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

Hi, good afternoon. I'm Samik Chatterjee. I cover some of the IT Hardware Networking Equipment name at JPMorgan. For the next fireside chat, we are hosting -- we have the privilege of hosting Arista Networks and we have with us Ita Brennan, CFO of Arista Networks as well as Charles Yager from the Investor Relations Group, here with us. Both -- thanks to both of you for attending the conference and participating and making the virtual conference happen. Thanks for doing this format.

Maybe if I can just start off you did report results recently, I think one question that we have been starting off with all our coverage company's is just to kind of set the background in terms of we've heard about supply chain constraints starting -- stemming some COVID. Just a background in terms of how was it in 1Q? What's the road forward in terms of supply chain headwinds? And we can take it from there.

Yes. I think for us the first impacts we saw were really began back in February and we are centered around China, whereas we saw companies, suppliers and manufacturers come back from their Chinese New Year breaks, et cetera. We saw some that didn't come back straight away. And then when they did start to come back, it was kind of a gradual ramp and a gradual return to work. There was definitely some disruption around the capacity, the levels of capacity that you saw there. And for us, that was mainly around supply chain and components. We don't have a large manufacturing footprint in China, but we do have some dependencies as you go deeper into the supply chain. I think following on from that, we obviously had other sources in other countries, et cetera. And as COVID started to impact other countries, there has been some revolving kind of impacts to different countries with different countries implementing mitigation, orders, et cetera, and us needing to respond to those, right? We've been reasonably successful at turning back on supply in a lot of countries given that we're supporting kind of what most company -- countries would deem an essential service. And so we has -- had success at kind of returning some of those factories to production, et cetera. It just takes time and time means no lost capacity as you go through the various scenarios. So that was kind of our -- it kind of evolved as we went through the quarter. We obviously had some inventory, et cetera, in hand that helped us through the quarter. But we definitely were constrained as we came through the back end of the quarter and that continues kind of into Q2. We're definitely operating as in a constrained environment in Q2. And we'll see kind of how things start to evolve then as you move further into the year. But -- so components, on the one hand, some manufacturing capacity, impacts and very much a moving target as COVID has impacted different parts of the world at different times.

All right. Any visibility on when we get back to a normal level, completely normal level on the supply chain side?

Yes. I mean, it's hard to predict what a completely normal level and when you see that. I think the team has definitely been taking actions that start to help and start to have effects. I think if we get stability on the healthcare side of things, you could start to see improvements into Q3, if -- given that there's some stability there. I think we'll just have to wait and see kind of how things evolve as countries start to reopen and start to allow people at our -- on a broader level to go back to work and what the impact of that is. But certainly, the actions that we've taken should start to help as we go through Q3.

Okay. Before I move on, I saw a lot of people dialed in kind of after we started just given the tight changeovers between sessions. So let me remind investors on the line that you do have the option of sending in questions through the Q&A feature, and I will ask it on your behalf as you send them in. So let me move forward Ita on that same line of thought. I think one thing that you mentioned in the last earnings call is your revenue outlook with the cloud titans have improved modestly. Not a significant change, but there's kind of that incremental positive view on that revenue outlook. Maybe help us understand kind of how sustainable you think that is? What's driving that? Just so that we can -- I think there was a general expectation that cloud is -- the cloud guys are pulling in a lot of equipment. So just help us understand what you're seeing from them at this point?

Yes. I think as we discussed on the call, we did see some strength from cloud in the first quarter. And we're working closely with them to understand exactly what their needs are and what their requirements are and working to kind of plan just exactly what are their priorities. And how do we respond to those and then lay out some plans for kind of future weeks and future deliveries. So we've gotten a little bit, I think, more visibility. It's still not by any means through the end of the year, but it does give you some understanding of what they're doing within a reasonable time frame. And so we look at that and based on that, I think we're happy to say that we think that flat to down guidance that we had at the beginning of the year. We now think about that as maybe more like flat on a year-over-year basis -- on a demand basis, for those customers. So I think we did see some incremental improvement. When we think about it in the context of the year, I think that's our view of -- there is an improvement there. We'll continue to obviously work with them and monitor that, but that's kind of how we see it right now.

All right. In general, what we've seen is led by the higher cloud usage that we've seen, I mean, we all are using video conference calls, et cetera. There has been some parts of the supply chain that have shown administrator kind of a pull-in from the cloud companies like Sovos et cetera. Any thoughts as to why networking equipment is in kind of following as closely as some of the other parts of the data center?

