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

Hello, everyone, and thank you so much for joining us. My name is Jim Suva. I'm the IT Hardware and Telecom Equipment analyst here at Citigroup Investment Research. We're very pleased to now bring this fireside chat with you for our Arista Networks, ticker, ANET. A couple of housekeeping items. First, please see Arista Networks Investor Relations website for the safe harbor and forward-looking statements as well as the risks and commentaries that they have on that website and a lot of useful information. Second, Citigroup also has disclosures associated with this. Anybody who is subject to [indiscernible], please ensure you have those research agreements. And any media or press, please disconnect. Media and press are not allowed on this and any media or press that we see up in the lineup, they will be promptly disconnected. This is meant for institutional investors who are Citigroup investment -- Citigroup institutional investors. I want to welcome the people here for Arista Networks. We have Anshul Sadana, he's the Chief Operating Officer; and for backup support, in case there's any details, we also have Charles Yager, Head of Investor Relations. Anshul is joining us as Chief Operating Officer. So maybe, first of all, before we get started Anshul, can you give people a little bit of background about yourself and your role as Chief Operating Officer and maybe a little bit of your career.

Absolutely. So I've been at Arista over 13 years now since almost the very beginning. We were a tiny company of about 12 people when I started, and I've seen the company grow and participate in every possible dimension of scaling our business, obviously, working closely with Jayshree, our CEO; and Andy Bechtolsheim, our Founder and Chairman, to define what we do with our customers in the business. At this time, responsible for various parts of the company, including all of the hardware development, the manufacturing operation side of things, the overall customer road map that we need to deliver, all of our sales support functions and any other go-to-market function that is involved in this space. I do work very closely with some of our top customers as well. So I sort of understand the market dynamics and where we need to be headed as well. Jim, I've been, overall, in the networking space for over 2 decades now, and hopefully can answer any questions about legacy and new.

Great. So why don't we go ahead and kick things off? And maybe talk about demand. This year was not a typical year. Can you talk about how you entered this year? And then the coronavirus pandemic spread globally and kind of how you sit there and see demand now in, say, the next 6 to 12 months?

You're right about -- this year is not a typical year. This has been a very unique year. First, with the pandemic, but also the fluctuations we've seen is people's expectation and whether there's volatility in the market and sort of transitions in the overall networking space. When we started 2020, we had some assumptions about what our cloud customers will do, we had assumptions what the enterprise customers will do. And then came February, March and things took a dramatic turn. We fared things would slow down. But surprisingly, the cloud market has actually held up fairly well, and some of the demand we've seen this year has improved compared to our original expectation. On the other hand, enterprise data center demand has been healthy, but new customer acquisition growth in campus vertical for us, as an example, has been taking longer because it's not easy when customers are not meeting you in person all the time. So those are generally the demand changes directly impacting Arista. But underlying our customers' business, I think many of our customers have been surprised on how strong the demand has been for them despite this COVID environment.

When we think about this COVID environment, has it changed demand, say, for example, to be more cloud-centric or more software-centric or a rush to the cloud or a bit of a pause of what historical road maps may have been from some customers?

I'd say, first of all, we should be all very proud, not just as Arista but as an industry that the Internet works. Without that, we wouldn't be here on this conference and other ways doing things remotely every day. And I think the Internet held up really well. Given the pressure that was put on it in the March time frame. What we've seen is a pull-in effect, and it's very hard for me to know whether it's a pull-in of the next 1 year full of demand or the next 3 or 5 or 10 years of the demand pulled into the cloud sector or some of the Tier 2 services. Simply because they're part of the digital economy, and it was very hard for businesses to function like they used to. So they all pivoted to getting access to resources or connectivity or services or apps in the cloud. So I think that has worked very, very well. We've obviously seen that in our results as well and improvements in the guidance. But that's really the biggest change we've seen that infrastructure investments have become critical and company is now valued even more than before.

When we think about the challenges across the situation for tech companies this year, there's been more than COVID. COVID is one of them. But also there was tariffs. There was sourcing of components, where they were assembled, where they were shipped to, import duties, who you can buy from , who you can sell to and all of this. Many companies have been adjusting their business practices to become more nimble. How has Arista changed or adapted its business practices? Because you are the Chief Operating Officer, so unfortunately, all this work gets pushed on to you, then you have to resolve all these different operational challenges. Can you talk about those?

Sure. Well, I do have a great team that lets me sleep at night. So I don't have to do everything myself. But you're absolutely right that the environment has been quite taxing. The number of changes that we've had to go through as an industry is beyond what any company can prepare for where you're trying to have a nimble supply chain and make changes to deal around tariffs or geopolitical pressures. But in addition to that, because of COVID, a way to think of this is that roughly in this calendar year, about 3 months of the output of the capacity, the world was really shut down, in different countries, different suppliers and so on. So we're trying to make do with the remaining 75% capacity to catch up to the 100% demand and not go back to any legacy supply chain that may be negatively impacted with tariff and whatnot. So the team has been working hard towards that. Being nimble in your decision making, being multi-sourced, multi-vendor, multiple locations at which we manufacture has all helped us balance all of that out. And then this year was unique. We had to deal with the simplest of problems where we figured out how to build our products despite all the COVID challenges and meet demand in many ways. But guess what, the company that manufactures the cardboard box couldn't get their permit from the local government to upgrade their company, and I had to go deal with those types of issues as well. So I think we've done well to navigate this. I think we have through the worst of the issues. And it's simply now gradually recovering to catch up to overall demand and build back a buffer.

