Arista Q3:2025 Earnings

Key quotes from the earnings call and letter to shareholders…

(note: bold and underline are my own)

“Arista continues to drive its 19th consecutive record quarter of growth in this AI era. We achieved almost $2.31 billion this quarter with software and services contributing approximately 18.7% of revenue. Our non-GAAP gross margin of 65.2% was influenced by favorable mix and inventory benefits. Americas was strong at almost 80% and international at approximately 20%.”

“Of course, we interoperate with NVIDIA, the worldwide market leader in GPUs, but we also recognize our responsibility to create a broad and open ecosystem, including AMD, Anthropic, Arm, Broadcom, OpenAI, Pure Storage and VAST Data to name a few, and build that modern AI stack of the 21st century. This stack includes the trio of compute, memory storage and a solid network foundation to run training and inference models.”

“Our stated goal of $1.5 billion AI aggregate for 2025, comprising of both back end and front end is well underway. We are now committed to $2.75 billion out of our new target of $10.65 billion in revenue, representing 20% revenue growth in 2026. We are experiencing momentum across cloud and AI titans, neocloud providers and the campus enterprise. The demand and scale of AI build-outs is clearly unprecedented, as we look to move data faster across multiplanar networks.”

“People and leadership are key to our success. And to that end, we announced Todd Nightingale as our President and Chief Operating Officer last quarter. This time, we want to celebrate the promotion of Ken Duda, our President and Chief Technology Officer, not only of engineering, but our top AI and cloud segment of customers as well. Ken, as many of you know, has been a champion of architecture, innovation and culture since founding Arista over 20 years ago.”

“One of the best things about working at Arista is getting to build some of the most ambitious networks ever built, ultra-low latency trading networks, global scale cloud networks and most recently multi-petabit AI networks. Our success in AI has many sources, the sheer power and performance of our hardware platforms, our innovations in fabric architecture, our AI-focused telemetry and provisioning automation, our reputation for the highest quality software and our leadership in the Ultra Ethernet Consortium, the UEC, and our work in Ethernet Scale Up Networking or ESUN. And most importantly, the way we partner with the world’s largest AI companies. Partnership has been key to our success over and over at Arista and the AI revolution is no exception. In addition to being a lot of fun, these partnerships benefit our company, both through the sheer revenue opportunity, but also in providing us with the opportunity to learn and innovate at the edge of what’s possible. We can then apply what we’ve learned to bring solutions to the broader networking market, helping a much larger and more diverse customer base, build the most advanced and reliable infrastructure in the industry. For example, our Etherlink distributed switch fabric powers some of the largest AI fabrics in the world. It’s also an excellent underlay for data centers of all sorts, providing a full line rate fabric with no hotspots at petabit scale for all workloads, including AI. Etherlink speeds are going from 800 gigabits today to 1.6 terabits in the near future while leveraging our EOS operating system and our NetDI diagnostics infrastructure for top hardware and software reliability. Arista AVA or Autonomous Virtual Assist, uses AI to help our customers design, build and operate their networks. AVA draws on both our internal knowledge base and also on the customers’ data stored in NetDL, Arista’s network data lake plus AVA has agentic capabilities to help troubleshoot proactively.”

“At Oracle AI World, Ken was invited to formally announce our collaboration with Oracle Acceleron. This builds upon a decade of partnership with Oracle, starting with our Exadata migration from InfiniBand to Ethernet for AI networks to RoCE, RDMA over converged Ethernet, and now multiplanar networking across cloud AI for on-time job completion in gigawatt scale AI data centers.”

“As part of our Leadership 2.0, we have built and focused a cloud and AI mission and organization, now led by industry veteran, Tyson Lamoreaux, reporting to Ken and Hugh. I am so delighted to formally welcome Tyson to Arista. Tyson, if you guys know him well, built the first cloud network for Amazon AWS in the 2000 era and pioneered the first AI network for a stealth sovereign AI company the last couple of years.”

