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

[Audio Gap]

[Audio Gap] forward-looking statements. Please read our SEC filings for more information about factors that could affect these statements.

Great. So it's been a busy morning for Equinix. And maybe we can start off with a few of the topics of today. There was an announcement indicating you're expanding the Equinix metal capabilities to enable more as-a-Service offerings and to include more hybrid multi-cloud architecture. Bill, maybe you could give us some of the details on what your goal is here and what this entails.

Sure. I mean, a lot of our metal strategy is all around getting sort of the right compute in the right place at the right time. And some of the -- what we announced was some capabilities with the combination of solutions from AMD, Ampere, Intel, Nutanix, but it's really all about deploying a sort of full stack accelerated computing platform with NVIDIA. And so with those abilities to have sort of the compute that different parts of the digital ecosystem want in the right place at the right time, it can enable some pretty interesting use cases. A couple of the ones that we included in the announcement that are good kind of general use cases of other adoption that we're seeing with Super League Gaming is a good use case. We think that as a digital service provider who wants to have infrastructure deploy, but they want to be using you there. If you think about it -- if your core competency is developing games, it's not about shipping servers and installing those servers. So it really allows these gaming companies to focus on their core competency, but still have the servers where they need them with the performance that they actually need. Another good example that was included was Solana, which, of course, is a blockchain company as well. So some good use cases. So good technology. We're going to continue to evolve the platform. Again, the idea is to get the right compute in the right place, at the right time, surround it with sort of an ecosystem of capabilities to unlock the kind of solutions that I've referenced with Super League Gaming and Solana.

Are these initiatives in direct response to client demand? Or are you anticipating needs that you think are going to develop over the next several years?

A little bit of both. And I think there's an interesting dynamic where if you look up and down, sort of the technology stack starting with even the silicon and going into the OEMs, a lot of these solutions that used to be delivered as an appliance or as a unit, that would be installed into a data center or in someone's basement, for example. They're now all looking on how they can change their model to continue to deliver what they've always delivered, but instead of shipping a box, being able to offer that more on an as-a-Service basis. So we are absolutely seeing customer demand from people like I mentioned Super League Gaming, their core competence is gaming. They want to focus on creating great games, not shipping equipment all around the world. So you have the confluence of good sort of demand from the customer side and then a supply dynamic where the suppliers need a new as-a-Service route to market. And we -- because sort of interconnection did see that in our locations, we're the natural place for that supply and demand to actually come together.

Great. And maybe just touching on a key theme here, there's [ displays ] coming on in 2022, have you been seeing any supply chain issues that could impact that server availability?

It hasn't been a walk in the park, but we think we're in good shape. We've done a lot of work to actually change our fulfillment, our supply chain model, where we now locally basically stage this equipment. And we're able to get it into the right market much faster. So we have a pretty good beat on what we think our demand for '22 is going to be. And we think we have more than enough capacity in place in order to fulfill that demand. So while I would be curious if I ever said that we have more than enough capacity and they're not going to be any issues, we think we're looking pretty good.

Great. And you seem to be making good progress with the full suite of Fabric, Metal and Network Edge. Can you give us any sense of what percent of your 10,000 customers are using each of these services and to what extent?

