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

Great. Good afternoon. My name is Paul Silverstein, Senior Analyst, Networking and Communications Equipment at Cowen. And thank you for joining us this afternoon. And I want to welcome Arista Networks' Mark Foss, Senior Vice President, Global Field Operations and Marketing; as well as Curtis McKee, Director of Corporate & Investor Development. Mark and Curtis, thank you for joining us. I've been watching you now for -- I think I wrote the first piece of research back in 2011, 3 or 4 years before the company went public. And the title was Products in Pedigree. The pedigree of course referring to Jayshree and Andy and Ken and Dave Cheriton and how that opened doors for the company. And it's been a great journey since.

Mark, given your responsibilities -- well, I've got a lot of questions for you, hopefully, some of which we'll get through. I really like to start asking you about as the company ages and looks to expand into new markets, both from a regional and from a product market, customer market perspective, where are the opportunities and where are the risks in the enterprise market in particular as well as outside of the U.S.? I think it's fair to say, historically, the company, understandably, as with all young companies, the focus first was in region in the U.S. And outside of region, good news, bad news, you're underpenetrated. But, of course, that creates opportunity. So let me pause and let you respond before I ask you some more specific detailed questions.

Sure. Yes. No, I mean Arista, as you know well, Paul, we started out focusing on the data center, data center switching in particular. And I started in '07, by the way, and I've seen the company go from 0 revenue to where it is now. But we focus on the data center with our strategy of a 21st century software operating system as our main differentiator, and delivering that operating system on off-the-shelf components and merchant silicon, and that's been our strategy ever since. So we focused on the data center and went from 0 to roughly just over $2 billion in about a decade. And as of recent times, probably in the last 2 years, our customers in the data center were so pleased with our software. And I can go over any detail on the details on our software. But in a nutshell, our software is a very, very high quality, highly programmable. It's a 21st century operating system. And they said, "We're so pleased with the performance of your operating system in the data center. Why can't we use this in the campus market?" So we thought to ourselves, "Now it's about the right time." And if you look at the total addressable market for the data center and the campus, they're about the same size. They're both -- depending on the analyst, you look at that, they're both in the roughly $12 billion range. So what we decided to do a couple of years ago is start investing in the campus market as well. And we started out with acquiring a company called Mojo Networks, which does cloud WiFi. And then we did some development on some specific campus PoE switches that provide power Ethernet for IoT devices or wireless access points for phones or whatever. And we started shipping those in Q2, late Q2 of 2019. So we had a nice campus portfolio then also leveraging our software. So I guess, the adjacent market that we've been going after for the last really year seriously has been campus. And our goal that we public -- stated publicly is to do $100 million in the first year, and we're on track to achieve that $100 million in the first year going after campus. But in essence, the campus market pretty much doubles our total addressable market.

And Mark, to be clear, correct me if I'm wrong, but Jayshree has said you'll get to that $100 million mark in the current June quarter, which is the fourth quarter of that 4-quarter progression, you all having only entered the campus enterprise market 4 quarters ago. At the same time, if I recall, Jayshree understandably acknowledged that the current pandemic, the crisis we're all living through, that makes it more of a challenge for all new entrants into any market, and that certainly applies to Arista in campus networking. It will make it more challenging for you to address IT staff. There are fewer people on-premise to speak with. And just like in my business, there's an element, as you well know, of wining and dining and the phasing that goes on as well as getting products into lab trials. Any thoughts -- I know it's still pretty early, but any thoughts on how you all are going to confront that or overcome those obstacles in order to continue to drive healthy growth in this very young business?

Yes. And that is a concern, right? I mean when we looked at our first 12 -- I guess, 9 months now of campus shipments, roughly half, if not a little over half, were actually new logo customers. And we were expecting that in the first year, almost all of our campus business would be existing customers. So it was actually kind of a pleasant surprise that over -- approximately half were new customers. And you're absolutely right. It is a little -- we have -- do have a little bit of caution over the next, however long this pandemic last because face-to-face meetings, in-person talks, everything that goes with it is very key to acquiring new customers. So what we've done is we've doubled down and focused on our existing data center customers who know us very well. We've also gone heavily into virtual events, I think, that everybody has. So we've stepped up our global webinars. We used to do like 1 a month, now we do 1 a week. We've stepped up all of our regional events. We have this event called Campus Builders, which we're doing -- in some weeks, we're doing 3 or 4 a week, and these are regional events. And interestingly, we are getting prospects on those virtual calls, which actually surprised me. So we're doing everything we can. But at the same time, you're absolutely right. It is going to be more challenging addressing this newer market during these times.

