Arista Networks, Inc. (ANET) Earnings Call Transcript & Summary

Good afternoon, everyone. Welcome to the session here. I have the pleasure of hosting Arista Networks for this session. And as you can see on the stage, we have Ita Brennan, who is the Chief Financial Officer; and Liz Stine, who is the Director of Investor Relations. Ita, Liz thanks for making it to the conference in person. Thank you.

I'm starting all my sessions with sort of 3 broad level questions that -- it's almost mandatory that you answer. So the first one, and I'm sure you've got this question multiple times today, looking at the demand trends, customer demand trends on your side of the business, how you're thinking about the likelihood of a potential recession here in the near term?

Yes. I mean I think calling a recession is something that we probably leave to some of you guys more than us. I think in terms of our business, and we talked about this on our earnings call as well, which was a couple of weeks ago, right, I don't think we've seen any change in trends in the business that would call that out, right? We're in a pretty unusual environment where supply is such a big part of the picture. It's resulted in kind of greater visibility on the order side, on the customer demand side. And we haven't seen any particular change to that. It seems like that it's pretty much status quo in terms of how the business is evolving. So I don't know that we've seen anything from a bottoms up. I mean, obviously, we watch all the same economic indicators, inflation, et cetera, et cetera. I mean, there's a lot of stuff tied up in that. But from a day-to-day perspective in the business, we haven't seen anything so far.

Okay. Let me just follow-up on that because I'm sure given Cisco's earnings print last week, there are a lot of concerns about or a lot of investors would want you to confirm what you're seeing with the large enterprises in terms of order trends. And is there anything to be concerned there or any slowdown there that you're seeing in the recent weeks in terms of the -- just the large enterprise vertical?

Yes. I mean, again, I think when we look at the demand that we see, we haven't seen anything different to kind of what we would have talked about on the call. And I think we were clear on the call that the demand environment remains strong. And really, it's all about kind of managing the supply side of the equation.

Okay. So moving to supply. One of the questions we're asking everyone as well as supply was already constrained, how much worse is the China lockdown, making it, right? And I think, again, just to sort of add the second question here, Cisco had a big supply issue with what seems to be power supplies being sourced out of China. Do you have any exposure even of low-value components of -- in the region?

Yes. I mean I think we're a little less maybe directly exposed to China. But the reality is if you have a prolonged lockdown in China that lasts for a long time that kind of impacts broadly. There's a lot of sub-component supply chains that will eventually get impacted. I mean we saw that kind of when COVID kicked in first in China, you saw -- you did see that even though you might have kind of distanced yourself from some of your direct manufacturing and other things that eventually, if it's a truly prolonged impact, it may well have some broader impacts. But I think so far, again, our issues around supply, we're more focused on kind of just these decommits of a handful of kind of smaller components. But again, if China were to be locked down broadly for a long time, that's going to show up in everywhere, right?

Yes. Any predictions around when does supply normalize eventually?

Yes. I mean, we've been kind of moving that out every -- it's hard to predict. I mean, certainly, our view now is that like this environment is with us through 2023. You saw us kind of do some planning around 2023 in the last quarter, place some orders kind of out into that time frame because we really do believe that there's no evidence that things improve materially. And there's still quite a bit of overhang from risk perspective be it China, and some other things. So we are planning and assuming that we're in this kind of constrained environment at least well into, if not through, 2023.

Okay. Okay. So thanks for answering those. I'll move to the more company-specific questions here. And before I do, I see some questions coming in already. But for anyone on the webcast, just a reminder, you can send in your question I'll open it to the audience as well in a few minutes. So firstly, what I wanted to touch on, Ita, here is you've guided top line growth to return to a more normalized run rate of about mid-teens starting next year. Can you just break that down for us a bit more in terms of how to think about the drivers? How do you really get to that 15% mid-teens sort of number between the -- maybe break it down by the customer verticals or what your expectations are there?

