Everyone, welcome to day three and the final session of Citi's Global TMT Conference. We have definitely saved the best for the last. My name is Atif Malik. I cover U.S. semiconductors and networking equipment names. It's my pleasure to welcome Chantelle Breithaupt, Chief Financial Officer, as well as Tyson Lamoreaux, Senior Vice President, Cloud and AI Networking at Arista. Welcome, guys.
Oh, thanks so much. Thanks for having us.
Yeah, thanks.
Great. I'll kick it off with my questions first. If you have a question, save it towards the end. We'll tend the mic, and you ask your question. Chantelle, deferred revenue has been a hot topic this year on Arista. Arista consistently achieves strong revenue growth, has guided to 40% growth this year, and yet the complexities of AI use cases have resulted in strong growth in Arista's product deferred revenue. What is the right way for investors to think about Arista's growth? Is it the growth in the revenues plus the change in deferred revenues?
Yeah, I think it's a great question. We're super excited to, in the last earnings call, raised our revenue guidance to be 40% growth, $12.6 billion. We're very excited to be able to deliver that kind of guidance. If we think about it, deferred revenue is a topic. I think that if we just level set what deferred revenue is, if we're talking about the product deferred revenue, that means equipment that's been shipped, invoiced, and cash collected, but we have acceptance criteria that keeps us kind of skin in the game with the customer to get to the kind of net results they want to get to. If you think about what goes into here, it used to be the cloud.
If you go back a couple of years, the cloud deployments were put in there, and now AI has basically put that on steroids from a complexity, to your point. From a how do you interpret it perspective, I think if you look at Arista, there's a few things I would keep my eye on. I would keep my eye on the P&L growth, what's happening in deferred revenue. But I wouldn't do a quarter-over-quarter kind of thing. I would do a trend over four quarters, maybe six quarters, because things will come in and out generally, because it's an acceptance-based criteria. I would look at the purchase commitments, and maybe what's sitting on the RPO table. All four of those things kind of give you a picture of what's happening and what's to come.
I think that from the perspective of deferred revenue as well, and sometimes people don't understand that some of these deployments can take 18 to 24 months till we hit the acceptance criteria. So it's important to understand the timing of how long this can be and when we would realize it from a P&L perspective.
Great. Chantelle, just for perspective, how far are we into this current cycle with deferred revenue, and how does it compare to what Arista experienced in prior cycles?
Yeah. Even though we're now going to be part of the S&P 100, we're only about 12 years old as a public company, right? We've had, I would say, two cycles, the cloud cycle and the AI cycle. What you're referring to, Atif, is during the cloud cycle, which is kind of the 2019, 2020, 2021, you saw deferred revenue raise because we had this deferred construct because we had the use case of cloud, 400G and then coming into 800G. Then we kind of went through a little bit of cloud, and then AI took off. This AI cycle is the second one. The cloud one was about three years. We're in, I would say, going into year two and a half to three on AI proper from a materiality perspective.
I think there's many years to come on this cycle, and so we'll have to see because we're talking about scale- up, scale- out, scale- across, training, inference, front- end, back- end. We don't see AI going away anytime soon, so this cycle could be longer.
All right. With that topic out of the way, the next big topic has been the supply environment. Supply obviously remains challenging. On your last results, you came out strongly on supply. Where are you seeing the most pressure in the supply chain?
Yeah. We started the year, and I'm very transparent, Q1 was not our favorite call because we did have to talk about the supply chain environment, and I think that's what you're referring to. But we wanted to be honest and transparent. A lot of this pressure is coming from right at the top of the fab capacity, and I think anyone in the industry should be talking about there is tightness no matter where you sit in that hierarchy. But as we work through to the Q2 call coming into August, we did find some vendor arrangements with fab capacity, memory. So those things, I think, we've sorted through 2026 and 2027. But you have Whac-A-Mole on some of the other component parts, PCBs, capacitors, et cetera. So there's things that are more peripheral short- term, and then there's structural things like fab capacity.
