Equinix, Inc. (EQIX) Earnings Call Transcript & Summary

All right. Thank you, everyone. I'm Sami Badri with Credit Suisse Equity Research. And right now, we have the Equinix team, we have Jim Poole, VP of Business Development, and we have Katie Morgan from the IR team. And before we get started, Katie, hand it over to you.

Thank you, Sami, for having us. For some of statements, we may make today are forward-looking nature, so please check out our SEC filings for all the risk disclosures.

All right. Great.

Jim, maybe to kick things off, could we maybe hit on some key operating and strategic priorities that you are working on at Equinix? And if you'd like to also throw in like if some of these key priorities have been very prevalent or topical of 3Q earnings call, that was very recent. You can also kind of weave that in.

Sure. So number one would be continuing the investment in go-to-market. So for us, historically, that's been adding more quota-bearing heads, obviously, but the more relevant part would be our channel motion. So for the vast majority of our history, we were a direct selling organization that has started to change. And so you'll hear in like the earnings transcripts that we're up to now 60% of new logos are through -- our channel partners. And so there's a lot of sort of automation that needs to be put in place to make that scale. So that's a big thing. So that would be number one. Number two would be sort of reach and scale of the platform from a geographic perspective. So it shouldn't surprise anybody that when we open a facility, we open it in proximity to large percentages of people, which generally then gets you GDP. And so we're on a constant basis, looking and evaluating whether we expand. So most recently, you probably will have noticed, we made a purchase from Intel in South America. So that got us into Chile and into Lima, Peru. So 2 more countries there. In Africa, several months ago, we made a purchase of a company called MainOne, which got us into 3 countries in West Africa. And as early as this morning, we announced that we are organically going into Malaysia. So sort of 15 kilometers across the bridge from Singapore, where we're the dominant provider there. So that will be our first presence there. But in those particular markets, South America, ASEAN and Africa are places we continue to evaluate on a constant basis. And then the third thing would be just in sort of our capabilities from a platform perspective. And so the 2 most relevant things there that we discuss are the adoption of Fabric by our customers as an automated interconnect platform to facilitate that sort of more dynamic environment that exists inside of our facilities. And the second thing would be our metal service, which is essentially based off of the Packet acquisition several years ago. And so that has continued to evolve. And so we're selling that across both enterprise and service provider verticals.

Got it. Thanks for running through that. On the third quarter conference call, management talked about record booking activity or just very strong enterprise activity. Could we just talk about how is the pipeline progressing today? And even what are expectations for 2023?

So yes, so we don't guide outside of what we just recently said in 3Q, but to reiterate Keith's point, it's strong, and so we feel very good about it. And so I'll leave it at that.

Okay. Let's talk about maybe something that's a bit more of your focus, which is 5G, right? And 5G and telecom service provider cores and upgrades and this has essentially been a theme for or at least mentioned as a theme for almost 3 to 4 years. We haven't quite seen the rubber really meet the road or hit the road really per se is now really about that time, we're going to see more of that.

Yes. I think what you'll see, at least in our space is a couple of things. And the whole 5G rollout, it's a lot more complex than the previous generations of wireless, which even they took multiple years to fully roll out. So the biggest thing that the operators have to deal with currently is what they refer to as kind of disaggregation, which is essentially putting breakout functions so that you can steer traffic off the radio network to an application execution venue of your choice. So that's how you get low latency. It's not magic, it's speed of light, right? And so the things have to sit closely together. In 4G, the breakout points are fixed, all the traffic is dumped into the public network, and that's not the purpose of 5G, which is, of course, to deliver business services. So the first thing they did was upgrade the radios and introduced new spectrum. That happened and everybody got a new phone. The second thing, of course, that happened was then doing what they call stand-alone, which was a software upgrade. And so now they're actually physically capable from an execution perspective to do those breakout functions in you could say, arbitrary places. Now obviously, you don't do it arbitrarily. And it would probably make a whole lot of sense that the first places you would do that are places that are large aggregators of enterprise and public IT infrastructure, which would be places like that. And so once that happens, then what you'll start to see is driving toward what the industry refers to as MEC, right, which is multi-edge access compute. So -- and that's where things like our metal starts to become even more relevant, right? Because the idea would be as a wireless business customer, I, as the operator, have no idea which counterparty they want to connect to using that ability to steer traffic. And so that ability to do that dynamically to get to the applications that, that user wants is something that we're heavily involved with them already on the wireline side. Most of the operators use our fabric service as their primary aggregation method into public cloud. And so that will continue to play out on the wireless side.