Yes. It's always been difficult to draw those correlations, right? And to understand what you see on the one hand in terms of news flow, et cetera, and then versus what other companies are reporting and how they're thinking about it. It's just been really difficult historically to kind of make those correlations. I think for us, it's much more important to work with these guys and understand what they need so that we meet those requirements and reflect kind of our understanding of that as we go forward. And visibility is still reasonably short-term with these customers. So it is a case of kind of working on reasonable time frames and understanding that. I don't know that we can make those linkages to storage or servers or some other parts of their infrastructure. It's not -- it depends on the capacity that they have. It depends on where they're seeing pressure points on capacity and so on. It's just -- it's very difficult to make those linkages from the outside.

Got it. Maybe looking a bit more longer term, did have very strong growth years when we look at 2017, 2018. Some of the moderation in that growth does suggest that cloud companies have been extending or pushing out the refresh cycle for networking equipment and typically probably kind of sweating the networking equipment assets a lot more than they used to. But again, correct me if I'm wrong there. I think the question would be kind of what have you seen on that front? Is there something which is kind of waiting for a upgrade cycle to really move into the next year? Or is this more something structural where the cloud companies have decided that that networking equipment will be on a cycle than in the past?

Yes. I mean, I think if you look back over kind of 2017, 2018, and the demand levels that we saw there, there was definitely some accelerated deployments from the cloud. And part of it is obviously just growth in capacity and growth in their businesses. Part of it was them establishing new footprint, and new locations, new geographies, part of it, just new use cases, us solving new use cases for those customers, et cetera. I don't know that we've seen any, I would say, generic change in kind of the useful life of the switches and stuff that's probably stayed pretty consistent, right? The technology refreshes have been happening on an ongoing basis, and we've seen those. I think the reality is this business has always had some volatility to it. It's just been to the upside historically, and this is probably the first time where we've seen them have some volatility to the downside. And obviously, that has a much greater impact on our business [ when there ]. But I think when you look at their businesses and what's happening and how they're growing and the role that they're playing kind of in the overall technology space. We're still believers that over a longer period of time, they continue to be meaningful customers, and they continue to spend and invest in in this infrastructure. It's difficult to kind of map the slope or the curve of that line in near-term chunks. But if you look at it over time, I think they are a part of the business that's performing well and that you can see that in their results. And they will continue to invest in data center and data center infrastructure. The volatility on a quarter-over-quarter or year-over-year basis, that's something that we have to navigate, honestly. But overall, over a longer period of time, it's still a very good business and a very good set of customers to have.

No, sure. Definitely appreciate that. Let me move to kind of specific to more the 400-gig upgrade cycle that everybody is very focused on. I think there's one line of thinking that we've been getting from some of the companies we're talking to, which is this COVID disruption can push out some of the timing from the cloud providers in terms of that upgrade whereas the other line of thought, at least from investors is why can't -- if there's a bandwidth concern or a capacity concern, why can't the 400-gig upgrade, be kind of pulled forward in that sense. Maybe what are you seeing in terms of your conversations with customers? We've generally heard that your peers are sampling with some of your kind of customers as well. So what are you hearing from them in terms of timing? Is there any disruption here in kind of the process itself just because of the COVID-led kind of the problems that we have at this point?

Yes. No, I think on the switching side, I mean, we've delivered products, and I think we're on track on our product development and product delivery schedules, right? I think the question is -- and Anshul had talked about this a little bit on a prior earnings call. But the first use cases for 400-gig are likely to be around DCI in combination with the ZR optics and some of the optics innovations that we see there. And the question is, how are those progressing? And really, when are you going to have a healthy supply of cost-effective optics to support this? So I think the switching technology, certainly from our perspective is -- has moved along and is on schedule. If there is a delay, and again, we didn't change our commentary, which has been kind of some revenue right at the end of this year and then ramping next year. So I think if there is some change in that, it's likely to be driven by some of the optics folks, and I haven't seen any real updates from them, at least that I've read around where they're at. But I think that's the -- the initial use cases are kind of dependent on those optical -- the ships on the optical side and particularly ZR optics, and that's going to maybe constrain our drive, just depending on how their delivery schedules are working out, what that timing is really going to be.

Got it. I mean, usually, we would take some of the audience questions towards the end. But as I'm reading through, and it's kind of quite relevant to the 400-gig topic. So let me just take it right now, which is the question reads, how is Arista thinking about the size of the 400-gig upgrade cycle? Curious how they're analyzing the TAM opportunity? What are the inputs underlying those estimates? Yes. That's kind of the question. And anything in terms of how do -- I guess, how to quantify the TAM is going to be primary question?

