Thank you, everybody. Welcome to the F5 Fireside Chat at the Goldman Sachs Communacopia and Technology Conference. I have the privilege of introducing François Locoh-Donou, who's the President and Chief Executive Officer of F5, and Cooper Werner, who's the Senior Vice President of Finance and incoming CFO. François has been at F5 since 2017. Before working at F5, he served as Chief Operating Officer at Ciena. Cooper has been with F5 since 2001, serving as the Senior Vice President of Finance for the last 12 years, and as I mentioned earlier, is slated to take over as CFO. My name is Mike Ng. I cover F5 and networking and hardware here at Goldman Sachs. We have about 35 minutes for today's presentation, inclusive of investor Q&A. First, François, Cooper, thank you so much for being here with us. It's a privilege to have you on stage. Thank you, Mike. Thanks for having us. Thank you. To start things out, François, I was wondering if you could just talk a little bit about some of the issues that customers are facing with multi-cloud environments, and how F5's portfolio of application delivery and security solutions are able to help your customers? Mike, if I could- Oh. Real quickly, just get the Safe Harbor on record. Sure. Please note that our discussion today may contain forward-looking statements, which involve uncertainties and risks. Our actual results may differ materially from those expressed or implied by these statements. Please see our SEC filings for more information on these risk factors. Thank you, Cooper. Mike, let's get to your question. So, you know, here's what we see that large enterprises have been dealing with for the last several years, and even as accelerating. The digital footprint for enterprises is growing pretty quickly. They have more applications, more APIs. They have increasingly more data that they have to deal with. And in the face of this, you know, continuously growing digital footprint, different solutions have been proposed to enterprises. The solution that has been proposed largely by cloud providers that have big weight in the market has been: why don't you move all of that to my cloud environment? And if you did that, you'd have all of your, you know, your data, all of your apps in a single environment. And I have built in that cloud environment, all of the services for your data and your applications that you need to serve to make your application secure, to make them work, to make them perform all the time. That vision of a single cloud, if I go back to several years ago when I joined F5, actually had a lot of traction in the market, and a lot of CIOs believed that we're gonna do that. If you look at what has happened, no large enterprise, or virtually none, I know of one, of the thousands of enterprises we deal with, have actually moved all of their data and their apps to a single cloud. And the reason people are not doing that is because they've realized that they need the freedom to choose the best environment for each application. All large enterprises are going to a different approach, which is: I'm gonna be in a hybrid and multi-cloud environment. The issue with that hybrid and multi-cloud environment is that it has attracted a lot of complexity for companies. They're in multiple environments. Ninety percent of our customers today are in multi-cloud environments. On average, they are in between four and five infrastructure environments, on-prem and multiple clouds. Historically, because of how the industry has evolved, in each of these environments, they have had to use different solutions. They would use, you know, native application services in a public cloud sometimes. They would use on-prem, you know, independent software vendors. They could use an edge or CDN vendor for part of their applications. And so they'd use different vendors for different solutions and different vendors in different environments. And that complexity is enormous for large enterprises, and in some ways it is untenable today because it's accelerating also with AI. So F5 has built a portfolio over the last several years with acquisitions and organically, that can secure and deliver all apps and APIs across that entire set of environments. And we're essentially building a platform that provides security and delivery for any app or API anywhere. And we're pretty unique in the industry in having assembled these assets and having assembled a platform that takes away that complexity and gives enterprises the freedom to be in any environment they want, but with the operational simplicity of doing things only once in an automated way, rather than having to stitch it together manually. That's great. That's a fantastic overview. To really solidify it for the audience, I was wondering if you could kind of walk through some of the product areas that you address and maybe talk about the competitive set, right? Like, what does it look like for traditional ADCs versus multi-cloud networking versus l ike, app performance, management, CDN? Just to put into context, who are some of the peers and competitors that you work with, and how you're differentiated relative to them? Yeah. So Mike, the first thing I'd say before getting, I will get to individual product areas, but I think the first thing that's important to understand about F5 is we understand application traffic better than any other company. That over the last twenty, twenty-five years, we have worked with thousands of large enterprises to make their applications perform, to make them more secure, and to do that at layer four to seven, which requires a lot of knowledge built up over time, with a lot of experiences with