So on one hand, we have a wireless kind of driver, which is 5G and this kind of cadence to make some changes to the network and make upgrades and improve latencies. And then you have the public cloud providers talking about dynamic slightly slowing down. How do we kind of put all these things into context?

Yes. I mean I think even with what we're seeing in the market today from a financial perspective, I think if you took the top 4 and taking out the currency issues. You're still going to see somewhere between $12 billion and $14 billion in revenue growth, which is not chump change on anybody's. So I think what you see is that in the hosting of application markets, you have extremes, right? On the one side, you have things that are completely consumption-based and virtualized. That's more the SMB market. We don't play in that end of the market. At the complete other end of the market, you have the wholesale market selling tens if not hundreds of megawatts to the hyperscalers, and that's a market that we only have exposure to through our xScale program, which is an off-balance sheet program. Where we're focused as a company is more on sort of that multinational company that's got $1 billion in sales or more and operates in 3 regions around the world, right? That's the sweet spot for Equinix. And the way to put that in context would be digital transformation for companies, tends to -- and we've got tons of data to show this is results in what we refer to as hybrid multi-cloud, which is just the truism that as an enterprise, I use multiple networks, I use multiple cloud providers, and I maintain some portion of my infrastructure on a private basis, and all 3 of those things have to interact seamlessly. And that is not a nontrivial problem. And so having facilities such as our IBXs that essentially have all those counterparties available and connected to things like our Fabric service so that you can then orchestrate that connectivity between those counter domains becomes hugely relevant to an enterprise as they go through that transformation process. And what we do for our enterprise customers is typically 1 of 2 things. One is that whole journey to cloud is a cost savings, right? And the component of what we do for that is fixed. So whether they put 1 meg through the system that sits with us or they put 200 gig to the system with us, their bill is the same, right? So it's sort of table stakes to play the game, right? But on the other side, mostly digital transformation is all about buying and selling. And so one of the most active dynamics from an ecosystem perspective is people inside of our facilities buying and selling services from each other without having to put anything on a public or a private network. They can just do that within the facility. They do it with very low latency and essentially fixed low cost so.

Got it. Got it. One thing about Equinix that's clearly evolving is overall network density, right? It's been shifting largely favorably for the company. But when we go from 2022 to 2023, what's the next stage of the evolution for network density?

Yes, that's a great -- so it's interesting, and a lot of it is tied to our evolution of the Fabric product, right? So the Fabric product started off as what we call the Ethernet exchange. So it was trying to link 2 networks -- Ethernet networks together that evolve to Cloud Exchange, where we started aggregating private network access into public cloud. And then we finally ended up with where we are today, which is what we refer to as Fabric, which is essentially arbitrarily any 2 counterparties within the platform. It doesn't matter if one is a network and one's a cloud. It's all an interconnection problem, and it's one that is a dynamic problem. And so from a network density perspective, what we do with the operators are increasingly is we're the primary third-party aggregator of access into cloud already, and we're starting to basically expand the number of use cases across the networks that they use Fabric for. So just to give you an example, you're an American operator, your primary PoP is in Sydney, in Australia. However, you sell something to somebody in Perth. I can either wait 90 days to order a type 2 circuit from the local incumbent or I can do that automatically in a couple of minutes from us. And once I'm done using it, I can then switch over to what I probably buy on a more favorable price basis from the other guys. So what we're trying to do with the operators is evolve their ability to be more agile by sort of taking our interconnection value proposition and tying it to what they can do as far as things like their 5G investments.

Got it. Got it. I want to shift gears a little bit and talk about power costs. So given the headlines been basically the most topical thing for 2022, at least, could you just walk us through or remind us about your power hedging strategy as a corporate? And then just maybe even specifically talking about unregulated power markets where power rates can be a lot more volatile?