Yes. Just some high-level stats. And Katrina, you keep getting me honest here. But if you think of -- sort of the most mature product we have in that set, you're right, the interesting set of the combination of Metal, plus Network Edge, plus Fabric. Let me just spend 2 minutes for folks who might not know sort of what those products do, just to remind them. So historically to access the Equinix value proposition, you would have to order servers, install those servers into one of our data centers, order some networking equipment, so routers, load balancers, et cetera, install those and then need to order physical cross-connects to connect to the ecosystem in our data centers that you care about. So you connect out to your Internet service provider, maybe your network, your MPLS, your network service provider, connect out to the cloud. So you have these physical boxes that would be in our data center. And then you have physical cross-connects that would connect all of that. What we've done is, we've basically virtualized our value prop. So instead of having to buy those things and install them in colocation, you can now use Metal for servers, so it's dedicated servers that you have. You can use our Network Edge product, which are branded like routers, load balancers, firewall, so think like a Cisco router or Palo Alto firewall, you can install that networking gear. And then you can connect it all using Fabric, which is our virtual interconnection platform. So what that enables our customers to do is, it basically have a virtual colocation deployment. And so with that context in mind, let me give you a little bit of sort of what we're seeing from adoption. So Fabric as an example, we have about 3,000 customers that's been out. I think we've had it for 6 or 7 years, has about 3,000 customers on it, continue to see very strong growth on Fabric that report out every quarter on. And then we're seeing that growth, not just who's connecting to it, but as we're seeing each customer that bought it -- if you bought it even 4 or 5 years ago, we're seeing more and more connections on that platform. And we're running -- I think we publicly said that that's north of 100 -- well north of $100 million and growing very well. So that gives you an idea on Fabric. Metal is more in the north of $50 million and also growing very well. Network Edge is the newest of those products. And it's a fair bit smaller, but it has the highest growth. So Fabric is the biggest and it's growing at a multiple of the base of our business. And then, the rest of the products are a bit smaller but increase progressively with their relative growth rate.

Great. And maybe just looking out over the long-term, if we're having this conversation in 5 years, do you think there will be a material component of your overall revenue stream? Or will it always kind of remain a niche subsegment?

I hope -- let me tell you why. So I think if you take a step back and you look at the amount of wallet share that Equinix is able to capture when we're just selling colocation. So when we're selling colocation, we're selling the space and power that the servers and network equipment go into. When we sell a virtual equivalent of that, and this is still at the infrastructure level, we're not fundamentally changing our DNA. We're just changing the substrate upon which we deliver our value prop. But when you are consuming sort of the servers on our Metal platform and Network Edge, instead of just buying colocation, there is materially -- because we're offering service that's higher up on the stack, the wallet share that we capture per workload is much higher. It's at higher margin and a better ROIC. So if -- to the extent that we can cannibalize around colo with this, awesome. And we think that there's a really good path here and we're actually spending a lot of calories evolving our go-to-market to be able to accelerate sort of the path that we see on all the digital products.

Great. Let's talk a little bit about the other announcement on the day and that's the expansion in Africa. You announced the acquisition of MainOne, which has 3 assets in West Africa and one under construction. How would you frame the Africa opportunity? And maybe you could also, in that context, discuss the risk that you're balancing when entering a developing market?

Yes.

Maybe I'll start with this one. When we look at emerging markets, we have talked about the African continent for a number of years. I'd say, Africa and India have always been top of mind for us. You saw us going to India this year with the GPX acquisition launching out of Mumbai. And then, with the announcement of a company called MainOne, we're going to now have a presence in West Africa. So they're across 3 countries, Nigeria, Ghana and Cote d'Ivoire. And if you look at the entire continent of Africa, a lot of times, we're looking at where the actual population centers are. So mapping out to where the population center is inherently, you're going to need connectivity there. And I think it's just a start of a broader story across the African continent. And then, Bill, I know you've been talking to customers too on their views of it.

Yes. I think you hit on the big one, Katrina. So we look for big pools of global GDP and GDP growth. Then, we want to have infrastructure that's typically within 20 to 30 milliseconds of latency of those. If you look at, in the Americas and in EMEA, for example, we have over 80% of the population in those continents is within 10 milliseconds of an Equinix data center. But when you look on the African continent, there were large pools of GDP and GDP growth that we weren't there. So we're -- and that's the same kind of criteria that a lot of our -- especially the most strategic customers use as well. So they're public posts from folks like Microsoft and Amazon, and they've spoken others about where they want to have their infrastructures. So it's a combination of just good demographic basics, accompanied with good strong demand from some of our largest strategic customers. And so if you take that asset and you couple it with our global selling engine, coupled with sort of the great demographics of that region, we think it's a really good combination for long-term growth. I mean, we're not going to flip it on and it's going to be -- do awesome immediately. These are long-term bets that we're making as well.

