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

[Audio Gap] into a degree enterprise. So I guess, Ita, to you, areas of surprise, supply chain headwinds limiting some upside to the quarter, but kind of what were your takeaways?

Yes. I mean I think looking back on the quarter, we had a reasonably stable quarter and probably better than what we would have anticipated as we kind of come into the quarter. We obviously finished ever so slightly above the upper end of our guidance range. I think on the cloud side, things played out pretty much as we had expected. We'd already kind of talked about cloud kind of having stabilized and maybe being incrementally better for the year. And I think we saw that continue on that path. Enterprise probably performed better than maybe we would have expected. We saw some good strength in the enterprise demand in the quarter, clearly indicating that some of these larger enterprise customers continuing to be -- to engage, continuing to deploy. So I think that was a net positive kind of versus where we might have expected things or thought things could be heading into the quarter. So I think all in all on the supply side, we continue to execute against the supply chain challenges. I think we're making improvements there. That is stabilizing. We're still somewhat constrained in Q3, and hopefully we start to come out from under that in Q4, is our current thinking. We're being pretty aggressive around sourcing supply, sourcing supply in the supply chain. And we'll add some -- you saw a little bit of this in the filings that we did for the quarter, but we'll start to add some purchase commitments and some inventory probably beyond our normal time range, just to give some comfort to customers as we navigate through whatever happens next on the COVID side of things.

Okay. Charles, just to weave you into the conversation, enterprise did a little bit better than expected as Ita just said. Data center, campus, a little bit of both. And why does Arista do well in the enterprise? I mean you guys have been chipping away at this for a while now, but just kind of what you guys saw in the quarter and differentiation?

Well, yes. On the campus, we did do our $100 million target, which was something we set out to do. And so that was completed in Q2, which is cumulative revenue for the last 4 quarters. I think enterprise data center continues to do well. I mean COVID is probably having more of an impact on the campus than in the data center. So I think enterprise data center has done well. And I think that when you're dealing with the larger companies and you're not quite as exposed to some of the industries that are more sensitive to the virus and the impact, I think that was probably why we did a little bit better than we thought.

Okay. So going back to last quarter, I guess this is -- we can start with Ita here. You upgraded your view on the fiscal '20 cloud titan contribution from flat to slightly down to flat. You reiterated that view again on the second quarter earnings. So what drove that change of view? What are you seeing as far as timing of new regions, new investments capacity? Kind of what gives you that confidence in how the second half could play out with that segment?

Yes. I mean I think we have seen that business, and we've started to see this even coming out of Q1, stabilize. And obviously we're working closely with those customers just around prioritizing supply needs and helping kind of make sure that we're organizing ourselves to best meet their needs in the near term and in those time frames. I think based on that, we feel like we can see kind of the stability and some incremental improvement, if you like, versus where we thought we were on that part of the business coming into the year. We'll continue obviously to work with them as we go quarter by quarter here, but this based on what we can see now, and with the kind of planning exercises that we're working through with them, I think we see that kind of incremental improvement.

I know, Ita, in the past, we've talked about how much visibility these cloud titans give you as far as timing of projects. And it can be all over the place, right? They give you a 3- to 6-month view on projects, but then all of a sudden, there can be kind of like this still and kill kind of capacity need at the end of the quarter. Has that changed at all as far as how they kind of project their need for your products during the quarter?

Yes. I mean I don't know that we've gotten any kind of -- we've always talked about this 1- to 2-quarter project visibility. I don't know that anything has changed tremendously around that. I think in the near-term planning process, there's probably a little bit more discrete work going on there in terms of just the mix of the products and the prioritization of the products, but has been helpful to make sure that we're using the capacity that we have, et cetera, to their best advantage, if you like, in this time frame. So I think there's been more of that. I don't know if there's any fundamental change to just the responsiveness that we need to have to their business demands over time. And I think that 1- to 2-quarter kind of project visibility is probably still the right way to think about it, but with an order book that's -- lead times that in a normal course of events is much shorter, right?

Okay. I think going back to some of the questions around Facebook last year and chip delays impacting the timing of spend at that account, I didn't quite walk away from the last earnings call really understanding kind of what the clear picture is with that account and what the server delays have meant for pull-through for Arista. So maybe you can just for the purposes of this call, just kind of take us through where you see that right now, or at least through the second quarter?

Yes. I mean if you go back to kind of the Q4 of last year, what was happening in that time frame, I think one of the changes that we saw was that they were pushing some server refresh activity out into 2021, right? And that was linked to their decisions around what chips they were going to use and deploy and those that are refreshed. So if you look at kind of year-over-year, that's obviously taken our business with Facebook down significantly on a year-over-year basis, and we've kind of talked about that previously. We continue to execute to that plan. There hasn't been any updates, I think, at a -- and any change to that strategy from Facebook. Certainly, there's been high-level discussions that things continue to move in accordance to that plan. We're not yet in the window where we'd start to see some of those things unfold at a bottoms-up kind of detail level because we're still kind of outside of that window, but best as we know, kind of that we continue to operate to that plan.

