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

I think for -- and I would broaden that to just semiconductor because I don't think the challenges are unique to Broadcom. It's getting wafers, getting assembly, everything is constrained. The downstream from all these suppliers. So it's -- it continues to be a challenge. Obviously, we have good visibility there, but we also think a lot about downstream and smaller components that we're buying today and making sure that we have them available as well.

All right. Just to wrap up on cloud before we go over to enterprise. Your commentary recently the last couple of quarters has been far more constructive, far more positive regarding what you call your specialty cloud or tier 2 cloud. Juniper calls it cloud majors. But this is a group that the investment community doesn't focus on nearly as much understandably as web scale. But if I look at the Dell'Oro numbers, there is big -- collectively, it's the big 4, 5 web scalers. And so collectively, they could have quite an impact. And what visibility do you have as to the improvement you've seen from them being sustainable and being meaningful going forward? What are you seeing from that group?

It's hard to pull them together as a group because they all have kind of individual strategies. There are some that we have quite a bit of visibility into and probably others less so. I do think that the same effect that we're seeing with cloud titans just in terms of the demand on their networks and more users, et cetera, and concern about continuity exists in this other segment as well, for sure.

Right. John, on the competitive front with that group, Juniper cited 100, 400 gig wins, all with those cloud that -- your specialty cloud, cloud majors. Any sign, any concern that they're having success against you and cutting into your opportunity?

Not really.

I know it's a big customer audience. I'm not sure you realize how big but...

Right. I think competitively, we see Cisco more in opportunities we're engaged with in multiple segments than Juniper. I can't explain to you why that is. It might just became -- we came from different areas and came to routing in a different way. I think in some routing use cases, we certainly see Juniper, but maybe less holistically in some of our data center opportunities.

All right. Let's go over to enterprise and start with campus, which is, I think, your biggest opportunity in terms of growth. You only entered not much more than a year ago. You drove $100 million as you said you would in your first year, which for anybody who knows, the life cycle networking that's quite an accomplishment. You put out there this $200 million number, 100% growth for this year. My short question to you is, what are the opportunities for upside? What are the risks of not getting to that $200 million?

Sure. I think it's important for investors to realize to how we're going after it because I think we're doing things a little bit differently. We're not viewing campus as some separate market segment. It's a use case within the enterprise go-to-market. We get asked a lot about go-to-market and coverage, et cetera. So our teams are very focused in the enterprise, I would say, in Fortune 2000 or even smaller customers that use their network strategically to drive revenue, financial services, media, entertainment. So that's been very consistent. What's happened is the portfolio has broadened, right? We went into routing so we could address interconnect use cases that with Big Switch and Awake network visibility and security detection and campus. I think since we've come out, the early days is -- can Cisco -- can you do use cases similar to Cisco, with data, voice, video? And I think we've demonstrated that. I think what we've done in the last year is really drive some differentiation in the offering. So in terms of risk and opportunity, I think continuity with our enterprise go-to-market is the opportunity. We have more insertion points with the enterprise now with a broader portfolio. Obviously, what people do in campus and whether they're going to invest continues to be a lingering question. But we do see people starting to come back to offices and kind of picking up where they left off on moving to IoT and need higher speeds. I can't trust the devices. I have to make some decisions.

All right. If I recall, you all have made clear that the pandemic, which was a concern for you roughly a year ago in terms of impacting your ability to grow, it sounds like you're beyond that, that it's not -- other than the supply chain constraints, perhaps, than it's not that pregnant if at all an issue at this point.

Yes. I think there's still some questions about what happens in the campus, and customers are all over the map. Some took advantage of the downtime to upgrade their networks because nobody was in the office. Others are still wondering whether they're going to have office space. So it's a pretty crazy collective. But I think given where we are, we feel really good about the opportunity to hit $200 million. We came out this year with a new modular switching product. I think it's a segment that's been largely ignored. And there are still customers that are running very old modular kits that have large levels of interconnect and high-rise buildings that are still interested in having connectivity for people coming to work.