We feel very well positioned. When we talk about purchase commitments now, that's a year lead time for chips going into Q3 next year. We are already talking that $9.6 billion at the Q2 end purchase commitments is already going into second half next year. So well positioned, but I don't think anyone's out of the woods when it comes to the full supply chain for networking data centers generally.
Can you remind us any purchase commitments you made to secure the supply?
We have grown. If you look three quarters ago, our purchase commitments was about $3.6 billion, and now we are at $9.6 billion at the end of Q2, so over three quarters, basically tripling it. That's leaning into demand we are seeing. So you should see it as a demand. It's getting some of these memory things sorted. Then there are some things that are a little bit longer than 12 months. From that perspective, we feel very good that the supply chain team learned a lot during COVID, learned a lot during the beginning of AI, got the contract manufacturing components sorted. I think it's a demand signal at this point, and we feel very confident about that.
Awesome. Tyson Lamoreaux, a lot has been said about the differences between AI traffic and AI networks versus classical cloud computing network, and Arista has developed its Etherlink portfolio to address the opportunity. What commonalities are there between Arista's classic cloud solutions in hardware and software versus the AI solutions?
Yeah. Chantelle and I were talking about this a little bit earlier, too, and Arista's heritage and foundation, we believe, really put us in a very well-suited position for the AI era. In a way, the company was born and bred to be prepared for the AI era. There are clear distinctions between cloud, traffic, architectures, hierarchies over subscription, interface speed, upgrade cycles, refreshes, and how the compute is interconnected. AI is much more demanding. The traffic profiles are significant. They are very highly coordinated. The reliability is essential at every level of the piece of the infrastructure because you do not have the ability, at least with current generation AI technology, to provide ways of working around failures as easily as you do in cloud type primitives.
The demands are just very high here, and I think the commonality for us from where the company forged its foundation was rooted in reliability, security, performance, and scalability. All of those dimensions are more critical for AI than they were even in cloud. I would highlight, too, in particular, that root around our heritage in hardware development, the ability to deal with latest generation, intercept latest generation technology, overcome very hard electrical engineering problems, signal integrity problems, power integrity issues, things that are exacerbated as you move up into faster link speeds, as we are now transitioning into the 1.6T generation. These are very difficult challenges. You are emitting a lot of heat because you are consuming so much power, so your ability to cool that and be power efficient is super critical. Something Arista has really built its heritage, again, in hardware engineering around.
That lends itself very well for our Etherlink platform and AI fabrics. The software is another piece on the reliability side and the security side that is essential. One of the things that has come to front of mind for a lot of people in the last year really is Mythos and Glasswing, and the notion of AI-driven attack vectors and an ever-increasing rate of security vulnerability discovery, fixes, patching, and this cycle that needs to basically run continuously now to keep your network secure. One of the benefits of our software is that it was built on a foundation of always staying up, so you can patch the software, deploy it, and you can automate the entire workflow seamlessly so that your data plane is never interrupted. The network stays up, traffic continues to transmit across the network.
That gives us a real relative advantage against competitors, as well as Open NOS and other type solutions. We feel very good about that foundation that we have been able to carry forward into the AI era and get leverage out of it. That reliability aspect. If you are talking about reliability that is a full point better than the next best in the industry, that 1% improved availability translates in a very significant way when you start talking about amortizing availability across all of your accelerators, which are the lifeblood of AI, right? This is the biggest expense. This is what you are spending the most money on. You do not want that capital sitting around idle. You want putting to work. You want your high utilization. The speeds and feeds, cutting-edge technology, reliability, and security story come together very well. So served us well in cloud.
I think it's serving us even better in AI.
Great. Historically speaking, you guys had concentration at two hyperscalers, and you've done a very nice job in diversification at the hyperscalers with potentially three or four this year, depending on how things shape. The hyperscaler concentration is improving. On the non-hyperscaler side, the neo-scalers and enterprise, can you talk about your adoption cycle over there?