“Spending time with customers has been a top priority for me since coming on board. And I’ve been so impressed with how strong these partnerships are, both with our long-standing titans and with our emerging customers. We’re deeply engaged with them on next-gen architectures for their cloud networks, front end, back end, scale up, scale out and scale across. I mean, really everywhere. It’s translating to a ton of wins, and I got to say it’s a lot more than I anticipated before I got here.”

“With Tyson’s credentials and a track record, Arista is really poised to address multiple facets of the cloud and AI innovation at a system-wide level converging silicon, hardware, software, cables, optics and racks as an overall platform. At the Optical Compute Conference, OCP, Arista unveiled its first Ethernet for Scale-Up Networks or ESUN specification, along with important 12 industry experts. While we began with 4 co-founders, we are now supporting and increasing to more people so that we can build the right interoperable scale-up standard. While there’s always white noise, Arista also continues to clarify our role in white box and how we will continue to coexist like we always have the past decade or more. The concept is clear. It’s all about good, better and best, where in some simple use cases, a commodity white box is good enough. Yet in other cases, customers seek the value of better Arista blue boxes with state-of-the-art hardware with built-in NetDI for signal integrity, physical, passive, active component and troubleshooting management. The best is, of course, the Arista branded EOS platform for the ultimate superiority.”

“We find ourselves amid an undeniable and explosive AI megatrend. As AI models and tokens grow in size and complexity, Arista’s driving network scale of AI XPUs, handling the power and performance. Basically, the tokens must translate to terawatts, teraflops and terabits. We are experiencing a golden era in networking with an increasing TAM now of over $100 billion in forthcoming years. Our centers of data strategy ranging from client to branch to campus to data and now cloud and AI centers is a very consistent mission for the company. We will continue to invest in our customers, our leaders, our partners and certainly most of all, our innovative technology.”

“Now on to the balance sheet. Cash, cash equivalents and investments ended the quarter at $10.1 billion.”

“Our total deferred revenue balance was $4.7 billion, up from $4.1 billion in Q2. As of Q3, the majority of the deferred revenue balance is product related. Our product deferred revenue increased approximately $625 million versus last quarter. We remain in a period of ramping our new products, winning new customers and expanding new use cases, including AI. These trends have resulted in increased customer-specific acceptance clauses and an increase in the volatility of our product deferred revenue balances. As mentioned in prior quarters, the deferred balance can move significantly on a quarterly basis, independent of underlying business drivers.”

“Our guidance for FY ’25 is as follows: full year revenue growth of approximately 26% to 27% or $8.87 billion at the midpoint. We are on track to deliver between $750 million and $800 million for our campus segment and our AI center target of at least $1.5 billion. For gross margin, the outlook is approximately 64%, inclusive of possible known tariff scenarios. We anticipate operating margin of roughly 48%, demonstrating Arista’s strong operational execution and scalable business model.”

“Our outlook for FY ’26 presented at our September Analyst Day remains relatively unchanged. Full year revenue growth of approximately 20%, now at a higher dollar amount of $10.65 billion, inclusive of both a campus target of $1.25 billion and an AI center target of $2.75 billion. For gross margin, a range is expected of approximately 62% to 64%, driven by customer mix. And for operating margin, an outlook of approximately 43% to 45%, allowing for investments in relation to achieving the strategic goals of Arista.”

“The breadth and depth of our customer interactions have never been stronger nor more exciting. In true Arista style, we remain pragmatic, yet are aware of the potential over the next few years.”

“There is no concern on our demand. I think the shipments and the revenue follows based on our supplies. So if we’re able to make the shipments, then the revenue, as you saw in Q2 went right — blew past any of our guidance, right? However, there are times we can’t ship everything despite the demand. And so you’re accordingly seeing that. I wouldn’t read too much into the quarterly variances. But I would say we feel — we have never felt more strongly about the demand aspect of this reflected in the continued commitment to 20% growth, even though the number keeps increasing from $8.75 billion to now $8.87 billion. So no change in demand, some variation in shipments.”

