uptake of the inference models? And what are the milestones that investors should be looking for as they try to gauge the timing and magnitude of benefit for Equinix, including what enterprises could contribute to this?

So first of all, we have the privilege of spending a lot of time with our customers. I meet a lot of customers every single week in different parts of the world from different industries. So you get a lot of anecdotal information around our customers' use of this tech based on those meetings, those conversations and the work that our technical teams are doing with our customers. I would say that we are like very early days really in this. If you think about your own company, probably like I have just described, many of you are making the move from proof of concepts that have demonstrated some business value to the implementation of a solution in real production environment.

And the vast majority of companies are no different to the kind of journey that you and your own companies are on. So I think it's very early to specifically call milestones at this point. But there's no doubt in our minds that the demand is extremely sustainable, and that we will see this continue to augment the broad base of workloads that we support in our data center environments. I think that 60% figure is an important indicator, but that was of our top deals. Today, we are working on the systems to enable us to capture that information all the way through the segments of transactions that we capture. And I guess we'll get better to let me when we achieve that. But today, it is early to call specific milestones, I think, but we can absolutely see architectural and infrastructure discussions, I mean, that take into consideration things like the fact that most customers will be multi-cloud. Most customers, particularly if they're regulated, will maintain a part of their data either with us, usually with us or behind their own firewall. Most customers will have complex requirements when it comes to managing the underpinning network and the connectivity that an Agentic workload may require. And data gravity and access to your customers are key components of being successful here. And I think these are all elements that play to the Equinix value proposition.

And then moving on to the financial guidance. Just given the bookings momentum that you were describing earlier, what are the opportunities to grow recurring revenue at the upper half of the range or better? And maybe also it's an opportunity just to bridge how the AFFO per share growth outlook has evolved from back from the Analyst Day, which is mid-single digits. Now you're in upper single-digit territory.

Yes. I think there are probably 3 reasons why you see a difference between now and Analyst Day, and then I'll get Phil to drill in with a degree of specificity. First of all, our bookings performance, our top line performance was better than we expected in the second half. And I think we have -- specifically as we ended the year, and I think we have built momentum and muscle memory into the system around that bookings performance.

Secondly, from a capital perspective, we were able to raise at a lower cost than we had initially envisaged at Analyst Day. And we also were able to utilize that capital very efficiently and capitalize it. So that also, I guess, was a second reason why you would see a difference in terms of where we are, and I'm completely having a blank on the third. Yes. And the third was the capacity that we were able to accelerate into the year, which I have already mentioned to you.

So I think those are some of the differences between where we were on Analyst Day in June and where we are now. And we're very pleased with that performance and very pleased with the momentum that we see in the business. And so working to ensure that we continue to drive for value creation and that we continue to drive as hard as we can towards the top end of that range that we guided to.

Yes. And maybe just to add on a little bit just on some of the numbers specifically. So on the AFFO per share guidance that we gave of 8% to 10% for the year, so 9% at the midpoint. That does include 100 basis points of benefit from the timing of a transaction moving from Q4 to Q1. So we want to make sure we normalize for that. But it's still, I think, to your point, Mike, 300 basis points of improvement relative to what we saw last summer, all to the reasons that Adaire kind of just pointed out and what's driving that. So we feel really good about the momentum there. And I guess just one last point that I would make around when you think about the underlying performance of the business, obviously, it all drops to the bottom line on AFFO per share. But really one area to continue to focus on is our MRR growth, our recurring revenue growth, which really talks about the underlying fundamentals of the business. And again, we've seen incredibly strong momentum around that in 2025, and that's obviously extending into 2026. So that's the key -- one of the other key metrics, I think, to focus on as we go through the year.

We have a handful of questions that are coming in through live Q&A. So we'll make sure to get to those. One of them is, are you comfortable with the return and control you give to capital partners at xScale? Will you adjust that in the future?

Sorry, can you repeat the first bit? We didn't catch that.

Sure. It said, are you comfortable with the return and control you give to capital partners at xScale? And will you adjust that in the future?

I think our partnership with our xScale partners is working exceptionally well. we continue to evolve and grow our xScale portfolio with their support. So no, I don't see that changing in the short term or medium term.

And so with xScale, maybe one other question is just an update on the pipeline that you're seeing for xScale and the regional expansion opportunities beyond the Hampton facility, which I think Equinix has described in the past is expecting to lease up during this calendar year.

Yes. We have a very strong pipeline for our xScale opportunities. It's a model that allows us to serve a very strategic set of customers. It's also important to remember that many of those customers also have a very strong position in our retail footprint. And that joint adjacency between xScale and retail is something that is a value add both for them and also for the broader customer base who require cloud adjacency for their workloads.

And I was just going to say that some of -- you've seen some of the land acquisitions that we talked about last year that we've done. Some of those we anticipate will be contributing to the xScale franchise here in the U.S. But I think it also points to -- we've got ambitions to do more in EMEA and APAC as well with our xScale business. And some of what we've talked about that we have acquired, we anticipate standing up additional JVs in those theaters as well.

And can you discuss how you look at the TAM expansion opportunity from going after what I think are described as more midsized deployments from your enterprise customers? And maybe dovetailing into the atNorth announcement, how does that play into your enterprise strategy? And maybe just help people understand how to frame that in the broader portfolio in xScale.

Okay. So when we look at our enterprise TAM, I think what you're referring to there, Mike, is that our customers are certainly asking us for larger footprints than we have been previously asked for in the past. And this is something that we factor into our acquisition strategy, both our land, our power acquisition strategy, our capacity development strategy and the design of our data centers. The latest data centers that we're designing are being designed to a higher density than the ones that we would have built, for example, 10 years ago.

We are still commanding high pricing and strong pricing against that demand backdrop. And I think that our flexibility to be able to support the multimodal requirements of our customers from those in -- who operate in the KVA space up to those who operate in the megawatt space is something that marks us as a little unique across the segment because we manage across 3 footprints from wholesale to large footprint all the way through to retail colo. In addition to the TAM at the enterprise level for large footprint, I also think that as we begin to navigate through our channel strategy this year that, that will unlock more TAM for us as we will have sellers on the street who are not part of the cost of sales specifically or the SG&A costs of Equinix, and that is something that we're actively engaged on this year looking at how we -- with some of the tools that we've developed, really activate our channel partners to sell and represent Equinix.

And then as it relates to the atNorth acquisition, Phil was very close to that, so I might allow him in a moment. But we're very pleased with the acquisition. We're very pleased with the partnership as it continues to evolve with CPPIB. We believe it's a very complementary acquisition. First of all, in a theater of operation that's extremely important to us. Secondly, with a company that has a 17-year history and shares many of the ethos and values that we do around power and the sustainability narrative in that region. But thirdly and more importantly, releasing capacity for us in EMEA, where we know we have a very, very strong demand profile in a time frame that will be critically important to us.

And then fourthly, knowing that as soon as we close this transaction, it will be immediately AFFO accretive. So a lot of very positive elements about the atNorth transaction, which, of course, is announced but still subject to regulatory process.

I just want to hit another one that came in. What is your exposure to the software industry? And how do you view the risks this industry is facing from AI?

So when we think about the software industry, I'm assuming that there's an underpin around some of the Software-as-a-Service providers here that would make use of Equinix as an environment. Our exposure is extremely limited. Less than 3% of our total MRR would come from companies in that segment. So our exposure is extremely limited. I think, again, just to reiterate, we have an extremely diverse customer base across all segments, across all industries, across all geographies. So we have very little risk concentration in either a single segment or a single customer across our revenue mix. I do think what's unfolding in the software industry will take some time because when you think about many of these solutions, they are the backbone of operations for many companies, and they are not something that can be simply just switched off or simply replaced. But to answer the question around exposure, it is less than 3% and speaks to the diversity of customer base that we enjoy at Equinix.

Great. That brings us to our rapid fire. So on what KPI would you pick as the most important for your category? And what is your expected growth in that metric for 2027?

Well, for us, AFFO per share is the most important measure, and I think we've already indicated our growth prospects.

And then just Will the data center sector have more or fewer or the same number of companies a year from now?

Probably roughly the same, but we do know that there are some that are looking to go public this year, mostly in the wholesale space, but we'll see how that pans out.

Terrific. Thank you.

Okay. Thank you.

Good afternoon, and welcome to the Equinix Fourth Quarter and Full Year 2025 Earnings Conference Call. [Operator Instructions] Also, today's conference is being recorded. [Operator Instructions].

I would now like to turn the call over to Philip Konieczny, Senior Vice President of Finance. You may begin.

Good afternoon. and welcome to our fourth quarter conference call. Before we get started, I would like to remind everyone that some of the statements that we will be making today are forward-looking in nature and involve risks and uncertainties.

Actual results may vary significantly from those statements and may be affected by the risks we identified in today's press release as well as those identified in our filings with the SEC, including our most recent Form 10-K filed on February 11, 2026, and our most recent Form 10-Q. Equinix assumes no obligation and does not intend to update or comment on forward-looking statements made on this call. In addition, in light of Regulation Fair Disclosure, it is our policy not to comment on its financial guidance during the quarter unless it is done through an explicit public disclosure.

On today's conference call, we will provide non-GAAP measures. We provide a reconciliation of those measures to the most directly comparable GAAP measures in today's press release on the Equinix Investor page at equinix.com. We have made available on the IR page of our website a presentation designed to accompany this discussion, along with certain supplemental financial information and other data.

With us today are Adaire Fox-Martin, CEO and President; and Keith Taylor, Chief Financial Officer. At this time, I'll turn the call over to Adaire.

Thanks so much, Philip. Hello, everyone, and a warm welcome to our call today. I'll start by saying on a personal level, how deeply pleased I am with our performance in 2025 and particularly in Q4. I Demand for our solutions has never been higher, and our teams have stepped up exceptionally well to capitalize on it.

To our employees around the world, I'd like to say a huge thank you for all you are doing to achieve our goals. I also want to thank our customers and partners for the trust they have placed in Equinix as we intensified our investment in growth, investment that is already paying off.

Our bookings have accelerated dramatically. Our recurring revenue growth rate continues to climb, and we are managing our spend with great discipline. All of these factors are combining to fuel an expansion pipeline and growth in AFFO per share materially ahead of expectations. Given our momentum exiting 2025, our confidence in our 2026 plan has grown. This is reflected in both our revenue and our AFFO per share outlook.

Our performance and our outlook demonstrate 2 things: The first is that Equinix is connecting the fastest-growing segments of technology infrastructure and the value of our platform is increasing with every connection. The second is that our execution continues to improve. This winning combination delivers superior customer outcomes and stronger shareholder returns. Not only are we on the right track, we're moving far faster along it.

Now our momentum is clear across several key metrics. Monthly recurring revenue, the most powerful driver of long-term value creation grew 10% in Q4 and 8% for the full year on a normalized and constant currency basis. And for bookings, the leading indicator for revenue performance, the story is even stronger. We delivered annualized gross bookings of $1.6 billion in 2025, up 27% year-over-year. Our Q4 bookings were $474 million, up 42% year-over-year and 20% from Q3, well ahead of our plan.

And we delivered record capacity to meet growing demand. including 23,250 cabinets in our retail footprint and more than 90 megawatts in our xScale business in 2025. We delivered more than 30% of this retail capacity ahead of schedule, which we believe will accelerate our growth in 2016 and beyond.

Now Keith will go deeper into our metrics shortly, including recommendations on how to model the Hampton transaction. As you saw in our release, the expected timing of this transaction shifted from Q4 to Q1, which, as we have shared in the past, simply reflects the fluid nature of ex-scale lease signings.

Before I turn the call over to Keith, I'd like to provide some additional perspective on why our business is performing so well. For starters, we are building on the strength of our fundamentally differentiated value proposition. Our decision to double down on the digital infrastructure and connectivity requirements of enterprise customers is fueling our success, particularly as they implement AI across their operations.

In Q4, approximately 60% of our largest deals were driven by AI workloads. That's up from approximately 50% earlier this year, a trend line that we believe will continue as we are really only at the beginning of this journey. Equinix is the neutral ground, where enterprise infrastructure converters, and we provide the essential layer of connectivity that make it all work at scale to unlock real business value. This is a long-term tailwind for our business, particularly as AI inferencing expands across industries.

With market-leading native cloud on-ramps, the largest global footprint across the most critical markets and the broadest enterprise customer base, we deliver what AI inference demands, network diversity, cloud proximity, AI-ready interconnection and low latency. These are structural advantages we have built over decades, and we believe they will continue to set us apart.

In 2025, we completed over 17,200 transactions, up 6% year-over-year with over 6,100 unique customers. Our Q4 transaction volume was the highest ever at over 4,500 deals with more than 3,400 unique customers. We saw an uptick in volumes for all workload sizes from single cabinet requirements up to our largest page configurations. And more than 60% of our existing customers added new services last year.

Now let me share some recent customer wins and use cases that really showcase what's driving this demand. Salesforce chose Equinix to create a private multi-cloud networking there for the engine inside their data and AI foundation. This is our largest global fabric cloud router sales to date. By deploying Equinix Fabric cloud router across 14 countries and 21 metros, we are enabling private network connectivity between Salesforce's presence in AWS, Azure and other cloud service providers.

Alembic, an AI-powered marketing analytics platform selected Equinix for the scale and consistency of our global operations and the richness of our interconnection ecosystem. We are working with Alembic as they deploy the NVIDIA DGX superpad with NVIDIA Grace Blackwell systems to expand their addressable market through distributed AI. Signet, a leader in AI-powered workplace safety and operational intelligence shows us because Equinix Fabric security connects their edge devices, cloud providers and customer networks. This enables them to deliver real-time AI-driven quality control for industries ranging from food processing to manufacturing.

Leading quantitative trading firm, Hudson River trading, selected Equinix because our global footprint and our advanced cooling solutions enable them to achieve the latency and the density requirements they need to power their next-gen AI trading workloads. And Fortune 500 multinational Honeywell Corporation expanded its relationship with Equinix because of the secure flexible solutions and global fabric connectivity we provide, including for key metros such as Shanghai, Tokyo and London.

This is the first of many projects driving the integration of internal AI applications and the wider transformation at Honeywell. These are just some of the use cases underpinning our momentum. Our progress is a direct result of the changes we have made through our Serve better, Solve Smarter and Bill bolder initiatives.

Starting with Serve better. Customers continued to secure both near-term and long-term capacity across our global platform. Our $474 million in annualized gross bookings in Q4 were supported by an incremental presales balance of $170 million. with more than $60 million occurring in the quarter. larger footprint requirements from enterprises and service providers contributed to this performance.

Our pipeline is strong. and we have already booked approximately 45% of our Q1 2026 target and signed an additional $100 million plus of presales as of today, already our largest presale quarter ever. As we accelerate growth, we are committed to a disciplined pricing strategy that's commensurate with both the strong and durable demand patterns we see and the differentiated value our solutions provide. This translates into our strong cash and cash return profile and the premium yields we secure.

We continue to underwrite our newest projects with these principles in mind. With Solve smarter, we are helping customers connect and simplify their infrastructure. Interconnection revenue grew 9% year-over-year on a normalized and constant currency basis. We added 7,800 net interconnections in the quarter, including our adjustments.

We also crossed an extraordinary milestone in Q4, surpassing 0.5 million interconnections worldwide. To put this into context, that's more than double our nearest competitor. As AI amplifies the need for massive real-time data movement, Equinix is delivering the connectivity infrastructure that enterprises depend on the most.

As part of our Build Bolder initiative, the work we are doing across our global development portfolio is strengthening our competitive advantage. In Q4, we delivered more than 12,000 cabinets to our retail business across key metros. Our development engine remains exceptionally active with 52 major projects underway across 35 markets, including 9 xScale projects. Since October, we have added 10 new expansion projects.

We also closed on a number of strategic land acquisitions last year, adding approximately 1 gigawatt to our powered land under control balance. This positions us well to meet long-term demand from both enterprise and hyperscale customers.

Our xScale business achieved a key milestone in recent weeks. In January, we contributed our Hampton asset to the Americas JV, an important first step towards deploying $15 billion of capital across major metros in the U.S. The Hampton facility will support approximately 240 megawatts of IT capacity when fully built out.

The signing of the lease for half of this facility to a hyperscale customer is expected in Q1, and we expect the site to be fully leased later this year. In addition, we have established a healthy leasing pipeline, demonstrating the value of our ex-scale strategy. Of the approximate 3 gigawatts of capacity that can be developed close to 1 gigawatt is currently earmarked for our Excel business.

Further, our land acquisition pipeline continues to grow as we maintain our focus on major metro expansion opportunities where we are confident of significant long-term demand. As such, we expect xScale will continue to contribute to NRR over the next several years as we execute our strategy. Overall, we're winning where it matters most. By connecting the infrastructure that powers the AI build-out happening across industries. And we are laser-focused on expanding our category leadership through disciplined execution that drives healthy revenue growth, margin expansion and superior shareholder returns.

I'll now turn it over to Keith to take a closer look at the numbers behind the quarter and our outlook.

Thanks, Adaire, and a very good afternoon to everyone. First and foremost, I'll start by briefly building on Adaire's comments. Simply, it was an outstanding close to 2025. From my perspective, Equinix delivered its best quarter ever. By far, closing the last quarter of the year with over $470 million of annualized gross bookings and more than $60 million of presales activity, well ahead of our expectations and certainly any prior quarter.

The magnitude of our quarterly activity both across a substantial number of diverse deals, but also with over 3,400 customers underscores that our strategy is working. -- and meeting the opportunity in front of us. Also, I'd say that is our retail business that's the standout, delivering record bookings across each of our small, medium and large size of deal categories. And alongside our strong sales execution, the teams continue to operate the business with great focus and discipline, and we're only getting started.

This created better-than-expected margins. We raised our capital more efficiently than planned and investor to utilize the cash more effectively, creating better-than-planned cash flows. Suffice it to say, our performance is creating momentum in the business to drive attractive recurring revenue growth and the scaling of healthy AFFO per share performance.

Now I'll cover some of the highlights for the quarter as depicted on Slide 7. Note all growth rates in this section are on a normalized and constant currency basis. First, as we indicated last quarter, our Q4 guide assumed the execution of a large ex-scale lease for 2 of the 4 buildings on the Hampton campus. As Adaire noted, this transaction is now expected to close in the first quarter.

We also had some onetime planned benefits, which will not repeat themselves in Q1. That being said, our fundamental underlying performance was meaningfully better than our expectations from top to bottom. Please refer to Slide 15 that bridges our quarterly revenues from Q4 2025 to Q1 2026.

Now let me move specifically to revenues. Q4 revenues were $2.4 billion, up 7% over the same quarter last year, fueled by continued strength in our monthly recurring revenues, which was up 10% over last year and as you likely noted, a steady quarterly improvement throughout 2025. Q4 revenues, net of our FX hedges, included an $8 million currency headwind when compared to our prior guidance rates.

Global Q4 MRR churn was 2.2% lower than planned. And for the full year, our average quarterly MRR churn was 2.4%. Looking forward, our teams remain highly focused on reducing MRR churn, which includes the development of AI predictive tools to help identify opportunities to eliminate or defer any MRR churn. Global Q4 adjusted EBITDA was $1.2 billion or approximately 49% of revenues, up 15% over the same quarter last year. Q4 adjusted EBITDA, net of our FX hedges included a $4 million FX headwind when compared to our prior guidance rates.

Our 2025 margin improvements reflect our continued focus on delivering higher operating leverage across the business while also investing in growth. As it relates to 2026, we expect to continue to deliver improved adjusted EBITDA margins while also absorbing both accelerated and increased expansion drag, the byproduct from our growth investments. Global Q4 AFFO was $877 million or up 13% over the same quarter last year, including the seasonally higher recurring CapEx spend.

Q4 AFFO included a $2 million FX headwind when compared to our prior guidance rates. And our nonfinancial metrics continue to demonstrate strong momentum across our core or key metrics of net interconnection additions, net new cabinets billing and MRR per cabinet pricing. Our net interconnection additions increased by 7,800, including both physical and virtual connections and also include our adjustments. As Adaire highlighted, we've now surpassed the 0.5 million interconnection milestone across our ecosystem, an unmatched competitive advantage, which has been curated over 25 years of history.

Turning to our cabinets billing. We added 4,300 net cabinets billing in the quarter, including cabinets from our Main One portfolio. Our underlying net cabinets billing increased by the highest level in 3 years, driven by strong bookings performance across each of our 3 regions. Our backlog of cabinets sold, but not yet installed, stands at a record given the bookings performance, especially in the second half of 2025.

And finally, we continue to drive attractive MR per cabinet yields, stepping up $65 quarter-over-quarter on a normalized and constant currency basis, driven by very favorable pricing backdrop, increasing power density and strong interconnection attach rates.

Turning to our capital structure. Please refer to Slide 10. As of year-end, our balance sheet increased to approximately $40 billion, including cash and short-term investments, totaling about $3.2 billion. Our net leverage was 3.8x our annualized adjusted EBITDA. We -- during the quarter, we issued $1.8 billion of senior notes at an effective rate of about 3.2%.

Throughout 2025, our diversified capital raising program allowed us to raise debt at very attractive rates, which helped us optimize our 2025 net interest expense.

Looking at 2026, we plan to continue to raise debt in Loracost locations, either directly or synthetically. -- including Canada, Singapore and Europe. And now looking at our capital expenditures for the quarter. Please refer to Slide 11.

Capital expenditures were approximately $1.4 billion, including our planned seasonally higher recurring CapEx of about $140 million. We opened 16 major projects across 14 markets since our last earnings call, adding important retail capacity to several of our key undersupplied metros. And we announced 10 new projects, which will be added to our global portfolio over the next few years. Revenues from owned assets are 70% of our recurring revenues.

Our capital investments delivered strong returns as shown on Slide 12. Related to our now 187 stabilized assets, revenues increased by 6% year-over-year on a constant currency basis. These stabilized assets are collectively 82% utilized and generated a 27% cash on cash return on the gross PP&E invested on a constant currency basis.

Now given our strong underlying Q4 results, our 2026 outlook is expected to be meaningfully ahead of our expectations that we shared with you at our June 2025 Analyst Day. Our strong pricing discipline, coupled with our best-in-class capital allocation efforts should allow us to generate industry-leading durable cash flows with attractive cash yields, thereby delivering revenue growth and long-term value for our shareholders.

Please refer to Slides 13 through 17 of our summary of 2026 guidance and bridges. Do note all growth rates are on a normalized and constant currency basis.