John, this is an important point you've raised. Feel free to disabuse me the notion, but understandably, you didn't enter the market all that long ago. You got to walk before you run. And that means, right now, you've got a handful or perhaps 2 handfuls of platforms that address in your binocular of the use case or the market opportunity, everyone to think about it. Cisco has got the equivalent and product families, as you know. So they cover the waterfront in terms of 92, 93, 94, 95, 96, et cetera, multiple iterations. And they've got 30-plus years of code in terms of the breadth and depth of that code. And that code takes time, just like the insurance provider business. You've, over the last year or 2, you've meaningfully enhanced the breadth and depth of your code, which enables you in turn to address more used cases and to be better positioned in the use case that you're already addressing. Are there like -- you just entered the market, so I trust it's not a meaningful limitation, but are there limitations in terms of the breadth and depth of your solutions today? What do you need to do? What are the most urgent things you need to do to expand that opportunity and to be more effective in competing against the opportunity you're currently addressing?

Yes. I think, first of all, just going back to your service provider analogy, the depth isn't as substantial as the various cases you have. And also when you get into the service provider, it's not only, do I have MPLS XYZ feature, how do I integrate into my OSS and my conformance to how I integrate everything else so there's some specialty functionality for different providers. In the enterprise -- in the campus, can I handle data? Can I handle voice? Can I handle video? That was step one for us. And I think, Paul, you and others had a lot of questions about whether we could do that successfully given the loads of features. And we've done that differently and in some ways, much more simply than the legacy. So I think we feel really confident in kind of that base use case. What we've announced in the last year with zero trust security is actually, I think, highly differentiated it from the competitive offerings. And being able to offer visibility into the campus, being able to see all the devices and what they're doing and being able to do some interesting things with group segmentation in a nonproprietary way that allows you to put devices and groups and set policy based on whether they can or can't talk to each other and do that within either our CloudVision or through APIs with third-party integration, like we've done with Forescout. So I feel really -- pretty good. There's obviously more products that can be built and more depth of the portfolio. But really with the anchors of the modular switch and the fixed switching, we feel like we've kind of set the tone.

John, your response raises multiple questions. And so I'm going to apologize because there's probably going to be several questions and one that I'm going to ask you here.

No problem.

Is there an opportunity at present primarily situations where customers want best-of-breed in winner diversity and are willing to incur the cost of bringing in or switching to new supplier? My last shot before Cowen, our CTO, used to love to say, I don't get up at a bed in the morning. So switch vendors for less than 25% price cost improvement, given the cost of retraining, given the risk involved and going to what have been a tried and few solution to the unknown. As great as you are, maybe it works, maybe doesn't. There's always issues, some boxes get deployed no matter how big the supplier, et cetera. Why don't I let you respond to that before I ask the follow ups.

Yes. I think that there are definitely folks that going to sweat the assets and stick with what they know. And the incumbent has huge share in the enterprise. There's no doubt about that. We've gone up against that before. But I do think with COVID, there was a disruption in the status quo. Disruption is good for a disruptor. I think we've seen a lot of people look at what we're doing, not just in the campus. I think it's important that they look at the whole solution and what we can bring to bear. And it's an unusual market. If you look at servers and storage, there's much more diversity of the vendors that people deploy in their IT environment. Somehow with networking, there's been this, I think, myth, that it has to be all the same thing to be consistent. And we're challenging that. Even things we've done with group segmentation, you can intermix Arista switches with an existing brownfield opportunity and introduce policy-based segmentation without going end-to-end. So we look for those opportunities to integrate but in a nonproprietary way with open APIs and work with third parties to put a best-of-breed network together as opposed to best-of-breed product.

Got it. I know you touched on this earlier, but from a go-to-market perspective, you're focused on larger organizations, where you probably already have presence on the data sort of switching side and where you can address those directly. So I trust the channel presence, it's not that you don't have a channel presence, but I trust it could also be far larger and higher quality, and it will be over time. How big an issue is that channel today? How much do you need to do?