applications. That's really the differentiation that F5 brings to all these product areas that we have. And because of that, we've been able to build this platform that brings the simplicity to operating across environments, that brings automation and a full security stack to support all applications. Now, if you go into individual product areas, i n the area of traditional ADCs, you know, we have now become the number one player in traditional ADCs, both in hardware and in software. And we have done that because of continuous innovation, and we continue to do that. We're gaining share. The competitive environment there has become softer for us, and so we are continuing to gain share in hardware and software. In the application security market, we historically were primarily a hardware software player. Recently, we have developed a SaaS platform to offer that as a SaaS form factor. We are an attacker in the market. Players there that have been in that market include companies like Imperva, Akamai, Cloudflare. We are, you know, increasingly competing in that market, and we're of course, we're starting from a small base, but growing faster than the market. Pretty excited about this net new opportunity for F5, and then in modern application environments, which is where our NGINX product competes, we have seen good growth because modern apps are growing, and our play has been to consolidate a lot of functionality into a single software pack, but if you elevate from these individual product areas where we have different competitors. You know, the ADC market that used to be defined, you know, several years ago by hardware appliances, it really has changed now. Right. Whereas today, companies want, in a single stack, things that used to be application delivery, but also web app and API protection, zero trust capabilities, multi-cloud networking, and increasingly, AI-specific services. And we have built a platform that has all five categories in a single platform. And so, yes, we compete in, of course, across individual categories, we compete with a lot of players, but more and more, the value we bring to our customers is the combination of these categories. That's great. I wanted to get a sense check on what customer demand looks like. During the most recent quarter, F5 talked about improving demand after this period of inventory digestion and customer caution and budget scrutiny. First, you know, could you talk a little bit about what gives you confidence to, you know, continue to have this outlook for a demand recovery? Are you seeing any specific verticals or customer types or regions that tend to be doing better than others? Yeah, broadly, we're seeing multiple signals, and I wanna preface it by saying that it's a little bit early to be making a call as we look into FY 2025. But there's a number of things that we're seeing. One is the close rates have been improving over the course of this year. They had been suppressed in FY 2023, so we continue to see relatively healthy pipeline, but the close rates that we're seeing weren't where we historically had seen them, and that's really improved over the course of this year. And we're starting to see, along with that, an inflection in the pipeline. You know, as we headed into FY 2024 or Q4, our current quarter, we saw an elevated pipeline. And part of that is on the Systems side of the business, which is really where demand has been suppressed, as customers have been sweating their infrastructure. Another indicator that we've looked at is just the aging of our installed base on the hardware side, where customers had, you know, continued to renew their maintenance for longer periods than they historically would have. And so we've seen that kind of start to level off, and in concert with that, we're seeing the pipeline specific to tech refresh opportunities start to go up. So these are good early indicators for us that, you know, things might be getting into a healthier state from the demand perspective. So, you know, we'll have more to say on that in October. The other thing then, on the Software side, is we've always had really good visibility on our renewals business, and you know, we've had very strong, not just renewals, but expansion as these software opportunities come up for renewal, but the new projects had been where we'd seen some challenges over the last two years, and we hadn't been seeing growth from new software. Last quarter, we saw new growth from that set of opportunities, and again, the pipeline there is also looking pretty good as we head into this quarter. From a geographic perspective, I wouldn't really call it any specific trends. We had a pretty healthy quarter in Europe in Q3, but I would say broadly, that stability is, you know, we're seeing across most of our major geographies. That's great, and I wanna come back to some of that renewals strength and software strength in a little bit, but before we do, could you just talk about, you know, who the typical F5 customer is, right? I tend to think of F5 serving a lot of large enterprises, service providers. Is that, you know, generally how you would characterize it, and, you know, do you see differing demand trends as you look at each of these customer types? Yeah. So I think you characterized it well. First of all, you know, F5 is a company that's focused on large enterprises. We do very little in the SMB world. We have about 20,000 customers around the world. Most of them are large enterprises, service providers, and government entities. So if I take these three segments, large enterprises, what we saw over the last 18 months was a lot of frugality, you know, reluctance to spend