Yes. I mean I think it's always difficult for us to do that on an overall industry basis, and we'll tend to look at some of the industry analyst views around TAM and TAM size and growth and product transitions, probably similar to what most investors are also looking at, right? I think our belief is that you will see 400-gig address certain use cases early in the cycle like DCI, like high-performance high-capacity use cases. And then as you see further product iterations and cost improvements, you'll see it broaden out and address more and more of initially probably the hyperscale parts of the market, maybe some very large enterprise customers and then start to expand out from that. So I think it is -- it will be something of a gradual transition and 100-gig will continue to be super relevant across most of the customer set for some time, right? And so it will be a gradual in our mind transition. The industry folks, it's actually in our investor deck, if folks want to take a look at it just in terms of how the industry analysts see that transition from 100-gig to 400-gig. But it does have something of a starting tail, and you'll see it broaden out as you get further technology revs later in the cycle.

Okay. When we think about the 400-gig upgrade cycle, you mentioned the initial opportunities will be in the DCI layer, primarily starting from a 400-gig. If I understand -- kind of based on the -- what the industry analysts have, it does look like Arista has a much bigger share when it comes to the spine and the layers within the data center and as you get closer to DCI or when it's typically been kind of the other vendors like Juniper who might have more share there. So when you're thinking about kind of the early share opportunities in 400-gig, do you see necessarily kind of being more challenging than the early 100-gig adoption in terms of gaining share? Or are you seeing this as a kind of a similar ramp in terms of market share for Arista as you had seen with 100-gig?

Yes. I mean, I think we're -- we like the products that we've delivered. We believe the products that we've delivered to customers are -- meet all of the criteria. It's a continuation of a consistent silicon, consistent EOS operating system. So for us, it's a very natural transition from 100-gig to 400-gig. And that gives you a lot of confidence that the products -- the performance of the products and the capabilities of the product. Charles, I don't know if you want to take some of the questions around DCI and when competitiveness, do you want to take that?

Yes. I mean, we have, over the last year, introduced 10 new 400-gig products. Some are based on local switching, some actually do the DCI switching and routing. And I think the beauty of it is that we have the same product that can do switching and routing with the same software. So that's a certain appeal to the cloud guys who don't want to have to necessarily switch software or operating systems when they're doing local switching or DCI. So I think we feel pretty good about our product offering in this initial wave. And then as Ita said, there's going to be cost reduction. There's going to be a whole new generation of 400-gig with the cost reduction, and we pride ourselves on being early to market with the latest silicon technology and fully qualified software to do that. So we feel good about the 400-gig opportunity.

Moving to white box, I mean, definitely, we're discussing white box less than we used to 2 years ago because their momentum appears to have slowed even with some of these Tier 1 hyperscalers. But one question that we kind of keep getting from investors is, what prevents companies like Facebook, Microsoft that are still leveraging a lot of the branded vendors in the future following down the same path that Google and Amazon have gone? We understand that white box momentum has slowed in the layers that they're kind of looking to displace branded vendors. But longer term, is there necessarily a different path for Facebook and Microsoft than like Google and Amazon?

Yes. I mean, I think you have to go back to kind of the history of how we kind of got to where we are in terms of the use of white boxes in networking, right? If you go back to when Amazon and Google were bringing their products and bringing their cloud platforms to market initially, the products that they really needed to deploy from a networking perspective just weren't available in the market, right? Then a lot of the initial investments that they made were really designed to enable them to bring -- to enter the market at the scale and the size that they thought they were going to need to do that, right? I think that's a desire, I think, for white box in a lot of these cases, it was always driven by just the enablement of the automation and control that they needed to have of their data center footprint in order to support the businesses that their businesses and their services to customers. I think as time went on, you start to see products like Arista's platform in the market, giving all of the control and accessibility, programmability, et cetera, that they needed and wanted to have from a vendor solution, right? And that that enables people -- companies' entering the market a little bit later, like Microsoft and Facebook, and others, sense, right, too, to rely on the vendor base and to deploy that that technology and drive the same level of control of their data center and of their services that that they really needed to have. I think that's been the motivation as you've seen this evolve. So there will be for various different reasons for different roles in some of these networks, you will see some white box solutions. You'll see certain parts of their network. So you'll have more competitive vendor solutions, honestly, over time. Our goal has always been to understand and embrace those efforts, right, and then help solve problems and solution problems for these companies in other layers and in more complex roles in their networks. So that's worked out very, very well for us, and we'll continue to do that. So I think the white box it doesn't go away, right? And I think we'll still be talking about it in 2 years’ time because it's an element of what goes into making these and driving these large footprints. But from our perspective, it's not a threat. It's something that we kind of work with and work in conjunction with and deliver the products that that really solution what these guys need, and that's where their -- that's what's driving their decisions. Charles, I don't know if you want to add anything to that or no?