That's helpful. [Operator Instructions] Charles, I think Anshul had some interesting comments about -- on the call about how cloud titans buying on demand a bit. They don't always are waiting around for the latest speeds, right? They don't have that convenience, right? They have to -- they see the traffic, they're buying capacity, right? And I guess in the context of the broader discussion around 400 gig, what did you think of his comments? And how does that play into how investors should look at the opportunity around 400 gig?

Well, I think with 400-gig, because of the price, I mean it's not -- I think it's about twice the price of 100-gig, 4x the performance. So it really serves its best purpose in the data center interconnect, where the lines are the most expensive, and they can get the highest return on that investment. So that's really being held back by the ZR optic. But it -- the 400-gig is just getting pushed out probably mid-2021 because of that. And 100-gig is still a growing market. And within the data center, when the demand is there, I mean these guys aren't going to wait until the next technology generation. I mean they have a demand. They have to stay competitive. So they build mostly to their demand, and they're buying what's available today. And 100-gig is actually very cost-effective for -- inside the data center.

Is there a crossover point around 100-gig, whereas the technology becomes more ubiquitous, the pricing obviously gets better for the customer, where as we wait for this 400-gig cycle sometime in '21, that there could be continued strong demand in the next 6 months for 100-gig? Is there some kind of crossover point where the pricing just gets really attractive and maybe, maybe these cloud titans start filling demand a little bit earlier than expected?

You can go back and look at when 10-gig went to 40-gig and when 40 went to 100. And I mean for 100-gig, it came right on top of the 40-gig. So the transition when that pricing happened was pretty quick. I mean let's face it, I mean if you can buy a 100-gig port for the same price as a 40-gig port, it's a no-brainer at that point. So when it becomes a no-brainer, that's pretty easy. The actual ratios, I don't want to pick a particular number, but right now 2x is probably not the transition point where it's a no-brainer. And so the cost of 100-gig is still very attractive for deployment.

Okay.

And I think Alex, you have seen evolution at 100-gig that's helping, kind of technology shifts that is helping the cost per port at 100-gig to continue to improve over time, right? We've seen new platforms, denser platforms, which continue to drive that 100-gig per port cost. And therefore that allows these customers, and that work was done in anticipation of needing to continue to deploy 100-gig until you got to a more optimized 400-gig per port. So there's a fair amount of planning that's gone into kind of this transition and understanding that it would start in certain use cases and then expand beyond that. So I think that's been part of, these our customers' kind of road map and thinking for some time.

Okay. No, that's helpful. I'm just going to go to the next kind of set of questions around the longer-term, mid-term and longer-term growth rates, and then hit some financials and go to some Q&A here. And I think Jayshree has talked publicly about the challenges of growing, given the just enormous cloud titan contribution you've seen the last couple of years, right? You just have this big chunk of revenue that is just hard to grow. By the way, it's a great problem to have, right, to be so strategic to some of the largest cloud titans in the world. So I guess beneath that statement, what are the goals to further develop alternatives to the cloud titans, right, outside of Facebook and Microsoft, and kind of where that stands today? And then what are the long-term goals for enterprise?

Yes. I mean, I think look, I'd be remiss if I didn't start answering this question by saying, look, we're not guiding or looking to kind of put a stake around 2021 yet, right? It's just too early to do that. We're not inside our kind of visibility window on cloud, and we're not -- let's work through some -- a little bit more COVID, so we can get a solid feel for how that plays out. But all of that said, I think we had some spectacular growth with cloud. And for sure, the cloud growth momentum has slowed, not just Arista, but just generically, their spending on the space has slowed. They will continue to invest in this space. And we think over longer period of time, they'll continue to spend money and increment their investment in the data center. And we absolutely want to be part of that, and that's a meaningful part of the business, right? It's been a drag in the last 12 months or so, just because we've been adjusting from that high growth rate, but I think over time, it's still a very important part of the business. It is important for us to diversify the business more. And I think we're doing that, right? And we've seen -- the enterprise piece has grown very healthily over the last couple of years, right? If we go back 2, 3 years, we were -- we were questioning, could we be successful growing that business in the face of incumbency, et cetera, et cetera? And I think we've done a good job of doing that. I mean it's now 35% of the business. And it's continued to grow at a reasonable rate. So if we were in a non-COVID world. I think for sure, that would be an obvious kind of growth contributor. In a COVID world, I think so far, it's still actually filling that role, and we'll just, we'll see how that plays out over time. We have campus. I think it's a $100 million business in the first 4 quarters. That's meaningful. If we can double that in 5 to 6 quarters, that's a meaningful contribution as well. Service provider, we'll see what we can do there. We've certainly stayed engaged there. We continue to make investments around that space because it's intuitively something where the technology is a real fit, and there is a strong overlap if you can make it work. And that's important, right? So I think the various pieces of the business are there. We've come through a year where for various reasons, some of those pieces haven't been working well. But I think going forward, some combination of those can give you a reasonable outlook. It's a little too early to try to call that and put too much kind of definition around that until we get through kind of closer to the end of the year here. But there are important building blocks there, and it's not just cloud. There are other pieces to the business as well. But cloud is an important piece of that, for sure.