Sure. And MainOne seems to have a larger fiber component than your traditional acquisitions that are more focused on space and power. How should we think about this initiative in that context and, maybe more broadly, the nature of the assets that you're acquiring?

Yes. I think -- so a couple of things. One, the one thing that we really love that they have is that cable landing station. So if you look at a fiber map of how networking and how the Internet of interconnection works on the continent of Africa, there are all these cables that actually go around the continent of Africa. And they actually have one of the key landing points for those cable landing stations. So we think that's going to be a point -- a good value of driving interconnection density for those strategic customers over time. Now we will continue to invest in -- the African market is a little bit different about the assets that you need in order to be successful. So that's something we're going to have to work our way through as we digest this acquisition.

And then, just on the customer front, I believe you're getting about 800 new customers. Would those include like all the large cloud service providers and companies that we're all familiar with or would it be skewed more towards local enterprises?

Katrina, you're going to have to take that one.

So I think when we look at it, I think it is skewed across enterprises, networks and cloud providers that you have cloud customers over there. I think they have a great momentum around network and cloud there. But as you can see, majority of those customers are net new, which implies that it's weighted towards more local, more African-based countries.

I would say that if you look at history as a guy like Mexico is a good example, when we do an acquisition like this, a lot of those cloud providers are [ treated ] like awesome, you're going to equinize this location. It's a standard global operating model. We know we can count on and it's sort of a fast follow action as we've seen that many times in acquisitions.

Yes. Yes, absolutely. And we tend to pull in the MMCs with us. So it's like Bill said, you're having these conversations. We literally are sharing math of where we're going and where they're going. And then when you make this announcement, you reach out to them and then they start seeing us as, okay, if I want to go to now Africa, I have a data center that I work with across the globe to be able to colocate with.

Great. It's great to have that absorption into and presumably greater pricing power than what MainOne would be able to attain on their own. And to that point, it looks like the EBITDA margins are about 38%, somewhat below Equinix's current level. So how should we think about the opportunity for improvement there?

Well, keep in mind, it's not -- it's a mix of colocation, fiber and other services. So I think it's a little bit too early to say necessarily how that would change. I think we're -- our goal is to drive the colocation piece. And as with any emerging markets, we have an in-progress data center. We're going to aim to fill that up with our multinational customers and expand that platform. But we'll have to kind of look at just how that all kind of fits from a financial perspective.

Sure. Then, looking at on-ramps around the globe, Equinix has about 40% of the world's hyperscale on-ramps, which is currently a great competitive advantage. I wonder how you think the value of these on-ramps is going to change when customers could direct -- connect the cloud directly in their cage or directly on-prem within AWS Outposts or similar product.

Yes, that's a good question. So I think when you're deploying something like Outposts, it's required for that Outposts to be able to basically connect back into the cloud -- into the home cloud. So if you're putting Outposts in an Equinix data center or any data center today, it has to be able to ping back to AWS for management plan and to be able to control that infrastructure. And one thing that's important for that calling back is, you want to make sure that connection is very performant. So it needs to be low latency because you might want to be doing like a highly performing database updates or something like that. You care about the performance of where you put that Outpost node relative to be able to call back. I mean, of course, also want that connection to be secure. So you want to be to be able to connect back to the cloud, if possible, not over the public Internet. But most importantly, you want it to be cost-efficient as well. And so you got to be -- where you put those Outpost nodes, you want to be able to connect back into AWS in the most efficient manner as you can. And so we see a lot of Outposts deployed in Equinix because we have those AWS edge nodes in our data centers and it's easy for them to call back into AWS for them to be able to connect back into their sort of core AWS deployment. And that same rough model applies for the other types of sort of on-premise solution stacks from the cloud providers. And so then -- but calling back, typically if you're doing that level of infrastructure, the speeds that you're going to be doing it can be really high. And so if you're using a network in order to do that, it can cost tens of thousands of dollars for you to be able to just do the networking. And the option to move that over to a cross-connect, that can cost you a couple of hundred dollars, is only possible if you're doing that in a physical location that also has the AWS edge node. So if we're not the biggest, we got to be one of the top sort of Outpost deployment nodes for AWS because of those dynamics. Because the AWS edge node is there, it's a natural place for you to put outpost because it's the most efficient, best performance, lowest cost way to connect back to the clouds. That makes sense.