Yes. I mean, I think just to clarify, we haven't really guided a 2023 growth rate, right? I think what we did at the Analyst Day was talk about a mid-teens CAGR from 2020 through 2025. Since then, we've talked about having a multiyear double-digit growth, but we haven't tried to specifically peg 2023 yet. I mean, obviously, that's because there's just so many moving pieces in the business, right? So I think we're working our way through 2022. We'll become more granular around 2023 later in the year. At this point, it's just too soon to kind of -- to try to call that. I think when you look at -- we had -- at the Analyst Day, we had talked about expecting kind of that 30% growth rate with pretty even distribution across the business with maybe cloud growing a little bit faster. I think supply is kind of constraining to that picture. I think if we had more supply, we could probably see cloud grow even a little bit faster than that. We could also probably see the enterprise grow a bit faster. But really, supply is constraining it so much right now, it's hard to kind of give a picture of what 2023 will look like until we work through some more of the supply issues.

Okay. Ita, so if I follow-up on that, I understand you didn't issue a 2023 guide. It was a 3-year sort of view on what the company does on the top line. But even that 3-year view, are you able to break down that mid-teens into sort of how you think about the components?

I mean, it's really difficult to put it into discrete buckets at this point because of the impact of the supply. I think we've talked about the demand in the business and the fact that cloud is healthy right now, and we have seen a resumption of kind of investment cycle with them. I think you've seen that in their CapEx as well. On the enterprise side, we see continued traction. We had some really good momentum there. We're seeing that continued. We've talked about the campus business kind of doubling this year. We're on track to that. And the provider business, when you look at kind of -- both of those are growing, I think the specialty cloud kind of maps to the hyperscale. They're investing in new technologies, and they're looking to kind of deploy equipment around that. And the service provider continues to grow at a healthy rate. So again, demand across the business is healthy. And again, the [ gate ], if you like, is really the supply situation.

Okay. Great. I guess this is more of a 2023 question for you, but maybe a bit -- any color that you can provide in terms of how we should think about pent-up demand or backlog going into 2023? And your visibility in terms of at least getting some supply improvement to digest that backlog in 2023?

Yes. I think we spent quite a bit of time today talking about kind of how do you think about orders in a time frame. We're in a period where really time is such a big driver of order numbers, right? And we've tried very much to focus on deployments and when will we see deployments actually happen as opposed to trying to count orders. When you've got a time horizon and lead times that's moving around, right, and we think that makes a lot more sense. At this point, we've talked about kind of having demand that's kind of starting to kind of settle in for some of 2023 orders. That's why we had kind of confidence to go place the purchase commitments for 2023 and to start to do that. It's not by any means done, but we are starting to see kind of some of the orders filling into 2023. So I think that's probably in terms of when do you see stuff be deployed, I think we are seeing, at this point, kind of orders filling in on the 2023 side. Again, not by any means done, but giving us enough confidence to go place purchase commitments, et cetera.

Okay. I just wanted to focus a bit more on the cloud vertical to start with. Historically, this vertical has been lumpy or inconsistent in terms of the spending pattern or the investment piece here. And do you see that changing in the future related to how they've sort of managed their spend in the past? I mean, one concern that we hear from investors very often is what gives us confidence that they can continue to spend at this pace or even continue to spend every sort of year on an ongoing basis. So what gives you that confidence as you sort of lay out the medium-term plan that this spend is a bit more sort of predictable rather than lumpy and inconsistent?

Yes. I mean I think -- look, we've said, I think, pretty consistently that over time, we believe the investment in their businesses continues to grow, right? And -- it can be -- the timing of it can be impacted by product cycles and investment cycles. If you think about where we are right now, we've come out of a period of lower investment than you had COVID. And now you're in a world where there's new product cycles and there's a desire to invest kind of further capacity needs in their businesses today, right? So I think this product cycle in reality is kind of going to be maybe a more elongated cycle because you can't -- it's capped by supply, right? So we're starting to see that kind of stretch certainly into 2023. I mean, beyond that, at this point, it's kind of too early to call it other than to say there's obviously future projects and other things that they want to invest in as well beyond that. But I think the visibility that we have now, and just how the demand and the technology cycle is kind of overlapping with supply. I mean, it is elongating that cycle, certainly well into 2023.

You're clearly sort of very relevant to the upgrade cycle for your cloud customers. I mean, you're thinking of more than 30% growth with cloud this year. At the same time, there has been a lot of discussion around share loss to competitors like Cisco. There was an announcement related to Nokia more recently with one of your competitors. I think what we're seeing is investors increasingly sort of evaluating where the competitive landscape is, is there more sort of shared loss that sort of -- even if it's not material or new competitors coming into the space that we haven't seen in the past, that starts to become more of a trend going forward with your cloud customers who you've been really successful with, right? So how do you -- how do we think about sort of you being able to maintain share with the cloud customers related to sort of the noise we've seen around a few more sort of suppliers coming in into that competitive landscape?