Right. And it's just that second tier, I think you call them cloud service provider, that's sort of second-tier contribution, right? They can be a big part of the revenue going forward, right?

Yes. And again, they had their own kind of challenges if you look at kind of the back end of last year, earlier this year. And again, that's stabilizing. And we saw some good performance in that part of the business as well in Q2, right? So I think we're starting to see the various pieces stabilize. And then we'll take a view on how kind of we think those can grow, but there's definitely a lot of work to do across those various pieces, including cloud, given that over time, they believe they'll continue to invest to grow their performance.

Okay. So let's start digging in some questions. I've got one for you, Charles. What are you seeing in terms of Microsoft SONIC adoption by third-party customers? How do you see that impacting demand for EOS-based products? Question is a good one, but it's a question that's been around for a couple of years now. Just -- what's the current state of SONIC, I guess, in the marketplace?

Yes. I mean Arista has been a supporter of open standards for many years now. We actually have a slide in our IR deck showing all the open standards that we've been supporting. SONIC is another one that's an important open standard. And so Arista offering its platform with a SAI interface that you can run SONIC over. It's just part of the evolution of us supporting open standards. That's one option if customers want to go that way. We offer lots of options. And it remains to be seen the actual adoption rates of that, but certainly something that we want to support and we'll continue to support.

But as far as impacting your demand, it's just not -- it hasn't materialized in a way that you guys are really focused on it?

Yes, and I think it [indiscernible].

I think it's [indiscernible], yes.

Yes, and it's just another lever in kind of one -- in these massive data center architectures, right? It's no different to some of the use cases where you have more competition, different solutions. It's just one more piece of the puzzle. But in terms of the overall footprint, I think there's more than enough for us to continue to do. And EOS is still important to a lot of the more complex use cases where we play today.

Just another question here, just about broader market share relative to Cisco and Juniper. It's a big marketplace, and there's still a lot of opportunity. Just maybe outside of the cloud titans because we talked about that competitive landscape a lot. But what is, what does the enterprise look like right now from a Cisco and Juniper's perspective as far as your ability to win, win rates and opportunities there?

I mean you've seen over the last 2 years or so that we've been growing that enterprise business in that part of the business at a pretty fast clip, right? And that's again, the strategy has been and remains kind of focusing on larger enterprises, enterprises where the data center architecture is critical to their business needs and where they have an investment in IT that supports their business. And that's where we'll prove them best, right? So we've seen continued, I think, market share gains, customer wins in that part of the business. And some of these, we've been engaged in for a long time, where it's taken multiple attempts, if you like, to win over that customer and to prove the value to that customer, because it's a big decision for a large enterprise to decide to start to diversify their networking infrastructure. But we are seeing an ability to -- with the right focus and approach, that you can knock down those larger enterprise accounts. And it's relatively early kind of in that journey in terms of finding those accounts and addressing those accounts. So I think that is a piece of the business that we're continuing to invest in, both go to market solutions, features-wise. And I think certainly, in a non-COVID world, you would say that could continue to grow healthily. And hopefully even with COVID, we're seeing that piece of enterprise kind of remain reasonably healthy.

And we're also offering a truly differentiated service. I mean a single operating system end to end with a single management tool is something that our competitors just don't offer. And I don't know if you saw the press release today, but we also announced CloudVision as a service. So you can now run that CloudVision in the cloud, which really gives you a lot of enhanced capability in terms of what it takes to manage that management tool. It's now just a turnkey Software as a Service managed in the cloud, which goes along with campus, data center, wireless and public cloud hosting. So it really is a differentiated service that our customers are finding very attractive.

So there's 2 sets of questions I want to finish up on. Ita, we'll get to you in a second around leverage and some of the questions I had for you on the earnings call then. But we're going to keep going on this question of software for a second with you, Charles. You guys have really -- have made an initiative the last couple of years to drive more of your software natively into the marketplace, right, whether it's the containerized product or products like what you announced today with CloudVision as a service. What's the longer-term opportunity for Arista to sell software directly to customers? I know at the outset of the company, there was an initiative to maybe just be a software-only company, and there were some changes there. But if we were at the Analyst Day and we're kind of talking about the long-term goal for software contribution, what would you guys outline for us?