because of uncertainty, reluctance to undertake new projects. In our case, there was also a digestion of the supply chain issues and hardware inventory. All of these, we are starting to see them abate a little bit, and we think that, you know, there's a reasonable chance that in 2025, we'll see more normalization of customer behavior relative to, you know, what we saw in 2023. Service providers, I would say similarly, service providers have been very effective at sweating assets over the last 18 months. We have seen green shoots of, you know, some of them, a minority still having really gone to the end of that rope and having to do things because 5G traffic is growing, wireless traffic is growing, private 5G is growing. And so those who are having success with 5G-type services really are seeing traffic growth and have to put capacity and expansion on it. So we're starting to see green shoots again there, and potentially normalization in 2025. And then in government, of course, it varies by country. But what we're starting to see a little more is the phenomenon of Sovereign AI. So government starting to invest in infrastructure to support you know, Sovereign AI in the countries, having AI infrastructure, sometimes doing it via large service providers in the country, but being prepared to invest to make countries AI-ready, if you will. Right. That's great to hear. And I wanted to follow up on what you said about service providers being historically effective around sweating assets. And I was just wondering if you could talk a little bit about that concept. You know, what drives the ultimate, for lack of better words, like, breaking point, where a service provider says, "Hey, like, I can't sweat this asset anymore." Is it, you know, the absence of, like, new features that you're delivering and a desire to, you know, get those new features and that innovation? Is it something else? No, it's generally actually simpler than that. Yeah. It's related to capacity. So service providers will, you know, take a platform, and they will monitor the utilization of a platform, whether it's hardware or software, and that utilization can start at 30%. Typically, when it gets to about 70% of the utilization of platform, it's time for them to add capacity because they always want to have some buffer for peak demand or issues. When service providers go to sweating assets, they can go and exceed that 70% and get closer to 80% and 90%. But then it becomes, you know, quite risky because they can find themselves in a situation where they have an outage, and they cannot deal with it. So it's really driven by capacity utilization of platforms. And also, in those environments, ruthless prioritization of where they actually put CapEx, where they absolutely, absolutely need it, versus areas where they would have put CapEx, but it would have been getting ahead of future demand. And they don't do that in this environment. They're really ruthless around prioritizing. Great. And maybe just translating some of that into the financials. You know, last quarter, F5 raised its fiscal 2024 revenue guidance to flat versus flat to down 2% prior. Mm-hmm. Could you just talk about some of the key drivers impacting that outlook? What gave you the confidence to raise the revenue guidance and how does, you know, the growing pipeline of deals impact it, if at all? Yeah. I mean, we're into our fourth quarter, so, you know, we have very good visibility into the rest of the year. But, I think the real strength has been on the software side, where we took the outlook up from flat to modest growth, and now we're seeing mid-single to high single-digit growth as the likely outcome. And I think a lot of that is really continued solid execution on the renewal base, that we've always had good visibility to, but seeing also really good expansion from that part of the business. And then, as we talked about just the environment, and we're seeing more new projects getting funded. And so we're seeing a little bit healthier demand for new software opportunities than we had, you know, expected going into the year. And so all that's feeding into the visibility that allowed us to kind of tighten up our revenue range for the year to the upper end of the guide. Yeah. And then, as you think to next year, there's an expectation of an acceleration in revenue growth, and it's encouraging that much of that will come from the software side. And I think you have very good visibility of that, just because of the timing of term contracts, right? Mm-hmm. Can you just talk a little bit about what we can expect in fiscal 2025? Yeah. I think from the renewals is the biggest area of visibility for us on the Software side, and it's similar to this current year, FY 2024, where we knew that the biggest part of the growth would come in the back half of the year. As we look ahead to FY 2025, we have a fairly substantial base of renewals that are coming up, largely in the back half of the year, and so that's really giving us confidence in the growth opportunities for software. And then, of course, we're encouraged by the more recent trends around the new side of the business on software. And then, of course, on the Systems side of the business, where we're starting to see customers move forward with some of those refresh opportunities. And then there's also the competitive market share opportunity, where it's a natural time. Customers are looking at their environments, and they're not just you know, refreshing on F5, but they're you know, they're in a similar stage with other hardware vendors. And so it's ADCs is a space where there are some operational complexities to