Sure.

Yes, I'm happy to take that. Thank you, Sami. So as you know that we operate across [ several ] markets around the world. And so we operate both regulated and deregulated power markets. On the regulated side of the equation, we are -- that's really just a take-and-pay environment. So that rate is set by the local utility or municipality. And so we don't -- we can't hedge in those markets. On the deregulated side of the equation, we do have the ability to go and proactively hedge our exposures. And it's very similar to what we do on the currency side of the house, where we progressively layer into our hedge position kind of on a rolling 8-quarter basis. And our goal with our hedging strategy is to really mitigate the volatility for our customers. So when they open their Equinix bill each month, they know exactly what it's going to be. We're not trying to pick where we think the spot market is going to be a year or 2 years from today. We're just trying to mitigate that volatility. And so as it relates to 2023, we continue to layer in our hedge positions. We're over 90% hedged globally. We're 97% hedged in EMEA for 2023. And what that is allowing us to do is to communicate to our customers in advance of raising prices. So we've really developed a comprehensive communication plan to our customers to give them advanced notification to be able to plan for those price increases for 2023.

Got it. And then maybe just to kind of refresh people's memories, you guys have already kind of sent out those letters to customers. And is there any kind of comments that you guys have made about acceptance of prices? Or once the customer sees it, is this kind of like expected, is it being ingested? Is there -- like could you give us any kind of update on that?

Yes. So I would say for -- just to kind of give some context, the way we communicated this out to customers, there's really 3 customer communications. The first one was to give customers just a general heads up that utility costs are going up just as you personally feel as a consumer, and then we will be raising utility prices as we look to 2023. The second one was to range bound it for customers to give them kind of the expectation of the power price increase, so they could plan in accordance with their budgets. And then the third was to give them their final increase, which will take place -- take effect on January 1. And I'd say, overall, we feel that the rates that we can offer customers are attractive, both relative to what they might be seeing the prevailing spot market as well as what they might be seeing from a competitive standpoint as well.

Got it. Got it. Going back to the third quarter results. Pricing was very strong in the quarter. If you were to describe why the strength was as prevalent as it was, is it pricing power? Is it inflationary drivers? Is it network density or requirements? How would you kind of characterize the strength of pricing in 3Q?

Yes, I'd say looking at 3Q, I'd say pricing remains firm across the business, and we do see that we have latent pricing power within our business. I would say, number one, it's putting the right customer with the right application into the right footprint. As you've often heard us talk about with our churn metric as well is that when we think about our approach to pricing with our customers, it's really on the value that we deliver to them and what they can extract as being part of the platform. And so absolutely, it's -- we've adjusted our list pricing higher to reflect a broader inflationary environment as well, just as we're seeing construction costs go up. And then we've also adjusted pricing on new cross-connects over December, but those really haven't flowed through for 2022. Yes, certainly flow through over time.

And then maybe we could think about if public cloud drivers or demand starts to slow down. What is that -- what kind of effect would that have on Equinix's business or even pricing?

Yes. I'd say, as Jim mentioned, you saw the hyperscalers are still adding a significant amount of revenue expected to in 2023. And I'd say we set it at a different type of workload with our customers. Typically, it's not surprising in a belt-tightening environment that we're seeing some customers pull back on their usage-based applications, but we typically don't really have much of that within our business. We're mostly recurring revenue model business and what we offer our customers is typically mission-critical in what they're trying to do is either revenue-generating or revenue-facing systems typically. So really just continuing to see them focus on digital transformations at large.

Got it. Got it. I wanted to shift gears to another topic, which is something that's very front and center, given what's going on in the broader world economy. And whenever you hear about technology slowdowns, the first thing everyone looks at is churn, right? Churn for Equinix has actually been fairly low for the last, I believe it was 5 consecutive quarters. It's actually essentially normalize below where we're typically used to seeing a churn come in. Is this a new structural norm for Equinix? Or is this like a post-pandemic kind of -- post-COVID-19 final normalization of churn? Or is there a lot of uncertainty with churn in general, but given your pricing strength, it's been able to kind of offset any kind of the churn dynamics?

