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

Thank you for joining. It is great to see an in-person crowd here, so welcome and thank you for making the trip.

Yes, it's our first one, exactly 2 years.

Yes, exactly. We might have been the last one, too.

That's right. That's right. That's exactly right.

Good choice. All right. All right. Well, I'm Erik Suppiger with JMP, the infrastructure analyst. And for this session, we have Ita Brennan, CFO of Arista. And so I will kick it off, but please do feel comfortable asking questions. The goal here is to make sure that you can take care of whatever questions or concerns that you have. So just raise your hand, and I'll be glad to welcome any questions from the audience.

So Ita, on the last earnings call, you had set expectations for 30% growth in fiscal '22. That would be the fastest growth you've had since 2018, I think. So comment a little bit in terms of what gives you the confidence to set that kind of expectation, particularly in light of an uncertain supply chain?

Yes. I think it all goes back to the building blocks of the business and looking at the building blocks kind of piece by piece. If you look at enterprise and financials combined, I mean, that piece of the business has been growing very nicely for us and very consistently. We shared that at the Analyst Day, but you could see that piece of the business has been growing at a 30% CAGR over the last 5 years, and it was 40% of the business on a revenue basis in 2021. So that's largely a case of share gains, both in the data center piece of the business and in the campus piece of the business, winning new logos, adding new logos. And there's just been very consistent execution there. It includes the campus targets that we set, which is taking kind of our $200 million of campus revenue in 2021 and growing that to $400 million in 2022. So I think that's one piece that has been executing really nicely. We continue to add sales, resources to support that, and we continue to see kind of payback on those investments and those additions as we go. If you think about cloud, which has been -- is obviously a very important part of our business. We're very closely aligned with those customers, but their spending has been somewhat muted historically, but they're also starting to return. And we're seeing some good strength in demand from that piece of the business with the new products. We're at the beginnings of some new product cycles there. So we also expect that to -- that piece of the business to perform well heading into 2022. And then you have our providers business, which is a combination of specialty cloud and service provider. And again, there, specialty cloud will tend to invest into new product cycles, et cetera. So we're seeing some strength there. And then the service provider piece is also growing, right? So there's good, strong demand across the business. And really, the challenge is, to your point, managing supply and supporting the supply to support that business. So that is a constrained view of the world based on supply. We spend a lot of time trying to rightsize that to the point where we feel good about kind of -- that we could achieve that, knowing that the demand is lining up.

Okay. So I want to go into a few of those. The campus is projected to go from $200 million to $400 million.

That's right.

Talk a little bit about what that is going to take from Arista in terms of sales force expansion, in terms of kind of new logo versus existing accounts? And what kind of visibility you have into that kind of demand?

Yes. I think the nice thing about the campus business to date is that we now can go to an enterprise customer with an end-to-end solution, right? Before we were going, and we were trying to identify enterprise customers that had a data center opportunity, and that was kind of where we could fit, and we weren't necessarily able to sell into routing use cases or into the campus. Now you can go to an enterprise customer and sell the entire portfolio, right? And your ability to engage in opportunities is greater. You're able to -- sometimes a customer will have a campus opportunity that they're trying a solution for and then you can cross-sell back into the rest of their IT spend. Even if it's a campus-only opportunity, these large enterprises can have a substantial campus business or will start in the data center. So Jayshree had said on the last call that 50% of the new logos in campus in the quarter were campus-led, where we were actually going into a new account and starting with the campus, right? So I think that's a very positive movement. It's the same sales effort across all of those products. And so that gives you some good leverage as well from the sales side. So we'll continue to add sales head count to address that enterprise market, and they will sell both switching, campus and routing use cases into those customers as well, right?

How much of that is new account business, if you're going to double? Put some context in terms of how much of that is new?

Yes. I mean it's hard to do that analysis. But obviously, we have a pipeline, and we're -- the most adjacent and obvious are existing customers as and when they have campus spend. They know Arista, they know the experience. They -- that's an ideal cross-sell, right? But we are seeing customers coming and starting with the campus opportunity as well, right? And we did see some good success with that in Q4.