Yes, that business is such a -- it's such a prize that everybody obviously wants to have it and have a meaningful share of it, right? So every time we have a product cycle, we go through kind of a strong impetus to push into that piece of business and more and more people kind of at least aspiring to that, right? I think our job and our goal as we go through these cycles is to continue to execute well. And when we do that, we're able to maintain kind of our preferred kind of supplier position with those accounts, right? It's not single source. It hasn't been single source. There are obviously other dollars that are spent there that don't come to us. But our goal is to make sure we get our fair share of that. And I think when we look at this cycle, we're pretty happy with how we came through that cycle, right? I think we spent a lot of time talking about kind of commentary, some aggressive commentary, et cetera. You don't walk into that business and take share lightly. It's something that requires a lot of demonstrated execution and then consistent execution. So for us, I think our focus is if we continue to execute, we have found those customers to be part of the business where we can earn our fair share, and we'll continue to do that. I think for newcomers into that space, it takes a long time to prove out technology. And that's really what it's about. And you don't get to grow that business until you've demonstrated that on a consistent basis. And I think that's been evident kind of over the last couple of cycles as well.

Okay. I know Anshul has talked about it in the past in terms of as you even see new competitors come in with your cloud customers, there's been sort of new use cases that Arista has been able to pursue to sort of balance out that share that you see as your fair share. Can you just dive into what are those incremental use cases that you're able to find consistently to balance out the share even as new competitors come in?

Yes. I mean, obviously, we've been doing a lot of work on kind of the routing protocols and the routing capabilities for the product for the cloud space, right? So we can continue to participate in more use cases on that side of the business. We talked a little bit about the 7800 being a good AI platform. That's again, certainly not done and baked, but it's another opportunity where you can start to see -- bring in the technology, having the platform be so robust that you can start to earn maybe some additional dollars that weren't being addressed by Ethernet -- by certainly by us in the past, right? So those are kind of the kinds of use cases where when you're performing well with the technical folks in these accounts, then you're getting that right to start to look at other stuff and start to participate in other areas. That doesn't happen overnight. It's a gradual kind of -- but you're exposed to use cases, you're exposed to kind of technology needs and then you can start to address those and grow that business.

Okay. One of the questions that I get on this front. If I step away from the market share discussion as well, is you had a big differentiation relative to your competitors in 100-gig, which helped in your success as well. As we go into 200-gig, 400-gig, the discussion of more competitors coming in with -- and getting share with your customers, does that -- what does that sort of imply in terms of differentiation of the product for Arista relative to your competitors? Is that technology differentiation somewhat compressing as we go to next-generation products? Like, that sort of more noise with more competitors coming in is leading to a lot of questions around how does the differentiation now compare versus 100-gig that you had?

Yes. I mean, look, I think you have to think about this in terms of everybody is going to compete for its business, right? So you're going to hear a lot of names around these accounts. That's just part of that ecosystem, right? At the end of the day, it comes down to who's executing, delivering the products on time, who's kind of allowing kind of them to scale at the rate that they want to scale, right? And I think we continue to do that, right? So there's always going to be noise around these accounts. If it's not branded name, it will be white box, it will be something else. And it is -- it's a vast and complex kind of set of footprint. Again, our job is to execute on our products and earn kind of our share of those dollars. I think we continue to do that, right? I mean we had lots of discussion coming into the cycle with competitors. I think we've come away very happy where we come out. Not to say that we got every dollar, but we got our fair share of those dollars.

Okay. Just to -- and you mentioned white box, just to round up the discussion there on the competitors, the narrative around risk from white box adoption has subsided quite a bit in the recent years, I mean, from what I can tell in my investor conversations. How are you looking at sort of white box share today with your cloud customers? And are you seeing sort of more of a stabilization? Or are you thinking increasingly about a reversal of share towards branded?

