Arista Networks Inc (ANET) Earnings Call Transcript & Summary

Good afternoon, everyone. Thank you for joining us today at Arista's Analyst Day event. Thank you. Over the course of this afternoon, we hope to share with you our vision and our strategy and our opportunity to drive innovation in this area of networking. My name is Rudolph Araujo, and I head Investor advocacy here at Arista. I want to thank you all for coming. I know some of you have come from far and wide. Today, you will hear from a number of our executives as we cover a number of key topics that we've prepared for you. We'll start off with Jayshree talking about our centers of data strategy, where we are on our Arista 2.0 momentum and where we're headed. We'll then have Ken and Todd talk about AI for networking and how we are uniquely positioned to bring the reliability of the cloud to the enterprise. We'll then take a quick break, allowing you to stretch your legs, grab some coffee. And then we'll head into talking about AI Ethernet fabrics and the power of Ethernet for AI and cloud networking with Andy and Hugh. We'll finally wrap it up with Chantelle talking about our financials. And then we'll have a Q&A panel where we'll have a number of our leaders up here to answer your questions. Before we go too much further, though, I did want to read our forward-looking statement. During the course of this investor event, Arista Networks' management will make forward-looking statements, including those relating to our financial outlook for the 2025 fiscal year, our longer-term business model and financial targets for 2026 and beyond, including revenue targets for certain market segments in 2026, our total addressable market and strategy for addressing these market opportunities, including the AI market, our drivers for growth and diversification, our investment and capital allocation strategy, EOS' architectural advantages and future evolution, product innovation, customer demand trends, tariffs and trade restrictions, supply chain constraints, component costs, manufacturing output, inventory management and inflationary pressures on our business, lead times, working capital optimization and the benefits of acquisitions, which are subject to the risks and uncertainties that we discuss in detail in our documents filed with the SEC specifically in our most recent Form 10-Q and Form 10-K and which could cause actual results to differ materially from those anticipated by these statements. These forward-looking statements apply as of today, and you should not rely on them representing our views in the future. We undertake no obligation to update these statements after this event. Before we get to the rest of this agenda, I did want to take a moment to acknowledge all the lives impacted by 9/11. It is now my great privilege to welcome our Chairperson and Chief Executive Officer, Jayshree Ullal, to the stage.

Thank you. I'm already on stage, Bill. So first of all, a warm welcome to all of you. It's always great to see you. And for those of you who traveled either from the Goldman Sachs Conference or from the East Coast, really appreciate it. I know it's so easy to just sit behind and watch this online. But for those of you who are doing that, we appreciate it as well. Today, I'm most proud to not only present our strategy, but to share with you the depth of our leadership team and all how long we've come. I talked to a lot of you who dealt with Arista back before we were public or when we first came to this new building and we were barely having any furnishings, and we had our first Analyst Day. And I think you'd all agree that we've come a long, long way. So today, I'd like to share some of the how we've come a long, long way. Some of the numbers and the market momentum and the revolutions and evolutions of Arista. What does centers of data and Arista 2.0 mean to us. Of course, no presentation would be complete if I didn't talk about AI centers and what we're doing in networking. Some of you have asked what's the blue box. And so today, I'll make a more concerted effort to share with you what that while we will coexist peacefully with the white box that Arista is here to stay with a highly differentiated strategy and architecture with our blue box. And finally, what you're probably most waiting for, I'll give you a preview and then, of course, Chantelle, our CFO, will share a lot more detail. Let's quickly look at the Arista journey over the last decade since at least the beginning of when we started shipping products in 2009. Many of you may know, we pioneered the leaf-spine. We started out as a low-latency company. And I still remember my early conversations with Nick Lippis. And I said, hey, everybody talks about access, aggregation core, but we're seeing a much more flat fat topology. And what do you think of that? And we had a lot of discussion back then actually on InfiniBand, which had this class architecture with leaf-spine. And I'm sure you'd all agree, Nick, especially you that Arista was the first to pioneer this leaf-spine architecture and bring this active end way. We started out with 4 and 8-way. Today, we are at over 576-way, which shows the massive ratings and intensity of bandwidth, telemetry, automation that we pioneered. And so this has obviously been the core of our existence and continues to be as we evolve in the next direction. If you look at how we went then in the 2015 to 2020 era, I would say we started then building a universal cloud network, UCN, where we moved from places in the network, which were different silos where each one, whether it was campus or data center or wide area routing or service provider, had their own operating system and their own network. So wouldn't it be great if we built a leaf-spine architecture where they had a universal spine, active, active and then any kind of leaf. It could be a compute leaf, a storage leaf, a campus leaf, of course, a data center architecture or service provider routing and a segue into that and of course, cloud. And this is something we have really gone from sort of the innovations on best-of-breed product to what I would call best-of-breed categories. And then finally, now, I think we are at this point where we're really building a platform. So not only are we looking to be the best data center company, but really the centers of data. And what does that really mean? What's the play on words? That networking is moving to something way more strategic than just connectivity. And today is at the epicenter of almost everything we do from a consumer to an enterprise, to a cloud to, of course, everything we do in AI. And we'll, of course, share a lot more of what we mean by that. We've got some old timers, present company included. So Ken and Andy started the company. So they're over 20 years with us. Thank you, both of you. And although I think Ken is glad that I don't anymore share the office with him because I spent too much time on the phone. But having said that, I think you're seeing a lot of the depth of our company still being very much in engineering and technology. We're a company built by engineers, often four engineers, which are our customers. But we've added some new leaders as well. I don't know if John McCool is here, but he's taken a nice break, but he will be returning back to help us, especially with some of our new acquisitions, in particular, VeloCloud. Mark Foss, Ashwin, Chris, Chris and Chris, as we call them, have all been with the company anywhere from 12 years to 18 years. So a lot of longevity. But I want to point your attention to some new faces. Chantelle, I'm not sure I can call you a new face anymore because you've been with us here in our second year, but it's great to have a great business partner. But on the right, Todd Nightingale, who many of you may have met at the Goldman Conference. It's an absolute pleasure to welcome him as our Co-President, along with Ken Duda. And Tyson Lamoreaux, who's now -- this is his first week, so be kind to him in the question-and-answer session, and he'll be heading up one of our largest and biggest strategic opportunities in cloud and AI. Tyson built the first cloud at AWS and also built the first AI network in a sovereign AI company. So it's really great to have this lineup here. And behind these leaders are an awesome number of folks as well as we probably trend to about 5,000 employees this year, maybe we'll cross it. So when I talk about these three categories of leaf-spine, places in the network to places in the cloud and now to centers of data, you can see there's been some pretty exciting numbers with it. As we built these best-of-breed products, we went from zero to crossing our first $1 billion in 2016. And then we went into the second and third billion right around the pandemic. And we did a little bit of the struggling in the $2 billion range around the pandemic and then in the supply chain crisis. But you can see the last three, four years, has just been absolutely stunning. And we've had that break away as we have gone from best-of-breed products to layering the right platform capability in the enterprise in the service providers, in the neo cloud specialty providers and of course, then in our very close titan customers, our AI and cloud titans. So we've committed to you that we will do $8.75 billion this year, a little earlier than we would have anticipated. But barring any execution on Todd and Mike Kappus on the shipment side, we feel pretty good about this number. And today, we'll share with you, of course, where we're heading in '26. But high performance also requires a measure on market share. These trends are still in '24, but I just got hot off the press that in 800-gig this quarter in Q2 2025, Arista is now emerging #1 despite the fact that there's a lot of Chinese competition and bundled competition and white box competition. So -- but this is a category of how we do with not only 800-gig, but 10, 40, 100, 200, 400-gig and including some 8. And we were always heading in the direction of overlapping with the dominant players on the port section, but it's really good to see how we are doing that both in switching ports and dollars. Probably never in the history of any incumbent company have you seen that kind of trend. It's been slow and steady, but it's, I believe, here to stay, witnessed by our real maniacal focus on that best-of-breed technology, both at the hardware level and at the software level. That's reflected also in our high-speed ports, as I said. There's really three sectors now. There's Arista, there's rest of market. This is where we would put in white box or NVIDIA or other customers. And then, of course, the dominant player of the switching and overall is still Cisco. But in high-performance data center, you can see that trends are really changing and changing very quickly. Now I want to kind of go back in time because once you get as old as I am, you end up being a historian and you'd tell stories about the past. I promise to tell stories about the present and future, too, but let's go back to the past. When Andy recruited me to this company in 2008, he said, do you think we can do -- and we were, of course, as close to zero as you can be. And he said, do you think we can do our first $100 million? And they said, well, we're at zero, and we are building a switching company. There's a lot of switching companies. You guys have to remember the playing field there. There was BLADE Networks, it was Force10, Extreme, Foundry, you can name them all, right? And so we -- in fact, it was commonly called Cisco and the seven dwarfs, if you all remember, right? So I used to take great pride in saying we were the tallest of the midgets. Now I don't want to talk about our height. I want to talk about our relevance, right? And I want to -- this slide basically describes our relevance because we were greatly inspired by an RFP we saw where folks like Amazon and Google were already building their cloud. But why is it a single vendor couldn't build what the cloud required, which was give me a nonblocking architecture. At that time, give it to me for 1 gigabit or 10 gigabit and give it to me at a very effective price per port, which at that time was $100 a port. As we have stepped to 40 gig, 50 gig, 400, 800, the $100 price is actually more now dwarfed by optics and many things. But the fundamental essence of scalability, predictable latency, highly non-blocking high radix architecture hasn't changed. So you can see here the switch capacity has just taken off from a terabit of switching to today, we talk about 200 terabits and going further. And then the programmability and bandwidth and latency on the y-axis requirements keep changing. So this RFP that inspired us greatly continues to inspire us just at higher and faster speeds. And it's important, I think, when you're a company to still have that substantive DNA on what you're built off and what you're made of. The last few years, although we're a great software company, we built great hardware. But hardware wasn't recognized, and now it's back. Chips and hardware are back. And our ability to bring the best of silicon, the best of platforms and the best of software is what's going to really drive and dictate how networks are built. Now I talked about this universal spine. And once again, it's so important to understand that Arista not only pioneered the leap-spine architecture, but actually pioneered the spine platforms. Many of you may remember the 7500, which at that time was the best footprint, the best power with the maximum density of 10 gigabit. We went from a 10-gig spine to in 2014 when we went public from a 10 terabit to 30 terabits. And then as we added more capability, we were pushing the envelope of compute and storage with the 7800. And today, the 7800R4, as we call it, is the platform for AI spines, compute, storage and this thing we call AI accelerators or GPUs. This progression and having these four generations of capability have been important. But by the way, every step of the way, we've had to develop more and more software features in EOS to make this hump. We couldn't take the same platform and just say, all right, it's good for the cloud, it's good for the rest. And so the Etherlink portfolio, as you'll hear from Hugh and Andy, and building very optimized features for the back end and front end has been fundamental. Now it's all about, at the end of the day, going from connectivity to managing data. And there really are different forms of data you're trying to manage. First, you're trying to make sure the data is highly available. It'd be great to just have two of everything, but the ability for us to do smart system upgrade at the leaf level, at the spine level, at the path level, at the supervisor level and really provide a copy of everything with one goes down is baked right into our software with our automatic fault containment and repair to our supervisor, to our line cards to the switches themselves. So that was the first goal. The second was not just zero-touch automation, but day zero, day one, day two, fully orchestrate. And I still remember when I was with one of the premium data center customers, and we were able to bring that time down from six hours to 20 minutes. That is still not doable with -- and if you don't do that, you're going to put 100 people to do that. So making the software reduce that kind of automation time has been even more important in the campus as you start to have this proliferation of devices, be it IoT, devices of different kinds, MAC addresses, no MAC addresses, are they users? Are they devices? Are they both? So that day zero automation that we stressed about with the data center is now coming to play to automate and provision and orchestrate with CloudVision across them. And finally, interact. Used to be we worried more about media and entertainment workloads and workflows. But today, we have to worry even more with our Etherlink portfolio about AI and how compute-intensive they are, how they have short-lived flows and yet they're very bursty. So the fidelity, the diversity and the interaction of these flows as you do the cycle time becomes -- the job completion time is fundamental, but so is the time to the first job completion, as Hugh will share with you. So our products, we now have over 50 products in our portfolio, a very complete one in data center, a very complete one in the routed WAN as well as in the campus now with the Velo heading to the branch center. The new ones that we added, particularly in the last year was the AI center Etherlink portfolio. And then all of the software services for segmentation, network detection, network identity, security, encryption as well as observability that we add into the network and they're not afterthoughts, but they're really part of this entire product portfolio. So when you manifest this into an architecture, and this is how really our centers of data come in, you have the leaf-spine architecture. The spine may connect to the WAN, the span may connect to data center interconnects. But now you have a leaf architecture with a single EOS that can have different personalities for the campus, for the data center to bring your wired and wireless together. And this underlay is not the only thing. Now we have a whole software EOS stack that's automating and orchestrating a suite of features. And here are examples of them, the zero-touch, the location identity with the platform we came out with call AGNI, the Arista Guardian for Network Identity, macro segmentation and NDR products our virtual stack that we introduced recently, where we said stackable should not be proprietary. It should be based on the same IP and Ethernet capabilities, so you didn't have to put a different physical capability. And you can go active, active and still bring a stack of 100 or 200 devices. And then finally, of course, the VeloCloud SD-WAN. On that overlay, we manage it all through CloudVision by bringing the streaming telemetry of every single switch that you can correlate, visualize, architect, manage the state of. And then finally, you'll hear today more about not just how Arista is great in networking for AI, which is our Etherlink products, but how we are using AI to diagnose, troubleshoot, correlate and bringing those LLMs to work into our network itself from Ken. So I know agentic AI is a buzzword. And today, our AI centers are largely in this phase, generative AI training of extremely large language models. This is the back end that's fueling our front end. But as we go forward, we see many more phases. We see bringing that AI into the edge with inference and really having agentic AI with all of the applications, and we'll be showing a demo shortly with Fred. So I think many people ask, is AI a bubble? Is it here to stay? I think AI is that the killer application that starts out like a mainframe in training and then permeates the entire network, including the enterprise over time. AI is also affecting why we did the VeloCloud acquisition. It is because of Agentic AI that we saw the pressure to put more bandwidth and put more capability into the branch with an SD-WAN. And so Arista is now bringing in Velo and our campus products, the campus center and the branch center for ease of use, for security, for quality of service into the same umbrella because a large campus and a mini me campus needs the same attributes, but they may need an MSP or a WAN across it, and that's, in fact, the goal. It gives me great pride to say that we're also heading in the right direction in many of the Gartner Magic Quadrants. Of course, undoubtedly, you would not be surprised to see that we are a leader in the leadership quadrant in the data center. Although ironically, it took us five years after we really were a leader in market share to get there, by the way. For the longest time, we were still an aspiring MQ player. And that's, in fact, where we are in campus, still aspiring even though we think we've achieved a lot of premium wins. It's only fitting to speak about 9/11 and all the things that got destroyed in New York. I'm reflecting now on a very major campus win where this brand-new building has come in that Arista is the smack in the middle of bringing together in that post 9/11 era. I can't give the name out, but these -- we're winning some of the most stunning premium enterprise names for especially wired and in many cases, also bringing in the Wi-Fi. And I believe our $750 million this year is the beginning of many years to come, particularly with the help of Todd. And finally, this is the Magic Quadrant on the SD-WAN, which has since moved to SASE. But a word on this. I think independent of how you look at the quadrant, you still need really good WAN to connect into your branch and that's where Arista will stay focused. We will work with market leaders like Zscaler and Palo Alto to do the cloud security overlay on that. So we won't pretend to be something we're not and fully take advantage of the Velo capability to be the best wired, wireless and WAN in a box in a branch. Now I said we will talk a little bit about AI networking and scale up and scale out Ethernet. And I wanted to quickly show you an anatomy of how this looks. So first of all, let's start at the bottom. You all know that we collect a bunch of GPUs as I'm starting to call more XPUs because they can be any accelerators. And these are the ones doing the massive crunching while you have a CPU assist for processing or indexing, et cetera. They typically today go through a PCIe switch and then they go to either an RDMA NIC or UEC NICs that are coming in for more packet steering and RDMA is becoming, in my view, over time, more and more optional. So where does Arista play? Arista's largest play today is as we bring this inerts of the AI rack to the outer world, where we connect with the back-end leaf-spine architecture. This is where we are strong. Now we do see the opportunity as this native scale-up starts moving from NVLink, there are many options today. There's a standard going on for UALink. There's a standard going on for scale-up Ethernet transport. And there is, of course, NVIDIA's own proprietary bus. The first thing to remember is this is a very fast, high ratings, low latency, almost simple network or bus. So the way Arista will play with this is enable more scale-up racks. All the EOS software that we do really gets ignited in the scale-out. But here it will be light fast -- lightweight, fast and move the packets and bits as much as you can. We're big fans, you can imagine, of Ethernet. I have never bet against on an Ethernet technology, and I'm not about to start now. It may take time, but we strongly believe the scale-up Ethernet will come together, perhaps under the Ultra Ethernet Consortium or as a spec. And products that are scale-up Ethernet is very possible in the 2027 time frame as the standards emerge. Now where does all this fit? A scale-up rack typically looks at 100 of GPUs or XPUs. Maybe it can go to 256, maybe it can go to 500, but that's the ratings. So it's really within the rack. And Arista is working with a number of our customers to make these scale-up racks possible even today. A scale-out rack, and this is where we're really smack in the middle of right now, can go from 1,000 to 10,000 XPUs. And then if you really want to build football fields of 50,000 to 100,000 GPUs, as you've seen me talk about the few that do, that's when you need a 2-tier, not only a scale up, but you also need a scale-out and very often a scale across because you're going across factories, across data centers to make this possible. So read racks where we could have co-packaged copper and really fast low-latency switches in the first category, read leaf spine in the middle, which we're doing and read scale across is where, in fact, the combination of our 7800 chassis and many types of leaves will really help us succeed. I'm going to skip a couple of these slides in the interest of time, and I know Hugh and Andy will cover it. But I do want to cover one thing that although we all get access to the same silicon, power is king and queen. Arista builds very optimized drivers, some of the best hardware designs. And you can see here, we can translate to $3.5 million a year or $15 million savings over four years, which is stunning. You can literally buy your switches free at that point if you get the right power, right? So Arista's maniacal focus, even on the same platform from the same chip vendor, Arista can design it better, and that's a big deal. Same thing with 800-gig linear drive optics, once again, we were able to save millions of dollars per year because you're getting rid of that DSP and you're co-locating the switch and the optics as closest you can together, the electrical and the optical together, and this is becoming very, very popular at 800-gig SerDes. I would be remiss if I didn't talk about the features that make a U.S. hub. We are working on a suite of features, not only on the back end that you've heard of like cluster load balancing, high availability, dynamic congestion control, but the front end is now getting stressed. And we need multi-tenant scale. We need encryption scale. We need streaming telemetry, a feature we call stands to make all this work. So the back end is putting pressure on the front end to make the AI network fully work. And therefore, the Ethernet portfolio is not just about high-speed hardware, but very rich features. Now most of you have been asking me about what about the white box Jayshree. You must be competing against them and you must be wondering how to challenge -- how to deal with that challenge. First, I want to take you to history. We have dealt with white box since the beginning of time. And of course, they'll exist. ODMs, JDMs, cheap, cost-effective platforms at 10% gross margin or whatever it is, may be good enough for someone. That's not Arista. Arista is trying to provide value and value means good price, good performance and good capability, right? Having said that, we have worked on co-development. So we don't look at this as do we build? Do we buy? We often work with our customers on a build and buy strategy so we can codevelop with them and make them better. And we've done that with SONiC. We've done that with FBOSS, and we've done that you can see across three platforms already, the 7388, the Meta X-Pack as well as the DS-7700 that was an Ethernet platform. To us, our strong partnerships are key to rapid deployment of our products and rapid co-development of their products. Both are very important for us. So why would they work with us if they can just go to a white box. So let's -- to understand that, I need to share with you what is the blue box philosophy, first of all. You have a NetDL foundation, that's our EOS stack. You typically on the hardware, have a switch abstraction interface. I'm not telling you anything new right now. These are all the things we have. You have validation tools that come from us or come from the customer. Often the customer deploys hundreds of engineers to do that or we do that for them in the enterprise. And we have a set of deployment guides and tools in the enterprise for us -- for our customers. What's missing, the missing layer that makes blue box really hum is this diagnostics layer that's getting more and more difficult between your software stack and your hardware. It's not just the drivers and the switch abstraction interface. It is the ability to create an environment and a foundation so that we can run any NOS, an open NOS like SONIC or FBOSS or an open NOS like Arista EOS. And it's this diagnostic layer that is fundamental and strategic to creating the blue box. So what is the blue box? It is a suite of features that we call net DI, network diagnostics infrastructure that allows us to work between that highly complex multilayer hardware we have and the EOS software layer to give enough troubleshooting, validation, signal integrity, L1 events for optics cables, L0 events for passive components like flash and memory and power supplies, deployment checkers, control trackers for our manufacturing teams and bring this all at massive scale. So it's a suite of functions. Actually, we've been working on for many years. It's a welcome secret. And we've had just on EOS alone, 30,000 mans, we're running at any given time, 300,000 diagnostic tests a day, and this is a large part of it. If you want to build anything at scale and not just build a throwaway box, the blue box is fundamentally enhanced by this firmware and diags layer. It's a huge work of art between our hardware, our troubleshooting teams, our firmware and our EOS layer. So we're very proud of it, and you're going to hear more of it. And so think of it as if you have a product, which is our hardware and if you have a NOS, open or closed and you have AIOps on top of it, the green layer is all the stuff behind the scenes that we are doing between silicon, power controllers, FPGAs, booting up things, control plane, data plane, management plane, how do they talk to each other at crazy speeds, high-speed SerDes and making that all work. In fact, as NetDL has its own state and base, this net DI has its own little mini one too to work with that NetDL. So it's a mini me of NetDL for the hardware layer and the L1 functions we have to do. So it's very foundational. And the Arista blue box complements the work we've done with State and net DI to really bring these kind of functions, tools, diagnostics, signal integrity, quality control, secure boot loaders, passive flash component management, active cables, optics, loop back management, and finally, dashboards for deployment. It's amazing. It's all the secret behind what we do in switching that most of you don't get to see, but most of us get to work on. So when you put this all together, I see Arista Advantage. You have the NetDL architecture, you have the diagnostics now, the NetDL coming in. You have the actual hardware that you see that's running all of this. And then increasingly with CloudVision, you'll see more and more AI-driven predictive tools that go beyond the topology and telemetry and observability we're all doing already to have natural language processing and queries for different types of events. I'm so excited about this that finally, I want to share with you that we are in a rare breed of companies. When I put this slide together, we were at $150 billion market cap. But whatever we are, it took us a record time to achieve our first $1 billion in 2016. We went public in 2014. And I think it's going to take us record time to achieve our first $10 billion. So our commitment to you for next year is $10.5 billion in 2026. which should be a 20% growth on extremely large numbers. Now what does this consist of? And why? Well, we think the TAM for this in '29 is north of $100 billion. So we better capture our fair share and keep growing from that point on. It would only be fair with all the innovation and technology we're doing. As part of that number, there's going to be two very fast-growing markets. The campus, which includes the branch, which we think is going to grow due to the addition of VeloCloud at 60%, and we're aiming to go from that $750 million to $800 million number this year to $1.25 billion. So try to add another $500 million there. Ambitious goal, and we're signed up to it. And then, of course, the AI market, which, as I described to you, has to now include the back end and front end. We'll be converging to grow that at anywhere from 60% to 80%. If we end the year at $1.5 billion, it's an 80% growth. If we end the year a little higher, then it's a 60% growth, but somewhere in the range that we're looking to achieve this number. So out of our $10.6 billion, two fast-growing markets will definitely contribute to achieve this. And I'm very excited because I think for the first time in a long, long time, we're seeing -- we're always worried about what's the TAM, what's the market, what's the acquisition we have to do. We're seeing something that's sustainable for multiple years. So -- and we're doing that with that foundational technology we have with net DI, with NetDL, with AVA. And for those of you who watch the video at the break, we're also doing it with a suite of partners that goes beyond NVIDIA. Last year, we announced the NVIDIA partnership, and no doubt, they're a market leader in AI. But it's going to take a whole ecosystem of innovators to do it, not just ourselves, but working with the LLM modules, working with other GPUs working with storage. So I would like to invite Fred to show you a little demo of the Agent AI and how...