Starting with revenues. For the full year 2026, we expect total revenues to grow between 9% and 10%, which includes a modest 40 bps attributed to the NR from the XL lease timing. We expect our monthly recurring revenues to grow between 8% and 10%, driven by strong 2025 bookings performance, including presales momentum. MRR churn is anticipated to remain comfortably within our targeted range of 2% to 2.5% per quarter.

We expect 2026 adjusted EBITDA margins to be approximately 51%, an expected 200 basis point improvement over 2025 and reflecting anticipated revenue growth and focused expense management. 2026 AFFO is expected to grow between 9% and 11% compared to the previous year. AFFO per share is expected to grow between 8% and 10%, which after adjusting for 100 basis points of ex-scale lease timing is 300 basis points higher at midpoint relative to our prior expectations from past summer.

2026 CapEx is anticipated to range between $3.7 billion and $4.2 billion including about $280 million of recurring CapEx spend. We have not included any on-balance sheet xScale spend as we do expect to be reimbursed for these costs as we transfer these assets into the XL JVs. And finally, we expect our quarterly cash dividend increase by 10% over 2025 on a per share basis. As a result, our total 2026 cash dividends paid will approximate $2 billion. So let me stop here. I'll turn the call back to Adaire.

Thanks very much, Keith. As this is my first opportunity to address all of you since Keith's retirement announcement, I wanted to take a moment to acknowledge his tremendous impact on Equinix over the past 27 years. His leadership has been integral to our success and has helped us lay the foundation for our future. I am particularly grateful for Keith's partnership since I joined the company, and I look forward to his continued support as a special adviser over the next year.

Our process for selecting Keith's successor is well underway and we look forward to updating you when we have news to share.

Before we turn to questions, I want to leave you with a final thought. Equinix is at the center of a historic multiyear infrastructure investment cycle. To deploy AI at scale, enterprises need to connect and manage increasingly complex and distributed technology ecosystems. Equinix is the neutral connector that unlocks business value for our customers. It is what we do best. It is where we have continued to focus and our focus is paying off. We were built for this moment, but execution is everything.

And we will continue to prioritize doing the most important things exceptionally well, so that we delight our customers whilst delivering structurally higher returns and AFFO per share growth for our shareholders. That's exactly what we did last quarter, and it's what you should expect from us going forward.

So with that, let's open the line for questions.

[Operator Instructions]. Our first question comes from Eric Luebchow with Wells Fargo.

Adaire, maybe you could touch a little bit on kind of the bookings momentum that you talked about that came through in Q4, and it sounds like you're off to a really good start in Q1. I think you said 60% plus of the largest deal came from AI workloads. Are you seeing more of these kind of coming from traditional enterprise companies adopting AI? Is it coming more from the hyperscale? Or you're just putting edge nodes in your facility? And as you kind of look out, do you think that is going to continue to rise throughout the course of the year.

Thanks very much for the question, Eric. Let me perhaps unpack a little of that stat for you. And I think the value that we provide is obviously something that's been amplified, thanks to the continued investment in this sector overall and the excitement around it. I think for us, the breadth and scale of our product continuum, they're very -- it's very uniquely aligned to meet this demand. And as I mentioned in my prepared remarks, and you just repeated there, 60% of our largest deals were driven by AI workloads.

Now during the course of the quarter, when I look at that 60% related to AI, interestingly enough, nearly half of them were deployed by noncloud and IT companies but they were deployed by companies in the retail, e-commerce, manufacturing, financial services and content sector.

So I think this demonstrates increasing enterprise AI adoption outside of the service provider community. We also had 11 liquid cool deployments in Q4, 5 of which were in our New York City facilities underpinning the requirements of our FSI customers to support elements, use cases like algorithm training in that sector.

So I think this also demonstrates continuing strong diversity, and from an interconnection lens, we see a very healthy growth in the AI service provider ecosystem, although it's still early days, given, I think, the breadth and depth of our established ecosystem density. So interesting to see that it was -- of the 60%, 50% driven by noncloud and IT providers, really demonstrating the growth in enterprise application of AI processes to business. And we see this as a continued positive tailwind for our company.

Our next question comes from Jonathan Atkin with RBC.

I was interested if there was any update to your multiyear guidance provided at the Capital Markets Day last June. And are those targets still relevant? Or should we be thinking about 9% as being the new baseline for AFFO per share growth?

Jon look, first and for foremost. Thank you very much for the question. I think what's most important is to understand just the underlying momentum of the business. As Adir highlighted both in her prepared remarks and certainly in the commentary around our booking activity. The business is performing well. We sort of said and we've said it over the last 2 quarters, and we're sort of -- we're talking about it this quarter. It's the execution on the top line is the management of the cost as well as efficiently raising capital and deploying it appropriately.

And so I think we're on the right trajectory. We're delighted with what we've delivered for 2026. I just feel it's a little premature to talk about 27 and beyond. But as you can appreciate, -- the momentum is windowed or back. And I would add to that, I think currencies are going to continue to be winded or back as an organization. So the combination of really strong performance, the ability to maybe invest faster than we were anticipating, as Ralph sort of bringing forward as many of the assets as possible. I just think we're in a really good spot.

And I think it's just a bit early to talk about 2017 and beyond. But I know you can do math very, very well. And so I'm going to leave it to you to sort of interpret both how we're performing this quarter being Q1 '26 and what that really implies is exiting 2026. So again, feel good about our position, but let's defer the '27 discussion until later.

Our next caller is Aryeh Klein with BMO Capital Markets.

First, I guess, Keith, congrats on your career and wishing you the best moving forward. I guess, just going back to the previous question around AI in that 50%. Do those deals look differently than different than some of your more traditional deals, be it from a size or location standpoint? And then, I guess, from an underlying demand standpoint, it seems like things have meaningfully accelerated since the Investor Day back in June. Can you just pass that, what specifically has, I guess, driven that acceleration momentum?

Yes. So let me comment perhaps on that accelerated momentum first, and then I'll address the second part of your question, which is around the characteristics of some of those AI deals that we saw in the first quarter. I think when I look at the gross booking outcome that we saw in Q4, in fact, it was an acceleration from Q3 all the way through to Q4, very much across the second half of the year.

And as we've mentioned, in Q1, we already see some early acceleration in the current quarter. So when I look at the characteristics of what is driving that, I think it's, in large part, 2 very specific but very different themes. The first is the external one, and I'll speak to that in a moment, and the second is the internal one. So if I think about the external one, we're seeing robust demand right across all workload types.

And what was extremely interesting about the demand profile more generally across our Q4 performance was that it was right across the different segments that we serve. We saw all segments and all verticals grow. So it was a very robust, broad-based demand across numerous different types of workloads. We, of course, saw specificity around AI opportunity where there was connectivity requirements for near metro connectivity requirements or connectivity requirements to unlock the value in that process.

So very robust demand across workloads, across all industries, across all regions and across all of the segments that we serve. Then there is the internal piece, which is the execution of the team against this particular opportunity that we saw in the market. First of all, we did accelerate some capacity into the year. And that was a helpful element in terms of our output and our performance.

But we also had a phenomenal pipeline conversion rate. So the team converted at around 49% in Q4 which speaks, I think, to the quality of the pipeline that we had coming into the quarter. And as I've mentioned previously, we've been putting a lot of focus on ensuring that we're forecasting the future pipeline with the same intensity that we're forecasting the current booking pipeline and really building our footprint strongly as we enter each quarter so that we know that we're executing on very highly qualified, very highly qualified opportunities.

Interestingly, it was also great to see some very firm pricing right across all of those segments and across all of those regions in Q4. So great disciplined internal execution that kind of met the market moment where the market was at -- and as it relates to the characteristics of the transactions that set in that 60% contingent, it was, of course, some of our larger transactions were related to our AI opportunity.

And I think one characteristic of all of them, I've shared with you some of the use cases already across the different segments that we facilitated this functionality into. But 1 characteristic was that we saw a 33% increase in density compared to the non-AI deals. So an average of about 10 kVA per cab for these transactions.

Our next question comes from Michael Rollins with Citi.

Keith, I also want to extend my congratulations and best wishes on your upcoming retirement. Maybe going back to some of the comments from earlier in the call, Keith, I think you were discussing the opportunity to improve churn. And curious, as you're preparing these new tools to help get that churn down, how much of this is in the company's control versus how much of it is just simply customer optimization of workloads that the churn will happen irrespective of how Equinix tries to get in front of it.

And then just a secondary clarification on the 6% stabilized revenue growth, can you unpack that a little bit in terms of what's delivering that incremental strength relative to the recent history.

Yes, Mike, this is Adaire. I might take the churn question, and then Keith answer the question around the 6% stabilized growth for you.

Let me unpack the churn commentary. So over the course of the past 2 quarters, we saw that our churn has set more at the lower end of the range that we guide to that at the higher end. And this is a combination of access to data and unpacking the data, the tool that we're working on and have some early pilots in place that Keith alluded to in his prepared remarks, the fact that we identify much earlier in the process, what we call ATR or available to renew. So much earlier in the process, the cohort of our customers that sit in that ATR category and then being able to deploy our customer success team onto that ATR cohort of customers in order to make phone calls much earlier in the process than we've been doing in the past and to facilitate not just a renewal, but even potentially an upsell opportunity.

One of the things that we saw in Q4 was that 60% of our existing customers added additional services from Equinix to their portfolio. And every time you have a call with the customer, you have an opportunity to share what we're doing around new services. So I think it's the combination of those elements, ATR, telemetry and visibility into our data -- the customer success team focused on this and tools that help us predict more accurately.

And there is, of course, as you quite rightly point out, a proportion of churn that is not addressable by us. And that is now something that we have visibility into, and it means we focus our resources, our efforts on the proportion where we can actually address an outcome for the customer and for the business.

Michael, first, thank you for the nice comment. I appreciate it. To respond to the question on stabilized assets and the growth rate. The beauty of what we shared, number 1 is that there's more and more volume going through basically the stabilized assets that isn't necessarily cabinet related because our utilization is 82%. So that's number one.

Number two, you got higher density per average cabinet. And so that that's working in our favor. Number three, which I think is really important. And then also sort of refers back to 1 of the comments I made in my prepared remarks, where we have some one-off benefits that don't necessarily repeat themselves at least quarter-over-quarter. And that's basically price increases, price increases relating to the fourth quarter. And so it's a combination of all those that really delivered a strong stabilized growth rate on an asset abate that's really quite substantial. And so feel very good about what we delivered. I probably would say, though, and to try and give you some steer if you will, as you look forward. We still generally feel good about the stabilized assets, growing 3% to 5%. That's the typical range.

We are going to have these these periods were slightly higher and sometimes we might be a little bit lower just based on the timing. But overall, we feel really good about the 3% to 5% growth rate attributed to stabilized assets.

Next caller is Frank Louthan with Raymond James.

So part of your footprint expansion was to be able to capture some of the increased power demand from enterprises and we've definitely seen the demand ramp up. So where are you today with regards to your ability to offer -- to meet that customer demand for folks needing incrementally new levels of power from enterprises. Have you caught back up on that? Where are you in that process?

I'll answer that for you. Thanks for the question, Frank. So obviously, power and the sourcing of power is a very significant factor for us as a data center operator. having power means that we're able to secure the compute and energy future of our customers. As we indicated, we currently have of developable land under control, and it's not develop a landfill stop. It is powered land or land that we are close to securing the power on.

So I guess we're not in the business of sortitiously buying a block land if we're not sure that we can power that asset. And so that means that as we look at our capacity and our build profile moving out, we are building against powered land portfolios, which, therefore, will enable us to continue to advance and evolve our footprints and our facilities to meet the density requirements of our customers.

As I mentioned, we saw already in the AI workloads that that we enjoyed in Q4 being 33% more dense than non-AI workloads, and we can certainly see that density increasing across our footprint. So we believe that we're very well positioned to address those requirements of our customers.

And Frank, maybe just adding on to what Adaire said because I think it's important is we have 52 projects currently underway. They're energized projects. And that -- I'm talking about generally the retail space. And as we shared with you at Analyst Day, whether you look at a DC 17 or our new Dallas build, they're coming with scale and size, but they're not so big that maybe there's excess focus on it. And so I feel really good. We talked about the 167 megawatts. And so it gives you a sense of we're building capacity in markets where there's a sort of broad need for that capacity, and they're the important markets, the Chicago, the New York, the Dallas is the Washington and you go around the world and think about all those critical markets and that we're trying as hard as we can to build out on that capacity. And.

So with the 52 projects currently underway, it sort of just -- it adds to what Adaire saying that we have the current opportunity and then we also are creating the future opportunity for ourselves as well.

Our next caller is Michael Funk with Bank of America.

Yes, Keith, first, congratulations there. over the years. So in the prepared remarks, you mentioned a disciplined pricing strategy. Keith, just curious kind of the magnitude of how much higher you can take pricing? And then you mentioned a minute ago, the 3% to 5% projected range for growth with some variance. Can you add any more comments around what would cause the variance, whether seasonality timing? And maybe a final part of my question would be, -- are you seeing ability to change contract terms on renewal, whether that's increased escalators or other factors?

Okay. I'll take the question, and Keith can add some components to it as needed. As I mentioned, we experienced very firm pricing throughout Q4 across all segments and for all regions. And we're very disciplined around the approvals and the approach that we take with our customers. because we recognize that in the platform that is Equinix, there is a range of differentiated value that allows us to accelerate the pricing opportunity.

Our pricing commanding yields as a result of our interconnection density as a result of our cloud on-ramps -- and of course, the metro locations, which becomes even more important when we look in a low latency world around certain applications of AI workloads. So I think from a pricing perspective, we're very disciplined. We're very focused on that, and we know that we have opportunity to accelerate that in that regard.

The vast majority of our contracts auto renew, and they renew with a particular pricing increment applied to them. But this is also an opportunity for us to have a conversation with our customers. And that's why the ATR program is an extremely important 1 for us because it does allow us to look at customer usage, not just of the space and power within our footprint, but some of the additional and incremental services that Equinix offers and then allows us to enjoy the price point that those services represent in terms of value for our customers.

Our next caller is David Guarino with Green Street.

Your stabilized cash gross profit growth has been excellent all year, is in 2025. And part of that, I know, was due to shrinking expenses. So wondering do you think that trend of reducing cost will continue or at some point, are you going to have to increase staffing levels if this outsized pace of bookings continues?

David, let me take that one, Adaire, jump in as needed. So one of the comments we made is that we're just getting started. I think when you step back and look at the organization, we're driving the top line. Ralph and his organization is doing a great job of managing the IV access to the gross profit line and then you have the rest of the organization. And today, for round numbers, you're 18%, 19% SG&A as a percent of revenue, right?

We have a stated goal that we really would like to get -- we'd like to improve that. And that's through bending the cost curve. It's not necessary -- it's not always about eliminating costs, it's bending of the cost curve and becoming more efficient. -- so our goal is to get to 15% over some period of time. And so the combination of managing into the different markets and also managing the spend while also investing behind [ Hermine ] and her efforts to, I guess, to create efficiency in our organization through processes through systems, through tooling -- it's a combination of all these things that I think can make a difference over the years and still early, but we're offering up this year a guide of roughly 200 basis points of improvement is coming from the top line, is coming certainly from Ralph's organization, and we're still investing in the business to develop future opportunity for us.

So I'm not necessarily sort of subscribing to what you said that we need to throw more bodies at it. We tend to be a little bit more head count dense than almost anybody else out in the marketplace. And I would argue that we get more leverage from that as we introduce more assets into our environment. You don't -- if you build another -- put up another building in Dallas, as an example, you don't need to necessarily go higher more SG&A to support that asset.

And so I feel very comfortable that when you look out that we can become an increasingly more efficient business. And consistent with the comments we made at Analyst Day that we're certainly on a nice trajectory, but we said 52% plus. And there's a very good reason why we put the plus at the end of the 52. So it's a journey that we're going to be on together, and I'm excited.

Our next caller is Michael Elias with TD Cowen.

And congrats on the results. question for you regarding the bookings. Obviously, great to see big bookings quarter in 4Q. What I'd like to get a sense of is how would you rank order the contribution to the bookings from the cabinets that you have coming online in capacity-constrained markets, obviously, brought on cabs and Nova as well as Frankfurt, both of which we're capacity constrained versus it being a structural acceleration in demand.

And really, what I'm trying to get at is how sustainable is the 4Q bookings level because that obviously has implications for forward revenue growth?

Okay. Michael, I'll have a stab at the question and then Keith can jump in if I -- if we need additional clarification on a couple of points. I guess, over the course of the last year, we began to show our annualized gross bookings to you as a metric. And the reason for doing that was to give you with the combination of the presales number to give you a sense of the momentum that we see inside the business in any given quarter.

And since we've done that, we've seen our annualized gross bookings moving upwards every quarter. And certainly, Q4 was no exception to that particular outcome. We have a very strong pipeline going into Q1. We've already closed 45% of our Q1 target. And we've had a meaningful presales experience in Q1 to date that has given us our largest presell quarter even though we're just halfway through.

So I think looking forward, our bookings growth will continue to be a very strong indicator of underlying help I think the best line of best fit is up and to the right, but that's where we see the demand. But obviously, there will be some variability quarter-to-quarter seasonality and other elements kick in. So that's something that we will manage as we look at it on a quarterly basis.

Then as it relates to the cabinets, Keith, do you want to...

Yes. Michael, I'll just add a little bit to what Aldaire said as well. I think our expansion tracking sheet gives all the people on the call and a pretty good sense of where we're making our investments. we need more capacity, and we're going to continue to invest more capacity. And as Adaire alluded to, we have roughly 3 gigawatts that we're considering over a period of time, while we're still building currently I think the combination of continuing to make the current investments while thinking longer term, while at the same time, demand shaping to markets -- and you've always heard us talk about the right customer with the right application going into the right data center, and that will continue to hold true.

But particularly in markets where they're constrained, not just for us but for the industry, we feel that we can demand shape that opportunity into other markets that are proximate or within the fiber route sort of environment that would make it suitable for our customer to consume -- and so I would just say that there's a lot of things that are going into it, different markets, they're going to have different sets of circumstances. But this is what we're focusing on as an organization, not only increasing the density, but making sure that we demand shape to the right markets in support of the customers' needs. And that also plays in a little bit to the pricing as well.

So hopefully, that gives you a bit of a sense that, yes, we understand that some markets are more constrained than others, but we're we're also going to build in adjacencies such that we can maybe continue to enjoy the opportunity that's in the marketplace.

Our next caller is Nick Del Deo with MoffettNathanson.

And Keith, congratulations on your upcoming retirement, and thanks for all your help over the years. At Dare, earlier in the call, you noted the strength in the interconnection franchise some of the big cloud service providers that you work with have announced and are developing products to help customers go to multi-cloud route I was wondering if you could talk a bit about the puts and takes of those efforts as it relates to your interconnection franchise.

Thank you for the question. I guess at Equinix, we have always understood and always appreciated the importance of the network. And I think some of the announcements that we've seen around connectivity is a validation of connectivity and the importance of the network as part of the broader AI ecosystem and landscape. And of course, we have continued to invest in our network products, in our interaction product portfolio, and it is a significant part of our footprint and a significant part of our global revenue.

We recognize that many of the clouds have made a cloud-to-cloud connectivity announcement. But this is a very simple use case. -- cloud to cloud is a simple connectivity use case. In reality, the reality of our customers is much more complex than this.

We've always believed in a multi-cloud hybrid world. And that requires a much more complex consideration around connectivity and networking strategies for our customers. And the interesting element for us, of course, is that for many of these clouds, they are very valued partners of Equinix and they use a large part of our infrastructure to help create their value proposition in terms of networking positions and network puts -- now we're always evaluating our strategy here and always looking at what is emerging around the ecosystem.

And that's why for us, we recognize the inherent value of our interconnection franchise. The role of fabric continues to evolve within that franchise as we see provisioned VC capacity stepping up quarter-over-quarter.

We have a range of very exciting developments around our footprint here, our product footprint here and developments that will simplify the networking journey for our customers, developments that will support as it relates to the augmentation of AI agents and AI workloads and development that will really enhance the optimality -- sorry, I'm stumbling over that word that our customers enjoy on our network.

So a lot happening in this space for us but we really see that many of the announcements really validate the role that the network plays in any AI ecosystem and is really, I think, a validation of the connectivity opportunity that we see ahead of us and is 1 of the reasons why we've continued to invest in this space.

Last question comes from Cameron McVeigh with Morgan Stanley.

I just wanted to echo my congratulations to Heath. And Secondly, you've spoken in the past about the shift from training to AI inference workloads. And curious if you have any updated views on the timing you've seen -- and then secondly, this may be a follow-up to the last question, but just how important interconnection offerings are to drying AI inference workloads from enterprises?

On the first part of your question, I think when you look at how our deal profile has moved through the course of into 26, an uptick of 50% to 60%. We can see this tailwind of opportunities perhaps margin earlier as an enterprise footprint than we originally thought when we presented in summer of last year. And so I think that's a really good news because it is taking the promise of AI and putting it into the hands of consumers of citizens and customers all the way around the world. And as it relates to the role of interconnection in AI workloads, I think there's a very significant element here for us to consider.

And in some ways, the sales force example and that I shared in prepared remarks is a really great example of where our connectivity capability was able to provide a very unique service to sales force around creating private network connectivity. And so we're very excited about the opportunity to serve our customers' evolving needs, needs in a complex hybrid multi-cloud world.

We think it's a very fast-growing space. We think it's something that would be additive to our colo capabilities because it will broaden the range of customers and that it will continue to strengthen our platform and our ecosystem. So this is something, as I've mentioned before, that we're very excited about the opportunity and see that we have a very strong competitive differentiation when it relates to others in the segment with this interconnection footprint.

Okay. Great. Well, thanks, everybody, for joining our conference call today, and have a great afternoon.

Thank you all.

Goodbye.

And this concludes today's conference. Thank you for participating. You may disconnect at this time, and have a great rest of your day.

Good afternoon, and welcome to the Equinix Third Quarter Earnings Conference Call. [Operator Instructions] Also, today's conference is being recorded. If anyone has any objections, please disconnect at this time.

I would now like to turn the call over to Phillip Konieczny, Senior Vice President of Finance. You may begin.

Good afternoon, and welcome to our third quarter conference call. Before we get started, I would like to remind everyone that some of the statements that we will be making today are forward-looking in nature and involve risks and uncertainties. Actual results may vary significantly from those statements and may be affected by the risks we identified in today's press release as well as those identified in our filings with the SEC, including our most recent Form 10-K filed on February 12, 2025, and on our most recent Form 10-Q.

