Good morning, everybody. Tim Horan. I am the Digital Infrastructure and Communications Analyst here at Oppenheimer. My pleasure to be hosting Stan Kovler from Extreme Networks. He is the Senior Vice President of Finance and, importantly, Corporate Development. We're going to be able to ask him all sorts of really interesting questions. As always, an exciting time in the industry and never lack of volatility in the stock, Stan. Thanks for that. Keeps us all busy. Yeah. It's small cap tech, so that's kind of the world that we live in. That is the world that we live in. The most important thing that matters, just what's the long-term growth rate and what's the long-term free cash flow, right, which we're going to get into. Yeah. Before we get into that, what's the biggest trends going on for enterprise networking at a really high level? Yeah. So, Tim, obviously, it's been an exciting place to be recently. A lot of our customers are upgrading to all sorts of next-generation technology. On the wireless side, you've got companies that are transitioning to next-generation wireless Wi-Fi 7, and that's bringing forth a lot more bandwidth, a lot more power, more deterministic connectivity for users, and we're taking that now to the next level. We've got some really interesting, unique antenna technology that we've developed and are introducing to some of our largest customers. Our large venue customers are really loving some of these capabilities that are producing really great results for them at the stadiums. On the switching side, it's preparing a lot of our customers for next-generation technology, and over the next three to five years, they'll see a lot more traffic flow on their networks from not just users and the prevalence of additional devices that come on the network. That's been the driver for us historically, but it's this notion of the agent-to-agent workflow that is on the come in our industry, and that will be another driver of growth and complexity. As networks become more complex, that drives the need for a lot of our automation software technology and the AI-based platform solutions that we're delivering to customers now. As you're well aware, we have a platform called Extreme Platform ONE that we developed and launched last year, and we keep upgrading this. There's another upgrade coming later this month, where it'll create even more automation and more introductions later in the year where it's not just focusing on devices anymore, but it's really focusing on making the network engineer's life easier and allowing agents to connect and work together and solve these problems more autonomously. That's a lot of what's happening in the network right now. We're mainly benefiting from driving AI for networking. It's not networking for the AI build-out, but it's allowing our users and customers to use AI to make their life easier in a very complex network world. AI, we have a lot of things going on. We have fundamental major improvements in networks just from Wi-Fi 7 and switching we can talk about. Then we have lots of new, hopefully potential use cases for the networks with AI- Right. whether the enterprises want it or not, because it seems like we're going from people-to-people communications to machine-to-machine, which could be orders of magnitude more traffic. Then we need ways to basically manage all this in a much better way, which your AI platform is kind of doing. Maybe we can just discuss each one of them, take those one at a time. What are you seeing in terms of just improvements from Wi-Fi 7 in terms of amount of capacity, latency, features, and use cases? I think the same, and if you can discuss some of the same thing on your switching side. What type of fundamental improvements are you seeing? Yeah. On the Wi-Fi 7 side, again, what we're seeing is just being able to deliver higher bandwidth applications to users, and in some cases where it was not tenable. Wi-Fi is just getting more predictable, it's getting more reliable, and it's getting to a much higher level of reliability and throughput for a lot of our customers, and that's what's driving the upgrade. A lot of our customers, they never went to Wi-Fi 6 or Wi-Fi 6E. These are customers that are upgrading from Wi-Fi 4, Wi-Fi 5, and making that leap all the way to Wi-Fi 7 now as the next generation is developing some of these capabilities. Now, the technology itself is interesting because for a long time, Wi-Fi operated at two bands, the 2.5 GHz and 5 GHz bands. As you get into 6E and 7, now we have a completely new, almost like HOV lane, if you want to think about it that way, with 6 GHz. That opens things up for other use cases and new technologies, new capabilities where you can run things on a clean spectrum. It's all unlicensed spectrum, obviously, because you operate in both your coverage is licensed and unlicensed companies. But it really opens things up where you can use that- I think we've almost doubled the amount of spectrum that Wi-Fi can operate on, and obviously the processing power is getting better and better, and some of the other underlying technologies. With Wi-Fi 7, are people seeing a quadrupling of, you think, actual useful capacity, or any way to measure that? Yeah. It just allows the users to do more video. A lot of what we're