Thanks, Jayshree. So what I'm going to show you here is a demonstration of our agentic AVD and how it can simplify workflows for our operators. So what I'm going to do is ask the agent here to add a new vast storage endpoint to our network. And the agent is smart enough to know what are our best practices, what are the features we need to turn on and make this storage network flow really well. But additionally, we've partnered up with VAs to be able to call out to their APIs and configure the storage device as well. So not only am I setting up the network, but I'm also going to set up my storage endpoint. We can do the same thing with another partner. So I'm going to have this also reach out to Pure Storage and create a virtual IP, so I can also set up a whole Pure Storage setup and configure the network as well. Now building this on top of our AVD framework gives us two really big advantages. The first one is that everything is built off of a data model. So this helps us constrain the LLM and reduce the chances of hallucinations when we're generating these configs. The second thing you can see here is that as it generates the configs, it also generates network tests. So that's sort of an extra layer of safeguards there that if we do, do something wrong, we'll catch it once we actually get into deployment. So we'll tell the agent now to go ahead and deploy those configs and then run the tests. So we've now pushed these changes out to our network, and we get our test results back saying that, yes, everything went well, so the network is fully deployed. And just to kind of double check, we can check out the vast dashboard, and we can see that our connection has been established. So very quickly, we've gone and deployed an entire new storage network and taken what usually takes maybe weeks or months, knock that down to hours and minutes. And what's more is if you're a storage guy or someone who's not necessarily a network expert, you're able to now configure and deploy things on the network without necessarily having all that expertise. Thanks.

So you see, I remember stressing over this when we were doing fiber channel over Ethernet and storage emulation, and it was days. It wasn't even hours. So thank you, Fred. This is a real demonstration of what we can do with AVD, which is our Arista Validated Designs. Before I end and transfer it over to our new President, I just want to thank you all. I think it's been an exciting journey, and I'm a worry what. I always worry about the next quarter and the next year. But I think what's going on here at Arista and as an industry has been transformational and will continue to be for many years to come. Thank you.