Equinix assumes no obligation and does not intend to update or comment on forward-looking statements made on this call. In addition, in light of Regulation Fair Disclosure, it is our policy not to comment on its financial guidance during the quarter unless it is done through an explicit public disclosure. On today's conference call, we will provide non-GAAP measures, we provide a reconciliation of those measures to the most directly comparable GAAP measures and a list of the reasons why the company uses them in today's press release on the Equinix Investor Relations page at www.equinix.com.

We have made available on the IR page of our website a presentation designed to accompany this discussion, along with certain supplemental financial information and other data. We would also like to remind you that we post important information about Equinix on the IR page from time to time and encourage you to check our website regularly for the most currently available information. With us today are Adaire Fox-Martin, CEO and President; and Keith Taylor, Chief Financial Officer. Following our prepared remarks, we will be taking questions from sell-side analysts.

At this time, I'll turn the call over to Adaire.

Thank you, Phillip. Hello, everyone, and a very warm welcome to our Q3 2025 earnings call. Equinix delivered a very strong third quarter. Our performance that continues to demonstrate our ability to rapidly invest in significant expansion whilst growing our top line and improving profitability. This performance was underpinned by 3 highlights. First, top line growth. We are seeing continued revenue acceleration, delivering MRR growth of 8% year-over-year on a normalized and constant currency basis. Further, we also achieved record annualized gross bookings of $394 million, a meaningful 25% increase year-over-year and up 14% over Q2.

Importantly, this accelerated growth comes from a highly diversified set of customers across geographies, industries and segments. Second, profitability. We again delivered strong adjusted EBITDA margins for the quarter, and AFFO was up 12% year-over-year on a normalized and constant currency basis. This was better than expected and reflects strong flow-through of our operating results, favorable net interest expense and timing of recurring CapEx spend.

As a result, we are raising our adjusted EBITDA, AFFO and AFFO per share guidance for the full year. Third, expansion. Given the strong demand backdrop, we are advancing our build bolder strategic move where our intent is to double capacity by 2029. We have recently closed on substantial land acquisitions in our greater Amsterdam, Chicago, Johannesburg, London and Toronto metros. Which will support over 900 megawatts of retail and xScale capacity. These results indicate that our strategy is gaining even more traction and resonating with our customers. as we continue to deliver differentiated infrastructure, products and levels of service. On the topic of customer resonance, we achieved significant momentum in Q3. Closing over 4,400 deals with more than 3,400 customers. This volume reflects continued demand for a wide variety of latency-sensitive AI and non-AI workloads, supporting significantly increased data residency and sovereignty requirements and delivering seamless connectivity to distributed data sources. Our rich ecosystems continue to proliferate across a variety of sectors. Including key verticals such as automotive, financial services, networks as well as cloud and AI service providers. Hyundai Motor Group, for example, runs its proprietary [indiscernible] cloud platform at Equinix. Using Equinix Fabric, Hyundai connects to multiple cloud providers in Asia Pacific, the U.S. and EMEA. This enhances customer experience and improved service quality for over 10 million Hyundai connected car subscribers worldwide. Zetaris, an AI data lake house platform provider relocated its AI workloads to Equinix. Using Equinix's distributed AI infrastructure, Zetaris is helping its customers develop agentic AI and other AI applications 6x faster and at 1/3 of the cost. ING is making a strategic shift by migrating its core banking infrastructure in Germany to Equinix, showcasing our ability to help customers meet strict regulatory standards and requirements. Nitori, the largest furniture and home furnishing chain in Japan with over 1,000 stores across Asia and the U.S. partners with Equinix to connect its Osaka and Tokyo operations with low latency to Oracle Cloud.

This helps them simplify their network for future expansion. And supports Nitori's growth objectives of tripling their branches worldwide. In addition, we saw continued momentum with key AI-related magnets and enterprises, including Ally Bank, Bristol-Myers Squibb, Nebius and Groq, amongst others. As I've shared in previous earnings calls, our strategy comprises 3 strategic moves orchestrated across the business, to accelerate our expansion, innovation and profitable top line growth. We continue to deliver strong results and see accelerating momentum against each.

The first strategic move is SAR better. As evidenced by our recent customer wins, SAR Better is rooted in delivering value to customers at every stage of their engagement with us. Customers are increasingly looking to secure both their immediate and their long-term infrastructure requirements. This robust demand profile resulted in a record $394 million of annualized gross bookings in Q3. For clarity, this annualized gross bookings number represents the bookings we expect to start generating revenue within the next 90 days.

Additionally, we have a presold balance totaling $185 million of annualized gross bookings. This presell cumulative balance will start generating revenue beyond 90 days. As of yesterday, we have closed more than 40% of our Q4 bookings plan. We have ample pipeline to achieve our Q4 bookings targets and to build momentum heading into 2026. Our second strategic move, Solve Smarter, is focused on simplifying the consumption of our solutions and extending the value of our leading interconnection capabilities. Our interconnection products had an exceptional quarter. We added 7,100 net physical and virtual connections in Q3. We bringing our total to more than [ $499,000 ]. Interconnection revenue grew 8% year-over-year on a normalized and constant currency basis. To $422 million, driven partially by a 57% year-over-year increase in our fabric bookings in Q3. We also added 2 new native cloud on-ramps in Barcelona and Dubai, adding to our market-leading share of native private cloud on ramps. These results highlight the critical importance of low latency and proximity to end users and our ability to deliver it as both enterprises and service providers manage their distributed architectures.

In September, we unveiled our distributed AI infrastructure solution. This includes a new AI-ready networking backbone and fabric intelligence software designed to support enterprise inferencing workloads. We showcased these capabilities at our first AI Summit, together with key partners and industry leaders. Including NVIDIA, Dow, Groq, HPE, Adobe, Zayo Zoom and WWT. Our customers and partners provided use cases to highlight how Equinix is uniquely positioned to comprehensively deliver on their demands and requirements at enterprise level. And finally, Build Boulder. Through Bill Boulder, we are both accelerating and innovating the delivery of capacity around the world and securing our future through strategic land acquisitions. As mentioned earlier, we are excited to announce that we have recently closed on land acquisitions in several high demand markets to serve customer demand across both our retail and xScale businesses.

This brings our total developable vocasity to approximately 3 gigawatts and nearly 50% increase from last quarter. These are the latest steps towards doubling our available capacity in the next 5 years. Since our last earnings call, we added 7 new projects, including our Dallas 12 development. which is expected to deliver roughly 3,700 cabinets or approximately 67 megawatts of capacity to this key metro. We now have 58 major projects underway globally including 12 xScale projects. 20% of our retail capacity has been considerably accelerated from the initial delivery date. We also opened our 77 market in Chennai, India as we continue to invest in this fast-growing region.

More than 75% of our retail expansion is in major metros, and more than 90% of our expansion CapEx is on owned land or where we have long-term ground leases. Our stabilized cash-on-cash return expectations for these retail expansions are approximately 25%, which is consistent with our existing portfolio. Our North American JV continues to show exciting progress with the closing of our Chicago land acquisition, which we anticipate will be contributed in large part to our xScale business in 2026. In addition, we are in late-stage negotiations for the lease of the entire capacity at our Hampton campus with potential xScale customers. The overall demand picture for our xScale business remains robust. As key players continue to seek capacity in major metros aligning with our xScale strategy.

As our results show, we are consistently delivering on the immediate needs of our customers and the expectations of the market, whilst expanding our salable capacity in anticipation of even greater sustained long-term demand. And we are doing this very profitably. We have been built for this opportunity, and we will continue to build for it.

Let me now turn it over to Keith to share more on the quarter and our Q4 outlook.

Thanks, Adaire, and good afternoon to everyone. Further to Adaire's remarks, the business delivered a very strong third quarter, which immediately translated into solid financial and nonfinancial results. Nearly every key metric was at or better than expected. Our sales team has delivered a record annualized gross bookings across our diversified customer base, resulting in very healthy net bookings in Q3. To further demonstrate the strong momentum in the business, Adaire highlight that we now have $185 million of annualized presales, which will be recorded as bookings in future quarters. Over 40% of the current presales balance was signed in Q3. As customers increasingly look to secure their future growth within our ecosystems. And we continue to deliver accretive value to the bottom line with AFFO well ahead of our expectations for the quarter.

So our strong performance in Q3, coupled with our strategic efforts to continue to secure land for future growth in our major metros is setting the stage for 2026 and beyond. And given our balance sheet is a strategic differentiator, it provides us with the flexibility to invest into a robust demand backdrop and the financial capacity to secure our future energy needs. Finally, our relentless focus on best-in-class capital allocation for our investors while delivering durable long-term value remains a key priority for us as we execute against our Build Bolder initiatives. As it relates to our nonfinancial metrics, they continue to demonstrate positive momentum across nearly all dimensions, including net interconnection additions, provision BC capacity, pricing, volume of transactions and cabinets billing. We added a healthy 7,100 net interconnection ads in the quarter supported by cloud and enterprise connectivity. Cabinets billing stepped up 2,500 cabinets led by strength in the Americas region. With our record Q3 gross bookings performance, we expect our Cabinets billing metric to continue to be strong in Q4.

Our global MR per cabinet yield stepped up $41 quarter-over-quarter on a normalized and constant currency basis, primarily due to increasing densities strong interconnection and firm pricing across each of our regions, consistent with the broader supply-demand dynamics in the marketplace. As Adaire highlighted, we continue to make great progress with our Build Bolder strategy. Our recent land purchases will support more than 900 megawatts of incremental capacity across our full product continuum once built out, meaningfully increasing the capacity to be developed across our portfolio.

The capacity to be developed now stands at 3 gigawatts positioning the Equinix to serve the expanding market opportunity across hybrid and multi-cloud and AI. Our investments remain focused on either enhancing our strong existing ecosystems or laying the foundation for both current and future AI inferencing solutions, which will be built on top of our highly differentiated platform.

Now let me cover the highlights for the quarter as depicted on Slide 7. Do note that all growth rates in this section are on a normalized and constant currency basis. Global Q3 revenues were approximately $2.32 billion, up 5% over the same quarter last year. Our recurring revenue growth stepped up 8% and which is underpinned by the continued bookings momentum of the business. As expected, our nonrecurring revenues moderated sequentially, largely due to lower xScale fees. Q3 revenues net of our FX hedges included a $9 million FX headwind when compared to our prior guidance rates. Global Q3 adjusted EBITDA was $1.15 billion or approximately 50% of revenues, up 8% over the same quarter last year. Q3 adjusted EBITDA, net of our FX hedges, included a $4 million FX headwind when compared to our prior guidance rates. Global Q3 AFFO was $965 million, up 12% over the same quarter last year and meaningfully above our expectations due to strong operating performance, disciplined balance sheet management and timing of recurring CapEx spend.

Q3 AFFO included $2 million FX impact when compared to our prior guidance rates. As expected, global MRR churn in Q3 stepped down to 2.3%, and we expect Q4 MR churn to be within our 2% to 2.5% quarterly guidance range. And now looking at our capital structure. Please refer to Slide 10. Our balance sheet was approximately $38 billion, which included cash and short-term investments totaling $2.9 billion. Our cash and short-term investments stepped down from elevated levels in Q2 as our capital and real estate investments stepped up, and we repaid $1.2 billion of senior notes. Our net leverage was 3.6x our annualized adjusted EBITDA. During the quarter, we issued U.S. dollar equivalent $500 million in Singapore denominated green notes at a rate of 2.9%. We also published our green bond allocation and impact report in September. Equinix has now issued approximately $9.5 billion in green bonds with $7 billion in net proceeds allocated to eligible green projects.

Turning to Slide 11 for the quarter. Capital expenditures were approximately $1.14 billion, including a recurring CapEx of $64 million. We opened 8 major projects across 7 markets since our last earnings call adding retail capacity in key metros, including in London, Miami, Montreal and Washington, D.C. We opened 2 new data centers, one in Chennai, India, the other in Monterrey, Mexico. Revenues from owned assets are 69% of our recurring revenues. Now moving to Slide 12. Our capital investments continue to generate strong returns. Our now 180 stabilized assets increased revenues by 4% year-over-year on a constant currency basis and are collectively 82% utilized and generate a 26% cash-on-cash return on the gross PPE invested.

And finally, please refer to Slides 13 through 17 for an updated summary of 2025 guidance and bridges. Do note, all growth rates are on a normalized and constant currency basis. For the full year, we're maintaining our underlying revenue outlook with a 7% to 8% normalized and constant currency growth rate. Our expected quarter-over-quarter MRR step-up is greater than $60 million. A significant year-over-year increase, highlighting the underlying momentum in the business and gives us the confidence as we look into 2026. And as we previewed in July, our Q4 revenue guidance also includes a meaningful step-up in nonrecurring fees attributable to the xScale business. As Adaire mentioned, our discussions with potential xScale customers are in their advanced stages. But as with transactions of this size and complexity, the timing of contracting can be fluid. Hence, the expanded revenue guidance range.

Given our strong profitability in Q3, we're raising our underlying 2025 adjusted EBITDA guidance by another $21 million. Adjusted EBITDA margins are expected to range between 49% and 50% for the full year. We're also raising our underlying 2025 AFFO guidance by another $31 million. AFFO is expected to grow between 11% and 13%, while AFFO per share growth is expected to range between 8% and 10% compared to the previous year.

And finally, 2025 CapEx is now expected to range between $3.8 billion and $4.3 billion including approximately $290 million of recurring CapEx spend.

So I'm going to stop here. I will turn the call back to Adaire.

Thanks very much, Keith. So in closing, we remain excited and optimistic about the future and our differentiated and durable market position. We are focused on executing against our Q4 expectations. And building our momentum for 2026, and we are very much on track on both accounts. Our intent is to continue to build on the outcomes that defined this quarter. Greater capacity, increased revenue and improved profitability and accelerate them as our strategy achieves even greater traction. We were built for this moment. And I am confident we will continue to make the very most of this opportunity in regards to both the long-term growth and long-term value creation for our shareholders.

So I'll stop here and open it up to questions.

[Operator Instructions] Our first question comes from Nick Del Deo from MoffettNathanson.

You have a very strong position with cloud on ramps, and that's obviously helped point a lot of enterprise business over time. you've been starting to land some neo cloud on-ramps and network nodes. Adaire, I think you mentioned [indiscernible] In your prepared remarks. I guess how strategic do you think these deployments will be relative to some of the more traditional cloud on ramps? And what are you doing to attract -- to actively attract AI magnets like these?

Yes. Thank you very much for the question. Yes, you're right. We have a market-leading position in native cloud on-ramps, which is, I think, a very important part of the connectivity narrative that we have for our customers, and we're in a position to add 2 additional on-ramps this quarter to our installed base. We also, as you mentioned, have a very strong presence in terms of AI magnets sitting inside the Equinix ecosystem. Companies that I mentioned in my prepared remarks there like Zetaris [indiscernible] , who is a GPU as a service provider in Germany, Block [ rockwitht ] Q out rider Nebius [indiscernible] to name about a few. Many of these neo clouds are using Equinix as a point of connectivity and a point of presence. And I suspect there's an element of a traction in terms of our 10,000-plus enterprise customers who are also making use of neo clouds for data storage and for connectivity. So our team, particularly our team in the Americas focuses on this aspect of our customer cohort and manages and engages those relationships appropriately in order to ensure that we have the right magnet representation in our ecosystem going forward.

Next, we'll go to the line of Aryeh Klein from BMO Capital Markets.

On the strength in presale activity, I think you changed your approach with sales not that long ago, enable them to sell capacity for their outcome delivery. Is that some of what you're seeing helping to drive the strength there? And then maybe on a related basis, there's a lot of capacity that's set to come online in some of your most important markets. How much pre-leasing actually you're seeing for those?

Aryeh, do you mind -- you broke up there just when you asked...

First part of your question. Yes.

Could you repeat that, please?

Yes, sure, sorry. Just on the strength in presale activity, I think not that long ago, you changed the approach enabling your sales force to sell capacity further out from delivery. Is that some of what you're seeing helping to drive the strength there? And then just on a related basis, there's a lot of capacity set to come online in some of your most important markets. How much pre-leasing activity are you seeing for those?

Okay. Thanks very much for the question. For the repeat, I think that you saw in Q3 in additional in addition to the annualized gross bookings of the $394 million that the team delivered. We also shared the cumulative total presold balance of $185 million of annualized gross bookings, which will be recognized in future quarters. This presales motion is a relatively newer motion for our core retail business, and we have recently extended the window for our sales team to be able to sell retail capacity ahead of delivery for the next 12 months. Previous to that, it was a 3- to 6-month window. This presales opportunity is something that gives our sales team critical capacity to sell into. And it actually is a degree of comfort for our customers because it enables our customers to know where their deployments will be placed when they need them.

And I think in the overall macro environment, with the demand continuing to outpace supply, we actually have seen the velocity of presales increase over the course of 2025, through to Q3. I think Keith mentioned in his remarks that 40% of our total presales balance was signed in Q3 of 2025. So providing these 2 elements, I think provides greater visibility to our investor community. That being said, when we look at the presales activity fairly evenly spread across the capacity that's coming online. We certainly have seeing very significant activity around locations like Frankfurt, London and others where capacity is in short supply, New York, et cetera. Yes. So definitely seeing the uptick in this, which is why we decided to share that data point with you.

Aryeh, maybe I'll just add one other comment, on to what Adaire said, I think it's also important to realize that part of the reason that we've sort of entered into another sales motion as Adaire refers to is the fact that you've got a supply and demand environment that has shifted. And we're -- right now, we're chasing demand. We're trying as hard as we can to bring new supply into the business. And so we want to make sure that our sales organization has the ability to have that visibility. At the same time, having Ralph and his organization, Global Design and Construction, accelerate as fast as they can, some of the builds that were out there to a quarter or potentially 2 quarters sooner than we originally had planned.

So it's the combination of basically -- demand-rich environment, and also as chasing the ability to deliver capacity into the market as fast as we can that really allowed us to enter into the sort of presale arrangement -- and so as Adaire said, the combination of the $394 million, the $100 million, $185 million, it just gives you a real sense of how much momentum there is in the business relative to where we were, not only 2 quarters ago, but certainly last year.

The one other remark that I think it's important just to reiterate, the presale relates to our retail footprint. So not to the entirety of our business. When we look at our ex-scale business, That, of course, is -- we're looking at through the lens of preleasing. So presale is entirely within the retail business. Thanks, Aryeh.

Next, we'll go to the line of Eric Luebchow from Wells Fargo.

Great. I appreciate you guys taking the question. I just wanted to touch on -- what you're seeing in kind of the pricing environment today, you said you're doing more presales -- are you seeing firm pricing or improving pricing just based on some of the capacity constraints we see in the market? And maybe just if I could squeeze one more in, I know you guys have kind of preguided to about a 5% AFFO growth rate next year. I know there's a lot of moving parts as we think into next year, and obviously, interest rates, financial costs looking a little better than you guided to. So any thoughts to some of the moving parts as we kind of roll our models forward to 2026?

Okay. Maybe I'll take the second piece first, and then make some comments on the remarks. Big part on your questions around pricing. So I guess, as we look ahead into 2026, we're certainly focused on our execution in Q4. So revenue is a very prime focus for us, ensuring that we have a very strong exit from Q4. We're certainly feeling very confident about the demand that we're seeing. And I guess our presale and our Q3 gross bookings is evidence of that demand. And Keith also mentioned the ability of our amazing design and construction team to be able to accelerate forward RFS states. We are monitoring and forecasting RFS with the same intensity that we do for revenue. And we did see a 20% acceleration on the 58 projects that we have underway. So the first, I guess, 2 elements of our 2026 color is this focus on revenue, this focus on RFS acceleration date.

We will also continue to focus on cost and how we are operating the business from an efficiency and effectiveness perspective. In order to ensure that we're delivering a very strong operating performance. And I think you can already see some very positive trends there as it relates to that. And then, of course, there is the capital model and the astute management of our capital and our CapEx requirements, which Keith you may want to comment on?

Yes. And so Eric, the -- as it relates to the balance sheet, the one thing in addition to all the good news that sort of -- Adaire was sharing with you there. The other part of the story is slightly different than where we were at our Analyst Day is that a rate of -- the cost to borrow, not only presently, but as you look forward is lower today than it was. So that's the positive news. Add on to that, our ability to raise capital in this environment across different markets and really get a fairly effective low cost of borrow. And as we look forward, we're probably going to continue to as I mentioned maybe in some of the last calls, continue to look at markets like Canada, Europe, whether it's synthetic or otherwise, we're going to raise more capital. And so we'll continue to drive down our over cost to borrow.

The other thing I would say is we have the ability. Again, we run the business on a global basis, as you understand. And so when we repatriate the capital into the United States, we get a higher return on that capital. And so the ability to sort of manage our balance sheet really effectively as well as the cost line that goes through the P&L is we're quite effective at that. And maybe the last part I would say is, as we've started to increase our construction in progress, no surprise to you, we're continuing to look at how we capitalize the interest expense associated with those construction initiatives, whether it's the prebuy or it's the development of land and you can see that over -- not only this quarter, but as we look forward, we'll spend more and more on land in powered land, and so we'll continue to capitalize interest into those projects appropriate with the business.

All right. So maybe let me just come back on the pricing question. We're certainly not seeing any dilution in our pricing, very firm.

Our next question goes to the line of Jonathan Petersen from Jefferies.

Great. Thank you. You talked about the 900 megawatts of land acquisitions that you did in Amsterdam, Chicago, Johannesburg, London and Toronto. Can you give us just some more details on if any one or two of those sites are particularly larger than the others and where you might be expecting to build ex-scale versus retail within those markets on that new land?

Okay. Yes, thank you very much. We were very excited about those land acquisitions. We believe they are very meaningful and associated closely with the metros where we have a lot of demand from our customers. So as you know, these recent land acquisitions have really brought our land under control. to a very significant level, enabling us to -- actually since last quarter to increase land under control to nearly by nearly 50%. And -- in terms of how we view the various different elements of the portfolio here in alignment with the long-term capital investments that we detailed at Analyst Day, we do plan to double our overall capacity, both inclusive of retail and xScale by 2029. And whilst we're not providing a very specific breakdown of the anticipated megawatt delivery between retail and xScale within these recent land acquisitions, we do anticipate that a significant proportion of London and Chicago land purchases will be earmarked for xScale business and will be contributed to the JV for which we will be compensated.

And I think our global design and construction teams also remain highly focused on delivering critical capacity across the portfolio. And so to some extent, the split of megawatts between retail and xScale is somewhat fungible because of the full product continuum that Equinix offers across traditional retail, larger footprint retail and xScale. And so really, we're looking to maximize the value based on the opportunity that we see for each of those land acquisitions.