I think I know the answer to the next question, but I will ask anyway. Is there any quantification you can give us? Jayshree put out that $100 million number that you're all on track to deliver over the balance of this year and through to calendar '21. Any thoughts for how far, how fast you could grow the business from that $100 million base recognizing that, again, COVID-19 has thrown a monkey wrench? It's going to slow down that progression. What are your thoughts for what you could accomplish?

Yes. It's difficult to say right now. Our goal is to get to be $1 billion -- have it be $1 billion market for us. We felt it would be -- it's a multiple-year journey. We're originally thinking 3 years, but we'll see how the whole pandemic thing pans out. But our goal is to get -- make this $1 billion market for us. Within 3 years, we'll see if that can still happen given the situation.

So you mentioned earlier that the data center enterprise market and the campus enterprise market, switching market about the same size. I would actually argue, if we take a larger perspective of campus enterprise, i.e., we think about not just the switching component, but also wireless LAN, security, automation, orchestration or visibility, application visibility, et cetera, much of which today you partner with for, albeit you now have your own wireless LAN having acquired Mojo, but there's obviously a significant opportunity. And you all obviously have extremely fine engineering talent. Your track record speaks for itself. I might observe that a lot of that talent actually came from Cisco, but regardless, and so the people behind the products. But my question to you is your cloud data center business, you came out with products in an architecture that was extremely adept, that was industry-leading and attacking that market. With respect to enterprise, I can see arguments on both sides. You have the people behind those products, but it's not nearly as straightforward. You've got a competitor that's been in that market for 20 years that has 100-plus different products, that has a sales force that dwarves yours, a channel that dwarves yours. And that's a challenge. The products alone, and correct me if I'm wrong, you've got about half a dozen today. The question I have for you is, how can you create differentiation? What's the differentiation you're driving the customers that have opted for your campus solutions to date? Is it just for the switching piece? Is it for switching wireless LAN? Why are they choosing you? What's driving that choice?

Sure. And I think the answer in one word is software. If you go back 10 years ago, we were getting these same questions, maybe even from you in 2011, about the data center business. It was exactly the same situation. And fundamentally, customers are tired of deploying 20th century software. They are looking for a 21st century operating system that they don't have to live in fear of doing -- of upgrading to the latest version, for example. Campus customers traditionally live on a very, very old version of software because they're -- they live in fear of upgrading to the latest version and then having a bug come and take down their network. So this obviously opens up some security issues for them. So with the Arista software, the way it's designed, our latest version is our best. And the software doesn't regress because of the way it's designed. So I mean, that's just one example of a situation where our software is truly the best out there. And customers are looking for choice. I mean if you look at the large enterprise campus segment, which is about half of the market, this is the Global 2000. That's about half of the market right there. And Cisco has about 85% or 90% share in that segment, which is abnormally high. But when talking to customers, they're really anxious to look for choice in that market. And this is why I believe one of the reasons why we had such a high percentage of new customers coming to us in the first year of our shipments because they're just simply looking for a good alternative out there.

Mark, any sense -- well, let me shift gears to your statement about software. Where is Cisco -- where is Arista in enabling full cloud management on campus similar to Cisco Meraki or Juniper's latest offerings? Is that fully enabled at this point?

Yes. We spent a lot of time integrating our CloudVision software with our Mojo Cloud management software as well. So that's fully integrated now. And we use the same CloudVision suite that we use for the data center and the campus. So we've kind of made them one. And customers like that, the fact that they could operationalize their data center or their campus in exactly the same way.

Okay. Mark, do you think -- it's a near-term question, and while it's not my particular focus, do you think you can grow your enterprise campus business in the second half of this year given those challenges in terms of go-to-market that we were discussing?

I think if we do the right things, like -- you've got to remember, we do have over 6,000 data center customers. Many of these are Global 2000s and Fortune 500s who have some pretty big budgets. So I think if we focus on those and get them to embrace our campus solution, it's -- I think we can do it. The thing about campus is a little more lumpy than the data center market. Campus, they'll upgrade it and then they'll walk away for 3 years. So you got to -- campus, you got to actually time it as well. It's not like everybody is just consistently buying campus products. They upgrade it, and then 3 years later, they do it again. So timing is also a big issue there as well.