Yeah. We're very excited about these other non-hyperscaler environments. We talked about, we do have two hyper 10% customers that you're referring to. We do anticipate a third and potentially a fourth. So that diversification continues. From a neo-cloud perspective, the one thing, back to Tyson's commentary, the neo-clouds are looking for a few things. What makes them competitively advantage is their token cost optimization, and so they come to us to say, "You have great experience both on AI front-end and back-end and cloud. Can you help us get to token optimization as quickly as possible with your vast portfolio?" All the different architecture designs that help them get to their token cost optimization.
If it's a fair fight in the sense of it's a best-of-breed conversation and it doesn't come down to commercial arrangements taking the customer in a different avenue, we have a very good chance of winning. Even some of the neo-cloud customers who have gone with a competitor, perhaps for different reasons, have come to us perhaps for their second installation because they realize the best of breed will get them to those token cost economics, which is how they stay in business. So we're very excited in both local and international growth from that perspective. Enterprise, I think, is just getting started. We have about 20%, 22% market share in data center. So even if we just continue doing classic data center, there's market share to go gain. But why else would enterprise want to accelerate their next refresh for us?
Back to the Mythos environment, if you think about where some of these enterprise players are, they are very nervous about if they're currently with many different operating systems, how they're going to get on this patching treadmill when it comes to making sure they're secure. We've had customers come to us that are new logos ahead of their refresh to say, "We're very interested in your EOS because we only want to patch once, and we want to make sure that it's done properly," because we have the lowest CVEs in the industry by a magnitude of 30x to 1x. The other reason the enterprise is coming to us is just total cost of ownership.
With one EOS, with CloudVision, they can get rid of the 50 to 60 to 70 people, perhaps 200 they employ, just to always be constantly updating, taking the system down, repatching. From a TCO perspective, enterprise is also very excited. Market share and TCO is the conversation there. Then you go to campus of the enterprise, right? Campus, we have 5% market share going from USD 800 million in 2025 to USD 1.25 billion in 2026. That growth rate's definitely larger than the market growth rate, so that's a share gain story. That'll be a long, slow rinse repeat that we're very excited. We win campus first deals, which is a great validation of our portfolio and our brand recognition. There's land and expand going back to the data center the other way.
There's lots to be excited about, I think, in those two categories.
Security seems to be a nice flywheel for you guys. All right.
Yeah.
Scale-across AI applications have emerged rapidly and are expected to be approximately one-third of your AI mix. What is driving the growth in scale- across, and what could shift the scale-across mix higher?
You want to start?
Sure, yeah. I'll start probably with the back- end of that, like what could drive it higher. Constraints are driving a lot of this in general, and there's some architectural choices, and there's some desire to distribute data centers either because you can't get the power and data center space in a single facility that you'd like, or you want some resiliency built into your infrastructure, so you choose to distribute. Another big driver here is actually coming around not just driven by constraint but enforced constraints, regulatory regime change, and top-down government initiatives. I'd highlight a good use case that really exemplifies it is if you look at South Korea, the government there is actually driving, by policy, distributed generation and local consumption of power.
By its very nature, what's going to happen to the data center infrastructure inside of South Korea is it's not going to concentrate into a small set of locations or cities where all the capacity is kind of all your eggs are in one basket. It's actually going to distribute around the country, and naturally, that is going to drive much more demand for a scale-across solution because your single largest facility is going to be capped effectively based on the local power available to it. As things federate, you naturally have to interconnect them and interconnect those clusters. That is something that can shift and change and shape the scale- across TAM quite a lot. I think I'll let Chantelle talk more broadly about the TAM and how we think about it.