Next, we'll go to the line of Michael Elias from TD Securities.

I'm going to see if I can press my luck here. I was looking at the municipal filings for the Manuka campus site. And from what I found in the minutes of the meeting, I think there was a talk about a split for that Chicago site, some of it being for xScale, some of them being for retail. I'm just wondering if there's kind of any direction you could give us in terms of what you're leading is for that campus for retail versus xScale. And then also as part of that, just curious, you talked about signing a bank deal in this quarter. We're seeing some large bank deals out in the market. I'm curious if xScale, you're still thinking about reserving that for hyperscale customers or maybe you'd be willing to take on some large enterprise customers within that space?

All right. So 2 questions. Let me address them one at a time. So certainly, the concept of the mega campus considers the possibility of jointly co-locating xScale with retail on a single campus. And that is certainly something that we are actively reviewing and considering as we plan out our mega campuses and something that I think will be a value add to all the participants, whether they are a participant in the xScale ecosystem or in the broader Equinix ecosystem. As it relates to transactions with the banks, certainly, it was interesting to see financial services as a very dominant sector in our Q3 results. And emerging requirements, for example, in series of operation like EMEA, where the door regulation is requiring a different level of resilience for financial services organization. This is also something that's helping to augment the demand that we're seeing in the Financial Services segment.

Certainly, we believe that our capacity is somewhat fungible across our xScale and retail footprint. And if there was an opportunity to service a large footprint through an xScale capacity that was available, then we would certainly work with our partners to ensure that, that would be something that we could consider for that customer.

And Michael, maybe I'll add on just one other thing. You made a reference to Manuco, which obviously, many hundreds of megawatts of capacity in that market. As Adaire alluded to, we look at the test fits and understand what is the best structure for for both the partnership and for the retail business because quite overly, we want a certain amount of capacity to come into the retail portfolio because we need that capacity in some of these larger markets. And so we're going to continue to look at that. The team, as part of any of these large campus type builds. We'll look to see how to bifurcate the land so that the portion stays in the ex-scale structure and then a portion of it can be owned by Equinix directly.

But again, I'd just say appropriately, you could appreciate that there is some of these larger markets, and Adaire alluded to London London and Chicago. But the other one that's out there is Amsterdam, which is in the greater Amsterdam area and something that we really are putting a lot of attention to. So hopefully, that gives you a little bit more color on the large-size opportunities that are in front of us.

Next, we'll go to the line of Michael Rollins from Citi.

So a couple of questions. First, with respect to -- there, you just referenced the larger footprint retail. And from the Analyst Day, that was one of the incremental uses of capital and investment for Equinix. So curious what you're seeing on that front in terms of demand? And are those the types of deployments that you should also see some significant pre-leasing as you get closer to bringing them online commercially. And then just one clarification. Keith, you mentioned the the width of the revenue guidance. I think this year, for 4Q is $120 million versus $40 million when you're coming into the fourth quarter of [ '24 ]. And just kind of curious like if you could frame the nonrecurring revenue toggle of what if all of that opportunity for xScale bookings gets pushed into 2026 versus maybe being able to get all of what you'd like to see in the fourth quarter of '25?

Could I go first?

Yes.

Thanks, Mike. So as it relates to large footprint, we actually had a very healthy mix in our Q3 revenue mix of large footprint deals but also a very, very healthy retail and interconnection uptick in the quarter. As it relates to the applicability of large footprint to the presale sales motion. Certainly, we're seeing -- customers who require capacity that is contiguous as part of that presales process because that's one way of securing that contiguous capacity long term, particularly in high demand markets.

And then the -- the second question, which is we really highlighted on Page 15 of the earnings deck. And I know some of you might not have seen the deck yet, but we sort of give a breakdown of bridge on sort of what's happening in. Clearly, there's a meaningful step-up in revenue between Q3 and Q4, and you can see that at 7%, and that also translates into revenue -- sorry, EBITDA profit, 7% quarter-over-quarter growth. embedded in that, in the prepared remarks, we sort of noted that $60 million plus will come from recurring. Again, a very steep increase from where we historically have been on a recurring basis, and it's very reflective of the comments that Adaire made on the strength of our booking activity and how quickly that turns into sort of a billing line. And so that deal -- that's sort of the $60 million as part of it. But going from sort of the normalized Q3 to midpoint of Q4, it's $153 million step-up. And so it tells you that there's roughly $90 million nonrecurring. Of that, of that, I would say, roughly 1/3 of it is traditional nonrecurring activity, 2/3 is this large potential transaction that we have been working on for months.

And so part of the reason we expanded the range quite simply is we have a certain assumption that's based in the fourth quarter here. We want to live within the guidance range that we're guiding you to. Our intention is, and we are highly confident that we can close this transaction in this quarter, but we wanted to give you the range on what would happen if it didn't close this quarter and it closed in Q1, which is not our intention. So again, I think it's just an appropriate way to manage the -- again, the large nonrecurring activity. I might leave you with one other thought. Again, Adaire sort of alluded to it already, as we work with great confidence with, again, potential party to this transaction, it is for the full build -- it is for potentially the full build-out, and that's 240 megawatts. Those 240 megawatts are split into 4 buildings at 60 megawatts each. Embedded in this number is roughly half of that. So you can get a sense that to the extent that we do the entire transaction, that sort of puts us in a different part of the range than where we typically try and run at midpoint or better.

So hopefully, that gives you the color that you need. And -- and because it was a complex response, we probably would happy to let you ask a follow-up question if you need to.

That was super helpful color. So Keith, thank you so much. But since you're giving me the extra question, I have to take advantage of it. So just curious, one of the things I was looking at was when I look at the Slide 7 -- actually, sorry, 8 and 9, and I'm looking at the regional performance. And then looking at the normalized constant currency type of growth rates in each region relative to the EBITDA and trying to think about margins, what's happening in the segments where it looked like margins accelerated significantly in the Americas. You had less EBITDA growth in EMEA and kind of middle of the pack in APAC. And is there anything unique about this quarter in terms of the way investors should be thinking about operating leverage or just continuing to expand margins as revenue grows in each of these regions?

Yes. Thanks for the question, Mike. I think one, it's a very fair question and and one that we certainly want to talk about. Adaire alluded to earlier in our comments, in addition to driving the top line growth, which you're seeing, obviously, the magnitude of increase in our annualized gross bookings. In addition to that, you can throw on the presales, the business is performing really well. But we are actively going after the cost. And the thing that's unique about the Americas, remember, we break our business into 3 segments. And embedded in the Americas business is basically all the corporate SG&A, which, of course, is a lot of the SG&A line that sits on our financials. And so by attacking the cost line, you're seeing a lot of benefit coming into the Americas region for those reasons.

The second thing I'll talk about is you've got some one-offs. And like anything, when you have xScale and you have transactions and particularly last year in EMEA, there's some transactions in the xScale space that didn't repeat themselves and so the profitability shifts between the years. But I would say that just fundamentally, there's one-offs in EMEA this quarter to the tune of roughly $10 million. And if you compare it to the same quarter last year, there was a $10 million benefit in the quarter. And so you have a $20 million swing. And that has a big enough move in the EMEA business to cause some movement around the margin line. Fundamentally, you can see when you look at APAC, you look at Americas and ultimately, when you take out sort of the seasonal aspects of power costs and some of the one-offs from xScale, the fundamental business, the profitability is increasing in all 3 regions of the world. And that's something that, again, Adaire has been pushing us on across the markets and also the corporate functions to make sure that we're as judicious as possible with our spend.

Next, we'll go to the line of Jim Schneider from Goldman Sachs.

A little bit of a left field question. Given the 12 xScale projects you're working on, maybe give us a sense about your level of confidence of power availability and scheduling for those additional projects? I'm assuming you feel relatively confident, but maybe kind of give us a sense of the time frame for planning those. And then any changes you're seeing in terms of your anticipated use of power, whether it's grid power or other alternative sources for those?

Yes. All right. There's a lot to unpack in that question, because I think -- there's a lot to unpack in the power and energy narrative around the data center space. And there's no doubt that this is a complex area, and the complexity is certainly growing by the day. I think Equinix has a number of significant benefits here when it comes to navigating power as a constraint in the industry. First of all, your 27-year history. Which has a very clear profile of how we use power, how we minimize and optimize our use of energy and also the relationships that we've developed over that period of time with all of the utilities who provide us service and partner with us from the grid. And this is something I think that's held us in good stead as we are looking at this issue holistically across the business.

I think the second thing that we can see is that there is a change in terms of the customer supplier dynamics between the data center and large NG users and the utilities overall in that, in many cases, for many of these land acquisitions there is a contract that you enter into in order to secure the load ramp that you have identified for this acquisition, and that often requires a CapEx investment in order to secure that power commitment. I think as Keith mentioned earlier, we're extremely fortunate to have a very, very strong balance sheet and to be able to meet those requirements in a very differentiated way. When we look at the land acquisitions that we've made recently for the land that's under our control, either the power for these lands is either already fully committed or we're in very advanced stages of discussions with the power providers because as I'm sure you know that we're not in the business of speculatively buying -- speculatively buying land.

As it relates to our 12 xScale projects that are underway, all of our current 12 projects all have power secured. So power is not a constraint on the 12 xScale under development scenarios. I hope that gives you a picture of where we are in terms of the 12 xScale projects, but also the broader power continuum.

Next, we'll go to the line of David Guarino from Green Street.

A question on the annualized gross bookings and now the new metric, presold gross bookings, I'm admittedly still trying to wrap my head around what a normal run rate each quarter should look like. So -- was this a one-off quarter? Or is this the normal run rate we should expect going forward? And then similarly, on that comment, Adaire, just to clarify the comment you made earlier. There is no price difference for customers committing to capacity 12 months out versus committing today. Is that correct?

Yes. Maybe the first part of your question, I'll try to address. Look, I think when you look at our bookings history and of course, this is a relatively new measure that we revealed on Analyst Day for the first time. And we did provide a little historical perspective when we provided that number on Analyst Day. You can see that it is a relatively consistent growth story across the annualized gross bookings measure. But in fairness, it is also a measure that could be quite volatile depending on how a particular quarter would play out. I think as it relates to our position as we move into Q4 and executing now as we are on Q4, already 40% of our Q4 budget and target is already closed at this point of the quarter. And the pipeline that we have as it relates to Q4 is a very strong pipeline. And so our standard conversion rates applied against that pipeline would lead us to believe that we will meet our Q4 budget.

So I think for the period of time that we have shown this data you can see a relative degree of consistency in terms of growth over quarter-over-quarter, but I would make the point that bookings can be inherently volatile. And therefore, it's not possible to predict that bookings themselves will always be up and to the right even though we believe that the underpinning demand in the market and the momentum in the market is up and to the right.

And David, maybe just one sort of add-on to Adaire's comment. To the extent that we don't have capacity in certain markets, and as we talked about, there's -- presale isn't in that number, that [ $394 million ] and so to the extent that we say we run out of capacity and given a sort of highly sought aftermarket, you might see a scenario where presale activity moves. So as Adaire said, you can have some volatility and we used to have more seasonal volatility. But because of the supply-demand dynamics, I think that's very different today. But that also presents a scenario where you could have maybe more presale activity than you would have basically annualized gross bookings. So there's going to be these trade-offs. And again, we don't -- can't give you a number right now, but I think to the extent that we see things happening, we're going to guide you to it. I'm going to say, "Hey, we're out of capacity in these 5 markets, we anticipate more more presales than we'd say annualized gross bookings, which turn into revenues much sooner than a presale item.

Next, we'll go to the line of Frank Louthan from Raymond James.

Talk to us about the uptick in the cross-connect revenue and the ARPU there. What's driving that? Is that more power densities or customer mix? And is that strength you're seeing in the Americas replicable in other markets and other areas of the world? Is that sort of a is that agreeing to see the lead there in the Americas?

Yes. Thank you. Certainly, the Americas is the leader in terms of the demand that we're seeing for our interconnection portfolio. We added 7,100 physical and virtual interconnects. And as a result of that, our interconnection revenue grew 8% year-on-year. The cohort of customers, you can see the technology and infrastructure providers, the cloud being the ones who are driving that demand, primarily in the Americas business. And I guess, as we start to see those organizations and segments proliferate into other markets, then we will see that continue to evolve and grow in other theaters of operation. But you're quite correct in your assumption that this was largely driven by our Americas business this year -- this quarter rather.

And was it power densities or customer mix? Or what's sort of driving that?

Customer mix -- customer mix.

And for our final question, we'll go to the line of Michael Funk from Bank of America.

Yes. Thank you for the question. So First for you, Keith. I heard the comment on the call about accelerating some of the builds and also your comment about capitalizing some of that expense. But does any of that change your thought or outlook on the level of AFFO or the shape of AFFO growth that you laid at Analyst Day? And then second, for you, Adaire, putting your tech hat back on, as you think about distributed AI infrastructure solution and other emerging solutions at Equinix, can you size the addressable market for us for those and how you think about it?

Yes. Why don't I take the first question, then we'll pass it back to Adaire. So as I said in my sort of -- in response to one of the earlier questions, the company has had the ability to raise capital at a lower rate than originally planned, even though the majority of it, we're going to do a lot of refinancing in '26, '27 and '28. The underlying assumptions have changed as well. And I would say, I don't want to take this as an opportunity to shift guide on, as you know, we'll spend time in February on what we're going to do for 2026. But it's fair to say that our cost of [ barrel ] is lower. We're getting good return on the cash that's sitting on the balance sheet, and we are capitalizing a little bit more largely because we have more construction in progress. And as a result, you should expect capitalized interest increase a little bit.

To give you a perspective in Q2, we capitalized roughly $14 million, if I got my numbers right, $14 million. In Q3, we capitalized $27 million. And then in Q4, the number will be somewhere between $20 million and $30 million of capitalized interest. And so that's a little bit higher than what it was before. But we're spending faster. We're accelerating faster. Again, we really ask Ralph to Raj and the teams to to really build as fast as they can. And so with the acceleration of the capital spend, you would see more capitalized interest into the balance sheet.

As it relates to the product portfolio, certainly, you can see this increase in our interconnection revenue up to $422 million in the quarter. And a very specific focus on fabric with a 57% year-over-year increase and actually continuing to provision at record levels, are now up to 110 terabits in terms of fabric provisioning. So I think all of this speaks to the connectivity narrative and opportunity that Equinix represents and the range of services that Equinix can provide to customers who are considering AI and non-AI-orientated workloads. The connectivity of the company is certainly the secret sauce and I do feel that there are opportunities for us to look at potential for this aspect of our product portfolio.

But I would say that connectivity is not just the only dimension around which people are considering Equinix when they look at solving for deploying inferencing and inferencing base workloads at Equinix. Latency is one dimension, but there are others around model provider flexibility around edge processing so that you're reducing the cost of bringing data back to the center. Around data residency and compliance so that sensitive data is being processed in [ situ ] and also around intellectual property protection. And all of these elements are considerations around the platform that is Platform Equinix when our customers consider workload deployment and workload placement with us.

So I think this represents a very viable opportunity for us to continue to evolve and grow.

Thank you, everybody, for joining the call. and go ahead. Yes. Thanks, everybody, for joining the call. Have a great afternoon, everybody.

Thank you all for participating in the Equinix Third Quarter earnings conference call. That concludes today's conference. Please disconnect at this time, and have a great rest of your day.

I'm Jon Atkin with RBC. Pleased to be spending the next 20 minutes on the fireside chat with Keith Taylor, Chief Financial Officer of Equinix. Welcome, Keith.

Thanks, Jon. I'm going to be making some forward-looking statements in all likelihood. So please do refer to our SEC documents. Now you can go.

Well done. So maybe just to kind of set the stage because there's a mix of folks in the audience, and you did have your Investor Day in June where you laid out kind of multiyear guidance. So maybe just to level set, remind us kind of what you see yourself doing over the next several years around top line EBITDA, AFFO and so forth?

Yes. So we did -- June 25, we had the Analyst Day. I think many of you are aware of it. Really, it was a story about growth. Investing for the future, taking up our CapEx, roughly $1 billion a year for the next 5 years, to $20 billion to $25 billion of investment over -- through 2029. Obviously, we felt revenue continue to grow with an aspiration to get to double-digit revenue growth.

And then AFFO. The underlying business is going to grow healthily. But with refinancing and the raising of additional capital, we have a little sort of -- we have a little bit of an investment horizon in 2026, and that sort of 5% to 9% growth. And then the dividend would continue to accelerate.

And I thought -- one of the things, Jon, I know you didn't ask for this, but I'm going to give to you anyway. I thought it was a very forward-leaning sort of investor-focused sort of presentation because I also said we're going to only raise debt. We're going to use our balance sheet and the cash that we generate, plus debt, and we're going to invest in this growth. And I think all of you -- or many of you probably already met Adaire. And if you haven't, she's a force to be reckoned with. Great lady, great CEO. And her aspiration is to grow as fast as we can.

But the challenge that we have is the time line to grow the business because it takes 2, 3, in some cases, 4 years to develop the capacity, the time horizon of when you invest versus when you realize the fruits of that investment are beyond the planning cycle. And so our aspiration is to grow as fast as we can with a margin that will be 52% or plus over that time period, by the time we get to 2029.

So your global. Have been for quite some time. A lot of it organic, some of it inorganic. But as you look at kind of the pie chart of Americas versus EMEA versus APAC, where you're allocating capital? Where you see the kind of the demand opportunities? What's different compared to today around geographic mix?

Well, I think probably for many in this room, I mean, the opportunity is immense around the globe. And we're investing heavily in the United States. But we also have some very critical markets around the globe in Asia and in Europe. xScale is a little bit different. At least xScale, our 2.0 exercise, is first focused on the U.S., but you'll see us talk about major metro -- or sorry, larger campuses in Asia and Europe, hopefully in the not-too-distant future.

But that's the appetite. It is a global opportunity. The markets are growing. We tend to be a little bit different than a lot of the other players, largely because we -- if it's not xScale, we're selling 4,000 transactions a quarter to 3,000 customers, or something of that order of magnitude depending on the quarter.

Where xScale is you're selling to a large hyperscaler and a lot of people are chasing those opportunities. We've got one project that's currently underway that quite openly when you step back, we thought it was sizable at the time, but it's really 240 megawatts. Seems like it's not that significant anymore.

Having said all of that, it does feel like it's going to be a really good market to build in, which is outside Atlanta. We're already preparing the land, and we're going to get the buildings up as fast as we can because the appetite for that type of offering is very real. But let me just stop there and make sure I answered your question.

Yes. So kind of still global and maybe opportunity driven in terms of...

I think it's opportunity-driven, but there's markets around the globe, as we all know. I think all of you know is they're very difficult to build in. And so you have to recognize that, that is the case, and Singapore is a perfect example of that. And we're building out our Singapore 6 asset, which is 20 megawatts, but we want more capacity in Singapore. That's almost SGD 1 billion market for us. And so because you can't build maybe at the rate and speed you want to, we're also building in Johor, in Kuala Lumpur in Malaysia. We just opened up our Indonesian site. We just opened up Chenai. We just bought a business in the Philippines. We're preparing land in Thailand, in Bangkok. So there's a number of things that we're doing that I think really matter to that part of the world.

We're investing heavily in Hong Kong. Although there's one building that we -- at the end of last year, we said that we're going to take a charge against. It was more because it was an old building that came with an acquisition, but we're investing heavily in the Hong Kong market because we think it's really important. Largely because Chinese companies coming to the Greater Bay Area of China, which includes Hong Kong. We think that's going to be an entree from Chinese companies coming out of Greater China into Southeast Asia and other parts of the world. And so we remain very excited about these -- the markets in Asia, and equally so in Europe.

And then there's emerging markets that we think are really attractive. And 1 day, I'd love to say that we have an asset in Riyadh. We're building out in -- where we have business in Abu Dhabi, in Dubai, in Oman, in Muscat, in Salalah. And so we're going to continue to invest in that part of the world. And then the U.S. is, I think, I mean, it's just a great place to build right now. And so it's an exciting time for the industry.

So you talked a little bit about growth indirectly through, say, tuck-ins or new market entry. There's also just pricing. There's less churn. Churn can be maybe at the lower end of the range. There could be revenue optimization. There could be head count in terms of solutions engineers and quota-bearing headcount.

How do you kind of view that all of those sort of things trending because they all do influence the top line as well as EBITDA? So a lot to unpack there. So start with sales...

We only have 20 minutes, so we're going to throw all in as much as we can anyway.

No, I think there's different aspects of it. Certainly, there's a volume play. And I think many of you probably feel that demand isn't an issue. And the real issue is supply. And so getting the supply into the market. So that's number one. So I think if we have the inventory, we'll be able to sell it, particularly in the markets that are very rich with opportunity.

Pricing is firm. Firm is maybe an understatement. Depending on the market you're in, pricing is strong. But then there's markets that tend to be a little bit softer based on the competitive dynamics and the supply coming into the market. And so each market has to be analyzed to optimize against the opportunity.

Having said all of that, then you have you have the ability to operate quite efficiently at the asset level, but then you've got all the corporate costs and the borrowing cost. And I'd say we're managing our -- we're managing our cost model very effectively. We're investing heavily in Harmeen, who is our new CDIO, and AI and all things system-oriented. That includes processes, and that will take costs out of our equation.

We've got a new CCRO in Shane Paladin. I think he has go-to-market strategy and all the things that we do to create demand and keep customers insight. Our assets are going to be important, including the AI initiatives that we put into understanding the predictive analytics on what's going to drive outcomes. And so between investing in those areas that we can drive down, I think, our cost to operate. And then we're going to invest heavily in our go-to-market, the front end of the customer-facing initiatives.

And at the same time, I said at Analyst Day that we're going to continue to raise capital, is going to be debt oriented. We're refinancing what we have. But the capital that we raised -- the last deal that we just did a couple of weeks ago was in Singapore. Again, USD 500 million equivalent at 2.9%. So well below the average rate that I assumed at 4.9% at the Analyst Day.

And since Analyst Day, we almost had a 100 basis point move in the 10-year treasury. And so we're going to be able to raise capital eventually in the United States, but the next port of call is going to be probably in North America, not in the U.S. And then we're going to go back to Europe because I think the opportunity is right. And so we'll be able to drive that. I think that cost to borrow further than -- probably lower than where we were guiding to, largely because the market, as I said at that time. If the markets allow or permit for it, will borrow money cheaper.

And then we're really undertaking a very strict review on capitalization because the market had asked us to do that, or the investors did. And so we're doing that as a team and looking at that. So a number of things are working very much in our favor.