Okay. Mark, beyond -- well, within enterprise, to date, that $100 million that you all are driving to, is almost all of that campus switching? Or are you also seeing success with the wireless LAN products before you get to the point that one is pulling in the other?

Yes. No, it's a combination of everything. It's a nice mix of the WiFi, PoE switches, non-PoE switches for aggregation. You have got service in there. You've got optics in there. You've got CloudVision in there, and then you've got other software in there. So it's a nice mix of all of those.

All right. In your sales role beyond enterprise, you spend a lot of time in enterprise. What are you most excited about in terms of market -- the incremental market opportunities that Arista has advanced into over the years, whether from a regional perspective, think about non-U.S. business, particular regions of the globe? From a product market or from a customer market perspective, what are you most excited about that can drive the needle over the next 2 to 3 years and over the next 12 months?

Obviously, campus, I'm really excited about. That's my roots and where I started, so I'm personally excited about campus. But I also ran Arista's international sales early on, and our strategy back 10 years ago was to focus on kind of the top 10 countries that had data center spend. So we really focused our efforts in those top 10 countries. Over the last decade, we've obviously expanded out. And I think we've sold into over 80 countries now globally, if I'm not mistaken. And we're just getting into -- we're putting salespeople now into countries that we haven't touched a lot before, like we recently added someone in New Zealand and within the last year. We added someone in Vietnam. So we're adding people in countries. And we're seeing nice opportunities in these countries that we never touched before, and it's just great to see. So there's lots of good stuff going on.

Mark, I trust we're all aware, fairly well aware of the challenges in certain countries like Brazil that I just mentioned where COVID-19 hit them later, but it's hitting hard throughout Latin America. And so when you talk about that opportunity, are you speaking more from a 2- to 3-year longer standpoint? Or are you actually seeing the opportunity notwithstanding the macroeconomic impact from COVID on so many countries throughout the globe? Or are you seeing that opportunity here now?

It's really hard to -- it's so early, but it's really difficult to kind of gauge it right now. But there is -- there are differences like, for example, like Korea, they are still doing in-person meetings. Restaurants have been opened the whole time. And we're doing proof of concepts in the Arista office there. So Korea is probably one outlier I could see, which hasn't really been completely affected by the pandemic. But other countries like Australia was actually pretty -- hit pretty hard. They're opening up now, but for 2 months, that business kind of shut down. So each country is different. But it's really too early to tell what the effects are, if it's going to come back quickly or if it's going to be a 2021 type of thing. It's really, I mean, anyone's guess.

And I trust that statement, it goes without saying, that notwithstanding that real time, in some cases, only a matter of days; in other cases, maybe only a matter of weeks, that there are certain states throughout the country that are opening up. Similarly, there are certain economies throughout the globe that are opening up. I trust it's too early for your statement, too, for you to see a meaningful improvement in demand.

Yes. It's really early. And we're seeing situations where customers are delaying purchases, and they're -- they want to hold cash and they're saying, "Oh, our campuses are closed for the next 6 months. We're not going to do anything." And other ones are saying, "Hey, now is a great time to upgrade because no one's there." So we're seeing it all over the map. But it's how this -- what the overall big picture plays out is still uncertain.

All right. And for those of my friends in the audience who are wondering when I'm going to get around to asking you about progress at Facebook, whether there's been any changes, I'll get to it at the end. But given your role, Mark, I want to stay on these incremental opportunities. I think Jayshree and Andy and Anshul have discussed Facebook and Microsoft at length, but we'll come back to that at the end, time permitting. If I look at service provider, including cable operators, traditional communication service providers, again, you all have great platforms. You added routing code onto the platforms. I think now it's 2-plus years ago. And on the one hand, you had some pretty good early growth. On the other hand, I think you all found it, for Jayshree and Anshul's public comments and your own public comments, that have proved to be somewhat more challenging than you would expect, not to say that you haven't done well. But right now, the business is not in growth mode. Any thoughts you can share on go-to-market challenges, what you're doing to improve the outlook for that business? I know it's only 10 -- not that much more than 10% of revenue that could be important middle growth.

Yes. We were successful early on with some service providers. And where we've been successful is if the service provider wants to embrace the cloud architectures and not necessarily have to tie those architectures back to their legacy. And we have had various service providers that have done that, and we've been very successful there. I think where we ran into some issues in the last couple of years is as we've gone out to the more international service providers, they have had more of a requirement to tie their new build-outs back to their legacy, which means that we would need to potentially develop legacy features for those tie-ins or new features to make that seamless. And that's where we've been having a few issues is just feature development to have the new architecture work seamlessly with the legacy. And some features will never do. Some we're working on and we have delivered over the past 12 or 24 months and are making things better. But I think that is -- the way I see it, that's been fundamentally the stumbling block that we've had there.