Yes. Good question on churn. I go back to the pricing comment where we're always really focused on putting the right customer with the right application into the right footprint. And so it really started at the beginning of our sales cycle where, in any given quarter, about 90% of our bookings come from the installed base. So customers landing and expanding what we're doing with us. And when we target new customers either directly through our direct sales force or through our channels, we have very developed propensity to buy models looking at if it's a global multinational customer. We're looking for those customers that want to be in either multiple metros or multiple regions or across all 3 regions with us. And we also, on the same side with churn, we also have a very predictive model and analytics around churn where we can -- we model that out and we look at that. So I'd say, overall, historically, as you've seen probably over the trailing, call it, 8 or 10 quarters, churn has trended lower. And that's partially attributable to when we acquired some of the legacy Verizon assets. We knew that there would be some churn associated with that. And so that's really lapped through the portfolio. And you've seen that in the Americas actually with the utilization, call it, over the past 18 months, stepping up from 72% utilized to 80% utilized, and that's just a reflection of the bookings momentum that we've seen within the Americas without having as much churn. So for Q4, we'd expect to be towards the lower end of the range on churn.

Got it. And then for the revenues that are churning off, typically, where are those workloads or applications or even business activities going to? Is this kind of customer rationalization of the footprint? Or is this customers disconnecting doing something different, right? What has been like the key source of like the reason to churn, at least out of the portfolio?

I'd say it's important to me with churn -- portion of our churn is where we call moves, adds, changes or delete. So it's just a reflection of that within our facilities, it's a living, breathing ecosystem where customers are constantly evolving what they're doing. So they may be adjusting what they're doing with us, for example, in Frankfurt because their end customers no longer needs that use case, but they also might be deploying with us, say, in London for a new application. So just a reflection of the constant evolution of our ecosystems.

Got you. Got you. I was going to shift gears over and talk about your services business. And well, maybe on service more like your Metal and Fabric business. So the big question I think people have is how deeply penetrated is the Fabric business into your actual customer base?

At this point, there are -- I'm trying to think with the note 20?

We're about 1/3 penetrated of our customer base.

Third penetrated?

Yes.

All right. Got it. And then when you think about -- so a third penetrated, just kind of rephrase the question I had for you guys. If you were to think about before Metal, right, there was a Fabric business before Metal. I think it was called ECX, right?

Correct.

When you look at ECX or Equinix Cloud Exchange versus Metal, right, is that 1/3 referring to from ECX or is that, in general, including ECX and Metal?

Yes, the way to think about that would be that the Metal service is addressable primarily through Fabric, right? We could do a cross connect to the platform, if that was a requirement by the customer, but most customers consume it on an automated basis. So a really good use case that we had recently made an announcement about that kind of puts it in context would be Orange Business Services. So traditionally, if you're a network operator, what you would do is buy 4 racks from us and then you go buy $1 million a kit and you'd wait 6 months to get it all done, and you would then have stranded capital for the time it took you to get the utilization rate of that stuff up to wherever you want it to be. That's just been the industry forever. Instead now, they can basically -- since they virtualize their entire software platform that faces their customers, they can take that software, they can drop it on the Metal. They've already been using our APIs because they are a customer on to Fabric. And so now essentially, they can do PoP as a service, and they can do that essentially as they sell. So we give them sort of an OpEx answer, a really quick time to market to satisfy customer demand. And then they still have the flexibility that if 3, 4, 5 years down the road, they wanted to hit an economic inflection point and go back to colo and populate it with their own gear. Well, it's going to happen in our facility because that's where all the interconnection happens, right? So that would be the way to think about it.

Got it. Got it. I wanted to shift gears. Actually, one question is when we think about 5G and the service providers, how much activity is 5G or service providers really using out of Metals -- at a metal and ECX. Are they the big users? Or is it predominantly coming from enterprises and clouds?