Maybe the last thing I would say, and I've said it to some of the people in the audience already. When the market speaks, whether or not you agree with our posture, the market has spoken. So we need to respond to the market conditions and what our investors are expecting of us. And Adaire is -- and Adaire and I are very, very much aligned on what we need to deliver for 2026, which was the low end of that range. And let me just leave it that. The message was delivered, and so we will respond accordingly.

And then customer retention, a couple of words on that?

Well, churn for us, it ebbs and flows. I've said a couple of things with our investors today. But suffice it to say, I would love to see us get our churn rates down. Largely, as your base gets -- starts to increase, keeping a churn rate of 2% to 2.5% is generally in that range. It's hard as the base gets bigger and bigger and bigger. And so our goal is to find ways to get that -- our churn percent down. But right now, I'd say we're not shifting our thoughts on that right now.

And I think the work that we're doing on our analytics and reshaping our go-to-market positioning, we -- just to give everybody a sense, our entire go-to-market, plus the support functions, are roughly about 3,000 employees -- 2,500 to 3,000 employees. And so you want to make sure everybody is selling from the right vantage point. And I think the -- obviously, there's work that we can do to make sure that we do better.

And then at the same time, work on the churn analytics and get the predictive behaviors of our customers analyze through an AI lens that will help us get ahead of it. And I think that's going to be an opportunity for us as well. Maybe not over the next couple of quarters. But as we look into '26 and beyond, I think it's going to -- it will pay dividends for us.

So as more workloads become associated with AI, whether it's from hyperscale customers that also reside in your data centers or enterprise. How does the latency -- the traditional kind of latency value prop of Equinix play into what you see as coming -- the coming wave of AI inference workloads?

Yes. I -- Look, I'd like -- I don't think I'm wrong in this assumption. I liken it very much to the advent of cloud and just where cloud was going. A lot of compute was done in locations. But at some point, you need to aggregate or deliver an availability zone, or aggregation node, or network node. And I think that holds true for AI as well.

The models will get trained. And whether they're sovereign, or foundational, what frontier oriented models, they will continue to develop over time across different markets, and across different regions of the world. They will eventually need to converge and inference will become a bigger part of it. And I thought we're pretty clear that we'll be the beneficiary of inference, less about training and in sovereign models.

It doesn't mean that we won't have -- we won't create value from that. I just think that's not really where our sweet spot is. And so having the diversity of network, the diversity -- pardon me, the diversity of cloud on-ramp, the cable landing stations, both origination and termination. These all play into a strategy and then you can get into satellite and the like that I think we're going to be the best representation of that opportunity as we continue to work with customers to distribute the AI opportunities.

And let me give you another example. I mean you've heard us talk about it. We're really excited about just what Groq's doing with us around the globe. And that's just -- that's the front edge of these opportunities. And if they get it right, then we got it right, because that is where inference will start to accelerate and there's more and more companies that will start to look more like them. Not to mention there's the GPU as a service, the private AI cloud. They're all the service providers that are going to have AI-oriented solutions. And it's the combination of all of those that I think will make a difference for Equinix because we are low latency. Because we have the on-ramps. We have the networks. We have the latency sensitive environments.

So it sounds like in the early innings of this journey, it's identifiable inference because Groq, G-R-O-Q, obviously, that's their business model. Inference as a Service. And you've got tech companies that are driving this.

Where is the enterprise, because you mentioned the analogy with cloud? So as enterprises start to do inference within their own environments, do you have a view yet as to how soon that might happen? Early days? Any early kind of reference cases that you can cite?

I think we feel 2027 is going to be a real turning point for the industry as it relates to inference. That all said, we have reference points whether -- it's the large pharma bioscience companies that are doing local AI, sort of AI, private clouds that are very much used GPU clusters to optimize basically research, or development and output. So you see that to other companies that are in the services space that are selling basically their services to enterprises.

Again, that's the front edge of, I think, the beginning of opportunity. You're starting to see liquid cooling applications go into our environment. We're liquid cool available, but you want to see the consumption of that type of architecture because that tells you the realness of what's coming into play.

And then the other part, I would say that not all -- not all -- not at all AI -- let me say this differently. Not all AI deployments need to be liquid cooled. A lot of it can be air-cooled depending on the concentration. And that's, again, an indication that we cater to a wide array of customers with different parameters that need to live in a co-located, or multi-tenant data center environment.

We're not selling to one customer that architects itself uniquely, that has to compete with everybody else in the space. And so that's why we have a different offering inside the retail space versus the xScale space, which creates a host of other issues.

So on xScale, you've got a lot of deep-pocketed competitors that are willing to take leverage higher that might be willing to build even bigger than, say, 240 megawatts.

So how do you sort of see the competitive landscape? Is demand such that you don't focus so much on the competition? Or is it a factor that you have to kind of contemplate?

Well, we -- I think it's fair to say, I mean, we're a public company, and so you have to live within the confines of what a public company can and kind of not do. When you have a shareholder in a private equity environment that is willing to take their tolerance up to 10x or 15x, that's okay because it's short-lived and they can monetize off of that. We can't do that in the public environment.

And so we focus on what we can do and so that's the comfort of realizing that we can't compete maybe head-to-head, but then we have our joint ventures and we have really strong partners. GIC is an awesome partner. CPP is an awesome partner. PGIM is an awesome partner. And there are other partners out there that we can potentially do business with. So it's the recognition that we will need to partner with somebody, to make sure that we can accelerate the growth opportunity in front of us.

And we're already between Hampton and then the other sites we're talking about. And they're only -- I'm only talking about three sites. That would be -- that in itself if we execute against that over the not-too-distant, that will deliver over 1 gigawatt of capacity. And that -- and we're only scratching the surface as a company.

So I think you're going to see Equinix start to accelerate. Largely because of how we're investing and what we're doing to shift the emphasis in the xScale environment.

Time for one quick question if there is one from the audience.

[indiscernible]

Yes, we are hosting AI Summit really to provide -- I was saying -- some investors, the community, the potential customers, the hyperscalers, the opportunity to understand how we're going to approach the AI initiative, and why we think we're going to be very relevant as we look forward in the future. And so some about technology, some about understanding just what we're doing and where we're doing it, and then also just the depth of our team.

Question at the end? Yes.

[indiscernible]

Well, the beauty is that we have the...

So question was about older carrier hotel assets, and how you future-proof yourself?

Yes. Look, I think, number one, those networks aren't going away. Two, we have Fabric, which basically connects all the sites together as if they're one. Three, the density factor in a carrier hotel is very different than basically a high compute environment. And so we have capacity that's available. To the extent we don't, we're building around it, and making sure that we connect it, connect the assets.

But simply put, we typically have built them in campuses, or approximate to a campus. And so we have the ability to grow and scale. And you don't need to, per se, future-proof it in the sense that you sell the right customer with the right application at the right center, and you're not going to put a large AI, or a cloud-oriented workload in a network data center. And so we have the ability to continue to sell. And our goal is looking for ways to augment those carrier hotels where we can.

So you mentioned Fabric. It's obviously not a very capital-intensive product. Just within that segment then, is it growing at kind of a steady pace? Or is there any reason to think that the growth rate around Fabric would either accelerate or decelerate?

Well, I think -- I don't think you should decelerate, but I think it is a steady rate, but we want to see it accelerate. And so that's part of our goal for '26.

Very good. Well, thanks for your time.

Thank you.

Kick it off now it's 12:45. So I'm Mike Funk, I head up the telecom, comm infrastructure, comm software research here at the bank. So great to see you all again. Really happy to have Equinix with us once again for our Global REIT conference. So thank you, Stu. Thank you, Katie. So Stu is SVP of Global Real Estate at Equinix. Is that right?

Yes, that's correct.

And Katie is from Investor Relations. Katie does have disclaimers. She wanted to read quickly. Just to cover the basis, then we'll get right into the Q&A. So thank you, Katie.

Thanks, Michael, for having us. Some of what we may talk about today is forward-looking in nature, so please check out our SEC filings for all of our risk disclosures. Thank you.

Great. Thank you for that, Katie. Maybe to set the stage, Stu. What do you do on a day-to-day basis in Equinix? What is the Global Head of Real Estate do at Equinix?

Yes. Thanks, Mike. Just -- I've been with the company for almost 14 years now, largely with corporate development and real estate for the whole time. My previous role, just as a way to set it up to answer your question. My previous role was running all corporate development for our EMEA region and before that for the Americas region. And real estate was embedded inside those corporate development teams.

What our leadership team has done is carve out real estate specifically from corporate development, make it its own global organization and then asked me to run that globally. So day-to-day, we're looking at how we execute our Build Bolder strategy, which is really how do we locate land and power in the markets we need our customers want to be. And just given that's the building blocks of our product offering, it's really, really critical. So I manage the real estate teams. We're looking at these transactions from, I guess, Tokyo to Montreal and everywhere in between.

And I do want to come to the Build Bolder strategy later because there are a lot of questions about the strategy but also kind of the market and where the market is going today. So I'm going to put a pin in that for a minute and maybe start higher level.

A key theme right now for data centers is obviously the demand environment, power availability and what operators like yourself are seeing in the current state of the market, whether it's on the demand side, the ability to access power to meet that demand. So maybe to set the stage with your broader views on demand, power availability and how you're meeting that?

Yes. Look, demand signals are good. We just -- at our Q2 announcement, Katie can give you sort of the data specifically around that. But we closed over 4,100 deals over 3,300 individual customers, $345 billion of bookings. And so just from those numbers, I mean, you can tell the demand side has been really strong. We continue to get really strong demand signals from our customer base. We have 10,000-plus customers across all of our 3 regions. And so yes, we are still seeing very, very strong customer demand.

And then, Mike, you're asking how does that translate into the pursuit of land and power. Look, it's definitely a challenging environment, but we've been looking at a power first strategy for our land strategy for about 3 or 4 years now. And so really, we're looking at acquiring land and we're acquiring power and solutioning the land around the power, power being the key ingredient.

And look, I think the good thing is, we've been in business for a long time. We've been in these markets for a long time. We've got good relationships with individual municipalities, the individual power companies, they know us really well. And so we've been able to leverage those relationships and make sure that we're doing the right land deals in the right spots for our customers' growth.

And then another important piece recently been pricing and re-leasing. And at least in the hyperscale side, the wholesale side, we've seen very strong re-leasing spreads. Some carriers we're talking about, 10%-plus re-leasing, [ active ] is like about 5%. What are you seeing for pricing dynamics and colocation in the U.S. international?

Katie, do you want to take this one?

Yes, happy to. I'd say we continue to see a healthy and firm pricing environment as measured by our MRR per cabinet yield. So we're able to generate across the portfolio. I think when we take a lens of pricing, we're always very thoughtful in terms of approaching pricing with customers, just recognizing that we are the premium price provider in the marketplace, but feel like we have market permission to do so based on the superior value that we delivered to our customers.

But also recognizing it's not, for -- as Stu was just talking about, it's not getting easier to bring that capacity online. So we want to make sure we're earning an appropriate return on our capital in terms of how we're approaching pricing with customers.

And our xScale portfolio, those are generally longer leases, right? So we don't have as much re-leasing risk given our xScale program is not that old.

Okay. That makes a lot of sense. I think in recent probably 12 or 6 months, there's some question about where demand is flowing specifically on the hyperscale would be xScale for you. We're seeing builds from AI companies out in place like Lubbock, Texas and Kettering, Ohio or places like that, nontraditional data center market. So can you just talk a bit about why or if location is still important in data centers. That was really the hallmark of the Equinix business going back a long time.

Yes, for sure. Look, we continue to believe that Equinix is playing the best hand in the industry. We've been focused on location for a long time. We've been focused on interconnection-rich locations and really curating a broad portfolio of customers so that we can build an ecosystem that's got durable value, not just for our shareholders but also for the customers, right? They want to be close to other parts of the ecosystem, other customers want to interconnect with them. And so that's still going to be a focus of ours.

We've looked at our xScale business as additional to our retail business. And we don't see us getting out too far out of these core markets, so that we can still take advantage of our interconnection and really create that durable franchise going forward. So we might not be in Lubbock, Texas, for example, but we still want to be close to core markets where we can connect back easily to our ecosystems.

Foundation model training is the first step for AI and then there's been expectations that inference is going to pick up the slack or at least be additive, right, in terms of data center demand, I don't think we've seen that shift yet, but part of the investment thesis for Equinix was that given the quality and location of your data centers that you would capture a large portion of that inference demand. I think it's a more robust debate today, if that's true or where the inference is actually going to sit. Can you give us your current view on why Equinix is well positioned to capture inference? And then what indications that you might be seeing that would support that?

Yes, we'll give Katie chance as well after I'll say a couple of words. Look, it's early days in this adoption. We liken it to the cloud adoption, that happened a decade plus ago. Again, early days, but certainly, inference is what we're focused on. That largely is latency sensitive. Those applications need to be relatively close to the eyeballs. You saw a similar thing when it came to some of these connectivity nodes and cloud on-ramps with the cloud deployments.

Just as an example, we have over 35% of the cloud on-ramps in the markets that we're in, and we're clearly the #1 provider there. And we intend to do that same model, repeat that same model for AI inference. And the inference is going to be latency sensitive, no question about it. You've got to have these models perform for the end customer and for the end user. And so we think we're incredibly well positioned with our current portfolio and the portfolio we continue to develop to be successful.

Yes. I'd just add on Stu, when you think about the enterprise, particularly in terms of inferencing like every enterprise today, like if I ask all you guys in the room here to like raise your hand like everyone is using some form of like agenetic AI, but really, enterprises are still trying to figure out how to incorporate AI into their broader just operations overall. And so our customers and if you ask anyone on your sales team like they will tell you it's part of every sales conversation they're having today, like enterprise is trying to figure out okay, like what data do I need for AI? Is the data structured, properly cleaned, is it in the right form? And then like how can I incorporate that in my operations?

And the key question everyone is trying to answer is how I'm going to monetize or generate a return on this investment over the longer term. So feel like we can help customers with that and the attributes that we've built over the 27-year history of our ecosystems of the cloud on-ramps, the networks that reside inside of Equinix feels like that gives us the opportunity to pursue that -- to pursue.

Yes. The last thing I'll say on this, Mike, is that our global reach is a real differentiator for us, right? We're in 70-plus markets. If you look at the last place that I was in, was EMEA, we've got 30-plus markets there. GDPR is really important. Where the data resides is important. And a company like Equinix, we have a natural advantage because we've really invested in our global reach, right, and global sales team, and we can operate in all of the markets that our customers want us to be in. And we continue to push out that global reach.

So we continue to think location is really important. Our customers are still going to drive where we end up being. We've announced new locations in Manila recently as well. So we'll continue to push that global reach out. It's been a real differentiator for us.

Can you translate all of that for me into the income statement. So how demand from inference is going to drive accelerating top line growth or even accelerating AFFO growth in your existing facilities, your same-store growth, for example, your occupancy is relatively high, right? I don't think you've seen a lot of increase in price in those. So is it primarily through taking more price from inference customers. So that requires some churning of legacy customers. What is your capacity to add capacity to legacy facilities that are bringing in more power or space. So how do you, I guess, translate the inference demand that you're talking about into greater top line growth, AFFO growth in the existing facilities.

Yes. I'll let Katie jump on that first question. I'll come back to your how do we look at existing facilities?

Yes. I would say to start, thanks Stu. I would say, overall, when you think about the inferencing opportunities, it continue to add to the ecosystem overall. And so in many respects, like we see AI as another ecosystem to cultivate inside of the Equinix platform and making sure we're cultivating those highly magnetic customers in terms of their networking nodes or things of that nature inside of our facilities.

And part of -- and you heard from this -- from us at Analyst Day back in June is as you think about our Build Bolder strategy that we've talked about, really is in response to customer needs. And as Stu mentioned at the top, like we have such strong customer diversity that we're not beholden to just one technology trend. So, yes, AI is very top of mind and very exciting, but there's also broader digital transformation. There's hybrid and multi-cloud happening inside of our facilities. And so it's just a continuation of building out those ecosystems and customer deployments to making sure we have the capacity available across the portfolio when and where we need it.

Yes. And just on existing data centers, we spend 3-ish-percent of revenue on our existing data centers to improve them to replace aging equipment, about 1% of our CapEx goes to that. But really, we're focused on developing the next generation of data centers to further our AI strategy. It's going to require a lot of capacity. And certainly, you've seen us announce quite a big expansion into our CapEx program, about $3 billion to $5 billion, right, annually is what we've announced.

[ Annually ] correct?

No, in terms of the CapEx that we outlined at Analyst Day back in June, the $4 billion to $5 billion, it's roughly $3 billion to $4 billion per year of CapEx to support our retail business with about 80% of that going towards expanding and bringing on new capacity and then roughly about $1 billion per year of land purchases and pro rata share of contributions to the xScale joint ventures.

Which is about 20% only, right? So a lot of this is we are still seeing -- and I think the message you should take away here, Mike, is that we are still very bullish on our core value prop, which is retail colocation, right? A lot of that capital is going to our retail business. The primary share of that capital is going to our retail business, which we still see a ton of growth there.

Does this shift probably more to your primary role. How do you evaluate for site selection, procurement? How do you manage these different pieces of your role and then selecting a slight negotiation. How do you manage to make sure that you have enough equipment on the procurement side that your build times are on time? And how do you manage all that?

That's a great question. Look, let me highlight a couple of teams that don't report to me, but I work very closely with. One is the public policy team. They're great. They work hand-in-hand with the real estate team to create -- as well as the local leadership on the ground. We have great managing directors in every one of the countries that we're in. And we all work hand-in-hand to work with local authorities, to work with the local power providers to make sure we've got those really good relationships so that we can continue to grow in those markets.

And the other team I'd like to highlight is our procurement team. I think we've got 1.4 gigawatts under control of forward equipment, right? We're pre-buying cooling and power equipment so that we can continue to deliver these sites on time in a predictable way for our customers, which is really, really critical.

So I mean, look, it's never been more complex than right now. But I think Equinix has great talent that we've grown up in the business as well as folks coming in that really know how to operate in a complex environment, really scale the business. And things like our prebuy program have been great for us. Again, like having control over 1.4 gigawatts of capacity just in the kit you're going to deploy, I think, is a really powerful message. Anything to add Katie?

I would just add on to, we also have a pretty holistic power team. So thinking just about power availability to make sure we have the power needed across the portfolio. We also have a dedicated team thinking about our sustainability efforts and renewable energy coverage, recognizing is very important to our customers and the communities in which we operate. And we also have a team that's focused just solely on like our power purchasing as well. And so having just that in-house expertise that we've built out over the course of our 27-year history continues to be a differentiator for us.

It actually blends on to my next question, which was how far before you break ground, you have to actually secure power? I don't know if either one of you have a sense of the sequencing of the timing around securing power versus breaking ground of the facility?

That's a good question. I don't know if we'd look at it exactly that way, Mike, but let me go back to what I said earlier, which is we're securing power first. In some markets, you can secure power. Independent land, a lot of loan markets you cannot, right? So we're locating where the power is and then solutioning the land around that so that when we come out of the ground with a project, we have the power solutions already. And that's really, really critical for us before we start breaking ground for a foundation. We know where that power is coming from. We've got it secured and that we're coming out of the ground with a project that can be powered.

And can you just remind me how much power you've actually secured today? How much you have contracted?

I don't know if I've got that.

In terms of what we shared at Analyst Day is through from a utility procurement perspective, there's roughly 1 gigawatt secured and then another 2 gigawatts either applications in progress that we've submitted. And then the team is continuing to pursue other additional opportunities behind that.

Okay. And then Katie, you mentioned it earlier, and Equinix has always been at the forefront of alternative energy and trying to go down that path. Can you just remind me how much of your power you're currently getting from alternative sources? And then what your target is and how much of that you've actually contracted?

Yes. It's a good question, Mike. So in terms of power, typically, we are tied into the utility grids across the markets in which we operate. And as Stu mentioned, we operate in 70-plus markets around the world. So power availability and the availability of renewable energy is going to vary quite differently across each and every one of those markets. And so from a renewable energy perspective, we talked about in terms of coverage, just given that at the size and the scale in which we operate, if we put solar on our facilities, that's not going to cover the entire load of the building. And so we typically look for -- our preference is for virtual power purchase or power purchase agreements with either like solar or wind farms, we're adding additional green electrons to the grid.

Recognizing that, that might not be available in all the markets in which we operate, so we'll use a variety of other different mechanisms such as energy attribute certificates or RECs to meet that need. And then from a holistic just power availability standpoint, we do continue to take a long-term lens on like all different types of power across the portfolio in terms of what we're able to support. And so there have been a couple of instances that we've actually done self-generation across our portfolio to meet the needs of our customers. So it's kind of a tool in our toolkit in the event that we need it, whether that be natural gas turbines or fuel cells. So we have a variety of different options and continue to take a long-term lens in terms of looking at other power options as well.

And I'll say again, we stay very close to our national partners as well as local partners when it comes to this sort of thing. I mean, the governments around the world are looking at what are the alternatives to their current mix of power sources, including SMR and other things to help fuel the digital economy. Governments around the world don't want the grid or any capacity conscious to be the limitation, right, on their own digital transformation because look, it's not just competitive inside the country, but they are competing with other pool markets, right? And so even at the national level, there's multiple countries that we're working with that are looking at how do we change the narrative?

Do you think SMR is a viable short term solution? Most estimate probably 5 to 7 years before the cost analysis makes sense for SMR.

Yes. Look, I think there's some challenges with SMR for sure. I mean you've probably seen Katie talk about some of our announcements that we've made. It's just one of the technologies that we're following and are looking at. I think there's challenges around SMR. But I think it's really going to be a mix of those technologies, right, that are going to be the solution.

Well said, Stu.

There's been a perception in the market that Equinix being the leader by far for years in terms of quality of location and your customers would effectively accept whatever price was out there because they had to be an Equinix facility, right? And the perception is that maybe the customer service relationship management wasn't maintained the way that it could have been during that period of time and that you've lost some deals recently or maybe the customers didn't come to you because of that historic relationship and the perception, right or wrong. I think there was a recent management addition, right? I think Chief Revenue Officer, added.

But I think in part was to help rebuild the customer relationships and management. And if I'm right, can you just kind of tell me what you are doing maybe to change that interaction, the go-to-market with your customer and change that perception.

Do you want to start, Katie, I have some comments.

Perfect. I'll start. I will say we did have a recent new addition. So Shane Paladin joined us over December as our new Chief Revenue Officer, backfilling a role that our CEO, Adaire was actually wearing the interim hat, as interim CCRO. I would say, at Equinix, like we've always put the customer at the center of everything that we do and want to be long-term partners to our customers, and that continues to be like our guiding light as an organization.