Yeah, I would be happy to. We are revisiting our TAM generally. At the last Analyst Day in October 2025, we gave you a $105 billion TAM. A lot has changed even since less than one year ago, so we will be revisiting scale- up, scale- across, and hopefully come back to you later this year with a revised TAM. We see that expanding for all the reasons Tyson listed. Also to provide a framework perhaps to help the audience just how the scale- across gets us to Arista being a good choice. We are very well-positioned. The customers, we feel, are looking for a combination of great software, operating system, hardware, and AI experience because this is not just a DCI, this is a DCI with a much more complicated structure and high large clusters, right?
If you have AI experience software and hardware, I think that is what wins in scale- across, and we do not feel our competitors have all three. Some might have DCI, some might have AI clusters, but we do not see any of them that have all three, and that is why we feel incredibly well-positioned for scale- across. Wait to see on the TAM, but if the TAM increases, we are very excited about that.
Great. Then on the scale- up, the market has been waiting for Ethernet standards, but a consensus is emerging that scale- up will be a huge opportunity. Broadcom talks about Ethernet being a leading candidate. How do you see the mix between scale- out, scale- across, and scale- up evolving?
Well, I can talk about the mix, and then maybe we could talk a little bit about scale- up, or perhaps your questions will lead us there. A lot of this depends on what timing we are talking about. From scale- up, we have zero in our guidance and zero revenue for Arista for scale- up in 2026. We feel that we will start to get some kind of maybe trials and pilots similar to our InfiniBand to Ethernet conversation later next year because the ecosystem has to come together. Ethernet scale-up is going to take some ecosystem parts coming together because it is a bit more complicated. We are talking 2028 is when we get the revenue there. That is why I say the percentage will depend on the timing. Scale- across, I imagine, will happen a little sooner because we will see that 2026, 2027, and 2028.
Both will have great growth trajectories, but the timing of it makes it difficult to answer your relative perspective. We're hoping all three are running very well for us when we get to 2028, when all three are online, let's say. We'll have to see the timing of all those three, how they show up.
Tyson, can you touch on the XPO roadmap, and when do you guys intercept that in terms of revenues, and how does that fit in this whole NPO, CPO roadmap?
Yeah. We feel very good about the roadmap in general for our products and the base building blocks that we have in place for advancing our product portfolio, generation over generation, and to intercept NPO and CPO. We are customer-driven in terms of our product roadmap, and so if customers want a CPO switch, they want an NPO switch, they have certain requirements that need to be fit, we're going to go build that switch for them. I think you've got some alignment right now with XPO and the scale- up timing. There's some coincidence to that, but they're also not just purely time coincidental. There is form factor and bandwidth density that comes with scale- up. As the Ethernet standards around ESUN are ratified, you get chip production. Chip production leads to ability to build a system. You're talking about very dense systems here.
We're going to build scale-up solutions to focus on customers who would like either co-packaged copper for lowest possible cost, but those are going to be constrained in terms of their reach because you've got a bandwidth over distance problem. The longer your link is, the less bandwidth you can attain over it. That gives rise very nicely to an NPO solution, which is what our XPO architecture is focused on. The ecosystem's coming together very well. The module makers are building all kinds of products, and they're going to come together very well in that timeframe. It's a product that we think it fits very nicely in our portfolio through optionality, serviceability in the field, things that our customers are telling us they want. It's also going to help them start to get into that multi-rack, high density scaling that they require that copper isn't going to reach into.
We see this as a nice two, three generation solution that lands in between copper running out of steam and when CPO can really be built at scale. But we expect some early adoption on CPO for sure. We are ready to build as soon as customers are telling us that they want it and need it. I think you will find a pretty common adoption there. Early adopters really drive the front of it. There is a whole ecosystem that has to be built out. It is a completely different manufacturing technology and architecture. The supply chain has to come together for it. Yields are going to be very challenging in it very early on.