It's predominantly at this point, enterprises and clouds because to my earlier comment about sort of where the wireless operators are from an industry perspective, they haven't -- they've just gone through the process of doing their 5G stand-alone upgrades and they haven't actually gone through the disaggregation step, right? So they haven't gone through the step of saying, "Hey, I put all the software and hardware in the place is necessary to create the steering function." So that's literally probably going to happen relatively rapidly over the next 2 to 3 years. But right now, what we're mostly seeing is when we do -- just to balance it out, the service provider use case that I gave you for Orange would be sort of the way a service provider would use us today. And that's mostly a wireline use case, but that could just as easily be a wireless use case. It's the same exact underlying infrastructure from the operator's perspective, right? So for us, it's more, hey, have we demonstrated the capabilities and have we sold that capability to the customer? They can decide whether that's appropriate for wireless or others. On the enterprise side, it's more an idea that says the normal behavior of our customer is land and expand, if you're an enterprise. So you start in 2 metros, then you're in 10 and then you're in 20 and so on and so forth. The gating factor to that process has typically been the CapEx associated with the equipment, right? So we can basically accelerate their adoption because they still have these -- the 80-20 rule of 20% of my workload has to run on different hardware and not in public cloud. And I want that infrastructure to be interconnected to public cloud, but proximate to public cloud. So now we can essentially -- you can think of almost Metal as a colo proxy. I can make it possible for them to land and expand in a far more accelerated rate.

Got it. Got it. I'm going to open up to the audience, if they want to ask a question about 1 minute. One thing I want to talk about is supply chains, right? So you started -- you not necessarily have seen supply chains ease, but you're starting to hear that supply chains are about to ease or just stop getting worse. How is that going to affect Equinix's business?

Yes, I'm happy to take that one. I'd say from an overall supply chain perspective, we continue to keep a very close eye on it. We're very fortunate that we have a procurement team in place, and we're able to leverage our balance sheet to make forward purchase commitments to make sure we have the critical inventory available for our builds as well as being able to procure for future builds as well. So continuing to keep our eye on it and on supply chains.

Got it. Got it. And then I wanted to open up to the audience. Does anyone in the audience have a question for the Equinix team? All right. Okay. I want to shift gears over to overall wage inflation, right? How has the company been dealing with CPI, income adjustments, salaries, all those are starting to shift, right, and all of those kind of put pressure on a lot of companies operating expenses. How has that affected Equinix?

Yes. I'd say overall, we continue to keep our eye on this inflation generally and feel like we continue to watch that across all areas of the business, specifically relating to wage inflation. I'd say we do look that we do hire in different metros around the U.S. in accordance with different cost of living standards. So it's something we do keep our eye on.

Got it. I wanted to go back to the topic of supply chains and mainly talk about markets that have seen some disruption. And a good example is, say, Virginia, right, where the local utility has sounded the delay of delivery of certain power requirements. How has that affected Equinix? And are you seeing that same issue appear in other markets around the world?

Yes. I'd say specifically related to Northern Virginia, we feel like we can continue to manage through that situation. We've been in the Northern Virginia, the Ashburn market for over 20 years, and worked closely with Dominion during that time, and we're also working closely with the data center coalition to continue to understand the situation fully. I'd say with the power equipments we have in hand and our ability with our retail model to move power around to different facilities, we feel like we can continue to manage through that. But I'll also add as you walk around the world, there's definitely markets that are continuing to think differently and how they want to allocate their resources or the utilities. And so I would add that having a very robust sustainability strategy when we show up to our communities is increasingly becoming important of how we shop to our communities.

Got it. Got it. I wanted to also pivot to the topic of digital infrastructure convergence. So the one thing we've heard about now is tower operators are proactively deploying edge data centers and these sites have not popped up. Some of these tower operators have 30 of these now at their tower sites. How does that motion by the tower cohort really impact Equinix's business?