I mean I don't think white box has gone away or changed, but it also hasn't progressed, right? I mean, again, we've lived through the idea that everything is going to be white box. I think Anshul's kind of discussion of make and buy is still the right way to think about this. The balance is kind of what these customers are looking for. They need multisource. They want to have some balance, but they also value the kind of the vendor and the contributions of somebody like Arista to that. So I don't know that there's been a huge change in kind of the status quo around white box. If you step back and look at it over a period of time, but certainly kind of -- it also hasn't kind of gotten -- taken more share in that place best as we can tell, right?

Okay. Let me take this question. Let's come in. And again, as a reminder, we'll just open it up to the audience in a few minutes if you have a question. So I'll break this question now, Ita, in probably a couple of parts just because it's a long one. So Liz, I'm wondering if you can elaborate on the enterprise campus go-to-market strategy. I think most potential campus opportunities are from the current Arista users in data centers. How do you identify the issues that can get fixed best by Arista's Cognitive Campus products? And let's take that and then we can probably move to the next few ones too.

Yes. So maybe I'll take the first piece and then Liz can take the kind of the solution piece of that, right? I mean, I think we've talked about and Jayshree's commented on a couple of calls now that kind of our campus new logos, roughly 50% are coming from non-data center campus customers, right? So we are seeing new customers in addition to kind of what data centers. So I think, yes, originally, we thought data center customers pushing to campus would be kind of the first leg of this growth story and that it would be harder to get new campus customers. I think we've been -- I think we're happy with kind of the amount of new customers that we're seeing. A fair amount of it is still direct, right? It's still through a direct sales model. We'll continue to add sales resources for the enterprise/campus business. to help us kind of get more coverage geographically. We'll continue to target accounts, et cetera. So that's one vector of growth, and that's probably the most important in the near term is continuing to land and expand where we have a data center footprint, win new campus footprints and then push back to, if they have data centers, back into the data center, et cetera. The other piece is channel, which is more relevant kind of for the mid market and downwards -- and again, that first piece is roughly 50% of that campus market, right? So then for the rest -- for the other parts of the market, I mean we are kind of developing a channel program and adding some channel partners to kind of go and address that part of the market. Again, it has to be somewhat I would say, sequential. You have to find the right partners. There's no shortcuts to kind of building relationships with partners and finding the right partners who want to kind of build a business with us. So we'll continue to do that. And we're doing that in conjunction with kind of that first step, which kind of helps that's what's driving the momentum today, and then we'll start to layer in the channel stuff over time. Liz, do you want to take the solution question?

Yes. I mean I think from a large enterprise standpoint, As Ita was saying, being direct. We're going in there with an engineering team with the sales team, it's a very technical sale, working with the customer to kind of understand the requirements of the network and then getting them experience and exposure with EOS. And I would say it's hard to sell quality because everybody is selling a quality solution, right? You'd be stupid to walk in and say your product was in quality. You want your customers to experience quality. So I think where we see the most success in these initial conversations with customers is looking for that maybe a small project, maybe it's a back-end storage network, maybe it's a high-performance compute that, maybe it's network monitoring to where they can experience the quality of EOS because once they experience that, and they recognize how vastly different it is from some of the competing products. They see the value in rolling that out in larger parts of their network. And the nice thing about over the last few years, we've expanded our enterprise portfolio. So we're no longer just going in and hoping for an enterprise -- sorry, for a data center opportunity. We're not waiting on that. We get to sit at the table for campus opportunities as they come up, but network monitoring opportunity, security, routing and data center and that just is broadening our times at that, right?

So Cognitive Campus, may be touch on that?

From a cognitive campus standpoint. So I think from my experience in working with these larger price customers, once they experience EOS and CloudVision, which sits on top of that, it makes sense to bring that out through the rest of the network. So one of our differentiators at Arista is that we have one operating system that runs on every platform. So whether you're deploying a switch in the data center or you're deploying a switch in the campus, and it's a PoE switch, it's running the same EOS and it's being managed and operated by the same CloudVision, right? And so that CloudVision software also gives you end-to-end visibility. So this is where your network telemetry information is coming. This is where your bug scrubs are happening. And so having that end-to-end visibility is really important to enterprise customers, right? Because it's, again, that kind of one tool, that one operating system across the entire enterprise rather than deploying point products that you have to manage separately and that you have to manage code upgrade separately and then configure separately, et cetera, that holistic approach really benefits our enterprise customers.