Beautiful. Now also as Chief Operating Officer, Jayshree, being CEO, she basically said, "hey, it's time to go into the campus switching market." Can you help us understand about the campus switching market, kind of what you're doing, what's your strategy of attack as COO to go into that? And does this coronavirus help you with that or pause it? Or how should we think about it?

Sure. Well, we're not as top-down a company. So this is actually a collective thinking and decision of the entire management team over several years to go after the campus investment, including a lot of demand and pull from our own customers. In Arista, this makes sense where you will love your products. Why can't I buy your technology also for my campus and simplify my life? So I think it's a culmination of various customer demand as well as our own readiness to go enter this space. To some extent, the campus domain is a leverage investment for us because it leverages our hardware technology, EOS, that's similar features, maybe with some enhancements to go meet the campus needs that we have been building. And I think that has been coming along very, very well until COVID hit. Now with pre-COVID, we had surprisingly, many new customers adopting Arista technology in their campus. And then these are customers who had never used Arista products before, but for the first time when they started, they started in the campus of all things rather than [ the ] center that you would typically expect. Post-COVID, it's been harder to meet all these customers in person, so there's only be some impact there. But our existing Arista customers have started to do a lot more testing and deployments of our campus products as well. So it's sort of been balanced off with the 6,000-plus customers that we have today.

And I believe your company has laid out some goals for the campus switching market. If my memory is right, and it's been a blur of a day, but was $200 million in revenues, was that kind of what they were talking about maybe in 5-ish quarters or something? Am I right on that? Or maybe you can correct me.

You are correct. Our first goal was to achieve $100 million in our first year -- first full year of the campus offering, which we did accomplish, and we announced that in our Q2 results. The next goal we have is to double that from $100 million to $200 million [indiscernible] 4 quarters, that may take 5 to 6 quarters, especially given the COVID environment. We are on track for that. So we'll come back and report that in time as well as we make progress and try and hit those targets. But so far, I would say the campus offering is well received by customers. There's lot in there, which is not just building the hardware, but also the automation suite that goes around it, including CloudVision. And very recently, we introduced CloudVision as a service that you can manage infrastructure, run your data center and campus networks all remotely, and that is being very well received by customers as well. So I think those are the anchors that will let us grow.

When we look at the competitive landscape, some of your competitors are talking about increasing their focus to gain share with the web scale customers, whether that's white boxes or some of the other. Is this a threat to your customer base? And what are you seeing as a result? Any pressure on pricing? Or how should we think about -- they're all saying we're going to go after the web scale customers.

Yes. Jim, I'm surprised they're saying that because it's not like they forgot about these customers for the last 10 years. They've been there every year competing hard to win, and you have to execute very well, not just on building a chip but building a good product, including the software, including doing integration and customization that the customers need and truly being honest and supporting these customers in making the right decision and deploying and solving their problems. And these customers have always been multi vendors. So I wouldn't be surprised if they give a minority share to some of these other players like they have done in the past. But I think Arista will do well. We are very well positioned with these customers. The cloud titans love working with us, and they're giving us more and more work to do and more and more projects that we're involved in with them. So I think we will be fine despite all the claims from competition.

Now when we look back to the year 2019 and so far 2020, Arista saw a meaningful slowdown by 2 of its web scale customers. Can you talk about that? Were they the same issues or different issues? And how should we think about is this pertinent or temporary?

Sure. So there were 2 slowdowns that people remember. Remember, there were 5 years of positive surprises prior to that. So this is a volatile market. And the -- in 2019, one of our large cloud titans didn't really slow down, but they shifted their spend from one quarter to another. So when you look at the full year 2019, that cloud titan customer did fairly well. They actually grew with us in overall spend, and we did well with them as well. So there was no major slowdown there, except for a shift from 1 quarter to the other. In the second case, there has been a slowdown, which we mentioned in our Q3 2019 earnings call in November of 2019, that one cloud titan customer did decide to delay their data center refresh and some growth, waiting for a new CPO cycle, and that investment really got pushed out to 2021. So 2020 is a slower year for that customer. And then 2021, they start again to invest and grow in that space. We're not surprised by that. Cloud companies are very large, and they're also quick in making decisions and changing their plan, so there is volatility. And we've seen 5 years of upside. And 2 quarters of weakness in this overall business. And we're not too worried in the long term, these things to even out. And they like us. They like working with us. So I think Arista will have a very healthy share of the spend from these customers.