“We do have a mix of product [gross] margin where it’s significantly below 60% with our cloud and AI titans driving the volume and higher obviously, for the enterprise customers. The average of which, together with services, is yielding that number. So when the mix tilts heavily towards the cloud and AI, you can expect some pressure on our gross margins. But overall, I think we managed it very well the manufacturing team, now led by Todd does a fantastic job here. So again, the discipline and mix plays well together, but I don’t think it’s any change from prior years where when we have a heavy mix of AI and cloud, we feel it in our gross margins.”

[This quote is about how Nvidia’s GPU server racks bring both the compute and the networking gear into the rack and how Arista is poised to respond going forward]: “Andy Bechtolsheim is personally driving along with the hardware team a significant number of these racks. I think at any given time, we have 5 to 7 projects with different accelerator options. Obviously, NVIDIA is the gold standard today, but we can see 4 or 5 accelerators emerging in the next couple of years. Arista is being sought to bring all aspects, the cabling, the co-packaging, the power, the cooling as well as the connection to different XPU cartridges, if you may, as the network platform of choice in many of these cases. So we are involved in a lot of early designs. I think a lot of these designs will materialize as the standards for Ethernet are getting stronger and stronger. We now have a UEC spec. You heard me talk about the Scale-Up Ethernet spec, ESUN, where we can bring different work streams onto the same Ethernet headers, transport headers, data link layer, et cetera. So I think a lot of this will be underway in 2026 and really emerge in 2027 as Scale-Up Ethernet becomes a more important part of that. In terms of how we will gain more recognition or revenue, some of this will be not the classic OEM model. There may be more the blue box JDM model where we work with them on IP and have reference designs and offer them capabilities well beyond the network. And — but many of them will also entail selling the network as is in these racks.”

“We always, as you know, coexist with 2 other types of competitors. One is the bundling strategy with NVIDIA and the other is the white box. So we have not seen any significant changes in share up or down at the moment, it’s stable. Having said that, it’s also a massive market. And we think rising tide rises all boats and this boat is feeling pretty good.”

“Demand is greater than our ability to ship. Lead times on many of our components, including standard memory and chips and merchant silicon and everything, it’s nothing like 2022, but they have very long lead times ranging from 38 to 52 weeks. So we are coping with that. And you can see Chantelle is leaning in and making greater and greater purchase commitments, we wouldn’t do that without demand.”

“I don’t like the word deceleration [of revenue growth]. We’re talking about big, big numbers here, guys. And I’m committing to double-digit 20 and above percentage [revenue growth], don’t call it deceleration. Call it variability across quarters, and demand is great. I just don’t know whether it will land in ’26 or ’27.”

“I go back to kind of how we started [with guidance] in early ’25, maybe even ’24. Part of our style is to not assume 100% of everything hits to get to a number, and we’d like to leave ourselves with some optionality. And so we’re putting some goals for ourselves with the AI. We’re putting goals for ourselves with campus. It doesn’t mean we’re not focused on the rest. But I don’t think it’s the right approach to assume everything is going to be 100% and leave ourselves exposed, and we’ll continue to update as we see it.”

“AI and campus are going to grow and do great runs for us as it should because they are 2 very large TAMs. Whether it is Ken and Tyson driving the AI and cloud TAM or whether it’s Todd Nightingale driving the campus and these 2 are going to grow substantially in double digits, right? So to your point, it doesn’t leave the core business with a lot of opportunity. But that’s not to say it may be flattish, it may be grow. It’s to say that our customers are putting more attention there and that the existing business, which is already on very large numbers, will have lesser growth. We don’t yet know if it’s flattish or single digit or whether more will go to AI. We frankly can’t predict the mix this early in the game on 2026, but we think we’re in for a great ride in 2026.”