I think the team has done a good job. We have to lead with the Arista differentiated message from our account teams. As you pointed out, these are customers that have long used the incumbent. So we really need to set the tone with the value proposition and everything we can bring to bear. In specific engagements, then a partner comes into play. Customers have partners they've dealt with for a long time that they want to use in the deployment, in the architecture. And as we've gotten bigger footprint, more of them are willing to invest with us to learn about Arista and come along for the journey.

I want to return to your comments about Awake and zero trust security in -- Arista often, over the years, has spoken about, you truly are differentiated from Cisco. And I trust you are differentiated. But from a different vantage point, I've long thought you, in large measure, replicating Cisco, just a more focused, younger version. And so maybe you won't end up doing everything Cisco is doing, but with the passage of time, and Awake is a good example of this. Previously, your mantra in security was we partner, we partner, we partner, best of breed, Palo Alto, whoever. And lo and behold, not surprising, I don't think it should be surprising to anybody, no shame, you buy Awake. And I trust you're going to make other security acquisitions in addition. And I trust you will make acquisitions to the extent you already -- you don't have it or don't have it in a full measure and net visibility, app visibility, assurance, automation, orchestration. I recognize, we have a lot of the stuff already. But my question to you is, aren't you going to continue? Won't you continue to strategically and organically to round out the capabilities and take this more of a full solutions architectural approach as opposed to just a switch, the traditional box approach?

I think we're down that journey, right? Even with -- before any acquisitions came into play, things we did around CloudVision and the management stack and the orchestration, but it all is aligned to making the network better. So there is some fundamental grounding in these pieces that we're adding. It's about running and operating the network. And also, despite Awake continuing to partner an integration with best-of-breed security providers, we don't see that as mutually exclusive. This is around network detection of anomalies. And we do have some capability that we can actually make that work better because we can stream from EOS more information given the sort of SysDB architecture we have and then just make it easier for customers to consume and deploy that. But again, we're very cognizant of the need to interoperate just given where we are from a share perspective, needs to work together with the incumbent and other suppliers and other third-party solutions.

All right. We've got 2 minutes. So we're really going to go into speed dating here, John. Routing, you've expanded that solution set. You're now addressing more opportunities, more use cases. Do you still have a lot more you need to do? If so, what's the time frame? Or do you think you now have plenty of opportunity in front of you? And what's new and different? And what can you share with us with respect to your success, whatever metrics [ for the year ]?

I think with routing, if you look at the different segments, cloud, we've done extremely well with the routing solution and what we had for a while. Enterprise, they just don't buy as many routers, but it's been very meaningful as a solution to be able to offer that as part of the data center because you have to route data center to data center or data center to hybrid cloud. Service provider has been the more difficult piece for us in routing. I think we did really well out of the chute with some opportunities that were more greenfield and less replicable. Ken and his team has added considerable functionality over the last 2 years, which I think is positioning us better. And also, as telcos kind of move from kind of traditional routing to more of a telco cloud approach that plays to our strength. So we would be excited if that timetable moved a little bit faster, and people are still predicting that's going to start to shift, but we'll see.

400ZR is your plus, it's on the verge. It's been on a verge for quite a while, but it finally seems or, hopefully, it will be real in terms of commercially launching, any thoughts you can share for your expectations? DCI, you've addressed -- you said a lot in the past in terms of being a big market opportunities, but what can you share with us? How far, how fast in terms of impacting the model?

I think it's kind of our thinking is baked into the model here, right? It starts with ZR. We see 400 gig coming in at the top of these networks and interconnect. We're pretty happy with the ecosystem as it's developing here. And we'll see how the supply ramps, I think, from early trials and qualification, things are pretty well behind us.

All right. Finally, strategically, where does Arista need or want to go in the future?

I think, continue this foundation of making the network easier to run, easier to operate and I think the announcements we made with zero trust security inherently more secure and making it more visible, so people can feel comfortable that any threats that come inside around the special -- people bringing devices and using at home into the office can feel secure about their networks and their IT environment.

Great. John, I want to thank you as always. Always great to have you with us. Great to see you. Stay healthy. I want to thank everybody out there again. I hope it's been a great conference, but if I could be any help to anyone on Arista or any other name, it would be my pleasure. Take care, everyone. Thank you.

Thank you, Paul.