I would say, as you think about our strategy that we outlined at Analyst Day, when we're talking about serve better to our customers, that's really making sure that we are delivering exactly what the customers are looking for and how we can be long-term partners and solution providers to our customers and having that holistic conversation with them of not just about their infrastructure needs today, but the infrastructure needs over the next 3, 5, 10 years plus and having that long-term customer relationship with them. Stum do you want to add on?

Yes. I'll just say like all of the execs on our leadership team, our C-level execs, the next row down like myself, SVPs, VPs in our company. we all meet with customers. I'm going to fly back home tonight to meet with the customer tomorrow morning. And I think that's really, really critical. So not just like our C-level execs, but guys like myself who are on the front end of the business who are candidly kind of far away from the sales motion, right, procuring land and working with the power team, et cetera.

But it's really important for all of us to hear what our customer needs are, where do they want to be? What value are they accreting out of Equinix? And so I would say like at least from my lens, Mike, like the customer is the most important thing. I go to customer meetings even in my job, you would think I'd just be kind of kicking dirt all day. But no, like I do customer meetings quite a bit to make sure that the whole company is really customer-focused and customer sensing. That's what I'd say about that.

And those conversations that you sit in, what's the common thread for what customers want from Equinix and where they're going?

Yes. Look, I think a big topic is new markets, like what new markets. Our customers want that global reach. They love the fact that we're global and they can have a conversation to us about Manila, right, or Joburg or wherever in some of the new markets that we've been in. The reason why we are going to those new markets because we're customer-led for those markets. And so a lot of the conversations I have is, hey, look, Equinix, this is where we need to be. Here's the application we want to put there. How can you help us, right?

And then it's really about like, okay, what does our entry into that market look like? Are we going to do greenfield there? Will we acquire a business in that market, et cetera? And then the other part of the customer conversation is around like what kind of deployments do they have? Who do they need to connect to? Where do they need to be relative to their existing data? How many milliseconds away? What is their latency requirement? And how does that interact with their existing deployments and their existing requirements?

And a little of a tangent here, but you touched on latency requirements. And in these conversations, do you see latency requirements increasing or decreasing? And then how does AI and then inference of the foundation models? How does it affect the latency needs of your customers?

Well, I'll say like more and more of the customers that we're seeing are saying, look, we need 3 or 4 locations. Let's just take Europe, that's where I last moved from. We need 3 or 4, 5-megawatt locations, and they need to be x milliseconds apart, right? That's a different conversation that we used to have, which was like, I need to be in Paris and Frankfurt with 250 kW. So that's really how the conversation has changed. They're seeing application level performance as the most important ingredient and how Equinix can really impact that.

And what are the applications and use cases? I don't know if it's about 10, 20 milliseconds of latency you're talking about, but what are the applications or use cases that are driving latency requirements to a lower number? Because historically, we thought silicon requires 10 milliseconds or less with things like autonomous driving, right? And maybe real-time 3D, things like that. Most of the applications, they could handle higher latency, right? So what are the applications that are driving your customers?

I think sometimes we can get lost in the AI over a little bit and what that means to Equinix. But look, we still handle billions and our customers still have millions and millions of digital payments, for example, right? Trading for New York City today. Trading is a huge part of our ecosystem here in the New York City metro area, right? So those kind of things are all very latency sensitive, right? I don't know, Katie, if any other examples come to mind.

No.

And then I wanted to shift to xScale for a second and what you think that the legacy Equinix portfolio or brand adds to the xScale portfolio, how bringing those 2 together maybe is greater than the sum of the parts.

Yes, for sure. I think that's really true. I think when we look at our Build Bolder program, we want to be doing things at a much larger scale. And part of that is co-developing. In my part of the business, there's no difference between xScale and retail, for example, right? My team is out there solutioning real estate and power for the entire portfolio. And so for us, it's really the ability to get some real scale out there, develop bold products for Equinix and be able to just become just a lot more efficient. But certainly, we've got our capital partners, right? So it doesn't sit on our balance sheet. And like Katie said, there's contributions that we make when we develop that into the JVs, et cetera.

And I think the market dynamic today, there's a view that the private operators, the builders operate at much higher leverage. Now it's off balance sheet for you, but some are operating at 10 or 15x net debt to EBITDA, right? And also maybe take more risk on where they're placing assets or just winning a higher percentage of these larger deals. So how does Equinix compete with that private company able to operate at higher leverage and maybe wanting to chase deals in the second and third tier markets, how do you compete with that? Maybe you don't. Maybe you feel your sweet spot isn't chasing the next big AI deal, maybe it's more leveraging your existing portfolio of customers.

So we're not out there like from that side of our business, trying to grab a bunch of market share and grow from an xScale perspective, right? We know that's not our core business. Our core business is what we're putting all of our capital in, which is kind of amazing if you think about it, like the capital numbers that Katie was talking about earlier that's going to our retail business, right? And so that still has a ton of growth to it. And so we're not relying to hit our growth targets merely on xScale and chasing those deals.

Yes. And I would just add, when you think about our xScale initiative, and we are certainly really excited about the xScale opportunity. When you look at our plans for xScale 1.0 and xScale 2.0 at full build-out. But as Stu noted, that when we think about xScale, it's really being able to have that full product continuum conversation with the various strategic subset of our customers and being able to service their needs on the retail side. So if they need a cloud on-ramp or a networking node, we can service that in the retail business. And then if they want to do something on a larger scale side when they say, hey, Equinix, we need 10, 15, 20-plus megawatts we can service them on the xScale side, but it really is very strategic subset of our customers and having that full product continuum to be able to give them with the right application in the right data center.

The reason we're in the xScale business is because it was a customer-led thing right? They want us to be there.

So there was command for it. Maybe they were asking you for...

Exactly right.

I'm going to open to the audience here in 1 minute, although Spector's team always has these Spector 3 questions that we ask each company, so I make sure I get those in for them. So first, either one of you, when the Fed starts to cut, do you expect borrowing rates for long-term debt to, and there's a multiple choice here, decline, stay flat or rise.

Give that one to Katie.

Put your macro hat on and...

I'll go with Option A, decline.

Decline, okay. The second question, Last year, the majority of companies stated they're ramping up spending on AI initiatives. How would you characterize your plans over this next year, higher, flat or lower.

I would say higher. We're absolutely incorporating AI into I know this is like multiple choice, sorry, into like our operations and just how we can become more efficient as an organization. So like a great example would be, we are leveraging AI to treat like digital twins of our data centers to help us [ manage ] the energy efficiency and just how we're deploying and looking at hotspots across the data center. So obviously, we are looking at ways to just continue to incorporate AI into our business overall. So I'd say higher.

Do you know who you're using for that? Is it Bentley Systems or Autodesk or I'm just curious from putting on software hat back on...

[indiscernible] Back to on it...

Exactly.

You're not the first person to ask us that.

Yes. It's full technology. Last one and then -- sorry, guys, you can jump in. Do you believe same-store NOI for your sector will be higher, lower or the same next year?

I think overall, we continue to see a healthy and firm pricing environment. So I would say -- I'll say higher.

Okay. So number one, thank you for entertaining me with that. And we have a few more minutes if there are any questions from the audience. I'll have to repeat them if you have any, please. No takers? Okay. Perfect. So let me just go ahead and just wrap it up with one more. Going back to Analyst Day, I think there was a little confusion just around some of the messaging and the impact from the higher interest expense, right? I think last quarter, there was some clarification offered around that. Can you give us the latest on that, Katie?

Yes, I would say -- thanks, Michael, for that. Yes, coming out of Analyst Day, I wanted to be responsive to the feedback that we heard from the Street. And so the Q2 earnings call in July was our opportunity to really update and provide you guys a little bit of clarity in terms of just the key questions that we've been getting. And so one of those was just our assumptions around interest expense. And as we noted on the Q2 call in July, embedded within the long-term outlook that we gave at Analyst Day, both for the refinancings and the incremental debt issuances, we're assuming that those are done at 4.9%. And that is a blended composite of continuing to access lower cost currency markets before returning to the U.S. capital markets.

Great. Thank you guys very much, and thank you all for attending once again.

Happy to be here.

Yes. Thank you, Stu. Thank you, Katie.

Thank you.

Okay. Good afternoon, everybody. Welcome to the Goldman Sachs Communacopia and Technology Conference. My name is Jim Schneider. I'm the data center analyst here at Goldman Sachs, and we're really happy to have Equinix here today. I'm joined by Steve Madden, VP of Global Technical Advisory, Equinix as well as Chip Newcom, Senior Director of Investor Relations. Welcome, guys. Thanks for being here.

Perfect.

Thanks for having us. Well, -- and quickly before I know you've got a lot of questions, just so we cover it from my legal team. Some of the things we're going to say here are forward-looking in nature. Please see our SEC disclosures for risks and uncertainties.

Very good. So we'd be remiss to start any place except AI. That's a key theme across the conference, but also a driver for your business. In your role as VP of Global Technical Advisory, you work with customers and partners to optimize their use of Equinix platforms. So you've got a bird's eye view to that. So maybe you start off by giving us some insight into how AI is shaping the customer conversations you're having and how that's evolved over time.

Okay. Yes. So I think that we've been on this journey of growing data, using data, doing machine learning, doing analytics, doing different things. And AI came along and kind of kicked all of that up a notch and accelerated the need to do more with data more quickly. And so a lot of our customers today, whether they're enterprises or providers, I deal with both, I'm either having a conversation with enterprises who are trying to explore the space, understand what this means to them, where do they get started? How do I get going in the cloud? Do I need infrastructure? Or I'm having conversations with new service providers and GP was the service provider. So how do they get into this business and where do they go? And I think that overall, a lot of what we saw in this trajectory is just being accelerated and amplified by AI, but a lot of what they're trying to do and solve has been challenges we've already had.

I mean given this intense focus on AI, can you give us an update on sort of the amount of AI workload demand you see within your data centers at Equinix and sort of how you're positioning the company to capture the growing demand for inference in particular?

Yes, I'll start, and I'll turn over to you. So I always thought why we set the context in that when a customer comes to us for the project, the project is going to require usually considerable infrastructure, connectivity, I need to manage data and have data ready and available in order to source and learn from it. I'm going to need potentially some GPUs or some AI accelerators. I'm also going to need other partners to consume models from. I'm going to be subscribing to data. I'm going to want to sell my data to people. It's actually a multi-location, multi-deployment discussion that we're having.

And yes, during that conversation, they might want to deploy GPUs in a particular location. So when we talk about AI and AI deployments, I get lost a little bit, do you mean large LLM training in particular facility? Or do you mean a project that's related to AI? So we have this conversation internally all the time. But I would say that all of our customers have it on their radar. Some of our customers are actively deploying in our infrastructure today. Some of them on the other end of the spectrum don't even have any plans to look at AI right now. It's just too far out for them. And I think that we're in a birds eye position that we start to see what the Peloton of the market are doing and where they're going with it, where the middle of the market is kind of sitting and exploring and where there's sort of this tail where nothing is really happening at all.

Yes. I mean I think to Steve's point, we are definitely seeing the leading edge in terms of what a lot of enterprises are doing. And certainly, there are specific verticals that are further along. So as we think about financial services and what customers are doing there, whether it's things around high-performance compute for capital markets or thinking about things like fraud detection. We've seen a lot of those use cases continue to proliferate and other places like within health care, thinking about a couple of health tech companies that are working on new drug discovery using artificial intelligence around some of their therapeutic data.

We're seeing both inferencing as well as training getting deployed into our data center. But the hard part is, to Steve's point, what is like a "AI deployment". In some cases, it actually can be relatively challenging to definitively say because, well, if it's an NVIDIA DGX pod at 120 kilowatts, clearly, that's AI. That one is easy to tag. But it might be networking associated with transiting data to a cloud service provider or a GPU as a service provider to start developing a model. It might be storage associated with what they're going to be doing within a data center, again, preparing for working with a third-party vendor.

So there can be a variety of different IT use cases going into an Equinix data center that may very well be associated with artificial intelligence, but it might just not be a DGX super pod where you're going to go and actually train something specific, which isn't to say we're not winning those too. We were the first company to be working. We had a press release with Block earlier this year where they were the first to deploy a GB200 stack within an Equinix data center for their own purposes for AI modeling. So we're definitely supporting that, but it's a snowball that is rolling downhill. It's not necessarily a massive avalanche yet.

Fair enough. I think at your Analyst Day, you noted that the mix of AI training versus inference is about 50-50 today that you're seeing. From your perspective, sort of how exactly does that manifest itself in the conversations you're having with partners and customers? And how the requirements from a facilities perspective change between the 2, if at all?

Yes. So for the first half, which is the inferencing -- sorry, the training part you said, the 50-50, you said for training. In a lot of cases, that's a group of enterprises or customers that have started exploring and built something or several things inside the cloud, but they have enough momentum now and enough demand internally, and they know they're going to need more of a consistent base infrastructure, they'll deploy their own DGX pods or their own infrastructure outside of cloud or cloud adjacent where they can start to see that more mature towards that's going to be where they're going to build a factory to start producing models.

And to the extent that Chip was just saying, there are certain industries where we see that more prevalent than not. On the inferencing side, it's a combination. And we sort of say inferencing is going to boom only because we're not -- when we say inferencing, we're not saying that enterprise is then going to take that model, deploy it at the edge and use it. Yes, they do, do that. But what we're actually seeing is those models are useful to multiple companies and not even just in their own industry, but also across industry. So yes, they go and deploy that model. They just start using it to make money or save money, but they also use it as a revenue stream and offer it to dozens of other companies to help them make money.

I'll give you an example. One company who is very much in terms of energy and energy tech built a model that helps building management systems be more efficient. So the building itself consumes less energy. They're smarter about how they turn things on, turn things off, but they keep everybody in the building happy. And that's like a very difficult thing to do, but it's real-time data, real-time analytics, real-time inferencing at the edge. How many companies do you think are interested in something that's going to save them 30% on energy in buildings?

Well, everybody with a building is going to care about that. But they don't want to all go and build a $100 million model. So instead, we're seeing a lot more of these models that are mature that solve a very horizontal use case are being deployed more prolifically and more availably today, whereas the bespoke sort of very proprietary models that are coming out for very fit-for-purpose occasions are coming a little bit more slowly. And so it is a bit of both. But I think that we're excited about everything being built in these factories is going to have to end up accessing and using and interacting with the physical world eventually. And it's going to be done at more horizontal scale than just the companies producing it, which is why we think inferencing is going to be so big.

Interesting. You cover a very diverse customer base. I'm kind of curious, -- are you seeing any kind of departure or difference between the technical and facility requirements for AI that you're seeing from hyperscale kind of customers versus ones you're seeing from enterprise?

Oh, yes. Well, hyperscale customers build everything at scale. It's kind of in the name. So they essentially design their own footprints and things for massive blocks of deployment chunks. Enterprises typically do the opposite. They really only want smaller footprints and they grow in different sort of chunk sizes. Where it crosses over with AI being the topic is certain infrastructure that either of them buy have particular requirements around density cooling and so forth, which I'm sure you all know. And so we need to make sure that when an enterprise is buying AI infrastructure specifically that's high density requirement, that we steer them to the right area of the campus or the metro that's designed for that.

So we have facilities already in all of our metro campuses that can accommodate those larger footprints. We work with them on which is it that they need to use. But we also have to have a conversation around what kind of cooling they want to use because there isn't a single standard yet. So there is a conversation to make sure we augment the environment to support that particular infrastructure, but we can do that. Whereas hyperscale, it's already kind of preplanned, predeployed, prebuilt into the infrastructure before we turn it over. I would just add that inferencing has a much lower requirement. So when I say inferencing at the edge, I'm not talking about needing massive megawatt cooling. It's much smaller and they're easier to accommodate in existing facilities.

And I'd add too, part of the genesis behind our Build Bolder strategy really is to what Steve was talking about, the amount of digital infrastructure that enterprises are looking to consume now, not just because of AI, but because of a diversity of different IT use cases. The deal that used to be 100 kilowatts is now 0.5 megawatt. And the deal that used to be 0.5 megawatt is now 1.5 megawatts.

And so part of what we're looking at is from a capacity perspective, given the demand that we're seeing and the demand signals we're seeing from our customers, we want to be building in advance of that because it takes us anywhere between 18 to 24 months to build the next facility or to build the next phase of the facility. And so recognizing that we need to be forward building in advance of the demand that we're seeing from customers, it's critically important for us to be able to be bringing capacity online.

Yes. From a constraint perspective, at the highest level, how do you see the next couple of years in terms of the industry and the power needs? I mean, are we going to run into a point where the industry just simply does not have enough power to accommodate all the data centers that are intended to be added? And how is that impacting the plans specifically for Equinix over the next 2 years?

Do you want to start?

Well, look, I think, one, the demand is robust. There's no doubt about that. And that's whether it's for providers like Equinix on the retail side, what we're trying to do on xScale side. And we're really trying to be able to serve the entire gamut of the data center industry. But the reality of it is there is constraints in the marketplace, whether that's power availability in the key metros where we're looking to operate, whether that's getting key mechanical and electrical.

And so a credit to our Chief Procurement Officer and her team, we've already got $600 million worth of capital equipment on our balance sheet to be able to help our forward builds. And so a big part of what we're doing is trying to lean in into our ability to forward procure into the fact that we've got a very large balance sheet to ensure that we're in a position to be able to build capacity. But I think part of what where we're differentiated is when we're thinking about building and bringing new supply online, we're looking to build it in major metropolitan areas.

So as you think about our Build Bolder strategy, the vast majority of our capital spend is going into those largest markets where we already generate over $100 million in revenue. So that's going into the Washington, D.C. type markets, into the London-type markets, into Tokyo. And these are markets where we've got large established ecosystems.

And so as we're thinking about power availability there, certainly that can be constraining. But we've been working through our corporate development efforts for years to make sure that we've got the land bank and the power bank so that we can continue to keep building. And also on the retail side of things, we consume our capacity in a much smaller way. We're not consuming in 50-megawatt, 100-megawatt chunks the way that you might with a hyperscaler. And so we're not necessarily having to go back and reload with the utility quite as frequently.

But I think the big difference, too, is that when our customers are coming to us for 0.5 megawatt, a megawatt, we can continue to keep building on a relatively consistent load ramp relative to when a hyperscaler puts out an RFP for 100 megawatts, they're then going to all of a sudden go out and talk to, we'll call it, 12 different vendors to see who will facilitate that. Those 12 vendors end up going to xyz grid operator and all of a sudden, it looks like there's a gigawatt and change worth of capacity needed when actually it's really 100 megawatts is the end demand need.

And so I think part of what we benefit from is that as we work with the utility operators, we've been very good consistent load ramp. We're the ones who we've got a high say-do ratio in terms of the actual load that we want to get and the load that we procure. And so as a result, we've been in a very good position to continue to drive our growth.

From a competitive standpoint, how do your technical capabilities and strategic partnerships sort of differentiate your AI offerings, especially when you're competing with some of the very larger competitive data center providers, ones that are private equity-based or otherwise?

There's 2 answers to that question. One is, in a lot of cases, when we're talking about AI deployments inside our retail business, our retail business is very ecosystem-centric. So when the enterprise comes in, all of the suppliers they want to use are all close partners of ours.

All the providers, GPU as a Service, model as a service, data service are already there, data management platforms, et cetera. And so it's really a matter of if I deploy this infrastructure, all the things I'm going to need to connect to wire up whatever it is I'm trying to do or even if I'm not sure what that is, it's going to change, it doesn't matter because everyone is here.

Whereas if you go to a wholesale location where there might be 4 or 5 customers in that whole building, you have to bring all that stuff with you or somehow get it to you. And so where the case is that it's more of a -- how do I exchange value, we're typically in the best position for that. If the requirement doesn't require that at all, I would start to ask why would you want to put it here then? If it doesn't require all of that, maybe you shouldn't put it in our data center.

And if it's really, I just want it nearby. So okay, it doesn't have to be in the central point where everybody connects, but I don't want to be too far away. Well, then we do have an aspect of our business where we do have capacity that we bring online, keep ready for that sort of thing. We call it sort of like a data hub adjacent where we can serve that capacity. But again, the proximity back into that infrastructure is really close. And so I just argue if the workload is either going to be revenue generating or the implications of cost is significantly lower because it's in proximity, you're going to want to put it here.

And if those things aren't true, and it really doesn't matter where it goes, we would advocate that you should just think of it that way and put it where it needs to go.

Yes. I mean that's almost why you hear us talk about almost like a mantra, this idea of right customer, right facility for the right outcome. And that's because, look, there's a real opportunity cost of our capital in the sense that data center capacity is scarce.

And so if we go out and sell a data center to an undifferentiated partner that's taking up half of the facility, well, that means that then we can't sell that to 50, 100 different customers, all of whom are going to interconnect with each other. And so as we think about how we're going to sell, we're very thoughtful in terms of trying to continue to curate these ecosystems. So seeding artificial intelligence and seeding all of the various different companies within that ecosystem, continuing to work with the hyperscalers, with the SaaS companies because as you think about our business, we're really a place where both buyers and sellers of digital services are coming together to connect to each other to enable the digital economy.

And we want to make sure that we're continuing to develop that within our facilities because, again, the secret sauce of our business, building a data center, that's not the secret sauce. If you've got a good general contractor, you can build a data center.

The secret sauce of our business is the ecosystems and interconnection. It's that 19% of our revenues. It's that 492,000 total interconnections that differentiates our business relative to others because that is how when you call me up on Zoom and you have to connect from Goldman's network to Zoom's network to our network and go across all the various different network service providers. Every single one of those is an exchange point of data where it has to go from one network to another, that happens in an Equinix data center.

So do you think it's fair to say the interconnect piece is still the most differentiated competitive advantage that you have? And do you think -- is that advantage increasing in importance or decreasing in importance as we move to AI?

Increasing.

I mean the amount of data, the amount of bandwidth connectivity, the amount of participants involved, the amount of people you want to exchange with, the amount of partners and providers you're subscribing from, the amount of people you want to sell it to are all just growing.

And it's physics, but it's not rocket science. If you've got massive amounts of data that need to be exchanged in the most efficient way possible, doing it in the same building on the same campus really is far more effective through volume throughput and cost than trying to run it across half the country. And so we're seeing that the density of population increasing is what's increasing the value of the facility.

Yes. And I think that's actually -- I love your -- it's physics, but not rocket science in the sense that the thing that hasn't changed with artificial intelligence is the speed of light. That has remained constant. And so latency is still a very real issue. And now for certain consumer applications, if you're going to be accessing an LLM on your phone, the latency of that doesn't necessarily matter in the sense that if you're paying a server in the middle of [indiscernible] that's okay because you to that server, 0.5 millisecond doesn't matter, a millisecond doesn't matter.