Yes. That's great. Maybe switch gears a little bit and talk about partnerships or somewhat related to that as well. Equinix now generates about 35% of their total bookings through its channel program. And this is certainly a broad reach. Many peers are trying to emulate the success that you have had with the channel. Maybe you could shed some light on the nature of these economic relationships that you have with your partners?

Yes. And Katrina, fill in some of the details here as well. But like, I think in general, we look at these partners as a sales force multiplier for us. And typically why I think it's easier for Equinix to do these types of solutions relative to some of our competitors is because typically our solution is a combination of -- it's where the digital supply chain comes together. So whether it's the network service provider doing last mile connectivity or the cloud provider providing core infrastructure or a systems integrator tying those pieces together, they are all at an Equinix data center. So the fact that we can -- we're the natural place for all of those things come together, make it easier for us to partner with any number of those and get more bites at the apple and have a better chance to win that. Whereas if you're just putting sort of a large deployment in an undifferentiated data center, there's not many ecosystem participants there. So there's just not many bites at that apple. So I think the fact that we're doing an ecosystem play is why it's sort of easier for us to do that. And then -- but we monetize that through a couple of different models. One is a true like sell-through model, where they take our capabilities, rebundle them, repackage them and then sell them out. And then, another is more of where -- it's obviously a referral model where they put it on, the Equinix brand still comes through in the [ selling mode ] shift it sell with where we sort of share the selling relationship with that -- the partner.

Yes. I would just add on that. The progress on the channel has been very exciting. If you look at where we used to be from a split, used to be about 95% direct, 5% channel. And so the fact that we can find scale with partners, it changes them all in terms of one, who we're able to go after. So getting the timing of when a customer is actually looking to migrate to cloud with a bigger package is a great way to go after that with these bigger resellers. And there's names like an AT&T of rising as well as the top hyperscalers that we work with around the go to market. So I think it broadens the customer set that we're able to go after, it's very efficient, and it's a great way to have both a very strong direct sales force. And we expect to continue to grow the channel as we go along.

Maybe just a follow-up there. To what extent are you able to mitigate competitive threats by creating partnerships with some of these companies that might be able to offer something similar to what you are offering?

I think -- go ahead, Bill.

Well, I mean, I think the biggest -- I think it starts with kind of where I was starting like, we're not creating like a one-off solution. It is -- the biggest competitive moat that we have is that we're enabling an ecosystem solution. And that's really hard for someone to have to replicate. We've spent 20-some years building those ecosystems in our data center. So in order to build the solution that we have, it's not just you're buying some colo, you're buying a place to put your digital infrastructure that's connected to your digital supply chain. So that sort of partner nature of what we're offering makes it -- creates a big competitive sort of moat around our business. And then, of course, as I mentioned earlier, the ability for those different partners to be able to bring deals to us and we get more bites to the apple changes the dynamic as well.

Yes. Just to follow-up on that, there does seem to be within this ecosystem an unprecedented level of competition with customers. Certainly there's a partnership component within the ecosystem. But also, if you look at any number of different angles, CSPs are major customers and also compete for enterprise workloads. Network carriers are major customers and compete with your SDN offering. Managed solution providers are customers that you compete with environmental, et cetera. I mean, even Equinix is a top 5 tenant of digital royalty and you compete with them for space and for -- directly for clients. And despite all of this competition, you've been able to drive for years growing AFFO at over 10%. But I wonder how long that can persist or if there's something that could change that dynamic?