And Mark, given that it sounds like to your statement, you're working on a number of the features, albeit, not all of them. In terms of the conversations you're having with those service provider customers, is there a view to sort of visibility to that -- the revenue from that customer segment stepping up in the next 12 months or so?

Yes. No, we've definitely had some nice wins, like in the areas that I work in. I still have some responsibility for some Asian sales. And then we had some nice wins at SK Telecom in Korea, which is their 5G network; Vocus, which is a Tier 2 service provider in Australia. We've had some really nice wins in these various service providers. So there's definitely traction going on. But as you mentioned, it's -- I think it's fifth out of our top 5 verticals, if you -- so it's not our largest vertical, but it's definitely over 10%.

Okay. One last question on the service provider front. Cable TV operators, particularly in the U.S., but also around the world, the architectures are changing, they're going to deeper fiber. Many of these projects are really not yet rocking and rolling in terms of meaningful deployments, but is there a meaningful opportunity there for you? Or is it the same issue that you just referenced in terms of feature sets?

No. Cable providers are actually -- they are actually some of the early adopters, and they're the ones that we're continuing to get in various areas of the world. I think one of the big things, trends that we're seeing is in media and entertainment as broadcasting goes from analog to Ethernet, there's a big opportunity there. So we're actually seeing media -- lots of strength in media and entertainment from the broadcasting companies that are moving to Ethernet for their infrastructure.

All right. We've discussed the enterprise campus. We've discussed now the communication service providers. How about the Tier 2 web, what you call your specialty web, which, correct me if I'm wrong, but those folks have been somewhere in the 10% of revenue range. I know you all have been not as crisp as I would like you to be, but I think you're around that 10% range. What are you seeing? The dialogue we've heard from you all over the past several quarters has been mixed. At one point, those guys were hurting for you and the tone and the substance of the comments were not very optimistic to say the least. But more recently, the comments from Jayshree, you and the rest of the management team, has been more constructive. What is your outlook over the next 12 to 24 months? What is going on for the customer segment?

Yes. Sure. I think in the last 12 months, with the exception of Q1, we saw some softness in the Tier 2 cloud providers. And I think there's a couple of things going on. Number one, they're moving a little more of their workloads to the Tier 1 providers is what we're hearing. And then secondly, what's also interesting about this market is it's a lot more lumpy than the Tier 1 cloud market. The Tier 1 cloud providers buy every single quarter. It's a pretty uniform amount of business every quarter. The Tier 2 cloud guys, some of these guys, they'll buy -- let's say, they'll buy in Q1 2019, and then they'll buy 0 for like a year, then they'll come back and they'll buy some more. So it's a little more lumpy and an unpredictable market. So we saw them come back. A lot of them simultaneously came back in Q1 of 2020. And it's -- it could be a function of the lumpiness, but it could be a function of them continuing to reinvest in their business, and it's just that. But we're still -- time will tell to see if this trend will continue or if we just had a really lumpy quarter with that.

Got it. All right. I got to ask you. I know I'm sure a lot of folks are wondering why I haven't asked yet with respect to what has been your 2 largest customers, Microsoft and Facebook. With respect to Facebook, where you all have publicly stated that they skipped a server cycle, and that resulted in some dislocation. Any insight? What's the latest and greatest? Or what you're seeing, what you're expecting from that customer, in particular, and Microsoft?

There's really not much change from what we stated last quarter or the quarter ago. I think we were saying 2021 was kind of when things were going to pick up there. And I don't think anything has really changed since then. So there's not really -- I hate to disappoint you, but there's not much to -- much else to talk about.

Asked and answered. Let me ask you about the rest of your "cloud [ twinzen ]." At the risk of being unfair, I'll observe, and I say this to Jayshree directly, you all have never -- in fact, I don't know that you've ever named exactly who's in that bucket. I think it's changed a little bit over time, to my chagrin. Also by extension, up until recently, we never really knew exactly how large that bucket was. Based on last quarter's disclosure, which I appreciate, it sounds like perhaps not surprisingly that the balance of your cloud titans, the Googles and the Chinese, and I don't think Amazon buys from any of your folks, but I'll throw them in the bucket, too, and perhaps Oracle, I don't know. But it sounds like those folks collectively are 10% or not more than 10% of total revenue. So not huge, but nonetheless important, especially from an opportunity standpoint. Any thoughts in terms of the potential for one or more of those to step up from, as best as I could tell, either being single-digit millions, in some cases, to tens of millions in others, but far off the hundreds of millions that you're generating from Microsoft and Facebook? Any opportunity, any prospect for one of these other to make that meaningful cost base?

