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

Okay. Great. We now have the green light from the operator. Thank you, everyone, for joining us today. I'm Sami Badri from Credit Suisse Equity Research. Today, we have the team from Arista Networks, including Ita Brennan, the CFO; and Curtis McKee, which is the Director of Investor Relations. Thank you both for joining us today.

Thanks for having us.

Great having us. Thanks, Sami.

Yes, absolutely. So really to kick things off. Probably first question is for Ita. And last quarter, you registered a record number of new customer logos, and you expect to double your campus portfolio by the end of 2021. So the go-to-market strategy has clearly started to pay off. Can you just walk us through your strategy for the enterprise and campus opportunity in a little bit more detail? And if you want to factor in the cognitive campus narrative as well that would also be very helpful for a lot of viewers.

Okay. Yes. Yes. I mean, I think the enterprise/campus. I mean, it's really, from a go-to-market perspective for us, a single effort, right? We've entered the enterprise space, I think, in a more serious way with kind of the 100-gig data center product introductions. That was when we saw some real momentum in the enterprise. And I think it's a combination of bringing the switch router capabilities to the product set in that time frame. And you now had a very robust switching and routing solution at kind of 100-gig introduction point, which I think was a big decision point for a lot of the enterprise customers. So we started to see some broader acceptance of the products in the enterprise, and that kind of drove us to make further investments there. And the approach has been very consistent, focus on larger enterprises initially. Enterprises where their business model is really dependent on their IT infrastructure. And that means that it has a higher priority that tends to be more investment -- more focus on kind of the robustness and the capabilities of that IT infrastructure. So if you look at kind of our approach, it's been very much look for larger enterprises with those characteristics. It was initially a direct sale, right? Where we were targeting fairly technical sales resources to those accounts. And as we started to get traction and started to see success around that approach, we continue to add sales resources to that. And if you look through 2018, 2019, we were adding probably 30% plus to our direct sales resources targeted at some of those larger accounts and then also expanding our coverage model geographically and regionally to make sure that we have coverage across the enterprise. And I think we've seen the rewards of that as we've gone through the last couple of years, and you've seen kind of that enterprise and financial piece of the business has become more important. That's kind of what gives us the right, honestly, to start to have a discussion around campus. And those customers who are now using Arista for switching and routing were interested in having an Arista solution for the campus as well. And again, that's almost like a land and expand continuation sale to that original direct sales model. I think what we're engaged in now is looking more at the channel and how do we move what has been probably -- maybe 12 months ago has been really a fulfillment model with the channel where we fulfilled quite a large amount of our business through the channel to more of a sales and a sales driven channel model. So we've kicked off some key partnerships, some channel programs that are targeted, mainly with the campus focus, but it doesn't mean that those partners won't also introduce opportunities for the data center. But really, with the campus focus, driving more channel programs, more training, more education, and leading those partners to build a business model that's built around Arista. And we're seeing some early, I think, progress there, but there's still work to do there, and we'll continue to drive that. So I think it's a -- and we'll continue to expand on the direct sales model. And I think that's worked very well. That technical approach, win the account and expand, the use cases within that account on the 1 hand. And now that we have some presence and credibility, and we have a seat at the table, start to expand into the channel and form closer kind of sales-type relationship with the channel players.

Got it. Got it. And then maybe just a double click on specifically that dynamic. When we think about your cloud-tieing customers, how have they been as far as revenue contributors to your campus and WLAN products?

Yes. I mean, I'm not going to kind of get into the customer breakout of campus, right? I think to say a couple of things. I think, obviously, those are customers that appreciate Arista and have large kind of campus footprints. But on the other hand, I would say that within that first $100 million marching to our second -- our $200 million target, there's a lot of diversity in there. It's not swayed towards a particular customer because you'll always tend to have an initial almost proof of concepts. With the campus products, where the initial revenues will be smaller and then we'll grow from there. So I think that's a pretty diversified group in that $100 million to $200 million today.