“Until now, the majority of how we’ve measured our AI success through our cloud and AI titans has been number of GPUs and how much are they installing and can we verify that the Ethernet network works. The majority of it [Arista’s AI business] to date has been scale out. First, I want to reflect that there are 3 big use cases sitting in front of us, scale up, scale out and scale across. Arista’s participation to date has largely been in scale out. So we’ve got 2 major use cases in addition to augmenting this, and that’s what makes the Etherlink portfolio that Ken described so eloquently so beautiful. Now how are these being built? Clearly, they’re being driven by large language models, tokens transformers, inference use cases, you name it all. So the influence is clearly coming from these players you named. But the way they are driving the infrastructure, and I can’t keep track of the gigawatts myself, it’s 10 gigawatts here, 10 there, 30 there. It’s adding up to a lot. But I can just tell you, no matter what it is, Arista has been looked at as a very important and relevant participant, especially right now in the scale out and scale across. We will participate in the scale up. It will take a little longer. Today, it is largely a set of proprietary technologies like NVLink or PCIe, and I think that will happen more in ’27. So that to say that as we get now confident about exceeding our $10 billion goal next year, we’re looking at our next goal of $15 billion [revenue] in the next few years. And I think AI will be a very large part of it, and so will be the companies you mentioned.”

“I’ll give you one example where they were just not getting their white box to work. These are AI mission-critical workloads. And we’re seeing a neocloud come right in with, in this case, non-NVIDIA GPUs, in fact, where they’re looking to deploy Arista with its excellent hardware. And at first, they wanted to do an open NOS, but now they are adopting a hybrid strategy where it’s not only an open NOS, but Ken’s EOS is coming to shine in its full glory in this use case. So in this case, I think it’s a Blue Box to start with, but it’s quickly going into a hybrid state of blue and branded EOS box.

The economics on that is not too different from our cloud and AI titans, generally speaking, although there will be scenarios, like you rightly mentioned, hasn’t yet come to play. But as we go to significant scale up volume, we expect more margin and economic capability coming together. In other words, the volume of these [Blue Boxes] things will be larger, the pressure on margins will be greater. So — but we will carefully have a mix of scale up, scale out and scale across to not affect the overall margins, but definitely take our fair share in that.”

“We were literally outside looking in at all these back-end networks that were largely being constructed by — with InfiniBand. We’ve seen a sea change, particularly this year, where obviously, more and more times we’re being invited to construct their 800 gig, last year was more 400 gig. And I think next year will be a combination of [ 800 gig and 1.60 terabits ] on the back end. The back end is putting pressure on the front end, which is why it’s getting more and more difficult for us to say, okay, what’s the back-end number that natively connects to GPUs and what is the front end.”

“I’d just like to point out that we’re seeing that Arista, I think, is the only successful vendor outside of China selling both front end and back end. And this is where our engineering alignment is so important because we can offer the customer a consistent solution across their entire infrastructure. I think this is a unique differentiator that will really help us succeed as these networks become more and more mainstream.”

“But look, we’re not religious. We jointly developed the DSF [disaggregated scheduled fabric] architecture with one of our leading cloud titans, Meta. And we’ve been selling the nonscheduled fabric for a very long time. So we’ve never been religious about this. And both are doing very, very well at our cloud titans and specifically the one we co-developed with.”

“You are right to point out that because we may not have the EOS layer [in certain Blue Boxes], we will take a lower margin on that. And that’s factored into our 2026 guide and mix. And we think the combination of the blue box and the EOS branded box, if I can call it that, will continue to help us thrive with a profitable and high-growing business.”

“In terms of the front end and back-end [of AI networks] converging, this is truly advantageous to us because the front end requires a massive number of features. It’s incredibly mission-critical and supports a whole variety of applications, not just the straightforward of demanding communication patterns of the AI back end. So we see that the — our ability to tackle both of them effectively is a significant source of strength and a real differentiator and something that’s not easy for competitors to replicate. If you look at NVIDIA, for example, the sales volume is small in the front end and Cisco is small in the back end. And so I think we’ll see that kind of convergence being beneficial to us.”

“And on the blue box, I’m not sure we model 3 to 5 years. But if I had to venture what I think the evolution of the blue box will be, I think it will be more significant in the scale up use cases where there’s a higher dependency on the strength of our hardware and our NetDI capability and a lower requirement for software.”

“With the cloud, we rarely got to talk about gigawatts and beyond. So much of them were multi-megawatts. So these are newly constructed AI build-outs as opposed to the traditional CPU or storage-driven cloud build-outs. Of course, they will have refresh too. But frankly, they’re not getting the attention. All the attention is going to the new build-outs for AI.”