Okay. And then your 2 largest customers, Microsoft and Facebook or Meta. First off, Meta had talked about some slowing in -- their guidance came down last quarter, and I think they talked about advertising was slower for them. There were some privacy concerns about new policies put in by Apple and the like. Is that something that investors should be concerned, could come back to Arista and the build -- their build-out plans?

Yes. I mean, it's always hard to talk about individual customers. I think if we look at that cloud vertical as a whole, you're seeing good demand indicators for them. We're seeing longer-term orders and forecast because you have to respond to the extended lead times on the supply side. So we're getting some more visibility. We're working very closely with them around deployments and when they plan to deploy the equipment that we're proposing to ship. We're working with them on priorities and what needs to happen first because it is a constrained supply environment. So I think we see good healthy demand in that part of the business, and it's kind of urgent to their day-to-day business, right? So that gives you some comfort as you look out through that window that this is real needs and real requirements that they almost have to do to support the business, right?

And is that across a number of the titans? I think there's 7 titans and -- it's -- the other players are also coming up in terms of their adoption here?

Yes. I think across kind of that cloud titan/specialty cloud, there's just a strong kind of investment cycle. Again, with these -- with some of the newer products kind of leading that, that we're seeing there. And again, it's more constrained by supply than anything else at this point.

Okay. In light of the conflict going on in Ukraine, I think you do about 25% -- last year, you did about 25% of your business in Europe. Any perspective in terms of, one, I don't think you have particular exposure or R&D or anything in Ukraine, correct me if I'm wrong?

That's right.

And any thoughts in terms of whether that poses risk to economic demand in parts of Europe that you serve? Or any thoughts or response to that?

Yes. I mean our direct exposure is really immaterial to that part of the world. I think we'll see if there's any kind of second order, third order effects over time, honestly. Even on the supply side, perhaps we'll have to wait and see. But I think for now, when we look at it, that kind of our direct impact is immaterial.

Okay. Let me check on any questions. Okay. Let's talk a little bit about the upgrade for -- upgrade cycle to 400 gig. It's been -- 400 gig has been out there for a while. It became much more economically attractive in 2021. Where do you think we are in terms of that upgrade cycle? And I noticed in one of your slides, it starts to show the next upgrade cycle, 800-gig starting to grow around 2023 or so. So where are we in terms of the 2 cycles there?

Yes. I think that chart that's in the investor deck is obviously industry analyst data. It's always maybe a little bit aggressive in terms of how these cycles ultimately roll out. We've definitely started to see -- what's been interesting about the 400-gig cycle is, obviously, the underlying silicon has been used to build 100-gig, 200-gig and 400-gig configuration products, right? So there's a slew of products that are being deployed kind of leveraging that -- the cost structure and the power capabilities, et cetera, that are being delivered with that silicon. 400-gig port configurations have tended to be more data center interconnect and some specific use cases. 200 and 100 will be more broadly deployed inside the data center still, right? But we're definitely seeing that cycle drives some increased investment. I think we're at the beginnings of that cycle. Supply, again, is playing a role in shaping kind of how long this is -- how long this cycle will be and how long it takes to kind of respond to these cases that are being -- where 400 gig is being deployed today, right? So we'll see 800. 800 will bring probably better economics again on a per port cost. So at some point, it will make sense to kind of deploy that inside the data center, et cetera, as well. But again, the capabilities will have to be there, and I think 400 gig probably runs for a while before we get to 800 gig.

Is the 400 gig still very dominated by the titans and the very highest-end accounts? Or is it becoming more mainstream?

I think the -- again, the 400-gig configuration products will tend to be high capacity where you need high capacity, where you're doing interconnect-type projects, right? Using that silicon in a 100-gig box, that's been broadly deployed today.

It is. Okay. All right. What about your software sales, like CloudVision and I think you've got several software products that layer on top of the switching? How would you assess your execution in terms of selling software? And what contribution can that become?