Got it. Got it. One kind of just, I guess, taking a 10,000-foot view in terms of the dynamics that have actually taken place, at least for the last 9 months is how has COVID impacted either the enterprise or WLAN sales motion or sales dynamics? Are there any kind of big changes you've seen happen either from an enabling perspective to the business?

Yes. I mean I think, obviously, our approach had to change, right? And the sales team's approach needed to change. You could no longer have kind of the same in person meetings, the same -- we used to do a lot of proof-of-concept activities here at headquarters with customers. So very quickly, the sales team pivoted to moving a lot of that online. And now we do a lot of our proof-of-concept initiatives virtually because that's what's required. And that's been quite successful, actually. I think what we've seen is, you'll get more meetings, probably better attendance at some of those meetings because it's less of a drag in terms of travel, et cetera, for our customers. So I think that's been positive. We were quite concerned, I think, earlier in the year that you would see a slowing of demand on the enterprise/campus side. I think on the enterprise side, what we've seen is that we've got the customer set that we have are continuing to move ahead with their investments. And that, again, I think it comes back to that idea that if the IT is critical to the business, it's even more so now. So I think we've seen that part of the business actually do well in this time period. Campus, we had talked about being a little harder to go make new wins on the campus side. I think what that's netted out to is maybe it's going to take us an extra 1 to maybe 2 quarters to get to that doubling to the $200 million. But I think with everything that's going on, that's probably okay. I think that's better than what we might have expected if you'd asked us back in March or April when we were heading into this.

Got it. Got it. So thank you for answering that. And I want to shift gears a little bit and talk about Awake Security and maybe Curtis, you could jump in as well. Ita, by all means, chime in. But the question is, can you discuss how Awake Security fits into the Arista networks' broader visibility strategy? And then maybe just a timeline for Awake Security's full integration into the Arista product suite would be helpful?

You want to take that one, Curtis?

Sure, sure, Sami. Yes, so think about Awake, they bring to Arista a network detection and response security platform. That can be used both for enterprises in the campus as well as in their data centers for detecting malicious behavior behind the firewall. And so looking at areas that are not so obvious, not malware type of detection areas. And so think of like how much network-based threat continues to evolve, and so the legacy of on-premise firewall equipment needs to also evolve and to looking at non-malware based threat equipment. And so really Awake has an AI-based threat detection system that's great for autonomous threat hunting. And so you think about Awake sensors sitting next to Big Switch platform store that are connecting to the network packet broker, and how the NDR platforms will connect up through to talk to like Splunk SIM equipment. And so this is going to be an interesting feedback loop for really enforcement techniques that you could use CloudVision in the future. And so this is really an interesting play for Arista to move into more broad-based security from some of our near-term adjacency areas within networking, network packet brokering and our cloud vision telemetry and visibility system.

Got it.

I think in terms of integration, I think we're in our first quarter of kind of really absorbing business and putting the pieces together. I think it's very intriguing what you can do with the data that we collect and that we manage with CloudVision and with the Big Switch kind of monitoring capabilities. And then on the Awake side, they were very excited to see how much data you're able to collect and what they could do with that. But there's still some road map, obviously, integration and parts that need to get put together around that, and that's happening now as we speak. But I think it's a good synergistic relationship between 1 for the networking, 1 on security. And yes, we'll talk more as we go forward in terms of what the plans are there, but that's all the work that's undergoing right now.

Got it. Got it. When you think about the data center side of your customer base and then the campus side of the customer base, which have been most receptive to the Awake Security offerings that you have now kind of brought in house?

Yes. I mean, I think it's early at this point to really get into which parts of our business is going to play well with, right? I think there's definitely customer interest across the verticals and understanding how we're going to bring that product to market as part of Arista. But I think it's just a little early to start to really get into the details of that.

Got it. Got it.