Yes. I think I would frame it maybe a little bit -- but I think it's how you frame the problem that I -- or the issue that actually it is not quite on. We're -- what we're focused on is not selling space and power, like our focus is to enable globally distributed hybrid multi-cloud infrastructure. And I think if you look at that on a spot basis, that is an enabling capability for a lot of the things that you just mentioned are considered competition. Yes, there are elements that are -- some of the tools that we use in order to enable that value prop are common across us, but we are fundamentally solving a different problem. So your cloud providers, really they're solving for an integrated infrastructure plus software experience. And having a place that can bring rich interconnection to them is accretive to them doing that. The network service providers are providing sort of ubiquitous connectivity that includes last mile and 5G and the fact that we can bring that globally distributed interconnection-rich infrastructure is accretive to that. So in each of these elements, yes, there are parts and tools that might compete, but the overarching solution is very complementary to what we're doing. It's like, I would say that Ford isn't competing with Goodyear tires on selling tires just because tires are part of the vehicle that Ford sells. So if you think of the solution as in aggregate instead of just what one little tool that we might use competing is, I think it's a very different story.

Yes. To that point, I think over time, it's become clear that you're growing closer with these other partners in your ecosystem, not more apart.

Yes. And I think the aggregate -- like those secular tailwinds that are sort of blowing all those businesses in the right direction is sort of how we're writing those same secure tailwinds as well, of course.

Yes. You touch upon migration from carrier hotels. Equinix has kind of cited intentions of owning more of your assets and leasing less. And you have initiatives to migrate some customers from lease space to own space out of carrier hotels. I think, New York, Atlanta and Chicago are some of those potential opportunities. Historically, we've always viewed the carrier hotel as the major try itself. So as you're going through this process, what gives you confidence that you will be able to migrate your customers with you? And that has a follow-up.

Yes. So obviously each market has slightly different dynamics, but we have a well-developed playbook, like you mentioned New York, that -- largely that story has already played out where Manhattan is not a great place to have data centers, like just the -- for price of real estate, getting additional power in, just the logistics of being in Manhattan and getting equipment in. It's not a great spot. And so really the interconnection is, [ instead of ] gravity has already moved out into Secaucus into New Jersey, and there's a playbook that we use in order to enable that when this happens is, you need to find who the interconnection points of gravity are. You need to provide them an incentive to move. You need to select where you want that new center of gravity to be in Manhattan to New Jersey, example, which we chose Secaucus. If you need to provide financial incentives, and typically you have to be able to offer this on a global basis, so you need to provide holistic enticements for those points of gravity to come to the new location. And then you have to stick with it. These types of migrations take about a decade, but there's a well sort of -- well -- and me coming on here and saying that where the hard thing -- our competitor, it's really hard for a competitor to do this because you have to have that ecosystem of partners that you can pull into the new location and you have to have sort of a global portfolio to have enough enticements to get them into that new location. But this is a dynamic that we've seen in Manhattan, also Secaucus. We've seen it in London, with Docklands into Slough. There's a couple of other metros. So this is a pretty well-developed playbook. But again, it plays out a little bit differently in each metro depending on what the dynamics are.

And in those experiences, how much has churn increased as you try to migrate the clients? I'd imagine, there's some component of them that just want to stay in the carrier hotel where they are already. Maybe you can give some color there.

There's a little bit of that. But honestly, typically what happens is, when there's a point of gravity move, it's typically from a place that is capacity constrained. So if you look at what happened in Manhattan, you look at what happened in Docklands, the enticement to be able to move to a location has a much more -- much bigger available footprint. The availability to -- so your customers that were in Manhattan who actually wanted a bigger footprint, but you couldn't sell it to them, when they move out to the new location, they're actually buying bigger footprint because more footprint is available. So in aggregate, what you might lose in churn is way more than made up for in what you -- sort of the additional footprint that you get in the destination as you move into the new sort of point of interconnection density.

Sure.

I would just add, it's such a long process. It hasn't shown up in any of our financial churn in any sort of material way. I think, again, it takes time to do the move. And I think it also depends on the factors, like 111 8th, not only are they out of space there, they're shutting down and changing that building over. So that inherently means any new infrastructure is going to have to find a new home, and it's to us to have the right strategy to capture that new builds and to bring it along over to our sites.

Great. Well, we're just about out of time. I want to thank you both, Bill and Katrina, for participating today. Really appreciate it and look forward to continuing the conversation.

Great. Thanks.

Thank you, Brendan.