Let me just finish off with the last part of this question which hasn't been covered yet in your sort of response. How big is the international enterprise campus opportunity for you relative to the U.S.? Like, how do you think about where the low-hanging fruit is or where the first approach is versus how do you think about the international opportunity there?

I mean, if you look at where we're adding sales resources more recently, more of those headcount are going into kind of that international footprint just because we're further -- we're more mature in the U.S., right? We've already invested significantly in the U.S. So clearly, we're -- this is always a staged kind of evolution. We focus on the U.S. first and then started to expand internationally. And we're -- so we're earlier, I think, in that enterprise -- international enterprise business. And so we are adding sales resources and kind of growing that business as well now with more of the headcount addressing that.

Okay. Got it. Let me open it up here for the audience. Just wait for the mic. Thank you.

Could you talk to supply chain, there's been weakness in consumer lately and that may alleviate some substrates? Are you guys seeing anything improving along the supply chain dynamic? And also because you guys have a contract manufacturer here in the U.S. so I don't know if that helps with the supply...

I'm sorry, I didn't get the last piece, Nick. Can you just the last...

Yes, you have the contract manufacturing here in the U.S. as well, too, right? So I would assume China does not hurt you, but could you talk to that as well, too?

Yes. So I think are we seeing the supply chain improve? I don't know that we've seen significant improvement, right? I think you're still very constrained, and you're still seeing some decommits. And we talked a lot about decommit, but there is kind of -- you are seeing kind of these decommits on, I would say, smaller -- normally components that you'd never worry about, but a handful of these components that are kind of disrupting your ability to scale, right? So even with kind of making the long-term purchase commitments and doing all of that planning, have a decommit against those, then it disrupts again, right? So we got some -- we saw some good traction from a ship perspective early in Q1, and then you saw that kind of get disrupted again by the decommits, right? So I don't know that we've seen -- we still haven't seen a period where you could get some consistent momentum. And I don't know kind of that we see a path to that. Yes. Like we said, we think this goes on now into 2023. For China, I mean, we have manufacturing footprint. I think it's in the [ K ], you'll see kind of we have Malaysia. We're in Mexico, we have some stuff still in the U.S. But again, that helps with the direct manufacturing piece, but China is going to be present in sub supply chains from a component perspective. So again, if you have prolonged severe issues in China, it will still show up in the supply chains of people who are not exposed directly to China. And that's something we have to keep monitoring and, hopefully, we get -- there's some relief on that. But if it goes on kind of for a prolonged period of time, I mean, that will impact us and everybody else eventually. And we saw that kind of in the beginnings of COVID when China showed up kind of in sub-supply chains, and it's just very difficult to have something not lead back to China eventually, right? Yes.

Just 2 questions. One on -- following up on the campus switching side. The fact that you do kind of have lack of supply and that visibility is now pushed out into calendar '23, does that delay at all kind of you expanding your sales force or your marketing? I mean, it's kind of worthless to sell something if you can't supply it for -- and so what I'm wondering is, like, are you trying to match up when to really ramp your sales force versus when you really have supplied for them to sell?

Yes. I mean I think for as long as customers are placing orders kind of with those lead times, which is what we see, right, I think you want to continue to kind of make sure you're positioned to address those customers, right? Again, if you go back to kind of the first question, if you start to see that change, then maybe you'd have a different view. But for right now, I mean, customers are -- you have to make those customer engagements, you have to win kind of those opportunities and then they're accepting those lead times. So I think -- so as of right now, I would say you continue to kind of make sure that you're doing that. I mean, again, things -- lots of moving pieces, so we'll continue to monitor that. But given what we see today, we would continue to hire those folks.

Second question, just going back to originally the China supply chain. Can you just refresh us what is kind of embedded in your June guidance as far as the impact from the China lockdowns?

Yes. I mean, our guidance kind of was in confidence kind of everything that we've seen through kind of through the call. Obviously, it's -- while we have some exposure to China, I think we understand kind of what that is and where that was at. I mean, that's not something that's new, right? So we are closely monitoring that. So obviously, our understanding of that is built in there. And then again, a view of this kind of decommit environment that we're having to deal with, and we put obviously a broader range on that, and we try to take a reasonable view of that as well. So, yes.