Relative to -- as you start thinking about the world of Agentic AI, where it's now machines talking to machines for things like building management or any number of other use cases, there, that latency becomes critically important because if you're connecting to multiple different data sources to be able to run a real-time inferencing at the edge deployment machine to machine, that needs to be in the major metro area where that application is going to be running.

Throughput. Volume of throughput goes up exponentially. Yes.

Fair. Maybe I'm going to shift to some technology questions. Power densities continue to climb. We've moved pretty quickly from 5 kilowatts to rack to 50 to now, I think some people are talking about 500 kilowatts per rack in the cases of NVIDIA's Rubin. So maybe what are the technical capabilities and limitations in your infrastructure for supporting these rather extreme demands? What advancements are being made to prepare for those future power cooling requirements?

Okay. Well, clearly, I think of it twofold. One is necessity is the mother of invention, right? And so when we didn't have customers showing up with 120-kilowatt cabinets, clearly, then the generic building infrastructure is fine. But when we started to see those requirements come in, we're already getting ready for growth and increasing kilowatt requirements, but extreme -- let's talk about extreme cases.

Well, then you actually deploy or build an area of the data center that's designed to handle that particular capability. You don't have to change the whole facility. You just change an area to cope for what that requirement is. And we've seen what those standards look like, and we got more familiarity around customers of what they're thinking deployed, what that looks like as a workload, how much more of that we think there's going to be. And we have designated areas to just take that under control and handle it.

I would also say the flip side is that there are chips and technologies and models and things coming out which require a fraction of the power of what we've seen before. So in a lot of cases, workloads are shifting to lower power alternatives that have the same performance, but don't have the same commitment to requirements. And so I think you're going to see a balance of both where we won't see it go up necessarily to 500 kilowatts of cabinet, but you will see it sort of start to normalize out a little bit, and you'll see this mix of what the technology types will change over time to a different mix of power thresholds and requirements. And that will also reduce the demand as well as improve the supply.

And going back to sort of what we were talking about earlier, this concept of right customer, right asset, right outcome, part of what we're doing is we're building new data centers is we are building to higher overall average power densities, but again, we're supporting a diversity of different IT workloads.

And despite what some vendors might tell you, the reality of it is the vast majority of IT workloads do not need to run on a GPU right now. Core networking infrastructure is still relatively low power density as is storage, as is any number of other applications.

And so when we're designing a facility, to Steve's point, we may have an individual data hall that is pre-provisioned for liquid cooling out of the box so that we're ready to go for a customer. But even as we're thinking about like take, for example, here in Silicon Valley, if you go to our Great Oaks campus down the way, down in South San Jose, our SV1 and SV5 assets, which are core networking assets, like as everyone is sending an e-mail or doing any Internet traffic here, it's probably either going through SV1, SV5 or SV8 here in the Valley. We're not going to put a super high-power density workloads in those facilities because that is for core networking infrastructure that is proverbially the beating heart of the Internet.

But what we will do is we're building other assets around that. So we'll have SV10 and 11, which are designed for more 5 to 6 kVA a cab. We're building SV18 across the road right now that's going to be north of 12 kilowatt a cabinet. And so depending on what the customer workload is, in theory, could we put a single cabinet at 0.5 megawatt in? Sure, it might be blowing like the sun, but you just have a roller skating rink around that basically.

Yes. As you think about expanding the footprint of your facilities sort of how do you balance growth in your sort of established interconnect markets with expansion into emerging regions? Are there any technical drivers that influence the international markets that you go into?

I'd say, certainly, as you look at our CapEx plans right now, north of 3/4 of our CapEx is going into our major markets. So those are the -- the 15-odd markets where we generate over $100 million in revenue. And so we're continuing to invest in those markets in large part because we see very consistent good fill rates. That's where a lot of our customers already are, and that's where they're continuing to grow.

Now part of our strategy is then to seed other markets so that they develop over time to hopefully become those $100 million-plus major markets. But there is an element of what drives us to go to new markets is largely speaking, our customers. I mean, case in point, we had been talking about getting into India for years and it had been a market that had been a target for our customers.

Well, it wasn't until we found both the right opportunity with the acquisition that we did a few years ago to get into and buy combined with getting a great leader locally that really then propelled us to get into that market. But most of our capital, as you think about our Analyst Day guide that we gave back in June, most of that's going to be going to our largest metros.

Yes. Okay. And then from a JV perspective, how do you think about -- are you shifting the way you consider what you do in a JV versus what you do on your own books these days based on any of these technical requirements? Or is it mainly driven by the kind of customers you're pursuing?

It's really driven by the kind of customers and the size of the deployments. So as you think about what we're trying to do in our xScale product, part of the reason why we set that up off balance sheet is, again, driven by the return profile that you're going to get when you're doing a hyperscale deal.

So if you're signing and filling up an entire data center, 30 to 60 megawatts going to a single customer, the return profile of that is completely different relative to the return profile that you're going to be getting in a retail colo facility.

And so as we think about our highest and best uses for our capital, that continues to be our retail colo facilities, where as you look last quarter, we generated a 26% cash-on-cash yield in the PP&E we invest, and we continue to underwrite projects we expect to be in the sort of mid-20s, around 25%. So we want to put as much of our capital to work on those assets as we can, recognizing that we still want to be able to support our largest customers, who are the hyperscalers.

And so being able to have a more capital-efficient way to do that, where we're putting in 20% to 25% equity. Our LP partners are putting in the balance of the equity. We can lever it differently. The benefit of xScale is it allows us to get a really attractive return on invested capital because of the fee income that we get and the leverage that we're taking, but do it in such a way where we're not swamping our balance sheet.

Interesting. And maybe -- and I think there's a lot of people who are interested in the supply chain dynamics here. So maybe give us an update on any long poles you're seeing in the tent at this point from a cost or procurement perspective? And are there any parts of that process that are getting easier? Or is everything just getting harder?

Well, I'd love to say it's getting easier. Unfortunately, given the demand profile, it's still relatively long. I mean, in some cases, you can be talking about several years to be able to get a generator for backup power. And I think credit to Ali Ruckteschler, our Chief Procurement Officer and her team, they had the foresight relatively early on in the pandemic to go out and start forward procuring for a lot of long lead items.

And so case in point, we've got $600 million worth of mechanical and electrical equipment that we've already prepurchased that's sitting on our balance sheet to be able to support our current CapEx plans for building our new data centers.

And so part of the benefit we have as one of the largest players in the space is we can very confidently use our balance sheet and lean into our plans where we've got 59 major projects underway around the world. And so we can very confidently go to xyz vendor and say, "Hey, we want to get this much of capacity. We know that we're going to be good for it because we've got all these projects that we're building" and then put it to highest and best uses.

Yes. And I think if you add the fact that we're so predictable and consistent with all of those suppliers, they're more likely to want to work with us because they know we're legitimate around what we do and we direct contracting.

Absolutely.

Yes. I just maybe close on a couple of different financial questions, if I could. Given everything we just said about sort of the exploding demand profile for data centers and everything else, I would have expected or most people would sort of assume that your utilization rates from a rack perspective are just kind of going to the moon.

But at least on the way it's disclosed, that's not the case in terms of like the overall utilization that you report. So maybe kind of give us some sense of the moving pieces that explain why utilization hasn't really gone like in the way that most people would have intuitively thought. And maybe what's the path to rightsizing that utilization trend? And are you seeing any kind of new disclosures that would help illuminate that from investors?

Yes. I think certainly, the one challenge, of course, is that utilization is a -- what a utilization is in a data center is a multidimensional metric in the sense that you've got the space capacity, you've got the power capacity and you've got the heating and cooling capacity. And with our cabinet equivalent metric, we're trying to sort of do all 3 in one metric.

But part of the challenge also is an averaging factor where as you look in our Tier 1 markets, we're actually relatively highly utilized. In our Tier 2 and Tier 3 markets, we actually have more capacity available because that tends to get consumed more slowly. And so part of what we're doing right now with Build Bolder is we are trying to rapidly build capacity in those larger markets because of the demand that we're seeing. But we're also then trying to be thoughtful to appropriately steer our capacity to the markets where we have capacity.

And so part of what the sales team is trained on is now going out to the customer and saying, "Hey, look, based off of all of your capacity needs, what you're looking for in terms of partners, what you're looking for from latency, have you considered these 3 markets?" Because look, anyone is going to be able to say, yes, I'd be happy to be in Ashburn data center capital of the world. But not every application actually needs to sit there. And so that's, again, part of this right facility, right outcome for the right reasons that we're going after is how do you steer to the elements of capacity, but then, again, continue to Build Bolder to bring capacity online in those constrained markets.

And there's the sawtooth.

Yes. And there's also a sawtoothing as well, where as we open up a new facility, inherently, that's going to tick down utilization and then you have to fill it back up and then tick down as you open up more capacity.

Yes. And I would also add that infrastructure is refreshing much faster than it used to. In a lot of cases, especially with the accelerators, every 2 years, there's this big bump in capacity performance. And that may or may not mean that they need fewer infrastructure from us to run the same workload or they double up. The technology is changing, too.

Fair enough. And then just finally, on the capital structure. I think at your Analyst Day, you expressed a preference for raising more incremental debt rather than equity going forward. But maybe talk about that in terms of the context of your overall preferred capital structure, how do you plan to measure sort of exert the various levers at your disposal, whether that's JVs, acquisitions, new builds and anything else?

Yes. I mean in terms of the outlook that we gave at our Analyst Day back in June, we did say our expectation is that in terms of funding our growth CapEx over the course of the next several years, that's going to be through a combination of our internal free cash flow generation after paying out our dividend, plus then moderately increasing our leverage.

So as Keith noted on stage, our expectation, again, back at our Analyst Day is we'll add about $8 billion of debt capital to the books over the course of the next several years through 2029. And we do have capacity to go as high as 4.5x net levered while still maintaining our current BBB+ investment-grade credit rating. And so we think that, that's the appropriate way to look at our capital structure. But again, relative to many of our peers, we actually have relatively modest levels of leverage.

But part of what we like about that is it gives us a lot of strategic flexibility where depending on as we see opportunities in the marketplace, whether it's M&A, whether it's leaning into Build Bolder even more, we have the strategic capacity to be able to lean into those types of opportunities.

Great. I think with that, we're almost out of time, but thank you both for being with us. We really appreciate it.

Great. Thanks, James. Thanks for having us.

Thanks for having us. Bye.

All right. Good afternoon, everyone, and welcome to TD Cowen's 11th Annual Communications Infrastructure Summit. For those of you who don't know me, my name is Michael Elias, and I am the commentator analyst here. For this session, we're joined by Equinix. And from Equinix, we have Raouf Abdel, who's our EVP of Global Operations. This is structured as a fireside chat. We have about 40 minutes for it. I have a ton of questions prepared. This one, I will, I promise you. I will open it up for questions. But with that, Raouf, thank you so much for being here today. I really appreciate it.

Thank you, Michael. Thanks for having me. Great to be here.

My understanding is that you have something to say before we get started.

I do. Being a public company, we have to start with -- I'll be making some forward-looking statements, and please be sure to check our SEC filings for any factors that may impact those forward-looking statements.

All right. Perfect.

Get that out of the way.

Done.

These two over here snickering because they weren't part of public and now we're private. We don't have to say that.

All right. Well, let's jump into it. Raouf, you joined Equinix in 2012. But from my perspective, you've had a major impact on the company based on conversations that I've had with both current and prior Equinix executives. So for those who are less familiar with you, could you just give us an overview of your time at Equinix and a bit about your path to your current seat?

Well, thank you for the kind words. So this month, I will reach my 13-year anniversary with Equinix. For nostalgic reasons, I went back and looked a little bit at where was Equinix in 2012. And we were just under $2 billion in revenue. Our market cap was $10 billion. And so in the span of that 13 years, we've expanded and scaled and we're -- revenue is about 5x what we were back then. And our market cap is depending on which day you pick for the stock price and which -- a few months ago was about 9x, and it's more like 7.5x right now, but we'll get it back up to 9x, I think. So it's been an incredible ride.

I've got some former colleagues in the crowd, and they're familiar with the journey we've been on. And we've had -- we've scaled a lot. We've expanded our geographic presence. We've expanded our employee base. We've expanded our data center footprint, and it's been a great ride. And I've had the privilege to lead our global operations function for about the last 8 years of that. When I first joined the company, I ran the Americas operations. Some of you in the crowd may remember Charles Myers. He was the President of the Americas at the time, and he's the one that hired me. He and I go back to our Level 3 days together. So we've worked together prior to that.

And then I ran that for about 5 years. And when we say operations at Equinix, it's really about the build and operate portion of the data center. So it's design, it's construction. I currently have procurement, energy, which I'm sure we will get to at some point in this conversation. And I also have the team that runs the data centers day-to-day, maintaining the facilities and the sites as well as supporting our customers every day.

All right. Well, you've been -- from what you just said, you've been in the seating your current seat for a while. But I am curious, what are the strategic priorities for you right now and for the business over the near to medium term? And as part of that, I'm just curious, where are you spending most of your time these days?

Yes. I would say there's been a bit of a shift for both the company as well as me personally. I would say I used to spend more of my time on the operate side of things. And I think anybody that's in the data center space knows that you live on the stability of your service and your product and that was not something you ever wanted to have disruption in. Most of our customers have mission-critical deployments inside our data centers. And so maintaining reliability was always paramount and remains paramount for us. But my recent focus has gotten a lot more to be on the build side of things, partly because we're investing at a higher level and partly because that space has become a little bit more challenging, both from an energy -- securing energy standpoint as well as supply chain.

So I spend the majority of my time on talent, energy, supply chain, and the strategic mapping and planning of what capacity we're going to build where I think you've heard about the initiative that we call Build Bolder in the company. And so I lead that for the company in terms of the planning, the strategy, the investment profile and probably most importantly, like where are we going to go? Where are we going to build, at what levels, at what pace, and what throughput. And then underpinning all of that is how we're actually going to execute to do that build.

Then we're going to have -- this is going to be fun then. So I want to talk to you about something that we started out today talking about on stage, which is deal sizes, Particularly, we've seen enterprise deal sizes increase. We've also seen, obviously, on the hyperscale side, deal sizes increase. We're going to put aside rack densities for a second.

But as we think about your Build Bolder initiative, one of the things that I think underpins it is that, "hey, look, that your typical enterprise deal has gotten larger. And as part of that, the construction motion and the standard capacity that you build needs to grow in size in order to accommodate that larger deal." I'm curious in your time in the seat, how have enterprise deal sizes evolved? And as part of that, what were the changes that you made to your standard design in terms of the number of megawatts per building or per phase? How has that evolved?

Yes, absolutely. So in my time, all of those dimensions have increased over the years. When I first joined Equinix, we were building 10-megawatt buildings. That was at the time, a really big thing.

Those were big, yes.

We wouldn't dream of building a 10-megawatt data center today. We're in the 30 to 60 sort of range for any given building. And then we're clustering them together in campuses to build multi-hundred megawatt campuses. And so we very much historically in the retail space were building just-in-time inventory and being very frugal with the capital that we are investing, and we are also expanding in many, many locations simultaneously, so you couldn't sort of concentrate that investment in fewer locations.

And so we were managing that capacity in call it, 4- or 5-megawatt chunks. And what we've been finding over the course of the last few years is by the time we finished the phase, we were selling it out, we were starting the next phase. And so we've changed that model a bit so that it's more capital efficient. It's easier on our supply chain. It's easier on our -- the people that are the projects. And so we're moving away from the notion of it's going to be just in time because it's really hard to manage that way.

And in the days of building a 4- or 5-megawatt chunk, selling a 2-megawatt to a customer right out of the gate, you're almost out of capacity the day you finish. And so certainly, in our really high-demand markets, Ashburn, London, Frankfurt, Singapore, we're building 10, 20, 30-megawatt phases, and the buildings are bigger, too. And when we say bigger, this is always a confused point. When we say bigger, we don't mean the building is bigger. In fact, the building is smaller. What's bigger is the power. But what's interesting is you still need roughly the same amount of land because the infrastructure consumes more space now. But the core and shell is actually, if anything, shrinking because the density is going up, and so you pack more power into that particular building.

So if we're starting with the building size, it's a 10 now, we're talking 30 to 60. We'll get to the phases in a second. But to that last point that you made about the footprint shrinking, obviously, the average rack density that you're architecting these data centers to is increasing. So as part of that, how are you thinking about the standard rack density that you design the data center to currently? And I'm more curious, how does that compare to what we saw before the beginning of this AI demand boom? I'd say, let's call it, end of 2022, first quarter of 2023, how has that average rack density that you build to evolved?

I would say 5 years ago, we were thinking about 6 kW a cabinet. I would say even before the AI sort of change and dynamic, we had stepped that up to about maybe 8. And now we're thinking 12-plus on average. And I think the average is a really important point because not every cabinet is going to be 12 or even the big numbers that we hear about, the 100 to 150, those are going to be a portion of those cabinets. And so what we sort of designed for is an average and the capability to increase that average with more infrastructure as needed to future-proof for those data centers. What's also interesting is in today's data center design, the ability to support concentrated heat loads is much better than it used to be.

I'll give you a couple of simple examples of how to articulate that. So in the days where we used raised floor, it was much harder to deliver a concentrated amount of cooling capacity to a given cabinet. Well, we don't use raised floors anymore and the industry pretty much is away from it. And I would venture to say we pioneered that. But in addition to that, the way a data center is designed today is the room is one giant plenum, as we would call it. And so what that does is gives you the ability not to have to manage hotspots the same way. In the past, you had to be really careful where you put hotspots. And in the days of raised floor, you literally had to redirect air. Well, if the whole thing is a plenum, it completely changes how you manage cooling and the ability to have 20, 30 kW here and 5 or 10 right next to it and the room can be balanced a lot easier with the type of cooling design and infrastructure that we're putting in data centers these days.

When you say 12 kW per cabinet, the first thing that comes to mind is, well, when I think of the super high densities that you hear NVIDIA talk about, right, obviously, this is much lower. But then the next thing I think about is, well, you're much better off stranding shell than you are having no space in the data center and just like running out of power. So from my perspective, it's -- is the reason why you're building at 12 so that is because you still see there's plenty of opportunity in that, call it, 6 to 10, that's going to be the standard, but you have the ability to flex up and new cooling solutions have made it easier for you to be able to support those loads. Is that the right way to think about it?

Absolutely the right way. And I would say to add to that, we have to have the ability to flex. But as with any business, you don't want to over-deploy capital and you don't want to strand capital. So we don't want to build for 30 today and have that be stranded. The other important dimension to what you're describing is if we get to the point that we're having to support throw out big numbers, 70, 80, 100 kW per cab, you're immediately moving into the liquid cooling space. In which case, you aren't relying on that air distribution that's in that plenum, as I described it, you're moving into liquid cooling.

And the beauty of today's design is it still leverages the same chilled water loop. It just it takes it from the central plant into a unit that's going to deliver a liquid versus taking it to an air handler that's going to deliver air. So a lot of the same common infrastructure. It's all about how you manage and flex that central infrastructure. That's really going to be the challenge that we're all going to face in managing increased density.

I want to dig in on that with you, and we'll get to it in a little bit. But I like this. This is fun. But I did want to talk to you about the phases. You made the comment that you're changing the way that you're looking at phases. We talked about the build sizes increasing. We talked about how the densities are evolving. But as I think about the capacity delivery schedule for Equinix, I build my model based on the number of cabinets and those are the phases that you have for your sites. So how are you thinking about changing either the quantum of cabinets or the quantum of megawatts that you deliver in a single phase?

Yes. So I'll start by saying we always think about power. We think less about it's this number of cabinets because a given phase is going to have so much power that it's going to be able to support. And how many cabs you can fit in that is a function of the density of those cabs is just simple math. And so we're going to increase the size of any phase that we do from a power standpoint. And I would say even in medium-sized markets where demand may not be the same, we're going to be 10 or higher. And on the upper end, we're -- in Ashburn, for example, which is one of our highest demand locations, we're building at 50 in the retail space at a time now. There's no -- it wouldn't make any sense to do 10 in Ashburn.

One of the things that I think about it, and I'm going to put this to you because obviously, CapEx has come into focus, right, particularly after the Analyst Day. I want to take what you said and presented to you, and I'd love your thoughts. So the way I think about it is that we've seen enterprise deal sizes increase. And it feels like they've been increasing at an accelerating rate in the last year. So if you're building just-in-time inventory, well, you're going to have to flex up in terms of your builds in the future in order to be able to handle continued increases in deal sizes. But then there's a second, I'd call it a twin engine. There's another dynamic, which is that, look, we're seeing the hyperscalers start to scale inference, right?

And I think there's an expectation that there will be a hybrid cloud model or a hybrid model just like we saw with cloud. And you need to be positioned with capacity in order to service that demand when it comes, and those two things are driving your CapEx higher. Do you think that is a correct framework for thinking about the Build Bolder initiative and kind of the increase in CapEx that you're forecasting?

Absolutely. I mean, for us, it's about all of the above. We're not counting and banking on any one trend to be the sole reason that drives our capacity needs. Obviously, AI is going to play a factor. We think inferencing is probably likely to be our sweet spot. But we also see a lot of digital transformation happening across the enterprise broadly, and that's what's been leading up to the demand inflection that we're currently seeing for the larger footprint. I don't think AI is that prevalent in terms of infrastructure deployment today widely across the enterprise. But it's coming. And obviously, we want to be prepared for it, and we want to anticipate it. And so we're going to lean into that. But that's not the only use case or scenario that's driving demand today for sure.

Okay. I do want to -- I want to transition a little bit, talk about xScale, some of the hyperscale trends. My understanding is that you oversee both the retail as well as xScale construction. Is that -- do I have that right?

Correct.

Okay. What I'm curious about is, are there any big differences that you'd call out in terms of the construction management process or the go-to-market in terms of development aside from larger, higher density, right? Any differences that you'd call out in terms of the go-to-market for building that capacity?

I think the biggest -- there's two big differences. One is just the sheer size and scale of that project typically. And the second revolves around whether there's a customer in the equation at the time we're building it or not.

Okay.

And when there is a customer, it's a much more demanding program management challenge. And we're -- I think what you're poking at, and I can sort of read between the lines, it takes a different caliber of program management talent and capability to manage one of those projects versus a retail project, particularly with how we approached it in the past. However, those two lines are going to merge because we're going to build much bigger retail buildings and phases. And so everything is going to be larger scale, which means we have to increase our program management capability. And we're adding lots of talent, lots of people in our design and construction team to prepare and anticipate for that. And we're changing something we call gearing. And gearing for us is the ratio of program management people to the project.