seeing from our stadium customers is that people are live streaming, they're recording video and sending video, and sharing where they are. It just supports the users and those types of experiences where you can be at an event and then share the event back home to people that are not with you, and then multiply that across, you're at The Swamp, you've got 90,000 fans that need concurrent access. The other trend that we're seeing is that in the past, not everyone was connected at the stadium, and you might have 20%, 30% of the people that have been connected at the stadium. Now pretty much everyone is on Wi-Fi at a lot of these events, and so you need that to support the users. That's at the user experience level. There's also, if you go to the events or I go to events, everything's getting more expensive. I think there's a lot of pressure on venue operators to make things more interesting, delivering additional experiences or driving more automation. It could be as something simple as ordering food or concessions to just multimedia activities. Also, facial recognition when you're entering a stadium for security reasons, for access. Video is becoming more important, security is becoming more important, and in order to support 4K video, then you're going over to the other side of our business that you started talking about, which is the switching part. That's driving a lot of the upgrades as well. Speeds and feeds are very much relevant in high bandwidth video, and you need to be able to deliver that. We're also seeing networking in a lot of places where it's never been. Every time you go through the supermarket checkout, you're on a network because the computer is networked back to the warehouse, and the processing, and all of these different systems have to be connected. When you go through the airport and you have the facial recognition taking a snapshot of you and recognizing you as a passenger that's allowed to board the plane- Stan, just back in the stadiums- Those use cases. While you're there, if you upgrade a brand new stadium, we'll upgrade a stadium, let's say Florida, with Wi-Fi 7. How many hotspots are you putting in? Thousands. Is there enough capacity for everyone attending to use it in a reasonable manner? That's exactly right. What's happening is that a lot of us are at the event and we have our cell phones. Wi-Fi is still a lot cheaper to deploy than 5G, and in most cases, we are partnering with a telecom operator, and they are sharing the cost of that deployment because for the operator, it means being able to deliver the service. Which, by the way, they can even white label as 5G, but on the back end, it's running as Wi-Fi because of the savings in terms of the deployment of a hotspot relative to a cell site. That's really the benefit that Wi-Fi still has. Even at the cheapest small cell, Wi-Fi could be three or four times cheaper to deploy at these venues. As you know, it's not consistent usage because you have an event, and it could be happening at 7:00 at night from 6:00 A.M. until 6:00 P.M. There's virtually no usage. So why deploy a lot of this expensive 5G technology at some of these sites? That's what a lot of the Wi-Fi use cases are doing for the customer. Well, you bring up a good point. The Swamp, they probably only have maybe 10 major events a year, where if you go to MetLife Stadium, it might be 80 major events a year. Yeah. It's very expensive to deploy all that 5G equipment with such low utilization, and Wi-Fi, to your point, is a fraction of the cost in many ways. Right. Exactly. We touched on the use cases, which was the second part of all this, but are you starting to see or what do you guys expect what AI. I guess facial recognition is obviously one example of it, and now you have automated concession stands where basically they look at your face and they will charge you based what you take off the shelf. But are you seeing other use cases, maybe machine to machine or AI-based, that are changing the usage on the networks? Yeah. The usage on the network, it is also happening in manufacturing. Everything is connected, and there is a lot of automation on the factory floor. A lot of that requires upgraded networking capability, and sometimes you also need a network upgrade not just because of the speeds and feeds, but also because of security. Our fabric technology, which is our flagship technology, security is paramount, where you want to make sure that every product line could have its own virtual network, essentially, so that you cannot cross and, worst case scenario, if you do get hacked, that the hack is contained in a particular area, and it cannot go any further. That has become very important. With all the geopolitical things that are happening, network security is becoming very important as well. That is an important use case. There is also the notion, from a use case perspective, of for regulated industries and for governments, both domestic and foreign, related to sovereign cloud, where your cloud deployments are not on public cloud infrastructure. But you are doing your network management and running your operations on protected environments. Those types of deployments are also driving a lot of upgrades and driving a lot of demand, again, for regulatory data and security reasons as