Yes. I mean, our approach has been very much -- there's a natural sales motion for customers and fulfillment model for customers around product and the operating system itself, right? And those go hand in hand. So we're not trying to disassociate that operating system from the hardware sale in any way because customers -- it's not something the customers -- that resonates with customers. It's not something that they are pulling.

Just to be clear, the operating system is clearly part of the switch when you buy, right?

It's part of the switch when you buy it. There's licenses and everything else that go with it, but we're selling the 2 together and people are requiring that when they make the purchase. So really, our soft business is kind of the value-add, add-on capabilities that we deliver. CloudVision, it's monitoring and visibility tools. It's the -- some of the new security-type capabilities and then just the suite of automation and data-driven kind of visibility that we deliver with CloudVision. And that's been growing roughly with the business. When you look at the metrics, kind of grew with the business in 2021. And again, it's targeting an enterprise. It's obviously -- more targeted towards the enterprise piece of the business. So we will continue to add, I would say, features and capabilities there. We've talked about leveraging EOS of the state-based data that we collect from EOS, combining it with some of the data that we get from the DMF visibility tool plus Awake and the sensors -- the data that Awake is collecting. And then ultimately, with some partner data as well to create kind of that net DL database of information that can then be used to provide more actionable outcomes for the customers. So I think that's the direction that, that is going. And again, it will be incremental. We'll take steps as we go here, but it's definitely valuable to customers. It's part of the enterprise traction that those capabilities are there and CloudVision resonates really well with the enterprise customers that we're winning.

Okay. Again, if there's any questions, feel free, feel comfortable raising your hand. Question here.

How much of a drag on gross margin supply constraint has been, and to that effect, how much of a benefit are you seeing from the price increase that you rolled out last year, and mention when do you expect this to grow?

Yes. We -- probably the most significant price increase that we did, we did in November of last year, but we didn't kind of try to reprice prior orders. So we won't see that really start to kick in until later Q3, Q4 of this year, right? So there's not that much of a tailwind, if you like, in the numbers from the price increases. On the cost side, I don't know that we've quantified it precisely, but if you look at the -- what's been happening to the gross margin, we've been operating in an environment where enterprise was 40% of the business, which is a heavy enterprise mix. And we've still kind of had gross margins in that 64% range. You'd expect that to have been higher than that -- significantly higher than that with enterprise being that larger portion of the business. So we're definitely seeing some significant impact there. And again, folks have been asking when does that come off? I wouldn't be in too much of a hurry to take that off because I just think it's going to go with -- we'll have to get through the supply constraints and then have some time for that to kind of readjust. So I wouldn't build kind of too much of a recovery on the cost reductions into the model in the near term.

Okay. You're -- in terms of the supply constraints, you had talked about inventories and purchase commitments in excess of $3 billion. That is proportionally quite a bit more than most any vendor I'm familiar with. I'm curious, do you look at that as a strategic advantage? Or how do you think of your inventory management or inventory sourcing relative to the broader market?

Yes. I mean it's hard to compare to what competitors are doing. I mean, obviously, we're growing a good bit faster. Our approach and thinking is because a lot of these products are at the early phase of their life cycle, that it makes sense to be a little bit more aggressive in terms of taking positions on purchase commitments because you've got a long life of the product ahead of you to consume that. And really, the cost of doing that is the time -- is the cash, right, on the time value of cash. So again, we're -- it's components, right? So it's not like we're planning to ship everything that we get. In 2021, we've been asked that question. These are really components that will obviously cover multiple years and will take time to be incorporated into products and sold. But the trade-offs are, you tie up the cash versus you've got more risk. You've got less ability to respond to upside should it come along and you're more exposed to expedite fees and other things if you don't make those commitments. It's still not a guarantee, right? Because we have seen decommits against some commitments, but at least if you're in line on a consistent basis, you should have some recovery. You'll lose the capacity -- manufacturing capacity for that window, but at least you have your next slot and your slot after that. So I think we're trying to be balanced. Again, it would be harder if these weren't newer products. But we think that's a good trade-off to have kind of -- to have that resiliency because it's not clear in our minds exactly when things start to return to normal.