And sticking still on this topic. If we look at the first one, you mentioned a shift in spending from 1 quarter to the other. Is the spending levels back to, and I'm not talking about coronavirus because that influences everything. But aside from the shift, is it back to the normal levels or higher or lower? Or how should we think about when they ship this spending? Was there actually impact to the level?

The overall 2019 spend for that customer was higher than 2018. So they actually improved their spending and overall CapEx as well.

Okay. And then on the second customer, did they typically do this waiting for a new CPU cycle? Is that pretty normal? Or is that kind of new?

That was new. And I think to some extent, this also reflects the maturity of their business where they're trying to squeeze out more efficiencies and would rather wait for a new cycle that can get then buy for the next 3 to 5 years rather than investing in something now they feel may not last the full cycle. But so they were able to stretch it out. And kudos to them for making such a bold decision, right? These are not easy to do at that scale. But I think technically, that is the right choice that they made.

I don't think any discussion would be complete if we don't talk about white boxes. So can you talk about white boxes? Because there are many people who think that this is really going to create pressure on the company's profitability, sales growth, and we just hear a lot about white boxes. So let's just go ahead and divulge on the top of our white boxes. There's a lot we can talk about.

Absolutely. First of all, when some companies talk about white boxes, and people think of these companies as the biggest consumers of white boxes, they forget that the same companies are Arista's largest customers. So if they were only using white boxes, they wouldn't be such large customers for us. They are very, very good customers. There's a lot of work we do with them. Lastly speaking, I would say the industry in this space has maintained status quo. Companies that have been designing their own switches including the ones, especially the ones that did this before Arista existed as a company or was a sizable company, those companies are continuing to build on their own. Companies that have been buying from the industry and from companies -- from windows like Arista are continuing to buy from companies like us. They have not gone back and started designing white boxes. In addition, there are a few in the middle that build on their own and also buy from the industry and are more focused on a co-development effort to a common design spec, and they are maintaining that status quo as well. And we've seen some of those announcements enduring from Arista at the previous OCP events. So despite the chatter on white box and so much talk about white boxes, the reality is that this market has maintained status quo for roughly the last 5 years or so. And what you have to keep in mind is these customers have a lot of money, they have needs for technology. And if their needs are underserved or unmet, then they will go a different route. But if you can meet those needs, then these customers do not have as much of an incentive to go invest in this space when, from an opportunity cost standpoint, they could be building other solutions to improve their own business. And that's where the partnership with Arista is key. We are very, very unique. This is not a play where you sell a box and you run away. You really have to be part of their network and help the customer build the best network and run it with them, support them, think about the road map, help them transition and then you have maintain a long-term partnership with them, which is where we are, we believe, with them.

What about pricing, though?

I would say that the cloud market is already pretty aggressive on pricing. These are almost cost-plus type of situations. This is not where you command a very high premium or a high margin. These are very competitive environments, and they have been like this for -- ever since Arista has been involved with them for the last 10-plus years. So I don't think the pricing dynamics have changed or are likely to change in a big way. There's always the need for efficiency. There's always a need to improve, and we have to stay focused on that. So it doesn't mean you just compress margins, it could also mean work along with the customer to figure out how to reduce their costs.

If we switch over to the coronavirus that we're kind of living in today, I'm doing this call from my house as are you and Charles, normally, we'd all be on a stage with bright lights and a packed audience in front of us. Instead, we have a packed subscription list who's all dialing in. Has the value of incumbency helped out Arista, meaning your customers just simply can't open the doors to entertain new products from competitors? Any thoughts about incumbency under coronavirus?

Yes. I would say, it depends on the customers' ability to switch. The largest customers have fairly skilled themes. And if they were committed to switching, they would. So the only way you can survive is by being ahead technically, not by being complacent because you're the incumbent. At the same time, I think some of our larger competitors have been enjoying incumbency in the legacy enterprise space. And those customers are more, I would say, risk-averse and will -- are unlikely to move quickly in the COVID environment. But the incumbency factor for enterprise, I would say, is more important time in the cloud. In the cloud, companies will move when they have to, and they're busy transitioning to some of our own next-gen products.

If we look at CapEx of your customers, and I know not all CapEx dollars are spent to you or away from you or not. But in aggregate, some of the large key cloud customers have been moving their CapEx numbers lower during the past few months. Should we be concerned about that reduction in their CapEx? Because, in essence, their CapEx is your addressable market that you're going after?

Sure. As you mentioned, the networking spend is a very tiny fraction of the overall CapEx. So the overall CapEx includes the very large day center facilities, all of the power and cooling infrastructure, all of the compute and stores they have to buy and the network that goes along with it. So the network spend is a lot less of the overall CapEx. So sometimes, even though the overall CapEx may go up and down, the network spend may not be as impacted. But having said that, some of our customers have been raising their CapEx budgets. And as their earnings are announced and some have lower estimates for this year with the comment that they will start investing again next year. So we are actually not as worried about the changes that all of you are now seeing, to some extent, some of that was built into our models because that's the guidance we were getting from our customers on directionally where they're headed. And we're getting ready for some next-gen upgrades that we'll start in 2021 with them.