Yes. I think that going exclusively white boxes may be a more risky approach in some ways, also very expensive approach for cloud titans because they have a tremendous development organizations that they have to put together. So having multiple suppliers, one merchant and one white box may be more cost-effective and competitive approach for this cloud titans.

Let me go ahead and take the second question that came in just because, again, this ties into the white box theme and is more centered on the U.S. software and the value-add for Arista through the U.S. software. The question reads, one of the themes this year at OCP is disaggregation of everything not just hardware from software, but even splitting traditional monolithic NOS into smaller components. If this trend continues, where does Arista view its continued source of value-add to the cloud titans?

Yes. I mean, I think we've always embraced kind of new opportunities to, be it open networking standards, be it support of -- we've supported SONiC, and I think that's a lot of what this conversation is about. We've supported SONiC in different use cases inside some of these customers already, right? So I think that still fits in the context of what we just talked about in terms of white box, honestly, right? There will be initiatives inside these large networks, where they want to drive different technologies and different solutions. And again, we'll embrace and work with those, and we've been doing that for a long time. We've been running U.S. with SONiC, we've been running, enabling SONiC on our hardware, et cetera, for some time now, right? And we'll continue to do that. And we saw some announcements again today, where we've continued to release products that support that. Again, these networks are vast, and there's many roles within these networks. And there's room for having different solutions, both on the -- from a software perspective and from a hardware perspective. I don't think the principal or the engagement changes, right. There are parts of the network where EOS and the capabilities of the U.S. and the combined solutions that Arista is bringing are very important and very relevant to solving those and there'll be parts of the network, which, honestly, are very competitive today already, where you'll see these companies choose to deploy different solutions. And again it's all targeted and ensuring some resiliency and multi-sourcing in their networks. And that's not something that we're concerned about, we continue to work with that. Charles, I don't know if you want to add anything on it?

I just think -- even though you have a lot of disaggregation, there's always room for innovation and improvement, and we keep striving to find ways to value add because nothing is really standing still. And the demand for improved networks will always be there. And so we'll always keep innovating even with disaggregation.

Got it. So let me move to the enterprise side of the business. And thanks for bearing with me for all these questions on the cloud business. One of the most interesting things that kind of as you announce your move into campus was the expansion of that addressable opportunity that you have, which now gives you much more opportunities for growth than you had just within the cloud data centers. So can you help us kind of just size of that opportunity, you steadily rolled out more products targeted at the campus. So how do you think about kind of the increase in that addressable TAM that you've been going after?

Yes. I think if you look at kind of the evolution of Arista's TAM over time, and again, we have some charts in our investor deck that that looks at this. I mean, we started out with a theory of cloud networking, and over time, really what we've been doing is extending that cloud networking capability through different parts of the market, right? It started with the hyperscale cloud customers, it expanded into kind of specialty cloud and other cloud players then into the enterprise. And we've seen some strong adoption in the data center of the cloud -- of cloud networking and cloud networking principles. And now into the campus, right? So it's really just a continuation of this extension of cloud networking across the principles that underpin that across various parts of the network. And as we've done that, obviously, you add to the TAM, right, be it new parts of the market that we didn't address before, like enterprise data center, routing was a part of that when we added, you had some time from routing. And now more recently, the campus market. I mean, the total campus market is a $10-plus billion market. We have a full set of products to address that. We're targeting large enterprises and the large -- the larger enterprise piece of that market right now. I think we're well on our way to our first $100 million as part of that. But obviously, that's very small relative size of that market. So there is a large TAM there for us to address. It won't happen overnight, right? It will take investment on go-to-market, et cetera, it will take time to -- for customers to make those decisions. But we've seen some really strong interests in the product set, and I think we're off to a good start, right? So it's been a gradual expansion of TAM, driven by just a very consistent principle of enabling kind of networking end-to-end, managing networking end-to-end, providing the tools to customers to be able to manage a single image across their network. And that's -- we've seen a lot of receptivity for that from customers as we've made these transitions across the market.

Got it. Let me just remind investors that you have any -- if you have any questions, you can send them through the Q&A feature. Just keeping on the enterprise business, Ita, for a bit more. In terms of -- as we now kind of are in a environment that is very uncertain, not a lot of enterprises know which way it’s headed or how strong the recovery will be, there is a general expectation of a moderation in spending from the enterprises. But at the same time, I think most of your exposure was to kind of these financial services firms where, I think, the expectation for weakness is much more modest than kind of the broader market. So just help us think about what you've seen till date in terms of weakness from -- or any order trends from your enterprise customers that might indicate whether it's kind of more resilient than you would have expected? Or is it more in line? Just help us think through that?