well. There has been a lot of. The carriers say you have to be on 5G standalone in these manufacturing facilities or stadiums because it is higher quality, better security, lower latency. I guess if you compare the Wi-Fi experience versus a 5G experience, is that true? It depends on the customer, but a lot of it is also just physically wired. It is not necessarily wireless. It is really wired capabilities as well that play into it and having everything managed from one console that gives you these capabilities when it comes to troubleshooting and automation and creating virtual networks or the network within a network. You need that cohesive solution that combines wired, wireless, and the management platform, where it is not just a discrete solution as let us just put some 5G SIM cards into places and run it that way. That is a really, really great point. Frankly, with Wi-Fi, especially in these venues, you are probably only going, like, 100, 200 feet where it is wireless, and it hits a wire right away, that hits your switching platform, and then it hits your cloud platform management system, Platform ONE, which is about to be, I think, Agent ONE. Is there going to be another add-on on top of Platform ONE, or are you going to call it just Agent ONE, or are you going to keep calling it Platform- Yeah, Agent ONE. Not to get caught up in the alphabet soup of the solution. The whole platform is called Platform ONE. Agent ONE is just our way of calling the next-gen solution. So we had a Gen ONE solution that was Platform ONE. It was more knowledge-based and people doing searches, just like how you and I probably use Gemini or Claude just to look something up and get a little bit deeper information on that, like knowledge worker or how-to. By the way, that was the first foray into giving people that additional comfort level of moving on from competitive solutions. Because one of the gating factors for us as a relatively smaller player is that there are larger, more established players that network engineers know how to use. With AI, you are lowering the barrier to entry, because now you are able to use AI to make changes on the network and help you where you do not feel locked into one provider as much as you used to. That is one of the things that is actually helping us in the market. Now, if I go back to your question about Platform ONE, think about it in terms of branding as the next phase will be Coworker mode, where a coworker will add more automation and capabilities to you. You can schedule tasks, and you can start to do some fun automation things on the network. There is another solution that will come later in the year that will get introduced in October, which is called the Operator mode. That starts to get into more of the agentic-to-agentic flow and automation of agent-to-agent workflows. It will also help and be licensed at the network engineer or at the enterprise level. That is no longer tied to a particular device and a device license for doing things. This is an add-on that is- What is the name of the add-on again, Stan? That will be Operator mode. Operator mode. Got it. Yeah. Platform ONE Operator. The Agent ONE is going to be a pretty big upgrade to Platform ONE. It will. Yeah. Because it's like when we first did some basic AI features, and then we all started to use Claude for our everyday use case that's now tied to, let's say, all of your Microsoft solutions. That would be the analogy, is that now you can do things and really dig in on a much deeper level and start to automate a lot of your processes. When is Agent ONE getting rolled out? It should be later this month for that. Congratulations. I hadn't really thought about the fact that your platform is also backward-compatible with all your competitors. It makes it much, much easier now for people to deploy and use it and make it backward-compatible. What do you think will be some of the low-hanging fruit? What will be some of the major productivity improvements they're going to see as a network operator? The biggest thing is going to be the troubleshooting and the speed with which you can reduce the mean time to resolution. That's the industry standard when we try to think about taking care of any network issue that comes up. It had taken a while when you do more of a manual troubleshooting process to look at the logs and see what events happened, what triggered some problems on the network, and to have AI do that and do those searches in an automated way, that will really improve the end-user experience if your networking team can resolve these issues very quickly. Just stepping back a little bit, the overall industry, it's a very cyclical industry, obviously. Do you think AI is going to drive just major enterprise network upgrades for a few years to take advantage of it? I guess related to this on AI, we talked about some of the use cases like manufacturing, obviously stadiums. Do you think they'll also look to bring a little bit more compute back on a localized basis? That's a big part of it, and that I think is a combination of what happens with campus and why the campus upgrade cycle will carry forward. I also think, in addition to AI, many of our competitors had multiple product lines, and they're still digesting multiple product lines. When