Yes. There's always more opportunity there. I think some of the ones -- some of the cloud titans that build their own stuff, there is an appetite oftentimes for them to do some peripheral projects outside of their core business where they are willing to go vendor it. And these guys are -- they spend a lot of money so these peripheral projects can be fairly sizable sometimes. So I think there's definitely opportunities on the peripheral side of these large companies. On the Chinese cloud providers, inside of China, they're pretty much by Chinese only. I think you've got Huawei, H3C and [ Luigi ], which have pretty much locked up the market inside China. Now however, when these guys build data centers outside of China, we definitely get invited, and we've definitely had a few wins there with a couple of the large Chinese cloud collectors for data centers outside of China. But inside of China, it's pretty much all Chinese, so plenty of opportunities.

In that position, your ability to sell to those Chinese operators, sort of the Tencent, Alibaba, Baidu, that's been the case for a while. You've had success outside of their China footprint. How meaningful an incremental opportunity can that be?

It's -- I mean, it's -- I mean, the lion's share of their business goes inside China. So if you look at their spend, it's -- the outside China spend is small compared to their inside China spend. So I mean, it's nice, but it's not a Microsoft or a Facebook by any means.

Got it. Mark, what -- where do you think, from your perspective, the company is most under-hitting? Where does it have the greatest opportunity?

I think there's a lot of interesting stuff going on in, like, software. Obviously, we just acquired Big Switch Networks, which really steps our monitoring capabilities. We're currently integrating that with the Arista hardware and software. I think just continuing to develop automation and advancements in just software services, I think that is kind of the biggest thing there and bringing -- I think bringing some of these software automation services to the campus. Because if you look at campus historically, I think Anshul always jokes that the greatest automation tool that campus administrators have is copy and paste. They really have -- don't have any automation. And I think bringing these automation tool sets that were originally designed for data centers into the campus, I think, is a really big deal. I think it's going to be really successful.

I just hosted a call with David Cheriton on that very topic of automation. And I agree with you that it's certainly one of the big ticket items that will improve customers' lots and represents a nice revenue opportunity. So on that point, CloudVision, it obviously adds value to the overall solution set that you sell. In and of itself, your ability to monetize that automation, in and of itself, i.e., above and beyond the fact that it makes all your solutions more valuable, any thoughts you can share?

Yes. No. We sell CloudVision as a subscription service, and it goes on top of the switches. And it's -- we don't -- it's not like we double the price or anything, but you could assume it's maybe about a 10% or 15% incremental revenue on top of the switches. Of course, that's spread out over multiple years because we recognize that revenue over -- it's probably average about 36 months of when we recognize that. But the CloudVision for our new customers and the more recent customers in the enterprise, it's been -- the attach rate has been very, very strong. We've been actually very happy with how CloudVision is sold.

All right. I think we've got time for one last question. I want to end on this. On the one hand, you all obviously made a very different in what hindsight being 2020 turn out to be a very smart architectural choice in focusing on software and letting the Broadcoms of the world worry about the silicon. And you've delivered tremendous value through that. So you haven't done a cookie-cutter approach, just the opposite. On the other hand, not surprisingly, we saw you acquire Mojo for wireless LAN. And the question is, over time, as you need new worlds to conquer, new market opportunities to expand into, is it a given that you will look more and more like Cisco in terms of not partnering for security or partnering for network visibility or app visibility, et cetera, et cetera, but actually, that you need to own those assets, and the only question is when or which particular asset, et cetera?

Yes. It's definitely part of our future strategy. I mean we have over $2.6 billion in cash. We will consider M&A when it makes sense. It's got to fit in with the others, not only the technology, but the company has got to fit in with our culture, right? There's got -- there's a lot of aspects we have to look at. And we've got to have -- it's just got to be the right fit from all angles. So we're not going to do this on a cavalier basis. But if it makes sense, we will not hesitate to do M&A.

Greatly appreciate you joining us today.

Thank you very much.

More success. Stay safe.

Absolutely. Thanks, Paul.

Thank you.