It doesn't. And I'd probably go back to my earlier comment about this step that the operators have to go through to sort of do the disaggregation and steer the traffic to where it wants to go, right? So you steer the traffic to where the business is, not to where you want the business to go, right? So the idea of sort of lift and shifting things into another place on an arbitrary basis. And to give you a context of -- because Edge is -- my joke is that it's the most misused word in networking technology today. But just to give you some perspective, Equinix's facilities, the IBX is in just the United States, sit within 10 milliseconds of 80% to 85% of the U.S. population. The average latency on a 4G network historically has been 40 to 50 milliseconds. And that has to do with that sort of fixed breakout scenario that all the 4G operators had to deal with. So right away as an operator, I can go from 50 milliseconds to 10 milliseconds. And the way I can do that is just by simply shifting the traffic and moving it toward an Equinix facility where, oh, by the way, all of my customers who are consuming my wired services also sit. And they're a very active part of our channel program. So we feel very, very good about the idea that says as Edge evolves, it is going to start to be beneficial to us as a business just in the sense of it's the logical place to go to first. Now this is not to say that eventually, will there be a compelling use case that says, we need 5 millisecond delivery latency across 100% of the U.S. population? That is certainly theoretically possible. I can pretty much tell you that every conversation I've ever had with any operator or anybody else in the application space, nobody knows what that is. And so until people understand that I wouldn't suspect that you would see people trying to figure out that particular problem. But our goal would be, as we learn with the operators, how to do that in the existing facilities using all the existing tools that we have already provided, we built essentially everything necessary to extend if necessary. And just to give you another sort of anecdotal thing, like just the idea of a modular data center. We actually have designed and deployed modular data centers, but we did it as a Swiss Army knife to solve multiple problems. So one thought we've had is that eventually that could be a 5G thing for more localized access. But right now, we've done it in generally, Italy, and we've done it in Bordeaux,, France's cable landing stations because the hyperscalers are huge investors in those systems, and they like to land them with us. And so from a supply chain, from all of the automation we use on Fabric and Metal in sense to allow the ecosystem to expand and potentially push out because that's what our customers already do on their own. We think we are very well positioned to be able to do that as that market develops, but I don't see it yet.

Got it. Got it. With this whole topic of adjacent asset classes and infrastructure asset classes, what has been Equinix's viewed? Does nothing really kind of compel you guys to change your business model? Or is it the economics don't work, what has been like the rationale to stay away from other asset classes?

I'd say we sit at a really unique spot at the intersection of technology and real estate and kind of sit at a very differentiated spot within the data center space. And so continuing to really focus on our sweet spot of highly interconnected retail colocation in the markets that we operate.

Yes. And I can give you at least another little just anecdotal thing is that digital infrastructure as automatedly, consumable components, things like Fabric, things like Metal allows all of our channel partners and managed service providers, network providers, systems integrators, everybody who does business with us to be able to take those piece parts and put them in whatever LEGO configuration necessary is to execute the business they're trying to do. That's a very powerful thing. I make the joke that it's like saying, point at Home Depot and say, that's where you buy your house, right? Well, yes, because all the pieces are in there. Very same sort of analogy. And we think that is a very durable, long-standing growth trajectory for the business that we're in.

Got you. Got you. I wanted to also talk about fundraising needs in 2023. And I think the genesis of the question is we are entering a world with higher and stickier higher rates, right? The data center asset class. I think this might be the first time where rates are this high for the data asset class to navigate. Could we just kind of hear about fundraising needs, if any, that Equinix has in 2023?

Yes. I'd say, overall, from a balance sheet perspective, we really view our balance sheet as a strategic differentiator for the company. As we ended Q3, we had $2.5 billion of cash on balance sheet. Our net debt to EBITDA was 3.5x. And I'd say we have one of the lowest AFFO payout ratios within the REIT space, which allows us to directly reinvest within the business. So continue to take a thoughtful approach to our balance sheet in terms of maintaining that strategic differentiation.

Got it. Got it. The last question I have before we wrap up actually is how do you guys generally think about share issuance, right? Is it -- in the last quarter, you did have an issuance of about, I think, $800 million. What was kind of the rationale behind that? And like, is that something we should just get used to seeing just because tapping debt as a source of funds has gone very costly? How should we think about equity as the alternative?

Yes. I'd say, overall, we continue to take a balanced approach of having both debt and equity within the business, and we have multiple tools available to us to continue to fund the business, whether that be our ATM program, which is very standard across REITs or just continuing to fund the business. So keeping a blend of both debt and equity within the business.

Got it. Got it. All right. Well, look, I appreciate your time. Thank you, everyone, for joining us today. Thank you again.

Thank you.

Thank you for having us.

Absolutely.