Yes. It's really full packet capture too, Sami. So you're going to see opportunities that are derived -- it's -- the more you put a cloud sensor from Awake across your network, the more clear visibility you're going to have within the whole enterprise network from data center to campus. So if it starts inside the data center, where you're going to start seeing them renewals and new areas for covering more and more parts of the enterprise network within the corporate IT environment. So I think it's got good opportunity to expand from one area to the next.

Got it. Got it. So I want to shift gears a little bit to switching on 400-gig. And then I just wanted to kind of get the latest from you guys as far as how have the early 400-gig cycle conversations and win's been given kind of like the kickoff that really just took off very recently?

Yes. No, I think to reiterate some of Anshul's comments, I think we've now gotten some good feedback from many of the key customers. Products have been in their hands for some time, and I think we're pleased with the feedback that we're getting on those products. You always feel that you've delivered a good robust product, but it's good to get some customer input on that as well. And I think we're happy with where that is at this point.

Got it. Got it. I wanted to just kind of dig in a little bit deeper on 400-gig potential opportunities or at least the wins that are coming in from a use case perspective. Where are the 400-gig switches going in right now? If you were to kind of highlight some use cases where they're kind of being selected, which would be the top use cases that are being used?

Yes. I mean I think it's still relatively early in the 400-gig cycle in terms of volume, right? I mean the initial volume will come, large volume will come from the hyperscale customers. I mean there will be wins and some early deployments, maybe in some large enterprises as well. But it's long been our view that you're going to see these 400-gig products deployed initially in kind of the data center interconnect type use cases in some of those larger network footprints. And that's going to be gated more by when those customers are ready to start to deploy that at a larger scale, and that's dependent on their supply chain and other things as well, right? So I think it's an interesting cycle. We've been saying this for some time now, but you're going to -- it's going to be a gradual ramp to volume in our view. And again, the focus will be on some of those data center interconnect type use cases initially. And then over time, it'll expand more broadly than that. So I think what's been critical from a strategy perspective is to leverage that silicon to drive a more dense power-efficient 100-gig configuration of ports, maybe even some 200-gig configuration of products, and then finally, 400 gig. And once you've got that suite of products, and we do, we have a slough of products to go across those different speeds, which we believe are kind of leading-edge in each of those speeds. Then really, it's the customer's choice as to when exactly they want to start making those transitions based on their internal, either architectures or supply chain. And the job then is more just managing the supply chain and making sure that you understand the transitions that you're planning to that. The technology spectrum is there. And I think that has been the approach for 400-gig because we've been pretty consistent for a while now that we think it is a gradual cycle that will ramp over time and 100-gig will remain relevant for a longer time period as we make that transition.

Got it. Got it. One kind of question on competition and the competitive landscape is, how do you view the competitors kind of matching up or trying to introduce their products into the market and specifically to 200, 300, 400-gig insertion opportunities. Do you see the competition inserting into comparable opportunities as Arista, right? Or do you see them kind of approaching the market from a different perspective -- going for use cases that historically Arista may not necessarily be inserting into?

Yes. I mean, I think it's not that dissimilar to what we've seen in the past, right? I mean, clearly, this is a part of the market that everybody would like to serve and would like to have a piece of that business. And the competitors have had to start the relationships with these accounts in different parts and different areas of the network. And they'll continue to have different play and different aspects of these networks as well, right? But when it comes back to, I think, back to the switching part of these use cases, I think we haven't seen anything that's that different, right? Everybody has put products in front of these customers. I think we're happy with our products suite. We really like the robustness of the products that you can deliver the speed with which you can deliver them because it's just a consistent EOS evolution with a consistent evolution of Broadcom's silicon, right? And that gives you a very robust platform on which to kind of scale products, especially when there's going to be multiple product iterations over time. I mean, obviously, if you're a competitor coming from the outside, you're responding to that architecture, and that's driving change in your product portfolio and your approach. Again, we'll see where this all comes out. But I think we're happy with the feedback that we've gotten on our products from these customers at this point.