Talk a little bit about how decommitments work? That's something that, I think, we're kind of learning about. Obviously, you make these purchase commitments, you make them long in advance. Then when do they give you a heads up they're not going to deliver? And how responsive can you be if they back off?

I mean it's usually going to be a surprise, something happens in -- probably in their supply chain, too, right? I mean everybody has a supply chain, right? So it's usually -- there's been some disruption somewhere in that chain, and they're not able to produce as much as they anticipated that they would produce in that time frame, and then they make a decommit to their customer, right?

How often or how problematic has that been?

I mean -- again, we talked about there being some of that happening in January on the call. So it's not -- I mean, again, the idea of having the purchase commitments is to help protect you from at least you get in line, but there are things that are happening on a regular basis kind of across this. But some you can recover from, some you can resolve and some just have an impact and you lose some capacity while you resolve them.

Okay. Coming down to last few minutes. So if there are more questions, again, feel free to raise your hand. Come back to the competitive dynamics around 400-gig upgrade cycle. I think you had very strong execution around the 100-gig upgrade cycle, and you've been able to maintain a leadership position. And you've expressed pretty good confidence you'll be able to do the same here. Are there different dynamics? Or is there anything -- kind of give us some context around your confidence in terms of leading in the 400 gig?

Yes. I think we were -- I think, measured as we went through the cycle as we normally are. There was lots of discussion around kind of why this one would be different. But I think when you look back now, I mean, it's been a very similar cycle of execution with customers, delivering good systems, products and making sure that they perform well and then earning your piece of that business, right? I don't know that there's been anything significantly different in our -- from our perspective in terms of how the cycle has gone, right? I mean these are technology buyers who test and validate products. And if the products perform well, and as expected, then you are in your piece of the business. So I think it's been a pretty consistent kind of next cycle for us in terms of how we've been interacting with these customers.

Okay. And on the campus side, obviously, you went from -- you reached $200 million in 2021, and you're projecting $400 million. So you're reaching some scale here that's got to be giving Cisco some concern, I would assume. What have you seen from Cisco in terms of changes in behavior or any response from a competitive perspective?

I don't know that we've seen anything different from them. I mean, obviously, they are refreshing that campus footprint. They have a vast presence in that market. I think we are probably the first really high-quality competitor in that large enterprise business, not dissimilar to how it was with data center at the time and how that data center business is evolving over time, right? Again, for us, we're selling an end-to-end solution. We're selling an experience for customers that seems to be resonating. I think the Cisco's approach has been different in that they have their subscription model, they're driving that refresh, et cetera. I don't know that we've seen anything different from them other than what they've been doing to solve for their own business objectives.

Would you -- does that subscription model that they're really pushing? Does that make sense for Arista?

Again, it comes back to -- we'll always be somewhat customer-driven, right? And we're not, at least for the large enterprise customers, the enterprise customers that we're selling, we're not seeing customers want to engage in that model, right? And if anything, they would prefer to be able to acquire kind of the software that logically runs with the switch and the switch doesn't run without the software. It's that -- it's less intuitive for them that, that would be a subscription model. They're happy to engage on a subscription basis on the value-add kind of CloudVision and some of those other products.

Okay. And then Broadcom has been your go-to for your network processors for some time. Given Broadcom has been -- is part of a much bigger organization than it used to be, it's still Broadcom, but it's gone through a fair amount of transition. Are you concerned about their level of investment in their R&D and their development? Are you concerned that they may lose their competitive position over time?

I mean I think we've -- this question was probably even more top of mind after kind of the Broadcom was acquired. But what we've seen is they've executed really, really well on the technology side of the space. And they have multiple product lines, multiple investments running at the same time. And the reason they have the position they have in the marketplace is that they're bringing products to market in a very timely manner. The combination of that plus EOS and the Arista software stack is -- makes for a very reliable overall system solution, right? So I think we're very happy with kind of how that partnership has evolved.

Okay. Well, I think that's it. So again, thank you all for joining. And Ita, thank you very much for the time.

Thank you very much. Thanks for having us.