Thank you all very much for coming. Really appreciate the chance to talk to you here. And what I'd like to focus on is AVA, our autonomous virtual assist and what we think can be achieved in AIOps, AI for networking. And of course, AVA is built on the foundation. So I'm going to start by talking about what that foundation is. Many of you have seen this before. But I'm going to say it all again because the foundations really matter. This is a structural competitive advantage right here in the architecture of the EOS stack. In the EOS stack, on top of the switch hardware, we have the net DI layer of diagnostic infrastructure that Jayshree told you all about. I'm not going to go into so much detail on that in this talk. But it's a lot of stuff, high-speed signaling, signal integrity, dealing with power and cooling, all the different scenarios, making sure the switch is really going to work under lots of operating conditions, dealing with single event upset when subatomic particles from outer space come in and hit the switch ASIC. Is your switch going to survive that? All of the sort of low-level hardware integrity validation, this layer of software is a major source of value of the Arista blue box platform regardless of what operating system is running on top. But in the EOS stack, naturally, on top of net DI, we run EOS, one operating system for all the use cases across the infrastructure. EOS feeds into our network data lake on top of which we run CloudVision and AVA, the Autonomous Virtual Assist. Having this consistent architecture across all domains of the network is a major competitive advantage for us. Because one OS, one architecture is just better for the customer. This infrastructure has got to work. Reliability is critical. And one of the main enemies of reliability and infrastructure is too many different configurations, too many different versions, too much complexity from having all that variance across your infrastructure. I've talked to tech leaders and networking infrastructure leaders, one of them at one of the largest banks in the country told me that from our competitor, he's running more than 200 different operating system versions. He has to track all of this, all the differences between them, the little flux and differences in the protocols, the bugs, the security vulnerabilities, what a nightmare. With EOS, you have OS to learn, one image to qualify. And this is actually really important, one API to automate against because one problem you have as an operator, you have all these different operating systems. They're all a little different and your automation systems have to cope with all those differences. But if you make it all the same, you do it right once, you do it right everywhere. It's just easier to configure. There's fewer mistakes. You get more reliable operations at scale. You can address more use cases this way. So having one OS across the whole domain is just better for the customer. But it's also better for us. Can you imagine if you're a software engineer, dealing with that many of different software versions, how do you test all of this? How do you make sure it all works properly? At Arista, we have one image to test. We run tens of thousands of tests every day, tests running fully autonomously. That means software testing software, 24/7. We can all go on vacation. The software is still being tested. This -- we're testing this against every hardware platform, every branch of code, all of our older releases, the work in progress, new features being developed, all being autonomously tested continuously. And this all comes down to this principle that is the development team at Arista that's responsible for quality. We don't say, oh, yes, we write the code, but these other guys, the QA guys, they're responsible for making sure our code works. No. The software developers are responsible. And when you give people both the mandate and the responsibility to take ownership of the quality of their code, you get better code, and that's what we've done. Sometimes people ask me, what's your evidence that EOS quality is actually that much better than your competitors? And I'd like to offer you a model here, which is the model of the iceberg of bugs. If you imagine bugs are organized into a gigantic iceberg, floating in the water. Poking out above the surface are the CVEs. It's like the tip of the iceberg. These are -- remember, not every bug is a security vulnerability, but every security vulnerability is a bug. The CVEs are simply the visible subset of bugs that are publicly reported, publicly categorized and classified. And so if you look at the size of the CVEs, maybe that says something about the size of the whole iceberg of bugs. Now if you look at the public databases, you'll see that Arista EOS has dramatically fewer security vulnerabilities than other network operating systems. Tip of our iceberg is 1/10 the size. What does that say about the size of the overall iceberg? Like I wish I knew how big their iceberg was. Unfortunately, it's not publicly disclosed, but I believe it's probably about 10x bigger. We've got 1/10 of CVEs, we probably got 1/10 the bugs of other types. And certainly, from talking to customers and talking to the field about people's experience, I think that bears this out as well. And we can talk about fancy features all day long. But at the end of the day, what the customer cares about the most is, is my network working? So that's what we care about the most as well, which is why quality is always our highest priority. Let me go back to the EOS stack and the software foundation. On top of the switch, we run a standard Linux release. Alma 9 right now is the release we're on. And on top of Alma 9, we run NetDB. This is an Arista database that contains all of the state of the switch, everything from hardware attributes, power supply voltages temperature sensors, fan speeds, control plane stuff, what's going on with BGP and MAC learning and IGMP snooping and management plane activity as well, network authentication and that sort of thing. All of that state of how the network is running is stored in NetDB across all the switches. And in about 2014, we suddenly realized, wait a minute, we've got all this information in the switches. What if we stream that all out of the switches continuously, generating a stream of updates indicating the state of each device into a common scale-out database, which we call NetDL. All of the updates stream out and NetDL winds up with a time series, a historical record of every state in the network across all of the devices in one place. And this state foundation is so valuable. Having all of your state in a common representation in a common infrastructure layer, it enables you for provisioning, security, compliance, telemetry, and of course, also for AI. NetDL contains actually multiple types of state. There's the low-level state about the switches, interfaces, counters, the stuff which was actually streamed out, the flows, the events on the switch, link traps when links go up or down or temperature events or things like that. But NetDL also maps all of those lower-level concepts to higher-level concepts, users, devices, applications, services and incidents. These are higher-level ideas that come from the ability to observe across the whole network and also bring in data from other sources, including vCenter, OpenShift, Kubernetes, DNS, ToS header inspection, OAuth, radio servers. All this information comes into NetDL along with information from the switches from whether -- regardless of where the switches are, they could be virtual switches running in the public cloud. They could be switches on your campus or in your data centers or all the way across the WAN service provider core across -- this is, again, the advantage of having a common architecture across every domain of the network. You bring all this information from all these different places into a shared common database that then supports end-to-end visibility, end-to-end uniform provisioning, a consistent treatment of network upgrades and software updates of security incidents, CVE handling, all of those things are unified. And of course, that same NetDL is the foundation for AVA, our autonomous virtual assist. Now I'd like to finally talk a little more about AVA. So AVA, just from the name, the -- as are very important. AVA is not a chatbot, okay? Chatbots just sit there and wait around. You type in your question, get back your answer, they go off and do something else. No. AVA is an autonomous agent running all the time in your network, always watching, always trying to understand what's normal, what's common? Why is that happening? How does this compare to that? Looking at events, trying to figure out what's important, what do I need to, what's changing? What might I need to alert the operator about? So AVA is autonomous in that respect. But also, I think very importantly, the second A, AVA is an assistant. The media talks endlessly about how AI is coming for all of our jobs. We maybe someday. But I don't believe this current generation of technology is taking away any network engineer jobs, not yet. It's not ready. What it is ready to be is a fantastic assistant that can help the network engineer, help the network operator deal with the complexity of their network, deal with the all the different tools that are available and all the -- just how hard it is to operate in the modern environment. So if you look inside AVA a little bit, AVA is constructed in layers at the lowest layer is NetDL, of course, NetDL is the state foundation for AVA. On top of NetDL, we run the AVA run time, which includes LLMs, standard off the shelf, a context engine that builds the prompting for the LLMs based on context elements that come from observing what's happening in the environment. A tool manager that helps manage all of the different things that AVA can do to get more information or even make changes to the network, talking to telemetry systems, obviously, CloudVision telemetry, but also third-party systems, policy and safety engine for the obvious reasons. And finally, an MCP client, so we can connect to arbitrary MCP servers within an AI environment. And then on top of the AVA run-time, we build specific agents for specific functions, state machines and prompting engines, history reducers that -- for each of the different areas of network operations. So ask AVA, this is basic question and answer based on our knowledge of what's happening in our documentation in the user's network from bug database entries, CVEs and tech support history. All of that is available to ask AVA. Monitoring, of course, always watching, like I said, AVA provisioning is an assistant for making configuration changes to the network. I think you saw in Fred's demo, an example of that. And finally, AVA troubleshooting. When things do go wrong, how do you put these pieces together, figure out what's happening. AVA troubleshooting is an assistant to help with responding to incidents of network issues. And so in summary, we have a multi-domain operating model across the entire estate from the cloud to the campus, everything in between with a single consistent OS, consistent state management, giving the customer consistent operations as a foundation for their environment. And now what I'd like to do is turn this over to Todd to talk to you about how we harness this architectural foundation and the shared elements across the whole estate and really focus in on our strategy around campus. So please, I would like to welcome Todd to the stage.

I appreciate it. This is an amazing event. My name is Todd Nightingale. I'm the COO here at Arista. I'm new, and I cannot tell you how excited I am to be here, of course, at Investor Day here, but most importantly, here at Arista. It's been a phenomenal experience to get to look across the whole business from the technology to the go-to-market and of course, the operations. One of the most amazing opportunities I think we have in front of us is the campus TAM. And I'm not sure we always get to talk about it as much as we'd like to, but it is really -- it's an enormous and profitable business for us. And we are, in so many ways, just getting started. This is a part of the network that for years was the largest spend across the industry. And today, more than ever, it is ripe for modernization and truly Arista's flavor of modernization, the kind of differentiation that we provide. There's a ton going on. There's an explosion of devices, thanks to IoT, but also a high diversity of smarter and smarter devices with more and more intense networking needs. There's phenomenal focus right now on spend and OpEx across every industry, and that's putting pressure on NetOps and the efficiency. And there is a real, real acceleration in the attack velocity. And when we're talking about CVEs, we have to talk about how we secure these networks and how we deliver Zero Trust networking. But all of this stuff, it really adds up to a need in the market for the kind of innovation that Arista has always delivered. There is no longer really such a thing as a network that is not mission-critical. We used to think of mission-critical networks for military sites and Tier 1 hospitals. But I assure you, every hotel whose WiFi went down when guests need it, think their network is mission-critical. Every retailer who couldn't take -- who couldn't process a payment believes their network is mission-critical. Every school during testing week, every university, every manufacturing plant, building for surge, every network, it's 2025. Every network is mission-critical. And now more than ever is a time for us to take the reliability that Arista has always been known for, the foundation of EOS that's made that possible and bring it to the campus network. This isn't new. Arista has been innovating in this area and pushing towards this surge in campus for years, and this has been our strategy. Delivering that truly always-on network. truly bringing mission-critical always-on networks to the entire industry. That is our focus in so many ways, that is who we are. And in the campus, that also means focusing on Zero Trust networking. Jayshree, I think, alluded to our strategy in the best possible way, providing the best-in-class networking security, whether it's on the firewall side, segmentation, the NAC solution we've invested so much in, but not forcing our customers and locking them into a SASE solution or identity solution that only comes from us, giving them choice to partner with a Palo or Zscaler or whoever they might pick. That strategy is helping unlock this market for us and so is a focus on Zero Trust operations. I'm sorry, on zero touch operations. This idea that you should be able to deliver a truly mission-critical network, a 49 and 59 campus network without armies of people, without having to be in the Fortune 100 with thousands and thousands of network engineers. It's building on the foundation of EOS to deliver that kind of reliability, exactly what Ken is talking about. But it's also building the technology we need to compete in this campus network where we are relatively a newcomer and reduce and remove all of those roadblocks so we can compete in every deal so that we can deliver Arista quality for every network. And that innovation has been going on for years here, and it's an exciting time right now because we're really starting to see that unlock. The WiFi portfolio, this started as an acquisition seven years ago of Mojo, it's been -- there's been phenomenal innovation and velocity here. And we are now sitting on one of the most complete portfolios in the world. We have not just indoor and outdoor APs, but high, medium, low offerings in all these areas, external antennas. No matter how sophisticated an RF deployment you want to put together or how simple you want that install to be, we have hardware offerings for you, and they run at the highest reliability of any WiFi on the network. WiFi is near and dear to my heart, and I'm telling you, I put this in my house. I have been compelled by this solution. Arista's switching is second to none. but it didn't go -- data center switching doesn't drop into the campus by itself. There's been an enormous amount of innovation across a campus switching environment, bringing EOS to the campus, and it is an incredibly powerful solution. We've always had SP and large campus routing, but the Arista acquisition really finally closes the loop and completes the puzzle. In fact, it's this continued investment that has filled in every hole in the campus portfolio and now leaves Arista with a complete networking stack. These are some of the key investments that we've made. AGNI is our NAC solution. For many years, we've been putting R&D into this. And AGNI is incredibly powerful. It provides network access control. It allows the Arista networking stack to leverage all of the security posture assessment from third parties and provide best-in-class network security using Arista technology. The Wi-Fi acquisition, I just talked about, and it's obviously incredibly powerful. But the Velo acquisition and bringing SD-WAN, it's an incredibly important part of the total solution. And it's important because while Arista has had best-in-class routing for large campuses, connecting headquarter sites from continent to continent, some of the highest performing routers in the world for service providers, et cetera, we have had a hole in our portfolio for the branch, for the small office, even the home teleworker and Arista fills that solution. It allows us to connect those branches over broadband, bring Arista technology into every single site and for customers who want to make a single architecture choice for the network, now no matter whether they're running at the largest university campus in the world or they want to deploy at the smallest branch office or coffee shop, we have a solution from routing to wireless to switching for them. It's an incredibly powerful acquisition. I'm super excited. They arrived the same day I did. It was like it's kismet. I love it. The VeloCloud acquisition also brings something special to our go-to go-to-market. We've been investing in bringing up a channel, especially this year. We're starting to see solid momentum in that channel, both systems integrator and service provider. We've been expanding our direct sales motion and it's amazing to see the momentum, especially in large strategic accounts, downtown major New York financial headquarters, like Jayshree said. The Velo team brings in an MSP motion. They've done a ton of their business traditionally through managed service providers who provide an all-in-one managed offering. And it gives us really two phenomenal opportunities. It gives us the opportunity to take those managed service offerings and bring all of the Arista technology through that and of course, to take the Velo technology and bring that through this kind of burgeoning campus channel that we're building today on the Arista side and have been building for years. The key here is this investment, both in leveraging the EOS technology for the campus and developing new technology on that EOS platform for the campus. One of the biggest roadblocks in these large campus deployments, especially in education, but any multi-floor office has been stacking. We've put a lot of investment in delivering campus stacking. Our version of that is called SWAGs. It's going to be coming out soon. This gives an opportunity to deploy the highest density sites by being able to stack not just one, but really not just one or two, but dozens of switches together, manage them as a switch and really have that cluster of campus switches operating as a single switch. It's been a competitive issue for years. Now bringing stacking to the campus at Arista is enormously powerful. And from someone with new eyes, it's amazing to see the speed of innovation velocity on EOS that made this possible. It's a remarkable innovation and removes an enormous roadblock in the market. But one of the cornerstone differentiating features of EOS in the data center has always been hitless upgrades. The ability to upgrade the firmware on a data center without taking any downtime. It's something that I had a hard time getting my head around when I first learned about it and to watch it in action is remarkable, so much so that I had to run tests to prove to myself it was real. Bringing hitless upgrades to both wired and wireless means that we no longer have to consider for planned downtime and outages due to waiting to upgrade as bugs and security vulnerabilities become too critical. Hitless upgrades means we can realize that promise of zero-touch operation that we can maintain the most secure software on campus networks around the world and that we can do it while maintaining perfect uptime, delivering on the promise of Arista, the most reliable mission-critical network in the world. We've seen an enormous amount of investment and focus on this concept of Zero Trust operations. And Ken mentioned it, CloudVision and the NetDL framework that's built upon is -- it's a remarkable tool that allows you to manage data center and campus network, something that no one else in the industry has, and I really don't think anyone else in the industry will have anytime soon. CloudVision and the whole suite of management products at Arista have flexibility. You can deploy them in air gap networks, but they deploy in the cloud with some of the simplest, most straightforward functionality. But NetDL and that idea of seeing the complete state of an entire network in a single platform gives enormous power to network operators today. But the opportunity that we have with AVA to use that data lake to deliver truly differentiated AI assist is phenomenal. So I'm excited to bring Ken back on the stage and give us a little sneak peek of AVA.