customers qualify vendors, they typically qualify certain product lines. This is very true in the government and regulated space, where you have to go through a public tender or request for proposals. As those competitors of ours try to get their other product lines qualified by these government entities and a lot of these regulated customers, they will run into having to open up these contracts for bid. When those contracts get opened up for bid, that gives us a nice shot on goal. With us and with our partner expansion plans and working more closely with the partner community, this is going to open up more and more opportunities for us. That's on the campus side. Your point about the data center side is also well taken. On that end, we're actually making investments this year. Some people, especially after earnings, have asked us, "What are you guys spending on when it comes to your OpEx?" Obviously, we're spending a lot of money to develop Extreme Platform ONE. We're also developing a lot of our data center solutions to take it to the next level, where, whether it's 400, 800, and then so on and so forth, over the next several years, we do feel like that trend of bringing compute and AI workloads on-prem will drive enterprises to invest more. Where we are well-positioned is that the management, like you said, not just our campus capabilities, but Extreme Platform ONE can be used to manage the data center switches that we'll develop. Now it becomes a single point where you don't have to swivel chair to multiple applications and multiple management consoles. It will be a de facto solution for you to manage your campus network, and then you can manage your data center network as well. By the way, several years from now, when a lot of this is more in motion, we'll have Agent ONE Operator that you can host on-prem. It'll be a very cohesive solution where you take the AI capabilities, and then you can host it on hardware that you'll buy on-prem from us and part of that solution. Do you think you're in a position now to really accelerate share gains? Who can you win share from and why, I guess, yeah? Yeah, we think so. That's what underpins our confidence this year. Again, another year we expect double-digit product revenue growth. We still have a ton of share from a lot of our incumbents. We're a relatively small player, and we have some of our competitors that own half the market and others that own 15%-20% of the market, and we feel like we're in a very strong position as the market opens up in the manner I described. In general, as a lot of customers have products that simply age out, end of life, end of support that our competitors are frankly helping us with and unlocking the market in that fashion. That's what we're going to be going after. And you've had some improvements with go-to-market, I know, and some new products that are targeted at some new channels. Can you describe some of those new products and packages and how they're doing, yeah? Yeah. In packages, Extreme Platform ONE is basically a packaging of our support and our AI-driven subscription management capabilities. That bundle which came out last year accounted for 30% of our subscription bookings in FY26. In Q4, it was all the way up to 50% of our subscription bookings. What's happening in that line, and you'll see an acceleration as we go through FY27, you'll see, Tim, that SaaS ARR will re-accelerate. But the offset is as we're combining these product lines on the subscription support, that discrete support line basically goes away. FY27 is a bit of a transition year for us as we get these customers booked, and there's a nice buildup of the deferred revenue balance that you saw in Q4, and that will continue forward this year. And we also show in some of our disclosures that the mix of our deferred revenue is really shifting towards subscription. 57% of our deferred revenue last quarter was already from subscriptions. So as the customer base transitions to Extreme Platform ONE, and that'll go from about 10% at the end of FY26. Well, we're driving towards half our base flipped over to Extreme Platform ONE by the end of FY27. By FY28, we're kind of done with the transition of flipping the customer base over. Then you'll get that benefit of driving Subscription support revenue off the balance sheet, and the growth rate could potentially double from where we are in FY27, because the implied growth rate for this year in our guidance is about mid-single-digit growth of subscription support. That will re-accelerate as we can recognize revenue off the balance sheet. We're going through a classic transition that many companies have gone through from support to SaaS. Some of that is what people are seeing in our results and outlook today. That's what we're trying to just- Once that inflection point hit, do you think where the growth it'll accelerate in that line item? Yeah, I think that towards the end of this year you'll start to see that inflection. You've had some price increases lately because the product has improved dramatically, but also some of your memory costs have gone up quite a bit. Can you talk about what you've done, and have they stuck the price increases, and do you think you have more pricing power? Yeah, we feel like the pricing power is there. We're