There are a lot of customers who are being pretty pragmatic about it, and in terms of their expectations of when they can intercept and what value it is going to create and when, and when the costs are going to stabilize and the volumes are going to increase. We like the roadmap we have right now, which is this continue copper as long as we can, bring NPO and XPO into the fray, build out a lot of the modularity in that generation to be able to then feed into CPO and then CPO coming on. I think, XPO early deployments second half of next year, really ramping in 2028, and probably 12 months behind that is where we start to see some of that CPO ramp occur.
Do you guys have a view in terms of the mix of ESUN versus UALink? Where is this market going?
Well, again, we are still learning about the TAM and the market because it is a fairly new thing. But preliminary views, I would say, just from a framework perspective, we see this starting off as three segments within scale- up. We see the proprietary, which you could put NVLink and UALink. We see probably some sort of commodity white box play for the lower value add needed. And then we see value add Ethernet, which is where we will play. Two things to think about. One is, if you have an overall TAM, how much does the Ethernet part start at? Is there eventually a journey similar to proprietary to open standard Ethernet? You could say NVLink or UALink to Ethernet. Is that a journey where that starts to shift over time? That is how we are starting the framework of this.
Great. Let me pause there and see if there are any questions in the audience. If you have a question, please raise your hand.
Thanks. You touched on optical a bit in your discussion of XPO, NPO, CPO. Is vertical integration something that you're considering given the rising importance of optical networking?
Well, we have deep conviction, I should say, and super engaged on making sure that ecosystem comes together. I think we're going to be pretty prudent about vertical integration, where we feel like we can really add value. I don't think it's really something that we want to do to go spend time, energy, effort, CapEx, R&D resources, defraying kind of our focus on building products just to be another commodity provider. So in a world where we're talking about XPO and XPO modules, we think that's a really rich ecosystem, and I don't think we really see a need.
But if a customer comes to us and say they really think that there's value in being able to come to a single source, they believe that they are going to gain significant benefit out of offloading a bunch of test, qualification, verification work to us rather than taking it on themselves. They want to kind of simplify their supply chain so they can focus elsewhere. We're going to be open to those conversations. That'll factor into, what do we stick on the roadmap, when, where and why. Unsurprisingly, we've built up a very competent team in the optical domain because it's so essential. Connectivity and connectors are so essential as we continue to step through the greater speeds and into next gen platforms. So we feel well positioned from an engineering product perspective.
If there's something to build there's a there there, like we can go do it, but nothing hard committed at this point. I think it's something we'll continue to discuss a lot internally. We'll keep listening to customers, and I think you'll learn about if we choose to do something, like when we're ready to launch it, because we don't tend to pre-announce anything.
Great. Question?
Yeah. Thanks. Taking the question. Can you talk to, I know it's early, but scale-up Ethernet. Your value add obviously is software, it's a big part of the pitch. Does the software requirements, do they go up? Do they change? Maybe just talk to that. Is the ecosystem different? What value can you bring to it? Maybe talk to that a bit.
Yeah, I think if you went back in time, maybe 12 months, there was a lot of momentum around the notion that scale-up is really just a big, fat, dumb pipe, and therefore there's not really much value to add. This is going to be a great white box play. I think there are white box companies out there who have definitely said, "We think this is a great opportunity for us." I think the reality is everyone's kind of recognizing reliability and security are just as essential in scale-up as they are anywhere else.
If you're losing GPU workload time or accelerator workload time because your network is down, your scale-up network is unreliable, you don't have the telemetry you need to debug issues, even after the silicon is shipped, and the ESUN spec is ratified, and you've got a purpose-built product in place, that's going to be a deal breaker for customers. I do think EOS and our software advantage is meaningful there. I think the feature set is definitely smaller. You're seeing a shrinkage in terms of breadth of capabilities. The question is going to become: how much do the features and differentiation matter? It's a breadth versus depth kind of question, but I think I'm gaining personal, deeper conviction every day that the security story is going to be really credible here for us and much better than a lot of our competitors.