Thanks, Todd. All right. So this is a very quick early technology demo of what AVA can do. And the scenario here starts with just a sort of a chatbot style interface. But again, it's not a chatbot underneath, but we do route the question to the right agent based on the content. And so here, the question is, what can you tell me about Alice's device? And you as mispell, one of the things I love about LLMs is they just do not care how you spill, okay? They just -- you can spell any way you want, they figure it out. So we typo in our demo and hey, it just works, so whatever. So what can you tell me about Alice's device? Ask AVA understands this calls for some telemetry, makes a telemetry query and brings up a bunch of information. I've -- you can see the actual screen kind of underneath. I've kind of called out what I think is kind of the key information in the larger window, so I try to make it readable. But here are some details about the device. Here's the host name. Here are the IP addresses that are in use. There's a MAC address. And here's how it's connected to the network and offers you some things you want to check on this or check on that. Actually, it can't reach the Internet. Okay. This is an incident. We create an incident record invoke, the AVA troubleshooting assistant. And the AVA troubleshooter wants to run the following action ping from this location to that location. And once the operator allows that action, we look at the resulting traffic and see that sure enough, nothing is getting through from the switch to the Internet. So it's not Alice's device that's the issue. The problem is actually wider spread. And then we -- there's a little bit of a back and forth here. I've kind of skipped some of the details. But after doing some other pings, trace routes, looking at some configs, troubleshooting AVA concludes, I've reviewed the access list configuration on a leaf switch that's involved in the flow. It appears there's an access list named rogue device list that contains a deny statement for the following subnet. Since Alice's device has the IP address it has, it falls within the denied range. This is likely the reason the device can't reach the Internet. and then there goes on to offer more things and there's a further conversation. The point of this is that the troubleleshooting assistant, I think, is going to really change the game for how quickly and easily people can resolve these kinds of problems. I want to leave you with here, the Arista way. Everything we do is based on our architectural foundation and our culture of innovation. The most important thing, again, is that thing at the top. My commitment to you and to all of our customers, we are never putting your network at risk so we can ship some shiny new feature sooner, okay? We're always taking the time it takes, whatever that is, to make sure that when we ship it, it actually works. Thank you very much.

Thanks, Ken and Todd. So we'll take a quick 15-minute break, allow you to stretch your legs, grab some coffee. The restrooms are right around the corner as well to your right when you exit the back doors. And then we'll be back for a deep dive into AI networking and how the power of Ethernet is transforming AI networks with Hugh and Andy. So a quick break. [Break]

What an amazing group of AI thought leaders. And speaking of thought leaders, it is my great honor to welcome on stage, Andy Bechtolsheim to talk about AI and Ethernet networking.

I don't need to tell you what a unique moment of history we're in. And wait a minute. This is the wrong slide. Sorry, guys, wrong slides. This is embarrassing. No, Analyst Day slides, not internal MDA slide, sort of this. How did this happen? Not possible. No, no, no. No, no, no. How did this happen?

Hugh, do you want to go up?

I can.

Okay. We'll do Hugh hover it next, and we'll reverse the slides.

This is advancing, but that is not advancing. Okay. Great. Hi. I'm not Andy. My name is Hugh Holbrook. I'm the Chief Development Officer at Arista. I mean it's really an honor to talk to you all. Thank you all for coming. I'm here to talk about AI and cloud networking. I spent a bunch of time both on internal development and talking to customers and in standards bodies, working on AI and from platforms and network design to software. And so I want to tell you about what we're doing. First of all, we've got the Arista Etherlink portfolio, which is really a suite of technologies, both hardware and software technologies to try to make AI better. And this is for all parts of the AI network, the front end and the back end, the scale-out, the scale up and the scale across. So it's platforms and software purpose-built for AI to try to make AI better, and that's Etherlink. In terms of the platforms, just on the hardware side, we have a range of switches from the 7060s, which are kind of targeting scale out and scale up. These are the lowest power switches with the lowest power, most reliable, lowest cost optics. We've got the 7800, high-radix modular chassis, deep buffered high featured, useful in the scale across and also in the scale-out dimension, then we have the distributed Etherlink Spine, the 7700. These are kind of the three major product families in the Etherlink portfolio. So one thing I want to talk about is the front-end network. So there's a front-end network and a back-end network is kind of like the world of AI networking kind of gets bifurcated that way. And the back-end network is the network that interconnects just the GPU to GPU connectivity. Today, it's almost all RDMA, typically rocky. So it's high-speed GPUs doing direct memory access to GPUs. That's the back-end network. Also in the back-end network is the scale-up network, which is just inside the chassis, just GPUs talking to GPUs inside the chassis or inside a rack. That's scale up. Scale out is the back end, GPUs talking to GPUs. And then there's a front-end network, which is kind of the lifeblood that feeds the AI compute fabric, which is, I would say, equally important and actually quite a bit more complicated than the scale-out fabric in terms of the functionality. So the front end is kind of the gateway, and it connects to storage, compute, cloud, WAN and both the back end, but also the front-end performance is critical for both training and performance. So if you look at like what the front-end network does is it connects the AI fabric to all these different things that are part of AI jobs, local storage, general purpose compute, cloud storage, the Internet, the corporate network. And I asked, of course, used AI to help me like explain some of these connections and why they're important. So the first thing I did was ask this query, like where is the storage for a RAG? Like a RAG is a database that is used as part of inference to get real-time data. And like the first thing you note is that like when I asked this question, it searches 94 sites. Like this is Gemini. It runs out and it searches 94 sites to in real time, get answer my query. This is I need good Internet access, right, for inference. Like I have to have solid Internet access from wherever I'm doing the inference from my AI cluster out to the Internet. And this is going through maybe it's Equinix, maybe it's going straight to Google, maybe it's going to Azure, whoever my provider is or maybe it's going across my internal network. Now the thing I was actually trying to do was like show this, which is like, well, regs are typically stored either in cloud storage or maybe in local storage. But if I'm going to cloud storage, my AI inference job has got to connect to GCP or Azure or Amazon via S3 to get to that cloud storage with all the protocols, security, VLANs, route advertisements, all of that stuff that's necessary to connect to those cloud providers. If I'm doing inference with KV Cache offloading as a technology you might have heard of, where I'm I run a query, I'm partway through it. And then like I pause and I get a cup of coffee or I just think a little bit and ask my next question, and that GPU is not going to sit there idle waiting for me to come to the next query. It has to get loaded with the context of somebody else's like Andy's next query or Ken's next query. And so there's a whole bunch of context and it's gigabytes and gigabytes of data that have to get paged out of the GPU and into storage. And that's going into local storage, which is typically not on the back-end GPU to GPU network. It's typically connected on the front-end network. And it's a storage cluster, which has different needs. The storage servers may have different requirements, different kind of switches are necessary. General purpose compute is super important for training. General purpose compute is not where you're doing the compute for the training per se, but it's where you're preprocessing the data. I have reams, and reams, and reams of data. I'm sure you've heard that like LLMs are trained on trillions of words of examples or trillions of tokens of examples. And like that data all has to get preprocessed somewhere using like tons of algorithms. And it's coming from the whole corpus of the Internet or it's my internal customer database or my engineering database or whatever I'm training or fine-tuning my models on, like that is going through general purpose compute. And then at the same time, my -- I may have data that I'm accessing. Maybe these are RAGs, maybe it's queries to my internal databases. Maybe I'm doing agentic AI where I'm going to like have an agent that is actually going and doing something inside my enterprise. It's got to reach out to my enterprise network. That's probably not in the same data center. I may well have confidentiality, security processes that require that data to be hosted in my corporate database or maybe it is just naturally stored there. It could be distributed across my corporate WAN. So I need access from the AI network to the corporate network. That may have security. It may be access protocols. I may have VLAN access. I got segmentation. I have gateways I'm going through. I've got firewalls. All of this is the purview of the front-end network. So in this kind of space where I have to access the front-end network, there's all these protocols that we've been developing for EOS that like are part of the solution for the front-end network. Multi-tenancy is super important. Like there are lots of clients on my network, and I have to keep their traffic separate. I mean this is obvious if I'm a service provider, like I'm a Microsoft or something like that, I have different customers. And of course, I have to keep their traffic separate. Security of what I'm sending into the AI, I cannot have -- Microsoft can't have their customers -- I don't know who it is, customer X and customer Y, Boeing or Walmart, like their data can't cross, right? They have to be kept separate, and they have to be kept separate all the way to the WAN. Back up a second. High availability, extremely important, keeping things alive, keeping access on my gateways, keeping all the connections running, availability of the AI traffic is important. all of this WAN traffic going to Azure, going to different gateways. I have to do route advertisements. I'm steering traffic. There's MPLS involved in many cases. This is because I'm accessing what the Internet is, what my metro network is, what my corporate access network is and my AI fabric is touching all of that. Confidentiality is important, and it can happen in a couple of different ways. It can be segmentation by separating traffic in different ways, and I'll talk about that. It can be encryption. It could be encrypted on the endpoints or it could be encrypted in gateways if my endpoints aren't doing encryption. Many times inside, inside the fabric, there's high scale, right? There can be many, many GPUs. So doing scale right, both route scale outside the data center when I'm talking to the Internet, policy scale, if I'm doing filtering is important. And then just like ECMP scale, if I'm going very wide because I'm building -- connecting a lot of GPUs through a lot of Tier 2 switches. And then observability at scale is important with all of this. So these are all the set of features, observability, telemetry, routing that come together in the AI fabric, like that there is a part of the network, that back-end network, which is relatively simple in terms of protocols, and there is sophistication there. In terms of routing, it's not the most complex routed environment, but that has to connect to the rest of the world. And that is not a simple connection. Segmentation is something I've talked about, and this is just keeping different customers, be they actually my customers or be it just different corporate clients that have different security profiles like my finance data has to be kept separate from my research data has to be kept separate from my engineering data. I may have siloed projects where I can't be crossing things. We have multiple tools for doing this. VXLAN EVPN helps with seamless provisioning, standing up subclusters of GPUs within a data center and then can preserve that segmentation in the scale across fabric going from data center to data center, if I have my job spanning multiple data centers, either for performance or for resiliency. We can do IP address segmentation. We have multiple techniques to do that, that we've been developing over the years, and we have customers deploying those to be able to like do the segmentation without having the overhead of a VXLAN header on the packets. We're using Dot1X, which is technology that we've been developing and have continued to develop for GPUs to be able to identify what job a GPU is from or is part of so that we can put it in the right network segment. And then encryption can be important in the network, especially if I've got data that is segmented somehow within the data center. But then when I leave the data center, I don't necessarily have the same control or the same comfort about the confidentiality of the data on those links, and I want to have wire rate encryption, many, many enterprises and larger clouds have policies that when things leave the data center, it has to be encrypted one way or another. And the way to guarantee that is to have link-by-link encryption. So this trusted segmentation is important as functionality that we've been building. So as I said, that segmentation happens within the data center, the secure connectivity happens within the data center, but larger jobs and larger AI clusters are now spanning a single building or a single site across a region because I can only get so much power, so much space in a single region. So it's not uncommon to have multiple data centers within a region talking together. And all of that multi-tenancy, all that segmentation to keep customer A from customer B or to keep finance separate from engineering has to be extended to the WAN. It also has to be extended to the other data center. It has to be extended to Amazon. Like I have to keep those separate paths of traffic going to Amazon or going to S3 all the way that I'm going to Azure, all the way that I'm doing it. The security, whatever security I've got, whatever confidential I've got, that has to be extended all the way through this scale out -- from the scale-out to the scale across fabric. I want to talk today about [ DANZ ], which is a secure traffic analyzer. It's a feature that we've developed that is doing telemetry at the host at the top of rack switch and at the spine. So we've got telemetry from an AI agent running on the host. We've got telemetry being exported from the first top switch and the last top switch. It's the sending host and the receiving host and also at the spine of the network that's gathering information about what jobs are running, what hosts are running, the performance, flows, pulling that data together, putting it in CloudVision, so we get data about -- I don't know if you can see the tiny little box here, but like about security applications, what tenants are running, what hosts are running, how they're performing, tunnels, et cetera. That all is pulled together, and it that as a foundation to build upon the secure segmentation or to extend the secure segmentation to do all of the traffic management that we need to do to build a reliable, solid performance AI network. detecting misconfigurations of the endpoints, doing a good job of load balancing, identifying congestion and hotspots, being able to report those, steer around them, being able to manage our tunnels, our peers if we're connecting to an Equinix or to another router in the Internet, detecting DDoS attacks, micro burst and then connecting to storage. So all of this is the foundation of the Secure Traffic Analyzer builds and then -- is built and then lets us implement these things to deliver better value. I want to talk briefly about platforms. Andy will talk about this. I'm going to go really fast. We have a suite of platforms for the Etherlink portfolio. And each one is optimized for a different role or a different set of roles in the AI fabric. And that diversity is valuable. We've got single-chip systems, typically one or two or maybe four RU for scale-up and scale-out fabrics that are optimized for power cost and speed. We've got the chassis for the largest Tier 2 networks. These can be useful as a spine in a scale-out fabric or as an edge device or as a kind of central layer in the scale across network. That scale across network is the network that connects multiple sites when I have a regional network of data centers put together. The [ DS Systems ], the 7700 for rapid, seamless AI back end. We have these deployed at some customers. It has the kind of attributes of -- the data plane attributes of a single router, a single device and the load balancing of that, but in a distributed fashion that can scale out to 256 or more top-of-rack switches and support networks of 4,000 and more GPUs. And then we have edge devices and a deep portfolio of edge devices that can play that role at the edge of the data center. These have deep buffering, routing, MPLS support, large routing tables, tunnel capacity, built-in encryption support for connecting the front end clusters to clusters together, connecting clusters to the cloud. And then we have more than one, multiple, shall I say, custom, and I can't say much about them, custom designs for scale up. And these are designs inside rack scale designs. This is an example of something that is representative of the kind of thing we're building. So we have a broad set of platforms that are optimized for a range of AI use cases. I want to talk briefly about the Ultra Ethernet Consortium, which is something that I was personally quite involved in. I chaired the Technical Advisory Committee. I'm on the Steering Committee. And the Ultra Ethernet published the 1.0 spec back in June. They're continuing to do work. But the Ultra Ethernet Consortium was a consortium founded by these 10 companies at the top, among them was Arista in the founding set of members. They are now more than 100 members. The goal of the Ultra Ethernet Consortium is to advance Ethernet in the service of HPC and AI. Like Ethernet is already very successful in AI and HPC networks. The goal is to take anything that makes it like that we can do to make it better and do that. There's more than 100 member companies, 1,000 participants. Arista is quite active in it and was a founding steering committee member. The Ultra Ethernet is supported by Arista switches. There's a lot that happens in the NIC in Ultra Ethernet. There is some functionality in the switch. Just an example here is packet trimming, which is functionality that we've put into our switch to support Ultra Ethernet. It's technology that will make the transport protocol be able to detect lost packets faster and more reliably. Ultra Ethernet itself, the Ultra Ethernet transport is a standards-based transport protocol for AI that does out-of-order delivery, congestion control and security. This will make things better. There will be other transport protocols as well. This is not the only one, but this is going to be, I think, one that will have some impact starting later this year and more so into 2026, I expect. Our 2025 products, specifically the 7800, 7280s also and the 7060s will have support for the Ultra Ethernet capabilities that are needed to run the Ultra Ethernet transport. We have a broad range of merchant silicon that we use. I'm not going to spend a lot of time on it. But again, like the platforms, and each of these is kind of foundational in the different platforms, and they're optimized for different use cases, highest scale, a perfect scheduled fabric, power cost optimized, super low latency optimized. And we use these judiciously when we need to, to build the best platforms that our customers want. I want to talk about a metric that I've started to call TTF or time to first job, which is a really important metric for our customers. Time to first job is what I consider the time between when all my equipment shows up on the dock and is ready to be built into a data center and when I run my first job. That is provisioning the switches, testing everything, testing the software, making sure that I'm getting good performance end-to-end, debugging cables, figuring out what fans aren't spinning, making sure that all the -- everything is inserted properly. It's a huge problem, and this is super, super focused for these customers because having those GPUs somewhere between the dock and the first job is just burning money, like it can be really expensive at a street price of estimated $30,000 just to pick a number, if I have 10,000 GPUs, it's $300 million. It's like a very expensive asset that's sitting there waiting to get it running. And there's like no budget once it's up for downtime. And the problem here is that like new switches come at the same time as new GPUs, NICs, optics, everything comes up together. And we need a super solid foundation. The features have to be ready. They have to be debugged when the silicon is ready. That is what we are totally focused on. That's what we've been very good at so far, and I think that's a compelling advantage that we have here at Arista. Things that we've done to optimize for time to first job, quality, just our relentless focus on quality. I'm sure you've heard Ken talk about quality or me or Jayshree, super important to us. Telemetry, just visibility, what is happening, what is going wrong? Why isn't this working? Super important to have the visibility, fine-grained visibility. We have many, many features that do that. The other thing we do is we provide deep sharing at the platform layer in our code across silicon families on things like optics and Fi management, platform commands and telemetry. This is something you cannot bolt on after the fact, and it results in better time to first job. And I've just got a picture that shows this. So if you look at the left, and this is kind of like how I always did things before. There's like the gray stuff is the software that's shared. It's like routing and BGP and SNMP, a bunch of shared stuff. And then there's a bunch of stuff that's specific to the silicon, programming TCAMs, programming apples, powering things on, managing the FIs, initializing silicon, pulling counters, like all this stuff. What we did at Arista was we said, you know what, there's a bunch of stuff in there that actually can be shared typically. Like this comes in an SDK from the vendor, but we can share more of it. Pulling counters, I don't need different code for different chips. Sure, the counter registers are different. But doing that efficiently, storing it in shared memory using DMA engines to pull the chips, like storing it, making it visible, we can do that in a shared way. And we did that across the board with fine grain state machines, programming the apples, programming the TCAMs. The very lowest level here, this white stuff needs to be unique. The registers on the chips are different. But we can share more of the code than what's typically done, and we can do it across vendors that compete with each other. And of course, like Broadcom and Marvell and Intel and Cavium, like they're not going to share code like their competitors, right? But we can inside Arista, we can share that code. And that gives us better tested code day one when we start deploying our switches because we already tested it on the previous platforms. Jericho2 can be shared with Tomahawk, can be shared with Trident. We can share code there. That gives us better quality, faster features, one team working on the same code base, and the result is better, and it improves our time to first job. So I think that is our -- what I would call our architectural advantage or one of our architectural advantages in addition to what Ken talked about that is like less talked about, but I think is equally important, honestly. And I want to just briefly talk on the blue box. This is my last slide, and then I'll hand over to Andy. The Arista blue box, as Jayshree talked about, is kind of -- there's a white box, which you know about. The blue box is the Arista version of it, and it gives you a choice of OS. I can run EOS SONiC, FBOSS or any of those or a variant of them or multiple of them, if I want. It's built on top of net DI, which Jayshree talked about, which is used to validate, validate not only the hardware, but the low-level firmware like the thresholds for power, the optics tuning, the programming of the fan speed algorithm. These are tricky system-level components that are hard to get right that can fail in the field. They can fail in the field years after, they can fail when the vendor makes what seems like a transparent change or a process change, and we need to be able to detect those support that. We're hardening components. This is all fully -- the blue box is fully supported by the EOS software team, hardware team, diags team. You've got the backing of Arista on it and the choice to run the operating system that you want on top of it. So Arista's Etherlink products are optimized for AI in many ways that I talked about. And I believe that we have an architectural advantage for AI that gives us an improvement on time to first job from features, quality, software architecture, blue box. That is everything I have. I'm now going to hand it back to Andy to tell you about a lot of the details underpinning what I talked about today. And what Andy has to say is like totally fascinating. So thank you very much.