We used to run it fairly lightly, and we're realizing that given today's complexity, I think this will resonate with most of the people in the room that are building. It is much harder to build today than, say, 5 years and much harder than 10 years ago. Everything about today is harder, getting equipment on time, failure rates, labor supply, productivity, permitting, regulation, design, everything about how we build today is harder than it used to be. And so for that reason as well as just we want to be on top of the programs in a tighter fashion, we're increasing everything about how we manage projects.

Where I wanted to go with this is, I'm curious, and I appreciate the answer in terms of, hey, on the retail side, the sizes are increasing, more akin to hyperscale, but then we're also seeing hyperscale. We weren't talking about gigawatt deals 2 years ago, right? So the whole continuum has moved up. I'm curious, as that dynamic plays out, do you think it's increasingly appropriate to maybe separate out those two functions and have like a dedicated hyperscale build team versus having a retail team?

Short answer is it depends on what layer of the organization you're talking about. And so what we do today and what we believe is the right answer, and that's not necessarily the answer for the ages, we'll keep evaluating is that the programmed teams are separate and dedicated per project. And we have a group of projects in any given location that have this dedicated team that's more capable as well as experienced with those types of projects on an ongoing basis because we don't build once we continue to build in any given market.

But as it comes up the tower of the organization, we think having it unified from a supply chain, from a contractor standpoint, from a planning standpoint, from a tool standpoint, from a sort of leadership and design perspective, we think it all still makes sense to merge at the upper layers of the organization. That doesn't mean you don't need unique capability on the ground, which is really where we're making the changes.

Okay. Perfect. For xScale, we've seen you move bigger scale, go into Atlanta. It seems like there are other markets on the horizon. Now that you're building these large hyperscale campuses, I'm curious how the engagement with the utility changes, right? Because when you're procuring 10 megawatts or 30 megawatts, it's one conversation. When you say 200, all of a sudden, it's a completely different world, right? So how has that go-to-market evolved?

Yes. We used to go plan where we wanted to build a data center, secure that land and then we'd make an application for that power. And when we were talking 10, 15 megawatts, it was a question of time and time was defined in a year or 2. But you could feel confident you get that power. Those days are gone for two reasons. One, we're no longer building in those increments. And two, the amount of power we need isn't sitting around on the grid. And so we are planning, and I think most people in the room that are doing data center development are ensuring you have clear line of sight to that power before you take down any land or plan any data center capacity. And that might mean a longer cycle, a longer planning horizon with the utility, and it may mean a different level of infrastructure.

And in the past, if you were taking down 10, 15 megawatts, you literally -- you basically had maybe a step-down transformer inside of your building and you had medium voltage delivered to your front door. It was connected to a transformer that maybe you owned, maybe the utility owned, and it was fairly straightforward. Today, if we go ask a utility for 100, 200, 300, 400, what you name the number, we're having to one plan differently. We're having to put different infrastructure in. We're building our own substations that we either own or we build and then turn over to the utility. And for the most part, we're connecting at the high transmission voltage level. So it's a very different proposition. And even then, we may not get utility grid fast enough. And therefore, we're looking at alternates such as on-site generation, either as a bridge or as the first phase and then wait for the utility for the second phase.

I think in today's world, you have to be super flexible, multipronged with your energy approach, and that's what we're trying to do. There's -- somebody asked me earlier in one of our private sessions is do you feel like you have it under control? And do you feel at ease with energy? And I was like, are you kidding? I don't think anyone can do it. There's not going to be a period. The same with supply chain, right? It's going to be hard work, grinding, managing. We've built a very large energy team within the company to do nothing but manage these various dimensions, variables, dynamics, changes is the -- I mean the days of I'm going to order something for utility and I'll wait 2 years and hopefully it arrives, forget it.

Yes, that doesn’t work.

We're structuring a special deal with that utility, and we're going to monitor every milestone along the way to ensure that in 2, 3 years, whatever the time horizon is, it's actually going to arrive. And if there are key milestones missed along the way, we're going to be engaging with the utility. And so it's going to take a whole different level of power planning and engagement and discussion. And public utility regulation is changing around. If you don't use it, you're going to lose it. And we're seeing that model emerge in a number of markets.

And so monitoring and keeping in close dialogue with the utility, so you understand what power they've actually allocated to you because there's -- you have to think about power, I guess, in the simplest of terms in two sort of relevant variables. One is the infrastructure that they brought to you has a capability to it, has a capacity to it. But there's power that sits behind that, that has to be managed by the utility. And if they're not planning for the infrastructure or to deliver the amount of power that's in that sort of connection, then you get out of balance really easily. And we've literally seen that.

And so you have to be in constant dialogue with the utility, "Hey, I've got a 20-megawatt connection. I'm using whatever, 12. Are you reserving the other 8 for me? Or here's my projection of when I'm going to ramp into that." So capacity planning for the utilities is no easy task these days. For the longest time, the utilities load look like this. And then with electrification coming to fruition, now they're seeing upticks that they have to get on top of from a planning and capacity standpoint.

I want to take what you just said, and let's take a step further. So as you think about the retail buildings that you're building, I appreciate that when you're doing 100, 200, 300-megawatt campuses on the hyperscale side, that's the conversation.

I want to touch on that point, but go ahead.

So one of the things that when we're talking about hyperscale, we hear mandatory minimum commits, clawbacks on power, cost and construction where you're -- to what you just said, you're putting the money up for either a transformer or maybe even some of the transmission upgrades. What I'm curious is, at the retail level, are you being required to do that in order to get the 60 megawatts of power in Northern Virginia? Because if the answer is yes, then that introduces obviously, a new set of questions, particularly if there are mandatory minimum commits involved.

The short answer is yes because of the increments you're talking about. So if you're building a 60-megawatt xScale or hyperscale facility or a 60-megawatt retail facility, the power dynamic to get power to that location is the same. The ramp-up might be a little different, but otherwise, the arrangement you make is one similar. This is actually a good segue to the point I wanted to make, which is if you look at what we're doing in Hampton, which I know you're very familiar with, that is a -- what we refer to as a hybrid location, which means we'll have retail and xScale in the same campus.

So we got 4 buildings to work with. The story of how those 4 buildings will be used is to be told yet. And we have complete flexibility. We may end up with 2 xScale, 2 retail. We may end up with 1 retail, 3 xScale or any scenario, any combination you can think of. We're maintaining complete flexibility with how we use that infrastructure.

So surprise, surprise. I was having trouble sleeping last night because I was excited about the conference. But one of the things I was thinking about is the Hampton site. And I was thinking, well, look, CapEx is obviously a focus for people. You have a JV partner that is funding, let's say, 80% or 75% in the new JV, 25% for you. Is there a world in which it actually makes sense from being capital light to lease capacity from -- that's built by xScale for use on the retail platform?

Yes. And that would be -- there are various scenarios out there. I'll paint the picture to the two most likely and easiest to wrap your head around. One is the entire property and the infrastructure and the buildings are owned and on the JV balance sheet, in which case, Equinix would lease a building or buildings. The other is when we go master plan, you'll see more of these cases coming, not to lead too much into it, but there's -- when we master plan a large campus, where we may buy a large plot of land, we may plan the upfront utility in a common way, then we may take that parcel and we may split it and some will be xScale on the JV books and some will be on Equinix's balance sheet. But master plan together, utility solution together and again, gives us utmost flexibility in terms of how we manage that and probably a little bit more capital efficient, too.

I mean that, to me, goes back to the core proposition or at least how I remember it being presented to me of xScale, which is that there is value in having the compute node sit on the same campus as the network node and the interplay between the two. It's like -- it's the same idea, just at a large campus scale is kind of how I'm interpreting it with some additive benefits of capital efficiency and you being able to have one go-to-market motion with the utility.

The utility, the supply chain, capacity management, all of those benefits would be there. The ability to attract more network providers to that campus. It just -- it's got the full networking effect proposition, the more we have multi-tenant data centers as well as hyperscale deployments all in that same general -- it's actually more operationally efficient, too.

One of the things we're looking hard at now is, okay, in this new world where I've got a campus that supports 240 megawatts, what does it take to run that? What is the operating model? How do we manage deliveries? How do we manage the infrastructure, how do we maintain it? For sure, I can't -- I'm not going to quote numbers, but for sure, the number of people needed on a unit basis is dramatically lower in that environment versus I've got a stand-alone 20-megawatt data center and the number of people needed to run that, again, on a unit basis is much more efficient.

Got it. I wanted to ask you, we've talked about Hampton. Have you secured power for that site?

We have. We have.

Okay. And is that from Georgia Power?

It's a -- so the way there's a co-op down there called EMC, I think is the acronym. And it's a regional player that fronts Georgia Power. And so it's a complex system down there, and people are going to run into this in numerous locations, the more you get out into the outskirt. But basically, they're the local distribution and what sits behind them are the transmission line and the generation. And there are three parties involved. You interface with the ultimate provider, which is EMC in this case, and they go and secure the transmission and the generation to deliver you that local power. So you strike a deal, but you have to be aware of what's happening behind that front line.

That's what I was going to ask is that my understanding is the cooperative, they don't procure a generator. They don't build generation.

They were building.

They rely on somebody else. Then the question becomes, do they have the power from Georgia Power?

Yes.

Okay. All right. That's great to hear.

We wouldn't do that deal if we didn't have confidence that the full system is...

Kudos. Okay. Cool. Now I want to transition in the time we have left. I want to talk about...

I think you misled the crowd. It's time for Q&A.

I'll ask one and then we'll do Q&A. All right. We'll make it fun. So here's my question for you. Rack densities continue to increase. One of the questions I get from investors is about the installed base of data centers. This came up in the enterprise panel earlier where we were talking about the potential need to retrofit. How do you think about that? And then I know that you've undertaken a project, I want to say this was 2 years ago to start getting your -- to get your existing sites ready to support liquid cooling. I also want to talk about that and how that has gone. So how do you think about retrofit? And then what are you doing in terms of supporting liquid cooling capabilities within the existing fleet of data centers?

Yes. So again, remember what liquid cooling is. It is taking the chilled water from a central plant and distributing it to it in a different way than just taking it to an area. And so the retrofit we're doing to enable liquid cooling is tapping into that chilled water system to extend it to offer liquid cooling. So literally, you have to just tap into that pipe. And as long as you have sufficient central plant capacity, then you can leverage that.

Now we can upgrade the central plant as necessary. But keep in mind, in the data center environment, broadly speaking, there is a lot of headroom in terms of what people buy and what they use. And so we're going to continue to push the boundary on what you can utilize. And in all the locations that we've described, we have spare chilled water central plant capacity to tap into. And if need be, we'll add more chillers, we'll add more, what have you, in cases where that's possible. But the retrofit isn't as dramatic as it sounds because we're just tapping into existing infrastructure and what we're really trying to do, and just to put it in context, is we're trying to monetize the pockets of space that aren't used. Most of these centers are fairly well occupied in the 80-plus percent range, right? So there are some pockets that...

Is that on a cabinet basis or on a square footage basis or on a power basis?

That is on a space basis.

On a square footage basis. Okay.

On a space basis. The power is actually lower than that because there's lots of headroom. People when -- the typical system is designed and engineered somebody is going to plan for, let's just say, I'm going to take down 100 kW. Well, I'm going to leave 20, 30, 40 kW of that for engineering headroom because engineers are conservative as well as growth. And so we monitor that trajectory and that load. And it's actually about half of what is consumed.

Okay. Well, I said I'd open up for questions. So any questions from the audience? Yes, please.

Thanks for the comments, Raouf. Great stuff. Curious, are you starting to see -- we're starting to see some of the utilities, Dominion, Oncore start these proceedings around large electrical loads, how they can impact the grid based on SAGs, based on the oscillation that we're seeing with these GPUs. And nothing's come out yet, but just wondering how you guys are thinking about that and how you guys are thinking about kind of incorporating that in terms of your offering and kind of protect yourself and all that stuff.

And just to be clear, that's -- you're talking about like load spiking with GPUs.

That's correct.

Okay. That was gonna be my next question, so thank you.

I think early days to understand how that's going to work in [indiscernible] whether if you had a center that's completely filled with that versus it's a blend of that plus other things and the load is somewhat stable and static. I think we're actually more worried about it down at the low -- the next level down at the UPS level because that's what's going to tie directly to that load. And so I think a lot of work is being done to evaluate how that's going to play out. And I'd say it's early days. I actually thought you were getting at a different dynamic, which is historically, what's happened is if you were going to take down a bunch of capacity off of the grid and incremental infrastructure was needed, the capital required to do that was spread across the rate base.

Well, I think what we're seeing more and more of is, well, that's not going to work. That's not going to fly. The rate base isn't going to just stand by for that. The PUC is not going to stand by for that. And so players in the data center space are going to have to step up to underwriting that infrastructure in various different ways, and that's playing out real time as well. So I think there's a lot of dynamic dimension to what's happening in the energy space right now, including how to manage the future world of spikes and variables and the like.

All right. Well, with that, I'm looking at the back of the room. It looks like we're out of time. Raouf, it was my pleasure.

Didn't leave time for Q&A.

I did. Finally, I did. Thank you very much, Raouf, for joining us.

And thank you very much, everyone, for being here.

Thank you. Thank you for that.

Good afternoon, everybody. My name is Brandon Nispel. I cover data centers and tower REITs for KeyBanc. Thank you for coming to the KeyBanc Technology Leadership Forum. This is a 25-minute fireside chat. We have Equinix with us. Tiffany Osias, Managing Director of xScale and Chip Newcom, Head of Investor Relations. Thank you both for being here.

Yes. Thanks for having us.

Thanks for having us.

Well, I'll do our quick disclaimer for everyone here and online. Some of the things we're going to say today are forward-looking in nature. Please see our SEC disclosures for any risks and uncertainties. So that out of the way, now we can actually talk to the fun stuff.

Great. Tiffany, I want to start with you. Managing Director of xScale. Sort of a new business for Equinix. Maybe level set us, what's your background? And what is xScale?

Yes. So my background, I've been at Equinix now for 9 years, and I actually came from our retail colocation business first. So those that know Equinix know that we have a couple of 1.5 decades history in global colocation services where we have data centers all around the world. And then, call it, 8 years ago or so, our largest, most strategic customers asked us to enter more of a wholesale space because they wanted to be able to -- for us to provide them much larger footprints. And that was the inception of xScale. So on the retail side, I think lots of customers, we have 10,000 customers. So lots of customers all occupying or colocating in the same building. Think on the xScale side, much fewer in number of customers and much bigger footprints.

What was attractive about sort of expanding into maybe more of the hyperscale customers that you couldn't address previously?

Yes. I mean the great part is that we were addressing those hyperscaler customers mostly from their network nodes. So Equinix has more than 35% of the global cloud on-ramps. So the hyperscalers were looking to Equinix to put in all those interconnectivity devices but hadn't been able to yet put their core nodes with us. And that's why we entered xScale about 8 years ago in EMEA and in Asia, in particular.

And I'd add to part of why we -- the genesis behind xScale was also about capital efficiency. And so the fact that within our retail business, we're targeting yields in roughly the 25% cash-on-cash ZIP code. xScale is different, but then when you add on leverage plus then the fee income that we get, it becomes a very attractive return for us and also in a capital-efficient manner.

Can you unpack sort of like where you stand today from an xScale standpoint? I know a lot of these assets sit in sort of a JV structure. Maybe help us understand what all of those JVs look like today.

Yes. So we have a handful -- over a handful of JVs operating today with capital partners, current capital partners of GIC, PGIM and now our newest CPPIB, where Equinix comes in with a 20% to 25% interest in the venture as a limited partner and then one of those other capital partners brings in the remainder of it. So as a 20% to 25% LP, we have equity investment and interest in it. And then the joint venture itself hires Equinix to then go run the venture, construct and operate and lease the actual asset as well.

And you've done different JVs, sometimes with the same JV partner. How do you look at sort of identifying different JV partners? Like what makes a good partner to you?

Well, I think it's like-minded companies who want long-term investments and want to help promote more digital infrastructure growth. That's what we've looked for historically.

Okay. Any consideration of like moving beyond sort of like just getting capital partners and more strategic partners? Do you view them as more strategic partners? Because you guys do a lot of the operations, right?

Yes. We do a lot of the operations, but they're also strategic partners. I mean, look, they're putting our trust, their trust in us. They're putting multibillions of dollars of capital to work, and they trust Equinix to go run these data centers appropriately. So they're still very strategic for us.

Okay. Let's talk about some of the developments around JVs and xScale. I think up until late last year, you primarily focused on xScale outside of the U.S. Late last year, you announced a $15 billion xScale JV in the U.S. Where are you with that?

Yes. So we're actively now -- we've announced one piece of land that we've closed on in Hampton, Georgia. We're actively looking at other campuses as well. But the ambition is to have a handful of multi-hundred megawatt campuses across the Americas region.

Okay. Multi-hundred megawatts. And how did you take sort of learnings from xScale 1.0 and bring it into xScale 2.0 into the Americas region.

Well, given that 1.0 is, call it, 8 years old, the densities and the size of the deployments were a bit smaller than where we are today. xScale 2.0, as I mentioned, they're going to be multi-hundred megawatt campuses and customers will be buying them in, call it, 30-plus megawatt deployments. where in 1.0, customers have the ability to go down to a couple of 5, 10, 15-megawatt deployments.

And is the -- when you develop a 100-megawatt facilities, is it one customer that's taking down that whole facility? Is it multiple customers? How do you sort of think about like pre-leasing some of those new facilities?

Yes, it could be either. It could be one customer could take down an entire campus or it could be that we end up with a multitude of customers.

And you've announced Atlanta, I think, as sort of your first campus. Where are you in terms of construction around that campus?

Yes. We're moving a lot of dirt around right now is what I would say. It's several hundred acres where we're leveling and flattening the land today. We have bulldozers pushing dirt every day and getting it pad ready.

Empower is committed to the pad.

Empower, we've been working with CEGMC, who's the regional utility that we're working with.

Okay. How do you think about sort of the development of other facilities across the U.S. Where does the land come from? Do you already have it in land bank and you're sort of trying to identify either the power and get all the permitting and zoning constructed? But where are you sort of on the other JVs?

Yes. Across all of our ventures, it's been a combination of either taking what we already had in land bank and contributing it into the venture to develop or it's going out and actively pursuing new parcels of land and then securing power for that. So we have both of those across all of the ventures.

And do you think about xScale in the U.S. as sort of like sticking to top markets? I know we've seen some deals with Bitcoin miners, right, that are doing 250-megawatt deals in the middle of North Dakota. What's the value of sticking to the top 10 markets versus move -- maybe moving beyond that?

Well, the way we think about it is generationally. We've already got a 27-year operational history at Equinix, and we plan to continue to be in this space for decades to come. So we're not thinking about meeting one trend in a moment in time. We're thinking about durability and creating a platform that can capture multiple trends over time. So if we think about that, a, we want durability. B is we have a flexible data center design that could capture AI training, AI inferencing, cloud or enterprise kinds of workloads. And if you're going to capture all of those workloads, you need to think about latency as well. So those top metros are still important to us. It doesn't mean we have to be in downtown Atlanta, like you're seeing play out here with this campus. We're about 40 miles away, but we're still staying very close to those metro centers because the interconnection back into Equinix is also really important for the success of xScale.

And Equinix has historically contributed assets into these JVs. What needs to happen to these assets that were probably traditional retail or enterprise facilities to turn them into an xScale property?

Yes. For those that we did historically, it was about having the right amount of capacity and the right kind of data hole that could be accommodated into a customer who needs that kind of space. Going forward, it will be a lot more about selecting the right kind of land that we would be able to develop on.

Okay. Yes. And I'd add, historically, the reason why we were contributing assets in as well is that we didn't want to slow down our momentum, right? And so even as we think about what we're doing right now, we would much rather go out and get started to be developing on our own balance sheet like we are right now, where we said at our most recent earnings call, we're spending about $450 million or we expect to this year on balance sheet for xScale but it's because we see the opportunity in the market, and we want to keep developing, recognizing at some point, we will move that into a JV and then get paid back for our spend.

And a lot of the -- is the development around hyperscale, is it more spec? Or do you have like an anchor tenant already sort of in mind for a facility?

Yes, we end up with both. We end up with both. We end up in a lot of instances where we do have an anchor tenant that has a start development, but we also have learned, having done this now for 8 years, that the closer that you are to a customer needing a service, the more important it is that you have it available on time. And many times, the price point has also increased at that point. So we do both.

As you've learned from sort of xScale 1.0 and move to 2.0, I would assume some designs and architectures of the buildings that you're constructing have changed. Can you help us understand like how you're designing the buildings today for the Americas JV versus what you might have done a couple of years ago? And maybe specifically focusing on like liquid cooling, power densities and just like overall resiliency of the facility?

Yes. I mean you've nailed on the things that are mattering most right now. So having such a long history with the hyperscalers and some of the other strategic SaaS operators that need this kind of capacity, we've been working with them for years on what their designs are going to look like. We do very regular design reviews, and we also go work with all the equipment manufacturers like NVIDIA and Dell and HPE and others.

So we know about when technology infrastructure is going to converge. Liquid cooling is certainly one of the big aspects of the design that's different right now in the new xScale 2.0 data centers across the Americas, they're coming up right away day 1 of having the ability to be air cooled in a narrow amount and liquid in a larger amount or the exact inverse of that, depending on where these customers are in their journey and transition to liquid cooling and then certainly density as well.

There's a great deal of flexibility from what we used to see in the, call it, 10 kilowatts per cabinet for these kinds of deployments all the way up to, call it, more than a megawatt per cabinet.

And what do you think of like construction time lines on -- let's talk about maybe Atlanta facility, but then establishing other facilities down the road?

Yes. I think first, it's about getting a pad ready, and that's a big part of the long component of construction is getting it ready for the first building. But then after that, you'd see the construction time line accelerate because you've done a lot of the upfront work that's required. So think like 18 to 24 months to get it pad ready and have the first building constructed. But then after that, you'd be able to see something pretty quickly.

Okay. Let's move on. How do you monetize xScale? I think that's one thing that I don't fully understand. What's sort of the monetization strategy of sort of the xScale buildings that you're constructing?

Yes. I mean there are 2 ways to think about that. One is when it's actually leased and a customer is in it and occupying it, the customer is paying for the service. That's one way you would monetize it.

Then that's like a recurring revenue stream.