If you look at the Open NOS ecosystem, you're subjected to long delays, self-discovery on vulnerability fixes, reviews, reporting regimes that are dictated by a governance model. There's no test infrastructure readily available to everyone. You've got to do it yourself. These are real challenges for anyone who wants to go down the white box route, no matter where they're going to deploy it. You've got to be really committed if you want to do that. If you think about the total expansion of the TAM and scale- up with Ethernet specifically over time, it's going to be a variety of different types of customers. It's not just the hyperscalers. It's not just the most sophisticated types of customers.
I think we're going to see that represent a lot of the customers we're working with now, and I think a lot of the things that are leading to wins for us now will lead us to wins in that domain as well.
Chantelle, last year I said you're making Ita proud. This year I'm going to say you're making Ita more than proud because in the last two years since you joined the company, Arista's operating margins have expanded impressively by 100 basis points. How do you think about the balance between investing in the business, innovation in high-cost commoditized environment? Are there areas that you're targeting for more operational efficiencies?
Yeah. Thank you. Let me start with thank you. I hope I have made Ita proud with the transition. I think that from the perspective of if you think about our guidance, we are guiding basically a 62%-64% gross margin and 48% range. That is almost a rule of 90. I do not know if we are looking for more operational gains. However, there is a lot of operational leverage which will come through depending on the mix of product. When you think about our 40% revenue growth this year, we do not need to scale the same rate in the sense of our operating costs. There is expansion, to your point. I think we will continue to see where we grow and how we grow and how much more operational efficiency we can get.
We do get a lot of leverage from the area Tyson has specifically with the kind of cloud titan space, neo clouds, et cetera. We will have to wait and see. Even if I keep the same ratios, we would like to be obviously innovation-forward. That is our goal. If you think about R&D is usually 8%-10% of revenue. If we grow 40%, that gives Ken a whole lot more just pure dollars to go play with, right? We feel comfortable that same sort of range, sales and marketing, 5%-7%, G&A is an amazingly 1% or less, because we only have one P&L in the whole company, so we are a very elegant business model. There is room for expansion over time. We will have to wait and see, but it is a great business model that way, I think.
On the topic of pricing and value-based pricing, how are you guys thinking about the value-based pricing in this environment where the memory costs are going higher and lead time is also stretching?
Yeah. I think that our philosophy is to be as transparent as possible. What do I mean by that? We try to measure a few times and cut once. We try to see what is going on with components, and memory being one of them, et cetera. Anything we can see, try to sweep it in and do one price increase, and we did that earlier this year. I am talking for the majority. We did not do just, "Hey, everyone gets 7%, and we are done," because that is very easy. We went through and said, "Okay, who has a higher percentage of memory in your BOM?
We're not going to price your backlog because you've already committed those purchase orders. We're like, "Hey, if you have something that's related to an increase we see, we can explain to you we have to raise your price increase to at least keep us margin neutral." We're not looking to accrete margin because we've seen that in some of the other people in the industry, the companies showing accretion. I don't think that's how we want to work with our customers. We'll be transparent where it's applicable to the BOM. We'll have the commercial conversation, and the commercial conversation will weed out how much of that increase we get to keep. We'll keep doing it that way. We'll see as we go into 2027 if something else is needed.
We just try to be transparent so that it's clear, the pass-through sort of mentality, I would say.
All right. Let's talk about the other businesses. At this conference last year, you talked about some of Arista's measured efforts to diversify revenue, including tapping into the federal sector and investing in Arista's international business. Can you update us on some of these diversification efforts?
Yeah. We talk a lot about AI and a lot about cloud titans and neo clouds, but there are a lot of people who show up every day to make sure that the less sexy, boring, slower growth things are happening. Part of this is international. Fed, I think, is very nascent for Arista, given where we are. Federal is an opportunity, not just in the Americas, across the world, I would say, from a public sector perspective. But when you talk about geographies like EMEA last quarter had 36% growth, which was a great growth quarter for Paul Jordan, the leader there. That's a combination of the large customers flowing through the region. But also when you get to that kind of growth, it's in region, for the region with enterprise, Neo Cloud, et cetera. We're seeing some good traction there.