Okay. Is this working? All right. So I want to talk to you about the truly extraordinary opportunity that's ahead of us. I mean you know the numbers. The CapEx numbers are going up every day apparently. But if you think about it, we're still at the early innings of this journey where not just the workloads are scaling and getting bigger exponentially, but the models themselves are evolving. And suddenly, it's not 10,000 GPUs per cluster, 100,000, but it's 1 million. And the requirements in terms of how you implement these very large-scale data centers, including how the network supports these very large data centers are really paramount. A few years ago, typically AI cluster was 4,000, 8,000, 16,000 GPUs. InfiniBand did that. Well, InfiniBand stops kind of at that level. I don't actually know a customer that's not planning on hundreds of thousands and in some cases, millions of GPUs that are tightly interconnected in a scale-out network. Now the second thing that's changing here is that the bandwidth per GPU or XPU is going up dramatically with each generation. So the numbers here are forward-looking statements, meaning the next version of XPU perhaps is a 12.8 terabit scale up and the 800 gigabit scale out. The one after that will double that. The one after that will double or quadruple that. If you do the math on a per sort of cluster basis, you're going from like 100 petabits to 100 exabits. That's a number that's 1,000x bigger within campus sort of wide data center, think multiple gigawatts. Now as Hugh was talking about, we do primarily use three silicon architectures to support these various requirements. Starting on the lower left is the Tomahawk Ultra, which is a brand-new chip that is the lowest latency Ethernet switch on the planet. 250 nanoseconds. So it's really custom optimized design for scale-up applications. The Tomahawk type of chip that we're shipping to Tomahawk 5 today, Tomahawk 6 in the lab, Tomahawk 7 is coming shortly that can span 10,000s of XPUs in a 2-tier network, which very efficiently. And the [indiscernible] architecture, which is by far the most scalability of all and we have successfully deployed in both the modular chassis form factor as well as the disaggregated switches. Now talking about what we can contribute here and what's really important to our customers is, number one, power reduction. So the power bills, of course, you know what they are. But the real issue is whatever power the network consumes takes away from the power available for the GPUs, which make the money, right? So basically, the network is for [indiscernible] on the delivery of the cycles that the power is supposed to pay for. So every 1% power improvement in the network means in a large data center, like one that has, I don't know, 100,000 chips, you get 1% more GPUs. So first step is the latest switch silicon is always more power efficient than the previous generation. And thus, there's this incredible pressure to get new silicon into the market in volume as quickly as possible in sort of a ramp that nobody has ever seen before. The second thing has to do with the transmission of bits. Copper cables have essentially zero power, but they only work within the rack. Beyond the rack, you need optics, and there's a lot of emphasis on how we can reduce power for optics. We have been an industry leader in promoting the adoption of linear optics known as LPO linear pluggable optics. And we now have multiple customers that have deployed these things successfully in volume. The short summary is the linear optics is 1/3 the power of a full retime optics in the next generation. And thus, you can get 3x as many optics if you go linear compared to the full time. Now another factor that's a little harder to explain is that the larger the radix of these switches, the fewer layers you need in a network, the less power it consumes. So we're all into these maximum fan-out radix kind of networks to accommodate most building kind of deployments in two tiers rather than three tiers. And the final one, liquid cooling saves power because there's no fans. It's between 5% and 10% at the system level depending on temperature, and that's an important step. Talking about liquid cooling, this is not a product announcement, but more like a directional statement of what's coming. We're designing 100% liquid cooled switches that plugs straight into an Orbi 3-style rack. You see on the picture on the right is the rear of the chassis, plugs in the busbar, the liquid connections, and that's it. It's like a line card in a big chassis. So a lot of our development is focused in this direction. In addition, we are designing fully liquid cooled switch racks that offer up to 32 payload of these fabric switches with patch panels and management switches and power shelves. And again, this is all optimized for the liquid cooled data center to enable high-density switch configurations. Separately, we have engaged with multiple customers on multiple projects on custom switch designs for their custom AI racks, which are customer-specific. And these are these very large 500-kilowatt going to megawatt class racks that have a lot of internal switches for scale-up in particular, and they use copper cables internally for the highest -- at the lowest power and highest bandwidth connectivity. So these projects we're doing in very close conversion with our largest customers and the whole focus is to minimize the design to volume deployment. One slide on pluggable optics. I don't know how much you have followed the market there, but there's this endless debate about what's the best optics and lowest power. Well, if you could eliminate the retimed DSP, it is lower power. And it's surprisingly also more reliable. There's few components that can fail, and we do see a lot fewer link flaps with linear optics than traditional retimed. But the most important thing is it is the lowest power optics. There is no other optics that is lower power, including co-packaged, which is essentially the same components, but just placed in the different part of the chassis. So what people like about pluggables is that it supports any kind of technology from multiple vendors, including future microwave ideas and slow and right optics and whatever comes next, whereas if you do co-packaged, you're tied into a vertically integrated single vendor stack. Now going back to the most important request from our customers is they're spending too much money and they're asking us to reduce TCO. It's true. So the way we can help is minimizing time to volume for next-generation silicon. We are all in on liquid cooling to reduce power, eliminating fan power, supporting the linear pluggable optics to reduce power and cost, increasing rack density, which reduces data center footprint and related costs, and most importantly, optimizing these fabrics for the AI data center use case. So what we call the purpose-built AI data center fabric around Ethernet technology is to really optimize AI application performance, which is the ultimate measure for the customer in both the scale-up and the scale-out domains. Some of this includes full switch customization for customers. Other cases, it includes the power and cost optimization. But we have a large part of our hardware engineering department working on the things on this slide, and I'm spending all my time on this topic here. And this is all the slides I had. So thank you very much.