Got it. Got it. I wanted to kind of take more of a 10,000-foot view with the next question. And then just talk about going into 2021. How do you forecast or anticipate the web scale customers or cloud titan customers performing a little bit or at least, I guess, you'd say, ordering a little bit differently in New Year in 2021 versus 2020? And if you were to think about the 1 big pivotal change that's taking place, if we were to do unpack what happened in 2020. We had COVID coming in with uncertainty, then all of a sudden, like we obviously saw the direct indexation of a lot of these hard times, positive work from home move. And then we now see going into 2021, some companies are calling out a very big positive inflection. I wanted to know where Arista sees this from an observations perspective going to 2021, at least how it impacts you guys?

Yes. I mean, I think we're definitely coming off a year where, for the first time, I think at a market level, we saw a slower investment level from some of these customers, right? I think at the back end of '19 and into '20, we definitely saw some slower momentum in terms of investment, and that impacted Arista also because we -- obviously, we have a great exposure there, right? I think what we've seen now as we get into our planning window with these guys for at least initial parts of next year or first half of next year, you can start to see that there is some increased activity there, and we are seeing -- we can at least see kind of activity momentum around some of the planning that they're doing, some of the product planning that we're doing with them, et cetera. So I think we feel, at least from our perspective that, that part of the business has stabilized and will grow into 2021. We'll see exactly how that plays out as we go further into the year. But I think at this point, we can see initial activities is good, and we've seen -- there's lots of planning happening, et cetera, around that part of the business.

Got it. Got it. I think 1 thing that was very interesting about the last quarterly call was the discussion on white box. And Anshul did shed some light on that. And I think the 1 kind of big question mark that investors probably had is, over the course of just a couple of months, the narrative really did flip from white box being kind of an overhang on the networking industry as a whole to all of a sudden, a potential tailwind to the Arista business model, or at least the Arista business as a whole. So I wanted to know just what was the kind of the big change that occurred, at least from an Arista perspective that changed the narrative so rapidly?

Yes. I mean, it's interesting, right? It's not -- I don't think the narrative changed so much kind of on the ground practically in terms of what was happening. I mean, it couldn't, right? If you think about product introductions and the length of time that it takes to work on a design and introduction of a product. You can't, that doesn't change in 3 months, right? So I think the dialogue that's been happening in kind of some of these forms around white box and the threat of white box was probably more significant, honestly, than what was happening practically in terms of how we were executing the business, right? We've been involved in kind of designs for products and delivering those products and planning for the delivery of those products now for -- if you think about 400-gig products, it's been multiple -- almost 12, 18, 24 months, right, of activity around that. So I think there was some disconnect between maybe the concern around what that would look like versus kind of what's been actually happening in the business and on the ground. So I don't think it was quite as big a shift for us practically when you think about that, right? With these engagements, we're ongoing, we're working with these customers. So I think the status quo commentary was much closer to how we perceived it all along. There were some drivers externally watching 400-gig market share and where 400 gig is going first, which is really being driven just more by the customers' strategies themselves that was causing some concern, et cetera. But on the ground, practically, I don't think there was that big a change in terms of how we were executing. I mean at the end of the day, it's not a slam dunk for these customers. And if you were starting today that you should -- the white box is better then using an Arista solution to address your needs, right? So there's ongoing dialogue there, but there's clearly a value appreciated on the customer side from having these co-development type relationships or having Arista as a vendor in the mix. And I think we continue to focus very much on proving that value repeatedly and customers that resonates with customers, right? So I don't think it was as bigger shift for us from a business perspective as it was maybe from some of the discussions that were happening externally.

And probably -- if I would add, probably more of us being more vocal than we have in the past, Sami, about how well we're doing with our cloud titan customers. So there was -- there has been a lot of noise about 400-gig insertion from other competitors coming in. And so we felt it was also necessary to reiterate why we feel so good about this space and our position.