We're winning new logos in enterprise and campus. We're seeing a lot of great Neo Clouds, for example, in EMEA. It's a focus of attention. We dedicate headcount and growth from a go-to-market perspective, lean a bit more into channel partners. That's a slow, steady growth. The rinse repeat for the APAC region, who also has a leader that's within a year coming from Microsoft. He's a great leader there, too. Very excited. I think it's, again, back to what are the growth trajectories for Arista. International is definitely one over the next 10 or 15 years that has a lot of room to grow.
Good to know. On the campus market, you have talked about having growing traction in that market, winning new customers. First of all, where are we in the campus refresh cycle?
Well, I think that there are some peers in the industry who have let us know there's a big refresh cycle going on, so we're very excited about that. We're only 5% market share, so I think you look at us going to $1.25 billion. I think we're coming in exactly when you're seeing a lot of the refreshes start. It's a five, seven, maybe nine-year journey for the customer. Coming in right when they're starting to do their next cycle, which is super exciting. That was just if we were talking about just the normal portfolio. You can layer on some of the conversations on them getting ready for inference, which we've talked about, right, Tyson?
Customers who are coming to us ahead of a refresh cycle or coming to us from the competitor because they're interested in how we can help them get ready for inference. They're worried about some of the security with the Mythos/ Glasswing environment. So many factors drawing them to us. We're winning campus first deals, like I was mentioning earlier. A lot to be excited about. We feel we have the right portfolio. VeloCloud's done very well integrating from an SD-WAN perspective, coming into a full- year now with the company, so we're pretty excited. That's a high volume, low dollar, multi probably decade growth.
Then I know Jayshree and Todd, they have talked about the go-to-market in terms of product versus go-to-market, and where are we in that stage?
I think that sometimes because we are 12 years post-IPO, we had a large cloud start to the company. We got into enterprise, and so we are not a heavy go through channel market to customers company. We continue to go direct fulfill through the channel majority-wise, and we are continuing to fund it that way. But back to the international, where it is probably most important, we have, with Todd Nightingale, doubled down on who are the 10 most impactful players we need to work with and how do we cultivate those relationships. So we are very excited by what we are seeing there, and I think you are just seeing Arista grow into their enterprise legs from that perspective. So it still stays in that 5% - 6% of revenue.
That gives us a lot of dollars each year to invest. So enterprise headcount or go-to-market, we have definitely been investing in, so more to come.
And then software as a percentage of the total business, how do we look at that number over time?
Do you want to describe maybe how the larger customers go through their software purchases? Because it does not really change over time, right? It stays at this 18%, 19%, because it is part of the operating system that goes with it to the customer so.
Yeah. For our hyperscale customers.
Yeah.
They have got a mix of things where they deploy an Arista Blue Box and they are going to run their own NOS or some kind of Open NOS for certain use cases that are better well commoditized, and then their mission critical, high capacity, high demand workloads, they rely on EOS. Part of that transaction includes not just software in form of licensing, but also support. We have single transactions that we do at the beginning of the life cycle every time. But as they are incrementally deploying and growing, with that growth comes additional license and additional support fees that come in. Then if they get to the end of the life, they want to keep the equipment around, there is support renewals that kick at that point as well.
It is a little bit more of a derivative is how I think about it than its own unique business model, and with independent drivers and independent growth factors. You should generally just, I think, and Chantelle can correct me if I am wrong, but I think you are just going to see as our numbers grow, the software revenue grows pretty linearly with it.
Yeah. It stays in this, to your point.
Yeah.
18% range, 18% - 20%. It's not that we have a software division that goes and sells it independently.
Yeah.
It's part of what we deliver in totality to the customer at this point.
Sounds good. We're almost out of time. Chantelle and Tyson, thank you for coming to the Citi conference.
Yeah, thank you for having us.
Thank you so much.
Thank you.