Yes. I mean I think combined between financials and other enterprise, right, it's a significant piece of the business. Now it's about 35% of kind of the demand -- the underlying demand of the business at this point. So I mean enterprise/financials is, it is a significant part of the business. It's been growing very nicely for us. I think Q1, and what we've said on the call is we didn't see a lot of impact in Q1 largely because I think on the demand side and some of the demand side, it likely even causes of impact. It didn't start to happen until later in the quarter, right? I think our sales teams, et cetera, have done a nice job transitioning to online sales and interactions with customers, et cetera. that's been -- I think that's been going well, right? But I think what you heard us say on the call is, we do have some concerns about as we move through the back half, the second half of the year, in particular, do you start to see some lengthening of time around decisions, around spending in that part of the business, and it's going to be somewhat gated by how the resumption of work and the resumption of the modifications of shelter-in-place orders, et cetera, take effect. But when we look at it and we say, we see the potential for there to be some slowdown -- potential slowdown as we go through that back half of the year. Again, we don't have perfect visibility. I don't think anybody really has a full understanding or visibility to what's going to happen there, but it is definitely a scenario that we've considered. And we talked a little bit about how we'd run the business if we did see that impact and managing OpEx through that period. And I think that's how we think about it, right? It's a period of time where you could see some impact in the top-line. If you do, we can respond through some of the spending changes that we talked about and still maintain kind of the focus on key projects and maintain the sales and marketing investments, et cetera, that we've made over the last couple of years and kind of ride out that period. Again, we don't have -- we're all looking at this in a future lens, right? But we do have some concerns about what that would look like or might look like.

Let me clarify on that front. One comment that you had on the earnings call recently because you kind of positioned the business to run or operate at 2018 levels. How much of that is a reflection of kind of the OpEx commentary in terms of running the business in terms of operating expenses at the 2018 level? And how much of that is to understand that based on your current visibility, if that's kind of the revenue, where you're thinking about positioning yourself in terms of the revenue outlook?

Yes. I mean, we're not trying to guide the year. We don't have the visibility to do that at this point. I think what we were -- what we did internally was to look and say if there was an impact, how could you run the business, right? And could we get consensus around what levels of OpEx would you need to maintain some of those key investments and initiatives that that are ongoing. And how much variability did we have in the model to allow us to respond if there was some change in the top-line and some impact from COVID. So I think the message was more that we could run the business at this level depending on what we see at the top-line. I think based on our experience, et cetera, we believe it is likely that there will be some impact from some of the disruptions that we've seen as we go through the second half in terms of exactly how that plays out. I mean, we don't have certain knowledge of that by any means. And there is a relationship, obviously. I mean, if things were on better to none, and unchanged, then we would resume some of that spending, all within the boundaries of our prior commentary around the operating margins and 35% plus or minus, right? We're not looking to drive a 40% operating margin in this time period either. So there is some connectivity between the OpEx and the revenue. But we're not -- we were not trying to guide a top-line number at this stage. We were more trying to explain how we would run the business and what would we do and what can we control, if there is an uncertain environment as we go through the rest of the year.

Got it. And just to kind of then as we start to wrap up here, the last question would be, if I keep that commentary about the OpEx being at the 2018 level in mind, you have also these aggressive targets on the campus revenue site for $1 billion over a few years. How does that impact kind of how aggressive you can be towards pushing towards those targets? And does that running at a lower OpEx level mean to some extent that you push out some of those revenue targets in terms of the timeline that you would have expected earlier to meet them?

Yes. I think we're well on our way to the first $100 million, and that kind of times out at the end of this quarter, and then we'll take another look at where we are then and then provide some further outlook around that. I think we've made some significant investments in the sales team. We've certainly expanded our coverage model, all in support of both the enterprise data center and campus initiatives, right. And we'll certainly do everything we can to keep those intact. We believe that regardless of whether there's some slowdown, et cetera, over time, those are very good investments for that business, right? So we'll continue to target those investments towards continuing to expand campus, the exact pace of that. I think Jayshree had some commentary on the last call that we need to be patient. We are seeing good success. There is good adoption of the products, and we'll take a shot at what kind of the next milestone should be maybe on the next earnings call once we have some more understanding and we've kind of achieve the $100 million.

Great. Just in the interest of keeping everyone on time here, given the schedule -- tight schedules, thank you both for participating at the conference. And thank you to everyone dialed in or listening in as well.

Thank you.

Thank you.