Got it. Got it. Yes, I would say that, I guess, in the investment community, we kind of saw the big flip, but I guess, it makes a lot of sense on your side, it wasn't as -- I guess, it's just not possible for these...

Yes. It's not possible.

So I wanted to shift the conversation a little bit more to the recurring revenue opportunities. And I was hoping, Ita, you could elaborate a little bit, and perhaps, Curtis, you can also kind of chime in with the security offering as well. Where do you see the key accelerators for software and services growth, at least as we think about 2021 and 2022, at least on the horizon?

Yes. I mean I think, look, our approach to this is that at the end of the day, 90-plus percent of our R&D is software focused, right? So really, we are a software company, and the investment and the emphasis is out of the software. There are certain pieces of that software that needs to be delivered in our minds and fulfilled and that customers want to have fulfilled as part of the initial sale, perpetual license sale. And then more and more over time, there will be other added features, added value that we're bringing. That will lend itself more to being solutioned as a SaaS recurring revenue model. And we believe that's largely customer directed, right? We're not trying to force that outcome in any way from a business model perspective. We still believe that giving customers that flexibility is important. So what does that mean business model wise that I think your focus from a recurring revenue comes from 2 things, right? One is there is a recurring, what we call A-Care revenue stream, which is -- it's maintenance and support, but it's access to software support as well, et cetera, and that's 20-plus percent of our revenue stream now, and it's growing very nicely, right? And that -- we believe that can continue to grow as you continue to expand your installed base and you continue to renew services on the installed base. That's a nice, solid, highly profitable business given the quality of the product. And then on top of that, you start to layer stuff like CloudVision, some of the Big Switch software capabilities. It includes kind of a wake we'll fit into that model over time. And the more enterprise campus business that we do, the more right we have to sell those add-on, value-add services on top of the core network operating system, which will still be perpetual, but it's still a software sale, right? So I think that's the -- that's how we think about it. So I think you'll see that services piece of the business will continue to grow healthily and it grew last year -- this year, I guess, 2020, even in spite of the fact that the product was somewhat muted, right? So it can continue to grow, I think, healthily as we go forward. And then we'll start to layer in some of these other software add-ons, which will be SaaS, naturally kind of a SaaS-delivered model.

Yes. And I would add to Ita's. As more customers adopt Arista in the campus as well as in their data center networks. Our 1 manageability plan, excuse me, CloudVision, we can deliver that as a service now too, Sami. And so that will have -- will be more sticky with customers that are looking for more telemetry manageability, observability monitoring of their overall network. And as I mentioned, he'll have more hooks to tie into newer software assets too over time.

Got it. Got it. I wanted to shift over back to hardware and talk mainly about the routing segment. And when you think about the opportunities that present themselves, kind of out of the core switching, but more on the routing side, what are some of the key customer verticals that are going to drive the routing product set?

Yes. I mean I think -- if you think about our routing journey, again, it started with kind of the 100-gig product set of that switch router capability finally be available. And it was targeted initially to the hyperscale accounts because they had a great appreciation of the advantages that they could get from having a merchant silicon-based switch router platform versus some of the legacy proprietary larger, higher cost kind of options that were there previously. So that was the starting point of routing, and that's driven incremental use cases across that cloud vertical over time, and we'll continue to do that as we get more capabilities on the silicon side and we add more software capabilities there. A lot of that software work has been, I would say, forward-looking from an architecture perspective. It was leverageable into the enterprise and financials vertical as well, right? I think we saw the concept of having a single hardware device that could operate as a switch and a router in the enterprise resonate well too. And in the provider's part of the business and in service provider part of the business, there were particular use cases where it resonated really well as well as a starting point to that service provider business. We've done very little business with service providers prior to that, right? I think the next challenge is to how do you break out of those very specific use cases in service provider and what does it take to do that. And we've been working diligently on the software side to add features and capabilities that help kind of translate maybe from an existing service provider footprint too. What we would consider to be an IT cloudified, if you like, network footprint? And what does it take to leverage what we've done for the cloud to make that happen, knowing that they're starting from a different place, right? And I think that's been interesting. There's been ongoing work there. We've kind of -- we retained the engagements on the sales side. We've done a lot of work around just getting some of the blocking and tackling contracts and all that step out of the way. We've had a couple of recent engagements where a, say, international service provider is ready to make that shift, and we've been kind of working with them to redesign that architecture. And that's probably something that's leverageable into other accounts, right? So I think we've made more progress. I think we're more prepared now. There appears to be some more receptivity on the customer side. But it's still work, right? And I think we'll monitor that and see how it plays out. I think from a business model perspective, I think about it as -- it's stable now, and it should contribute. Exactly how much is something that we'll see as we move through the year. But there definitely seems to be some -- we're making some progress there.