Hello, everyone. Very nice to see you. Always a tough act to follow, Andy, but I'll give it my best shot. So thank you for being here for the last part of today's official agenda before the panel. So the whole thing I'll focus on in these slides is giving you a different perspective on where we see momentum for Arista. And so just -- we were talking at the break actually and some people are mentioning, do you ever step back and just realize how far you've come as a company. And coincidentally, we do have this slide here to show the momentum since the IPO in 2014. You can take a look through. But I think that to say there's 52x market cap, and that's actually higher if you take it today versus the August kind of cutoff point we did, 15x times the TAM. So very excited to see the results to date. But what I would like to leave you with on this slide is to leave you with the thought of the tenacity and the conviction for Arista to execute when they have the intention when it comes to pure-play networking. Just a quick summary of some of the key financial metrics for you over the last five years. You can see which ones you resonate to. My two favorite children in this slide or if you look at gross margin and operating margin, you can see even though you have the fluctuation and the volatility in the gross margin during different times, depending on what's happening with mix and inventory, you still see our ability to deliver operating margin expansion. And I think that's a good testament to the fact that we have a very efficient and effective business model that we're very proud of. So now let's get into the different aspects of building momentum. The first slide I have for you here is to kind of set the foundation when it comes to TAM. So Jayshree had mentioned at the beginning, the TAM in the sense of crossing that $100 billion mark as we get to 2029. You can see in the last two years, we've had a 75% increase in our TAM, very much bolstered by the AI conversation, but equally so in kind of the campus branch segment of the TAM. And so together, we're looking at $105 billion by 2029. Super excited by this foundation to give us the growth foundation that we need going forward. So now let's talk about building momentum in cloud and AI. I'll start with an external view first and then talk about Arista's ability to deliver in this market. So I like this 650. AI is such a big space. There's all kinds of projections that you kind of need a framework to gravitate to. One of them is the 650 group framework I like that talks about different waves of how AI is going to come in. And you can see kind of the dollar values coming in going from the foundational models and content creation to agentic AI growing to $1 trillion between '25 and '28 and then wave 4 being autonomous transportation and robotics, humanoids, another $1 trillion 2027 onward. So even if these are indicative of the opportunities, we can see that Arista is very well set up to play in this space. And I'll give you three reasons why I think Arista is very well positioned, perhaps uniquely positioned to maximize our potential in this space. You can also just go more specific to the cloud kind of data center infrastructure CAGR for the CapEx being about 16% '25 forward. So pretty robust numbers no matter which ones you look at. But the three reasons I'd give you why we're uniquely positioned to take advantage of this space and to do very well. The first one is all the stuff that was spoken by Jayshree and Hugh and Ken and Andy in the sense of our product portfolio, our software capabilities, our net DI, scale out, scale up, scale across. Very excited about what we have to offer from product and solutions. The second one is Andy talked about the thought leadership when it comes to minimizing or optimizing the total cost of ownership for the AI data center build-outs. He talked about the silicon. He talked about liquid cooling, linear optics, rack density and optimizing Ethernet fabric. So very much a playbook we can use with our customers. And the third one is that great set of AI partners we announced this week in the sense of working with them in the community. So three really compelling reasons, I think, to say why Arista is going to do very well in this space going forward. So now we can switch to equally as important, how we're going to build momentum in enterprise. This one is not as cyclical. It's not as volatile, not as big. It's a little bit slower to grow, but quite a steady Eddie in our portfolio. Very excited for Todd and his ideas he shared with you earlier on campus specifically. You can look at our expansion of customers. If you look at the last 12 months of customer growth over FY '22, some pretty, I think, demonstrated results in the sense of growing customers internationally. You can look in the sense of we're starting at a pretty fair space when it comes to share, when it comes to campus and enterprise generally. So we feel we have lots of share to go get. And the way we're going to do that are the 3 growth drivers you see on the right: Acquiring new logos; land and expand; and then this whole new AI use case when it comes to enterprise and Agentic AI. One of the things I want to do is give you -- sticking to enterprise, is to look at kind of what's an outside-in feedback that we received, and we're super proud when we get these kind of sentiments shared back to us. This one is from the Gartner Group. And it looks at -- on the left-hand side, you can see it talks about the size of companies that are considered, the industries and the geographies, so very well represented across many different domains. But the right-hand side is the thing I think we're most proud of. It talks about ease of use of doing business. It talks about data center and the wireless. So a wide breadth. So I thought this slide was just very indicative of the things we're very proud of because we only focus on networking and try to do it very well. Again, sticking with enterprise, to give you a different perspective of why we're confident we can continue to take share in this space is from the 2 data -- 2 examples we have with customers in the enterprise space. Customer 1, a U.S.-based, very large insurance company. And this is a story of going from data center to campus. And you can see the journey since 2015 to now, just the stickiness with the customer, continue to take share of wallet. And if you look at the 2025 to 2015, 29x expansion of their spend with us. Then you go international. And here, we have a campus to data center win. And to me, that's really great news because it validates, one, our brand; two, it's international; and actually, a third one is it shows that our product portfolio is there to serve some of these largest customers, this being a widespread financial institution. So very happy to see that, and we're 6x expansion share of wallet over 5 years, and we'll continue to see what we can do to help serve that customer. Third space we're building momentum is the -- what we call specialty providers. In this category, you have the telcos, the SPs, you have the streaming services, you have the networks. And now we have this great benefit of putting Neoclouds in this space in this category. You can see on the left some of the cloud -- Neocloud, excuse me, growth drivers, very specific curated needs. You can see the great CAGR predicted by the 650 Group, 28% CAGR this year forward. So lots to consider in that category for us. When Neoclouds have the ability to have open, best-of-breed conversations in their RFPs, we absolutely want to play in that space. And the thing they come to us the most for is to take our experience with the larger Cloud Titans and how can they have that experience, how can they have those outcomes with all the things mentioned before me. So I think we're uniquely positioned there to have those conversations, and we're excited to continue to do so. Now switching to the other part of the P&L. So those were to cover how do we get to the top line numbers that we've been discussing. So here are some -- just some insights, I think, to margin drivers, which are some of the conversations we have. So if you look at operating margin, one thing we're very focused on is our commitment to innovation, quality, reliability. So we'll always be in this 10% to 13% range percentage of revenue on the R&D side. We're very committed to that. On the gross margin side, we've had lots of conversations last year, this year, what's driving your gross margin. If you take a look back over time, you can see that E&O as a percentage of revenue has fluctuated from 1% to 6%. Some of it was during COVID, but some of it is inherent in our business model. We have 1-year lead times and 2 quarters of visibility. So we have to lean in at any point and predict educated guess in the sense of what we should be doing, and we don't always get it right. So E&O is a part of it. We have this 1% to 6% range. I just wanted to provide some insights because we've had some dialogue on that. And then the last one is just the customer mix, which is the second category of what affects gross margin the most, E&O and mix. And mix, just to be clear, it's the mix between the 3 categories: cloud AI, enterprise and specialty providers, but there's also mix within those categories. As hard as we try to keep it simple, there is mix within them depending on macro environment, the use case, et cetera. So just to give you some thoughts from my perspective on how that works as we talk about the guide at the end of this presentation. Capital allocation framework, not much changing here. No need to change what we think is not broken. From this perspective, we have organic investment. We have the share repurchases. We have marketable securities investment, and then we have what we call tuck-in M&A. Hopefully, you've seen, during the last 12 months, that we do lean in for share repurchasing when it's the right opportune time, and we've done so. You've seen that we do, do tuck-in M&A where we think it's strategic with the VeloCloud acquisition. So we are demonstrating, and we will continue to, but no major change in this so far. Also want to talk about building momentum from the kind of the inside the company-scaling processes perspective. So I just wanted to bring some insight to all the things we're doing from a supply chain resiliency because I think that's important and Todd coming in working with him and his team and the CISO from a security perspective. But looking at how do we have optionality in the sense of who we work with, what kind of breadth do we have in that kind of supply chain from the vendors. We look at the location and make sure with all the different tariff scenarios, we have mitigation from a geo risk perspective. And then the last one is actually just the security aspect. Arista is in there in all the different steps to ensure that there is no issue when it comes to our customers receiving our products and services. So now we can talk about building momentum when it comes to outcomes. And I know Jayshree mentioned earlier the FY '26 outlook. So there you see that 20% growth at $10.5 billion; very, very interested and excited about that. Gross margin, we're going to keep it at 62% to 64% based on the mix that we know and the other drivers in the sense of inventory. So 60% to 60% -- 62% to 64%, excuse me, in the sense of gross margin outlook. And then OM, 43% to 45%. We have a lot of new leaders coming in to join our team, a lot of thoughts as to how we scale the company. So we're going to leave some room for Todd, Tyson, Hugh, Andy, et cetera, to decide what we need to do to keep that top line growing. And we'll continue to update as we go through until the February call to see if anything changes from that perspective. Very proud of some of the highlights in the dials below that you see: 10,000-plus customers; 87% Net Promoter Score, which is a fantastic way above the industry average; and then you can see our share in countries shipped, too. So just a blend of metrics to give you an idea of kind of the breadth that Arista is getting to back to the first slide where we came from 11 years ago. From building momentum on the long-term model, from this perspective, so from '23 to '26, we're committing to that 20% CAGR with the guides that you've seen. '26 to '29, now we're talking a little bit of law of large numbers. If you take that 15% growth to '29, we're talking $16 billion plus. So the mid-teens, we don't feel comfortable at this point. And we'll see as it progresses, what that mix will be. Gross margin is a fairly wide breadth of 60% to 64%. That's to give us some room from a mix perspective, mostly, given all the AI conversations we're having, what will be in our world when it comes to '27, '28 and '29. Operating margin at 43% to 45%, if we do find a different way to scale the company to leave some room for that investment. And then percentage of revenue, back to the point. We'll always be in that 10% to 12% for R&D, keeping sales and marketing at 5% to 6%, and then G&A between 1 point and 1.5 points. So hopefully, that's helped to demonstrate the momentum on the top line between the outside opportunity and our ability to execute. It's given you some insights in the sense of how our gross margin moves across the different elements. Very excited about what that does to our long-term model and all the opportunities there. So we're grateful for the opportunity and very excited about the possibilities. Thank you.

Returning to the stage, Rudolph Araujo.

Thank you, Chantelle. We'll now proceed to the final item of our agenda today. We have our panel for you to ask questions of. For this session, please raise your hands as we go through the -- the hands are already coming up -- as we go through the questions, and one of my colleagues will come find you with a microphone. We ask you to limit yourself to one question so that we're respectful of all of the other folks that have their hands raised. So along with Jayshree, Ken and Todd and Chantelle, please welcome on stage our newest executive, Tyson Lamoreaux, who is our Senior Vice President for Cloud and AI. We'll now proceed to open it up. Our first question comes from Ben Reitzes from Melius Research.

Mic is coming to you.

It's a pleasure to be here, and thanks a lot of the question and congratulations to you all. Seeing it from the IPO, this is a really neat day. A lot of your peers have actually talked about -- this for Jayshree. I don't want to talk about the numbers you put out, but a couple of your peers have talked about acceleration for next year. Hock Tan has been very upbeat about his business, which includes XPUs. Jensen can't really accelerate necessarily from where he is, but his numbers are huge, but he's talked about a 50% kind of CAGR for the CapEx. And then networking is going up as a percent of the overall spend of compute and networking. So I'm wondering, are you seeing all these trends. I know you typically give a conservative guidance. So qualitatively, though, this networking acceleration, is there potential next year for networking to accelerate? And if it does so, is there any reason why you wouldn't benefit and whatnot? Just parsing that through, are you seeing the same things in networking becoming more strategic and that potential?

Hey, Ben, that's got to be the longest record question I've been asked.

That's it.

Okay. So the short answer is we're absolutely seeing a lot of momentum in our business, in particular, in AI, in particular, the combination of the back end and the front end. Timing is the hard thing to predict, right? So we see this as a multiyear phenomenon. This has been a critical, crucial year for production of migrating from InfiniBand to Ethernet. We're seeing that. We're also seeing the advent of not just scale out, but scale across and then scale up coming in probably in the '27 or '28 time frame. So when you add all that up, there is a level of -- what's the right word, buoyancy and excitement and enthusiasm. But we want to stop short of that and be realistic about also how much of that will translate to numbers in '26 versus '27. Is it a 1-year phenomena or a multiyear phenomena? And I would say, Ben, it's a multiyear phenomenon. So when Broadcom and Hock Tan express enthusiasm on a chip, we've already bought the chip. We made the purchase commitments from it. Then we get it a year later, then we translate it into systems, and that translates into customer revenue. The delay from his enthusiasm to customer revenue can easily be 18 months to 2 years, right? So that's another important thing to remember. So look at this as a multiyear, really transformational piece. One thing I can just say personally is having been here the longest along with Ken, we're not living quarter-by-quarter right now. We are getting at least a 6- to 12-month view, and that's a good feeling. And we believe a lot of that is because people have to plan their AI centers well in advance, especially the power and the space and the building. And that's -- so absolutely, I think '26 can be a good year. And you would normally argue the law of large numbers. We shouldn't get too ahead of ourselves. But I think we are going to continue to experience not just double-digit growth, but at least mid-teens growth, and we'll see. It might get better.

Our next question comes from Amit Daryanani from Evercore ISI.

I guess just a question on the AI infrastructure side. A lot of the cloud companies are looking at different ways to deploy disaggregated network fabric, and there's a scheduled approach that I think helps or favors Arista more prominently and a non-scheduled one. Do you think cloud customers like Meta, for example, will skew one way or the other? Just talk about the pros and cons around that, that would be helpful. And then, Chantelle, how do I think about your deferred number in the context of this 20% revenue growth? Do you see that continuing to increase? Or is the growth really going to come from there?