trying to be smart about pricing. Right now, we have a nice expanded advantage over some of our competitors that increased price three or four times. We've only done it twice over the course of the last year. We're also guaranteeing price to customers that register their interest and register their deals with Extreme through October and November. I think that kind of gives you a clue of our near-term positioning. After that, we'll explore what we do with pricing relative to the component pricing and component costs and freight costs and things like that that we're seeing out in the market. We're staying very close to all of those trends right now. It's a pretty dynamic environment when it comes to pricing. Having the supply, securing our supply chain, especially for memory for the long term, and also knowing where our pricing is for memory this year, that allows us to be very smart and tactical about how we think about pricing. Great colors. The last round, I think about two months ago, that pretty much stuck. That wasn't a problem on pricing? Yeah. We've seen pretty stable discounts. For FY27, we have some incentive plans for our channel and salespeople to hold on to discounts and make sure that the pricing that we are increasing is sticking. We are at the industry level in an industry that discounts very heavily, as you know. What the list price is, it's not like buying a car, MSRP is $50,000, and you're buying it for $49,000. This is more like MSRP is $50,000, and someone is going to be paying $20,000 for a product. It's a very heavily discounted product. It's just the nature of how our industry is constructed, and the industry leader essentially set that construct because the channel is a multi-tier channel. We have a distribution channel, we have a partner channel. That allows the channel in that two-tier construct to price up to the customer. Absolutely. Makes perfect sense. The other big trend in the industry, I think, is kind of bundling a lot more services together. How do the customers feel about that, and is that another form of a price increase on customers? Yeah, it's a great point. The bundling can be kind of what we're doing in terms of support subscription at the network layer. We find that some of our competitors are also trying to bundle unrelated things together. So we serve a very particular market need, and everything that we do is focused on networking. When you start to bundle things in that are more security-oriented pieces, but unrelated to the network, then customers have to make a choice because they have to choose either best of breed or bundled solutions. We know that customers choose a lot of bundled solutions because obviously we've got companies with large share in our space. So it works for some customers. But when you try to bundle too many unrelated things, then it'll drive the customer towards more pure-play, best-of-breed solutions. That's where we can come in and partner with some of the other security companies or other hardware companies, and our channel partners can construct a package for customers that meets their needs. Very helpful. Thank you. You hit also on you guys have the memory supply. Do you think some of your competitors maybe are short inventory? In some of their product lines in some of the cases, they are. We are taking full advantage of that by having better lead times. Some of our competitors have challenges in their lead times getting certain products, and that is certainly helping us in terms of additional share gains. With memory prices going up and prices going up, what do you think is going to happen to the buying dynamics? Are people saying, "Look, I will hold off and I will upgrade the networks in a year or two when prices come back down," or are they worried price is going to keep going up and maybe they are accelerating their upgrades now instead of waiting? What do you think is the dynamic? Well, the dynamic that we wanted to create with this price guarantee is to allow customers to plan on a more natural cycle. We do not want a feeding frenzy where everyone buys over the next month, and then we do not have any demand over the next year. It is also not great from a supply chain planning standpoint, where we cannot supply everybody tomorrow. There is certain lead time to building products and making sure we can meet customers' needs. We think that the guarantee helps from a planning perspective. We also remember the lessons from the previous supply chain-driven cycle where backlog just went to untenable levels, and then it was just an unrealistic amount of backlog that customers put in orders that they did not really need. We are trying to manage a lot of these things and use the lessons of the past to guide us going forward. We usually put backlog numbers into our 10-K, so as our 10-K comes out later this month, you will see that backlog is at very reasonable levels. Yeah, it is improving somewhat, but it is nowhere near the levels of what you have seen in the past. We feel pretty good that buying is more at a steady pace, and that we are seeing measured growth, which is how we like it. Just on the backlog point, do you force customers to put down any kind of deposit when they order? It turns out the backlog was pretty meaningless, right, three years ago. Is