Got it. Got it. So we have 2 questions that came in from the audience. One of them is very broad. One of them is less broad. The first 1 is about 5G infrastructure and then the kind of internal discussions that Arista has had regarding addressing that opportunity. How big of a lift from a customer data center infrastructure perspective does Arista believe 5G will be as networks are densified and built out. When you think about your forecast, does that factor in a big 5G build out curve or not really? So this is the first question, which is 5G interested demand. I could layer in the second one, if you guys could decide how you guys want to answer them in what order. The second 1 is, you still have cash and you're generating free cash flow. Is there appetite for further M&A in the foreseeable future, at least in the next 12 months? Those are the 2 main questions I just got.

Okay. So maybe we'll take the second one, first, right? I think we have -- we are generating a fair amount of cash. We've been returning a fair amount of that cash as well, right? I think year-to-date, we'll have returned close to 70% of our free cash flow, our cash from operations. On the M&A side, it really is -- it's a lever, right, that's there. We have some clear criteria around how we think about what works and what doesn't work in terms of just how the technology needs to integrate, the importance of the single operating system. So looking for things that are adjacent that meet those technical and cultural criteria. I mean, that's real work, right? That's part of -- Curtis's day job is looking for those opportunities, looking for things that are adjacent, But at the same time, obviously, expanding the business and expanding the business potential. I think it makes a good example of something that really doesn't work right for that, and we'll continue to look for those, right? Back to the 5G question. I think from our perspective, we're not selling a proprietary 5G equipment, but we do believe that it can drive maybe more movement in terms of some of the service provider and other areas of the business just because it will drive, hopefully, more IP-based networking architecture, and that gives us a better opportunity to play there. It's been something that's been moving relatively slowly, I think, versus people's expectations. I think we've been talking about it for a long time in this forum. So we'll see how that plays out. But I think it's a net positive without us being like a specific player in those proprietary products themselves. Curtis, I don't know if want to add anything?

Yes. I mean, by all means, Sami, it's going to play each part in digital transformation. As 5G networks continue to -- start to be built out more aggressively, I'd say, over the next couple of years. And so that's going to improve more bandwidth access to the campus, which is going to drive more and more demand for higher bandwidth networking equipment, like the new Arista flagship 750 platform and also just more layers in the network. And so I think from an edge standpoint, you're going to have to continue to build out the network, which -- and build out the package core to support higher density and more bandwidth going in, that should be a tailwind over time.

Got it. Got it. I wanted to segue back to Arista specific in some of the past quarters and years. And more specifically, if we look at the company on a regional basis, and we look at European revenue generation, some of the quarters have seen declines, at least for the past few quarters, and I just wanted to double check to see how you're approaching either the dynamics in Europe or just the go-to-market strategy in Europe? And maybe if you want to unpack that in data center versus campus, that would also be helpful because those are probably 2 completely different dynamics at this point in Europe. So any color with that would be great.