Yes. Any other questions? Okay. Ita, I wanted to move to pricing, and you talked about pricing a bit on the earnings call and you've tried to sort of take pricing actions a bit less frequently than some of your competitors, in the number of times that they've raised prices. I mean, how are you sort of seeing -- How are you seeing customers respond to your pricing actions? And is sort of your approach to pricing sort of helping you gain any share in the marketplace?

Yes. I mean, it's always hard to compare to, again, what the competition is doing because these pricing actions, it's hard to tell if they apply to the whole business, parts of the business. Is it kind of all of the business, multiple times? I mean, it's just really hard to evaluate that from the outside. I think our approach has been, and it's somewhat driven by our customer set, right? But we did have substantial price increase in November of last year. That was negotiated, right, with customers that had to be negotiated with customers. We got pretty broad acceptance of that. We did not reprice backlog. So we're starting to see kind of where we should expect to see the benefits of that. When we ship kind of those new orders at that point, which is probably Q4 of this year, we should start to see some benefit from that pricing action in Q4. Jayshree talked about a pricing action that we were in the middle of when we did the call. Again, the goal is to scoop up as much cost impacts as we see and expect and kind of put that together into that price increase, and we're in the process of doing that now. And again, we won't reprice backlog, so that will take some time for that to flow through as well. But we'll start to get some pricing benefit really for the first time in Q4 of this year.

Okay. And have you seen any pushback on pricing from your customers? I mean, I would imagine, if at all, that would be more cloud. But like what are you seeing in terms of customers responding?

No. I mean, I think the reason why we've done it, I think, more kind of less frequently is because we have spent quite a bit of time exposing to folks what's driving those costs -- those prices increases and the cost increases that are driving those. And so we did get pretty broad acceptance of the need to kind of have a price increase, haven't exposed kind of the cost structure. And we've got a pretty broad-based acceptance of that. I mean in this world customers understand they want the supply and they understand that these are actions that you have to take to get that. The decommit issue has been a little bit more difficult because it's showing up and you're having to respond to it. You've already given them a commit some time ago. And that's why that's a little harder to pass on because it's kind of unpredictable, et cetera. But the sustained kind of plan, full cost increases and price increases, there was pretty broad acceptance of those.

Last couple of questions. One, you did guide down gross margins in the near term on account of expecting more open market purchases for components that you have to make. I mean, as an investor, how would you think of -- sort of how should investors think about balancing growth in the year versus profitability? And I mean I would think that if you were not to make those open market purchases, that basically goes into your backlog for next year, not shipping to the customer. So how are you thinking about sort of striking that balance between growth and profitability?

I mean, it all comes back to the customer, right? The -- that 200 to 300 basis points impact that we talked about is really responding to decommits, right? And it's -- we're -- if we can find those components and again, you have to exercise good judgment, you can't be ridiculous, right? But if you can find those components and it helps you solve for a commitment that you made to a customer 9 months ago, they did everything that you asked them to do, they accepted the price increases, they gave you the forecast, they gave you the long lead time, then we will do what we can do to solution that for those customers, right? I think that's really the driver. And if kind of in the near term, that impacts gross margin. as long as we can see that systemically it improves once there's more supply, I think we're willing to do that, right? Just because customers have -- I think, have come to the table and been part of a collaborative process, and it's very disappointing that we have to decommit on that. And even though we're being decommitted to, I mean, that doesn't solve the customer problem. So we're doing everything we can to try to solve that customer problem. And we're willing to kind of incur some incremental cost to do that in the near term. It eventually gets solved when these suppliers add some capacity, which kind of happens maybe later in the year and likely probably beginning of next year. But I mean it will eventually get better with the added capacity.

Last one that I'll squeeze in. You have a strong balance sheet, good demand visibility with your backlog how are you thinking about sort of using the current market conditions to be more opportunistic with your buyback? Like, would you be sort of open to it or how you think about capital allocation?

Yes. I mean, look, we have -- we're using the balance sheet on the purchase commitments for sure, right? We are -- we've talked about kind of the base level of the repurchase plan being kind of offset dilution and then we'll be opportunistic as the opportunity presents itself. I mean, I don't think there's anything new to add there. I think that's kind of our approach. That's been our approach.

Great. We've run out of time, but thank you both for coming to the conference. Thank you, everyone, in the audience as well. Thank you.

Thanks, everybody.

Thank you.