I mean I think that the software business has the opportunity to grow and become a larger percentage of the revenue. It's probably more so in the enterprise because the enterprise is looking for more of those software services. CloudVision is a perfect example of an additional capability that you can sell to an enterprise. But I still think that integrating the software with the hardware is the bulk of the business and probably will be for the foreseeable future, but there's certainly a lot of opportunity for growing the software as well.

Ita, do you want to add on to what the software contribution could look like over time, or what it could mean for margins over time? Or is it still a little bit early for that?

I mean, I think it's a little bit early to start to put hard metrics around that. I think that the key is, we're not going to try to carve out the operating system itself and try to have that be a separate software sale, because it really is integrated and customers view it that way and buy it that way. So really, when we talk about software, it's going to be more where are we selling kind of add-on features that make -- the customers are willing and want to pay for, because it's adding real value to them. And that's CloudVision, that's some of the management tools. It's like the routing feature set that we did before. So there is absolutely a focus, I think, internally as you think about more enterprise accounts, more solutions for that enterprise market. And that's where you can actually start to grow that business. So there's definitely a focus on figuring out exactly what those next feature sets should be, and then how do we monetize that. Big Switch was part of that. I mean they had already a software monetization system that had a lot of the tools around how you do that. So as we integrate that with Arista products, that will help kind of drive that initiative as well. So I think it's definitely something that we're focused on, and certainly something on Jayshree's kind of top 5 list in terms of how we drive the business going forward. And we're picking up pieces of the puzzle to kind of make that more meaningful as we go forward. Maybe it's a little early to try to put discrete numbers on it.

Okay. Just to finish up with you, leverage into fiscal '21 and just sort of OpEx investments and how we should think about -- you guys always still over-deliver on the leverage. We've gotten to know this about the company at this point. So looking at it as far as where you need to make OpEx investments next year and kind of how the Street should be thinking about leverage, without really understanding what the top line should look like, how should we be framing that in our models?

Which is always a little hard to do, right, to do one in isolation of the other. I think our long-term model still sits at the 35% operating margin, plus or minus, right? That gives us lots of scope to do anything that you can contemplate from an investment perspective in the near term that -- and still achieve that number. So that's a good, solid kind of baseline number. I think we guided 37% operating margin for Q3. Thinking about it in time of Q4 and beyond, I think that's a good way to think about it. We will start to add back. We have talked about being able to run the business at 2018 OpEx levels. I think as we see the top line solidify a little bit better than that, we will start to add back in some investments. We've already kind of turned back on hiring for R&D. We've turned back on hiring for targeted sales headcount adds, et cetera. And so we'll kind of manage to that 37% plus or minus kind of in that near-term view. And I think that's the way to think about it as we head into next year too is, we'll look to make incremental investments as, and if the top line is there. And sometimes that will get ahead of us, et cetera, which is kind of what has driven some of the outperformance in the past. But I think that's a good kind of target margin to think about in that time frame.

Okay. So I just wanted to finish up, and we've been asking this question to other panelists in fireside chats here over the last few hours. What surprised you about operating the company during COVID? Whether it's collaboration tools, travel and marketing expenses or just efficiencies, good or bad, what's been kind of the internal takeaway so far, we're only a couple of quarters into this?

Yes. I mean I think the sales activity and the level of sales activity that the team has been able to maintain has been a pleasant surprise, right? I mean that was probably the area that we were most concerned about. I mean you can think about the software engineering team, and we can pretty much work from anywhere. And yes, we do something on collaboration, but they're used to working in these tools. They're used to that environment. I think the one that we were concerned about more was the sales and marketing efforts. And the team has done a really good job, Chris and his guys, of driving as much interaction with customers as possible. Maintaining those engagements, maintaining those activities. And while stuff like campus may get impacted just because of -- it's not top of mind necessarily in the same way, given where we are with COVID. But the rest of the business has continued to execute really well. And that's not to say that these guys wouldn't love to just go back to seeing their customers face to face. They would, but they've done a really good job of making the best of the tools that we have, delivering content, driving those engagements. So I think that's been very positive. And then I think overall, everybody understands this is -- we have to make the best of this and be as productive as we can be in this environment. I think there's definitely parts of the business that would like to go back to something in person, especially around where relationships are important, et cetera. But we'll have to do that very carefully and safely, as we navigate through this, right?

Okay. That's great. Well, we're up against it. I appreciate your time. Thanks both for joining us today, and have a good rest of your day with us. I really appreciate taking the time here.

Thanks, Alex.

Yes, no, thank you. Thanks very much for the time.

All right. Have a good rest of your day. Thank you.