there a way to make the backlog more real? That's exactly right. I mean, if you just put in an order, right, it's like they put in their orders. That's why we want to make sure that this gets controlled. A lot of it happens, like I said, through two-tier distribution channel. We're not seeing like, for example, this is not a deposit on a new car, where you put down $1,000 or $100 on an R2 Rivian, and you're going to get it in six months. We want to meet the demand where it is today. We're not letting customers give us a small deposit and then take delivery in March of 2027. That's not the customer behavior we want to drive. Got it. Fair point. A lot of enterprises have been blowing through their IT budgets because of AI, token maxing, I guess. Do you think that's also maybe impacting enterprise upgrades? I know they got to go together, but like- It's a great question. It's a great question, and I can put my finance hat on for a second and tell you that it comes from two different parts usually. Because when I pay for cloud usage for our employees, that comes from our OpEx budget, and it's the direct manageable spend that I manage from OpEx. Then if my hardware team says, "Listen, I need a new network switch, I'm doing an upgrade," or, "We need some equipment for our labs," then that becomes CapEx. Those are two somewhat different budgets. I'm watching my direct manageable spend and OpEx a lot more closely than I am CapEx because it's different dynamics, different useful life, different depreciation, and how I think about my P&L. I think the difference here is that when you are buying a lot of this hardware and networking hardware and tapping into that CapEx budget, it has a slightly different effect than the OpEx budget. That is a really good color. It is a chicken and the egg, right? Without the network, you are not going to be able to use AI and vice versa. What you just described on the CapEx versus OpEx, you think that is a real thing for companies as you go in and talk to companies? I think so. Again, we see it in our own operations, so we can see it like as a guide post for how others are thinking about it as well. There is your CapEx budget for hardware, and then there is your OpEx budget for day-to-day use and equipment. Everything you can kind of tell in this cycle here, obviously we went through a bit of a COVID cycle, had a bit of a lull. The industry has been cyclical historically, but AI seems to be a very different animal in a lot of ways. We can just basically automate and we can have robots all over the place. Those stadiums will have robots in all sorts of different forms, manufacturing, self-driving cars, you name it. Where do you think we are in this cycle, and how much longer can it last? Yeah, I think the use cases keep growing. To us it feels like a long-term cycle, both in terms of the hardware, but then also, like you said, we'll need a lot of automation. It drives the need for the hardware investment, and the hardware's going to live for some time. Also how you run the hardware and automate a lot of the processes is becoming important as well. A lot of installed base will need to get flipped over. Then there's the greenfield growth that you talk about as well. It feels like a pretty nice cycle to go into and gives us a lot of confidence. We had our investor day last year in November, and we talked about 7%-9% product revenue growth. We're exceeding that. Even if you think about 2027, we're operating at above that. We told the market. It feels like a little bit of that super cycle to us. Again, for us, it's a function of how subscription support transitions from 2026 to 2027, 2028. It does feel like a nice long-term cycle of growth, and the top line is translating to 20% plus growth for us. We get some nice operating leverage on top of that. Do you think the subscription and support, it's improved so dramatically that I'm going to make it up, maybe I can cut my network personnel in half, right? Maybe there's a lot to- It's a lot of cost avoidance. Yeah. Cost avoidance. I do not know that you will cut a lot of people on day one, but as you grow and as your needs grow, it is that cost avoidance story where you do not need to add personnel as your business and your network expands. You think that is meaningful enough that you can get a lot more pricing power on the SaaS side of things? Yeah. Because you are saving relative to a full-time employee that you might need to add, and the incremental cost of paying for the AI is a fraction of what it would be to make these human capital investments. Yeah. Oftentimes we hear from many companies that for every dollar they're spending on AI, the customers can save $10 in labor or other costs. That's correct. Do you think it's in that type of range? Yeah, I think that's pretty reasonable. Well, Stan, I always love talking to you. Exciting times. I really appreciate the time. Anything you want to mention in the last minute here that we didn't talk about or just the last pitch on the stock here? We feel really good about where we are as a company. Feel great about our outlook. Thank you for hosting us. Really appreciate having you and the strong interest that we're seeing in our story. Thanks for all your help, Stan. Good luck. Thank you. Take care, Tim. Bye guys.