Well, I think when it comes to deferred, which is a very common topic that we have a lot of dialogue on, not much has changed in the sense of the mechanics, but what has changed in the sense of what's going into it are these large -- remember, it's use cases, new products, new customers. And so if you think about the new use case being AI -- and Jayshree just mentioning this 18 to 24 months kind of time frame, it's going to take time to work through. And so the growth could be next year or the year after in the sense of things even coming through for this year. So that's the way to think about it. And then you have what's going in on the balance side, the P&L side will be what's accepted by the customers in that time frame. So of course, some of it's in the 20% guide for next year. It's in our 25% guide for this year, TBD on what we see on the acceptance side.

All right. I can talk about that.

Yes. I love it, if you can. Please go ahead. Is your mic working?

This one works. If we think about disaggregation, scheduled in the network, scheduled on the endpoint, I think everybody in this room can appreciate the amount of ongoing research that continues in AI. What becomes the hot thing for training a model or pushing more workload to inference kind of quickly dies. New techniques are coming every day, and traffic patterns are shifting rapidly, and the semiconductor space is seeing increased levels of competition. NVIDIA isn't going to be the singular solution forever. You've got kind of a lot of investment. And this is, again, a multiyear kind of long game. It's going to -- we're going to see things pay off at different points in time. I think the thing that Arista has going forward is this view of the solution as a holistic thing. And products, software, systems, teams, technical capabilities that can meet customers where they are. I think that, that is a differentiator for the business is the quality of the people and the way in which we engage in these deep technical partnerships. And so is it going to be scheduled fabric? Like we've got a product, we're shipping that. I don't see that slowing down. That's a very successful solution right now. A lot of customers love it, are continuing to buy it, continuing to invest. Andy and the team have invested in hardware innovation to continue to push that forward in novel and unique ways, been there first and are going to continue to push that. I think when we start pushing out to the endpoint scheduling, again, it's the same kind of notion that the fabric needs to be there, the scale up and the scale-out needs to be there. The network doesn't really change. I think those are just as equal opportunities, if not bigger opportunities for us because of the way we're going to work with customers, and we're going to prop them up and take our lessons learned from working with the big folks, driving the innovation, but cascade that across the entire, what I would call the AI practice in general, enterprise, service provider, cloud, it doesn't matter. I mean there's a lot of commonality there that we're going to drive leverage on.

I'm going to give him an A for doing all that on his fourth day in the job. Congratulations.

Thank you.

And just to add to what Tyson said, to answer your question, we co-developed with Meta, the scheduled fabric DSF7700. So naturally, it's being used quite a bit. But it's -- there's going to be different use cases. And some of them, it's going to be a simple single switch implementation that needs no scheduling, and then the operators have to figure out how to schedule it, right? And then there are use cases that Tyson, when he was a former customer once deployed, we absolutely need the scheduled fabric. Otherwise, you're going to need tons of resources to schedule the fabric with people, right? So it's really going to be depending on what your philosophy on this is. I'll buy a cheap switch and then throw people at it or I'll buy a slightly premium switch and throw less people at it. So I continue to see that those 2 architectures are living long and it really depends on the use case.

Next question is right next to Amit, Aaron Rakers from Wells Fargo.

Aaron Rakers at Wells. I want to ask about the numbers a little bit, right? If I look at the numbers you gave out there, $10.5 billion, I'm getting wrong, $2.75 billion of AI, you've got $1.25 billion of campus. If I look at that relative to what you guided this year, it would imply that non-AI, the non-campus number looks more flattish. So I'm curious, is that just conservatism? Is there an element of mix within AI that is hard to kind of discern? And then Jayshree, why do you think scale up is not a '26 story, more of a '27? Because it seems like you alluded to in the presentation a couple of them that you've done some custom work with some customers. So why is that timing, not next year?

Two questions, Aaron. Let me answer your first one. Any time you have enthusiasm in fast-growing markets, those are taking over. And we think our run rate business may not be negative, but it will be slower growth. Have we modeled it exactly correctly? No. But we're not expecting the rising tide for all boats. Some of the boats will just bop along. We're not being conservative. We're being realistic and some of them are going to grow faster. And it is true that we don't know quite how to count the front-end AI revenue. We've always struggled with it. But if it has a lot of AI traffic, it's going to go in the AI bucket. If it's a more traditional classic use case, it's going to go in the cloud bucket. So maybe we're not -- we don't exactly know until the year progresses. What was your second question?

Just the scale up, the timing.

Yes, scale up. So I think there's -- as Andy would have pointed out to you, there are many projects we're working on in scale up. But the most important thing I can tell you on scale up is the fundamental low-latency chip requirements would be the next generation of Tomahawk or Ultra that is still in labs right now. So the first will be just pure availability of these chips will be 2026, right? So by the time we put it on the board and make it available and work with the compute fabric, validate a rack, figure out if it's co-packaged copper, retimer, SerDes, optics, et cetera, it's going to be December 32, 2027, okay? So that's the reality of how things work with the software and everything. There's a second reality, which is standards itself. Some people may just go with it like they go with proprietary NVLink, but there's still a lot of confusion on scale-up technologies. You've got the NVLink, you got the UALink and then you got the Etherlink, Arista family of products, right? So I'm really counting on Hugh and the leadership from UEC and the entire consortium to define this better. I think Broadcom has done a fantastic job of putting out the scale-up Ethernet spec. But realistically, all these things take time to sort out, and hence, my view on '27, right, and '28 and beyond. But I fully agree with you that there'll be a lot of proof-of-concept trials earlier than that.

The next question is from George Notter at Wolfe Research.

This is George Notter. My question is just on the enterprise initiative. As I kind of look at your expectations going forward, think about some of the margins, I'm kind of wondering if there's like a big push into the channel here. Obviously, the channel is fairly expensive in terms of margins. And it's been something I think you guys have looked at and wrestled with for years and years. I'm just wondering if there's any kind of a pivot here and how you think about the channel.

The wrestling is real, but it's top of mind for sure. We've had kind of an early days channel effort for a little while, but at the beginning of this year, put out a program that has gained real traction. And we're starting to build some momentum there, not just in processing deals through systems integrators, but actually seeing that demand gen come from the channel and true deal registrations. And the real power of that, and I think the value to us, is using this as a force multiplier on our sales team so that we can keep our cost of sale down and just put so many more feet on the street. And that new logo deal registration is incredibly valuable to us for exactly that reason, helps us keep that cost of sale where it is while driving up the coverage of the sales team and letting us start to approach accounts below that Global 2000 that we cover so well direct. As you start to creep down there, it's true that there's a little bit of margin that hits the channel, but the discount -- benefits in discounting tend to more than make up for that. So we still see the enterprise business as margin accretive channel.

Just as a follow-up. Is this like a full frontal assault on Cisco everywhere and anywhere through the channel? Or this is measured certain partners, certain channel strategies? Or it creeps over time in terms of your intention and how deep you get into the channel? Or how do you think about this pacing?

Go ahead.

Go ahead.

Cisco looks at itself as a technology provider beyond networking. And so I don't know if you want to talk about a full frontal assault, but there should be no network where we can't provide a best-in-class solution at Arista. And any way to reach those customers, we're going to go after that. And if that's a full frontal assault, then yes, that's what it is.

Yes. And the only thing I want to add was outside the enterprise, just back to the margin question to your question, George, is that, yes, so some is enterprise. It still stays within that 5% to 6% range of revenue, so more dollars but same percentage. There's actually probably a bigger range on the R&D side in addition to the enterprise Todd talked about on all the other innovative things we'll need to do to ensure we hit that top line just to kind of round out the margin differential.

We'll move right next to George, to Samik Chatterjee from JPMorgan.

Maybe if I can ask you on blue box, and you tried to highlight the sort of what you're trying to do on that front. But how should we think about the relevance here in terms of what's the mix today of that opportunity? And as you look forward, is the relevance of that product going to increase? Is it more required to compete with white box? Or actually regain share to white box? And just a quick one, another one, just in terms of the partnerships you highlighted, OpenAI, Anthropic, are those opportunities for blue box or more broadly sort of across the portfolio?

Okay. So first of all, we will always coexist with white box because there's a business model there, whether it's 10% margin or just a basic ODM where they're not looking for the premium, the value, the features. So this is not an attack on white box. This is how do we coexist with white box. But as you know, there are a number of our customers who just love our performance, love our features, but might need the flexibility of not always requiring every bell and whistle of EOS. And particularly in some users, like if it's a leaf switch that's just connecting general purpose compute or use case is very simple and they're just doing some layer 2 functionality. And so we have already installed blue box with NetDI with some of our largest Cloud Titans in certain use cases. But we can see that expanding to some of the maybe specialty Tier 2 providers where they want to play with SONiC or they want to do EOS and SONiC and see in their labs and see what the delta difference might be. And so to be able to do that without making an either/or decision and saying, okay, I'll do this. But I have a hybrid strategy where if I don't want this, I can always load EOS. I think it's very powerful. So they're not giving up something to get something. And obviously, the pricing won't be as cheap as white box, but it won't be as expensive as EOS either. So the choice model, the economics model and also the total cost of ownership, if you add the CapEx and OpEx will be very favorable to the blue box for simpler use cases. That's kind of how I see it. And you asked the question on the partners. Obviously, you wouldn't expect these partners to stand up and say nice things about us if they weren't working with us. So every single one of them is working as an ecosystem partner with us, and we expect to do more and more with each one of them.

We'll go to Meta Marshall from Morgan Stanley.

Great. Maybe one topic that wasn't touched on today is just a lot of talk in the atmosphere about OCS and just kind of how you see kind of OCS versus your opportunity and kind of the opportunity people have been talking about there. Maybe I'll just stick to one question there.

Meta, I was asked to specifically not speak on that topic. This is an Arista Analyst Day. So we'll stick to Arista topics. But look, listen, I think you all know, Oracle is part of our Cloud and AI Titan category. They've been a very important and strategic partner and customer and will continue to be. And we look forward to a multiyear vibrant partnership with them. I'll leave it at that.

Next question is Michael Ng from Goldman Sachs.

I wanted to ask a little bit about the traction that you're having with Neoclouds. Chantelle, during her presentation talked a little bit about there will be a point in time when those conversations will occur where they're seeking best-of-breed. What milestones are you looking at to measure your traction with these types of customers? And just a quick follow-up, if I could. The gross margin range over the long term of 60% to 64%, a little bit wider than you gave at the last Analyst Day. Maybe you can talk about some of the thoughts in providing that wider than historical guidance.

Yes, sure. Happy to help. And if anyone has any comments, jump in. So from the Neocloud, perspective, I think there's a few things we're doing intentionally. One is to ensure we're going out proactively to understand where the opportunities are from a global perspective because a lot of these conversations are global and international. And fortunately, we have some coming to us, first, because they understand, and Tyson, you could probably speak to this, all of the great experience that we have with the hyperscalers, they want that. They want to understand Arista, how do I get that quickly? How do I get that outcome? How do I get that performance?

Yes. If I can jump in on that. I think that's exactly right. What I see what I'm talking in Neoclouds is the vision, the desire to operate at cloud scale without having built that 1,000-person team building full custom management solution. So they need help with the automation. They need help with the Arista validated designs, with the CloudVision framework because they don't have the time to build all that themselves the way the hyperscalers have. So we're seeing very good traction there.

Yes. I think the only thing I would add on that is if you look at it then through a kind of a similar lens to full frontal assault, I think we want to win every one of these deals. We have the stack to do it. And there's recognition of that, I think, amongst all the folks who are investing in AI, whether they're Neoclouds or even enterprises. And it's for all the reasons Ken talked about, but it's also all the hardware innovation that was talked about earlier. I can say from firsthand experience, making selections around who your suppliers and partners are going to be. For me, in prior lives, it has generally been pretty easy working for big companies, and you take that experience and it's accretive to future experiences. That doesn't go unnoticed amongst industry players. And I think that helps build momentum. So it's the credibility of the team that's doing a lot of work here. And I think people are getting wise to it. I think we're evolving kind of how we're thinking about it and engaging here. And I think, as I said, we want to win all these deals. So we're going to go out and get after it. And I'll let Chantelle talk about the margins?

Do you want to take that question? Okay. So I think for the -- so the margin range you spoke of, the 60% to 64%, that's absolutely to allow ourselves room, at this point in time, this far away from that time period, to allow us to see how far this AI mix will be cloud, Neocloud, AI mix in that number. And that's exactly what that's giving us room for, nothing more than that. It assumes the current tariff scenarios, et cetera. So it's really just that mix. How big will AI be cloud AI because that has a different mix than enterprise AI in those years.

We'll go to Tim Long from Barclays.

Sorry, sticking with the AI kind of related to that last question, and maybe I'll go back to Meta's question. The -- very impressive growth, could you talk a little bit about kind of distribution and how diverse that revenue outlook will be next year? Obviously, you have a few very large Cloud Titan customers. Just curious when you're looking at the growth into next year, do you think the AI bucket will be more diversified? And I think the question before, just related to that, curious your outlook on optical circuit switching and the impact on the spine part of the network? If you could touch on that as well.

So I'll take the first part. Jayshree, you want to take the second. So for the first part, the diversification, I think your question is which kinds of customers will be delivering that AI target of $2.75 billion in 2026? That's going to be a combination of the ones that you know the large hyperscalers. But if you recall back to our last earnings in our prepared remarks, we talked about 25 to 30 enterprise Tier 2 AI customers, and that list is growing. So it will be part of the hyperscalers you know, and part of this growing kind of enterprise Tier 2 market, and we're very excited about that. And then Jayshree, do you want to take?

And I do want to add that while these 25 to 30 will be meaningful, the large hyperscalers or Titans, as we call them, are going to be significant contributors to that 2026 number because they're building such large clusters. Coming back to optical switching, gosh, it's been around for a long time, and we're certainly aware of one customer that has deployed that. But we're not seeing it as mainstream. In fact, I would go the other way and say customers are looking for pluggable choices, whether it's pluggable optics or pluggable copper and not locking themselves up into one type of technology. So we don't see that as the mainstream way to scale up or scale out. We do see electrical switching going all the way. And then when they're trying to determine distances and flexibility of Layer 1 capabilities, certainly different types of copper and optics come in, but I wouldn't say that's a predominant architecture.