Yes. I mean, I think the first thing to kind of maybe unpack is just how the geographical split actually works from a disclosure perspective and cloud plays a big part in kind of how those numbers move, right? Because depending on where our cloud customers are deploying their equipment, it's categorized on a geo basis based on the ship to location. So if a cloud provider is doing a build-out in a region, then that region is going to look stronger. And then when they flip back to doing something in the U.S. or somewhere else, it kind of offsets that again. So some of the volatility that you're seeing in there is coming from just that cloud shift in those cloud deployments and when they're deploying where, right? If you strip all that out, I think what we've seen is we have seen some gradual growth in, say, EMEA and APAC. We saw some volatility, I think, in the prior quarter. Mainly probably COVID related, where we saw some opportunities kind of push out in time. Some of that came back in Q3. Our approach has been somewhat similar in that the product plays and is accepted in the same way. The market requires a little bit more channel. So we're probably more channel-oriented there just because it's more diversified, you need more support locally. There's language, et cetera, that makes that easier for the channel to do. So it's a business that has a little bit more channel engagement already, if you like. It's a coverage business. I mean you need to have people in region and country, right? Otherwise, it doesn't necessarily work. So we've been building those sales teams. I think Chris is definitely focused on targeting key accounts there and adopting that same approach there, kind of once you have the coverage model, you still need to have that focus on big bets and on big accounts. More of the business is probably in the hands of service provider, that would necessarily be a case in the U.S. So that's a slightly different dynamic as well as means that you have to focus on those multi -- those international service providers. So I think there's a lot that's the same, but you still have to be cognizant of kind of the regional and country differences and respond to those. And I think we're investing, and we'll continue to invest to do that. And like I said, some of that international service provider focus, et cetera, is part of that, right? And then large enterprises, financials and other pieces as well. So I think it's definitely work in progress. We're more mature in the U.S., there's no doubt. Maybe on the channel side, we're a little more mature internationally than we are in the U.S. just because of how it plays out. But I think there's definitely work to do there. But the numbers overall will move around a bit just because of the cloud impact as well as what's happening in the in-region business.

Got it. Got it. And then kind of just to conclude. Going into 2021, we know you've kind of given some soft or qualitative guideposts in terms of how we should be thinking about the business trajectory. If we were just to talk about maybe OpEx. What should we expect the OpEx levels and kind of the dynamics there? What should we expect from like OpEx and margins at least for 2021? And you could stay -- stick to qualitative or kind of soft as well because I know you're not very explicit in both the year kind of lapse. So any kind of color around that would be helpful.

Yes. I mean, I think assuming that we return to growth in the ranges that we've talked about, I mean that obviously helps a lot with funding. Some of the investments that we want to make and targeted investments around the sales and marketing side. We'll continue to invest on the R&D side. But once the top line is growing, that definitely helps to fund that. I think what we've said is we see the kind of the 37% operating margin that we guided for Q4 is probably a good way to think about it in the near term. Reserving the right if there was something discrete or significance that we wanted to invest in, that we could go back to more of a long-term model, which is 35% operating margin. I think gross margin, I still think the 63% to 65% as a bound, is a good way to think about it. Obviously, we've been doing well, kind of at the midpoint and above of that, but it is still a customer mix dynamic so there could be quarters where you'll be at the lower end of that range, just if cloud is stronger and there'll be other quarters where we'll do better than that. So I think over time, that kind of -- the midpoint of that range is fine, but we could move around within that range over time. So I think that's how we think about the business. If we're growing at the rates that we've talked about, confirming consensus, plus we may be a little hot today at the 38.5% or whatever it is that we've been operating at. But the 37% plus or minus is -- seems reasonable in the near term. And then with the right -- reserving the right to kind of revisit that 35%, there was something incremental that we really wanted to invest in.

Got it. Got it. Well, Ita and Curtis, thank you very much and the Arista management team for participate in our conference. I really appreciate the time of you taking for this fireside chat. So appreciate it, and hope everyone has a great day.

Okay. Thanks Sami. Thanks for having us.

You too, Sami. Thanks for having us.

Bye-bye.

Absolutely.