We'll go to David Vogt from UBS.

I have 2, if you will. Can you -- Jayshree, can you talk about -- Andy talked a lot about power consumption from your customers. What are the practical implications from a competitive dynamic perspective? From a power perspective, if we go from 3 tier to 2 tier, pluggables to CPO, LPO? Kind of talk about where you competitively see yourself performing relative to your peers? And then the second question for Chantelle. You mentioned very small contribution from these 25 enterprise customers. But over the longer term, what are the road maps -- what's the hurdles or what's the road map for enterprise to be a much bigger portion of the mix, particularly on enterprise AI, not on the campus side, but more on the AI side?

Do you want to do that first?

I'll do the first one, yes. So you should think about a 3-tier network being roughly in physical space and power consumption, 20% to 40% less efficient than a 2-tier using the scale-out solutions that Arista has. So that's a big deal. Back to optical circuit switching as well, the customers want to spend every penny that they can on GPUs and nothing else. And so the pressure is there. Power optimization, floor planning, space planning, density of network. And that is where Arista does really shine. I mean I think you can go across all of the competing solutions out there, and you can drive and fine-tune, but like LPO is not broadly qualified. You want to do white box and you want to do even a small cluster. Those savings are meaningful for those customers, and they're waiting on the ecosystem to catch up on the software side to be able to enable them. So we invest a lot in qualifying optics, pushing ahead, driving all that efficiency. The 20% to 40% matters that those are deal-making decisions, right? They look at those and they say, okay, that's a difference maker. So we think we favor -- or we fare very well against competitors in that regard.

And just to add to that before Chantelle goes, in my LPO slide I showed, should you, on top of reducing the tier, add linear drive optics and get rid of the DSP if the distances allow, there's another 20% of power savings of $3 million to $5 million a year. Add that up, it adds to a lot of money, right? So -- and that's based on a -- I think my slide showed it was 1,000 switch configuration, which is quite small actually in the large scheme of things. So you can get a real collapse of OpEx by reducing tiers and a real integration of optics to save money and power in a very significant way.

And I think that -- at least I'll start, and I'm sure Todd will have a view specifically on campus and AI. But generally, for enterprise and AI, I would say what we hear from our customers are just some growing pains. There's definitely intention even so far as some mandates. You hear some CEOs and some of the enterprises mandating an AI outcome or an AI implementation. So what do we hear and see? We hear and see, is it on-prem or off-prem? Is it in the cloud? Are we going to have it here? Is it training or if it's inference? Is there a specific ROI that the Board is looking for? And all those things just take growing pains take time. So we see intent. We see mandates. We see that there's a little bit of confusion, congestion on how to make that happen. But there's definitely intent in the conversations. If you listen to -- I'll give you some examples in the education university systems, big adopters of AI. The banks are definitely -- you guys probably know this more than I am. The financial institutions are looking at AI. Health care is looking at AI. So there's definitely intention there. But I do think that there's a little bit more scrutiny on the ROIs, on the use cases, probably expectations in some of their future years, what they expect to be cost out with employees for AI. So I think that's where the rubber will meet the road. But do you want to talk about at the edge, maybe?

Well, I would just say on the enterprise AI deployments, we are engaged in many, many of those conversations, but largely, the enterprise is making a build-versus-buy decision. And these are still early days. They can start off by buying AI capacity from the Titans, from the Neoclouds and make those decisions later on. We plan on being the premier solution for them whenever they choose to get going. But they're making a build-versus-buy decision right now.

One quick comment I can't resist on the 2 tier versus 3 tier. This is a topic we've been wrestling with since, I don't know, 2010 or something like that. I mean our success in the hyperscalers is very much related to the success we've had building these very wide, very flat networks because 2 tier is so much cheaper than 3 tier. And so if you get a high radix switch, if you get the modulars that have the internal fabric, our distributed Etherlink switch and the rapid reconvergence routing stack, these things work together. I think Arista just has -- I mean, AI, of course, on both front end and back end requires the same architectures. And so it fits really nicely and something we've done for many, many years now.

We'll go to next to Tal Liani from Bank of America.

I'm not going to ask you about the margins. I want to ask you about blue boxes. Does it open opportunities with customers you are currently not serving? Meaning customers who are buying white boxes, philosophy for white boxes, can you penetrate to these customers with blue boxes? And second is how much of an opportunity -- you touched on it, someone asked here on blue boxes. How much of an opportunity for blue boxes you see outside of the big Cloud Titans? Is it at all addressing second-tier baby clouds, even the enterprise? Or is it strictly or mostly for large cloud?

Thanks, Tal. And you can always ask me another question on margins. The -- I think the way to look at blue box is there's no doubt that if people have the staff, it's a lot easier to do, right? No question about that. So it's naturally appealing to our large Cloud Titans and AI Titans for very specific use cases, and this is why you've seen us largely be deployed there. I think there is an intriguing element for some of the smarter enterprise staff or Neoclouds where they want to try that for choice of flexibility where they want to sort of lean in on their expertise. They may not have the stuff to do that, and we are starting to see that. It's not meaningful yet in numbers, but it is an important sort of innovation area and sector for us so that they have choice of flexibility. I have, in my mind, a very good example of a customer who started with the white box and couldn't get it to work and has now adopted our blue box. And it all happened in a span of 6 months. And it will probably show up. Now they're in lab trials, and we'll see the deployment next year. So when it happens, it can happen either because they are already an existing customer. But in this particular case, it was a call to me directly. They said, we can't get this to work. Can you help us? We helped them. They jumped into it immediately. So I think there's a recognition that Arista is only doing premium, and white boxes are only cheap, and nobody knows of us yet in this hybrid state, and we're hoping there'll be more and more of those customers and use cases.

Is there a target for the contribution of blue box?

Is there a target for the contribution of blue box? Is that what you said? It's in my $2.75 billion target. How much of that it will be? I don't know yet. We'll see. Please refer some customers to us.

We'll go to Simon Leopold from Raymond James.

I wanted to ask about your purchase order commitments because over the last several quarters, they've been growing significantly. It doesn't look like we're in a supply chain crisis. So how should the analysts think about it? I think in the most recent quarter, it was up more than 70% year-over-year, and you're only forecasting merely 20% growth. How do we align these commitments? How much of this is safety net? What constraints are you facing? How should we think about those numbers?

Yes, it's a great question. Thank you. So I would say some of it's a safety net, some of it is to ensure we have the capacity, but that's not the majority. But you have to think about the -- and I think, Jayshree, you were referring to this earlier. The purchase commitment to transaction to acceptance and showing up at Arista can be a few year journey. So is it indicative of future transactions? Absolutely, else we wouldn't be doing them. But the timing of it, you just need to be a little cognizant of and the sense of it's not 12 months. We're talking 24, maybe 36 at the outset, but it's a longer lead time into ours showing up as revenue.

That's a very good point, Chantelle. In addition, I'd just add that besides the multiyear cycle, so that's not a 1 year. Many of our components have greater than a year lead time. Many are brand-new products, and we can't buffer them enough as the customer suddenly wants them. So we're trying to lean into more satisfaction of lead times, particularly in the enterprise and campus. It's easier to plan in the data center because we know they have something going, but it's harder to plan in the enterprise and data center. So you'll see us leaning in on more investments, in purchase commitments, both in the AI, where things spin up suddenly and in the enterprise, where there's more of an expectation of shorter lead times.

We'll go to Atif Malik from Citi.

I just have a kind of a dumb question. When Broadcom talks about SUE, or Scale-Up Ethernet, and you guys are talking about Etherlink, they both have ether in it. Can you talk about the difference? And are they trying to take the customers a different path than you guys?

Yes. So Broadcom's SUE, Scale-Up Ethernet, is a concept spec on how Ethernet can be scaled, and we are a huge supporters of that. Arista's Etherlink is our AI portfolio for scale up, scale out and scale across, so we can optimize all of our features and our hardware for Etherlink. So it's our branding name for our portfolio of products using Broadcom chips and Broadcom implementation.

I want to comment on one specific technology, which is the distributed Etherlink switch, which is a scheduled cell fabric. And that makes use of a very different technology than Scale-Up Ethernet. So Scale-Up Ethernet is all about optimizing latency, minimizing the packet sizes, cutting the delay down to nothing because the back and forth between these components within the GPU chassis is everything. But the distributed Etherlink switch is all about scale out. It's all about how do I get the most nodes connected through a shared fabric without creating any bottlenecks. So they're actually very different technologies, but addressing different parts of the AI use case.

Any other questions in the room? Right over there?

This is Yang Pu from BNP Paribas on for Karl Ackerman. I have a question about your CPO strategy. So I think you recently discussed maybe probably last week, you discussed that you are agnostic to CPO, and you can provide it if customers want it, which I interpret that as you have the ability to make CPO switch using merchant ASIC probably from Broadcom. I believe some of your peers are already doing that. But just before Andy was talking about LPO and alluded that CPO switch are vertically integrated. So can you maybe help me understand that how you -- what's your CPO strategy? And what's your progress on that? Are you making any samples?

We thought we'd bring Andy back to answer the question. Go ahead, Andy.

We are nonreligious about CPO, LPO, whatever it is. But we are religious about one thing, which is the ability to ship very high volumes in a very predictable fashion. So to put this in quantity numbers here, the industry expects to ship something like 50 million OSFP modules next calendar year. The current shipment rate of CPO is 0, okay? So going from 0 to 50 million is just not possible. The supply chain doesn't exist. So even if the technology works and can be demonstrated in a lab, to get to the volume required to meet the needs of the industry is just an incredible effort. I mean if you think of the math on how many modules to make every day, every hour to hit this 50 million going to 100 million quantity, it's surreal, right? And it works in the context of pluggable modules. The industry on this chipset is not ready for this. Okay. I'll leave it at that.

Thank you, Andy. Well said. And I think the other piece I'd add to what Andy said is, look, we're all about standards. If you look at today's CPO implementations, you got one version from NVIDIA, you got another version from Marvell, you got a third version from Broadcom. So that's what CPC, co-packaged copper, and pluggable optics give you and deliver all these capabilities richly without locking yourself up into one implementation, which our customers care about. So we'll embrace them all when they get a little more mature and ready. But...

And when customers want them. I mean, like...

There you go. That's a minor problem.

The customers aren't really asking us for them. They like our solution. 200 gig SerDes is fine. We can scale it. It's reliable. They know how to operate it. I think that customers struggle with the CPO paradigm because from an operator perspective, the -- what we call the blast radius, right, the unit of failure and the size of that failure increases. And there's genuine concern about that. So I think a lot of the early CPO interest is still in the trials, understand it. The supply chain has to catch up. And if the customers get comfortable with it and say we want it, obviously, the rest of the industry is going to get behind it, including us.

And I have a follow-up here, if I can. So I'd like to ask about your '26 target from a different angle. You have this massive $4 billion deferred revenue, including $1.8 billion product deferred. And let's say, just the product deferred gets delivered in 18 months and the service get deferred in 3 years. So if we do that, you already have like 20% revenue growth locked in.

Are you assuming all $4 billion comes out in '26? You should.

Well, if we just think like the...

100% is not going to empty out, right? Just so be thoughtful about that.

Yes, right. Not emptying out, but say product deferred is...

There'll always be -- I think the important way for you to think about this is deferred will come out, deferred will go in. There will always be a net deferred. And I realize you think it's particularly high now. And if we emptied all of that, why isn't the number $15 billion or whatever you want it to be. But let's be realistic, right? Let's be realistic that there'll always be a deferred short term, long term. And let's be also happy that we are agreeing to grow at $10.5 billion revenue 2 years ahead of schedule. If we can do better, we absolutely will, which is a sign of more customer momentum. But I think that's too early to tell when '26 is 4 months away and entire '26 is 15 months away. So we'll see how it plays out. But deferred will come in, deferred will go out next year and the year after and the subsequent years.

We have got time for one last question.

One last one. Okay.

I get the last question? It's got to be an important question.

Better be a good one, Nick.

All right. So first of all, really great basket of innovation. So really congratulations to like all of you guys. But I do have one question on the NetDI in AVA. I think it's like one of the best innovations that I heard today. But the question is, why centralize all that data? Why didn't you do an Agentic model where you put agents into the switches and have AVA interact with the switches versus centralizing? I'm sure you have some really great reasons for it, but like that's just curiosity.

Yes. The reason for that is because the decision-making around what an AI agent is going to do depends on the full network state. You just don't have the perspective from one device to know what's happening in the overall network and what actions make sense to take there. We absolutely give AVA tools to reach into a switch. In fact, you saw in the demo where one switch was sending pings across the network to another endpoint. So we -- the AVA agents are able to take action from the point of view of a given switch when that's what's called for by the overall scenario. But actually executing out there isn't needed or helpful because of the fact that full network context is required to make good decisions.

And Nick, just to add to that on NetDI, NetDI is fully distributed. It's not just centralized. So every switch has an expression of a diagnostics infrastructure. And then in an aggregate, we can manage it. So I think you need to think of it as centralized management, distributed data and control. And so there is a lot of mini NetDLs or NetDBs sitting in the switches and NetDIs sitting in the switches. But to take any kind of sensible action, you've got to do it across the entire scale of the network. Thank you, Nick. It was only fitting that you and I talked about Leaf, Spine, and now we end with you on beyond the Spine.

All right. Thank you, everyone. That concludes our panel, and it also concludes our Analyst Day.