changing into a new competitor or customer. But with you know, some of our customers, as they're evaluating refreshing their estate, whether it's F5 or a competitive estate, that's a natural time where they would look to move ahead with a new vendor. Great. And just, so for fiscal 2025, I think much of that growth is gonna be driven by that software acceleration. Could you just frame, like, within software, you know, how you think of the different business models and, you know, what gives you confidence that we'll see that uplift in terms, based on what you've seen in prior renewals? Yeah. So the first one is, within the subscription business, we've got really good visibility as to what customers are doing with F5 and what their rates of utilizations are. So, you know, we talked about True Forward model where we uplift the customer's existing contract to the current level of usage. And so we know the timing of when opportunities are coming up for renewal, but we also know what customers are using. And then the third component of that is the opportunity for expansion into new services. And we've seen you know, we've had a really good track record with customers when they go through that renewal cycle, evaluating more and more of our offerings from our portfolio. And so just based on the execution we've had on this side of the business, that gives us a lot of confidence in the growth opportunity there. That's great. If we could move back and t alk to some of the products. You know, I wanna start with ADCs, where you know, that includes traditional ADC, but also the cloud-based one with the acquisition of NGINX. You know, my understanding is that F5 is the leader in ADCs, and, you know, the innovation is what really drove that market leadership. But could you just elaborate a little bit more on, you know, what you had mentioned earlier around competition and ADCs and, y ou know, what drove the positioning there? Yeah. Yeah, I think, Mike, the you know, I think we are really truly redefining the market for ADCs. It used to be, you know, ADCs was a hardware appliance that delivered, you know, primarily load balancing and other application delivery capabilities for global applications. But if you look at and so the way the market evolved was, originally, it was hardware appliances, and it was companies like F5 and Citrix and a couple of other small players that were in the market. And with the arrival of the cloud, the ADC-as-a-service market emerged, and that ADC as a service was primarily native load balancing capabilities in AWS and Azure and GCP and the major cloud, and they were consumed by people building their application in these clouds and taking advantage of that. When you look at where enterprises are today, neither a point solution that is just a pure hardware/software appliance or an ADC as a service is really the answer that people are looking for, right? Today, people are looking for something that's way different, that's a platform, that's way different than either one of those things. And the big differences are, number one, the stack of services that people are wanting in a single stack is way bigger than what was in old ADC. So it's in that stack. It's application delivery, it's web app and API protection, it's zero trust services, like authentication, privileged user access, those kinds of things. It is multi-cloud networking because they are in hybrid cloud environment and want to connect all that. And it is increasingly some specific services for AI. So those five things have to be in the stack that customers want today as an ADC. That's big difference number one. Big difference number two is, they're also today looking for a platform that is going to work across any form factor, whether it's hardware or software or SaaS, and whether it's in public cloud, on-prem, at the edge or anywhere. And so the work we have done over the last several years is really to build that platform and really redefine the ADC market to be what enterprises in the hybrid multi-cloud era are really going to need to secure and deliver their applications. Right. So if you're hybrid multi-cloud, and most enterprises, if not all enterprises, are at this point, like, you just can't rely on the, you know, native solution that's provided by the cloud provider, right? Because you have to worry about all the traffic elsewhere. Correct. If you rely on that, then you're gonna have a point solution in this cloud. You're gonna have to train a team in this cloud. You have another point solution here, you're gonna train a team. You may have a different security stack in different environments. If you have new vulnerabilities you need to deal with, you're gonna have to patch it manually in each of these environments. It takes weeks. It adds a lot of manual complexity. And so what we're seeing is a lot of customers have kind of incrementalized themselves into that situation. Right. And it is only when, you know, it gets to a very complex state that people step back and say, "Should we do things differently?" And so what we're doing is working with customers to get ahead of that problem and say, "You are on your way to a pretty complex situation." We have an affectionate name for it at F5. We call it the ball of fire. We say, "You're on your way to having a ball of fire, and we've built a platform that gives you the flexibility you want, but also gives you the simplicity that you need to automate your environment, secure all your applications, do it once and not multiple times." And that's I think ADCs of the future look more like a multi-cloud platform with all these capabilities than point solutions that we've had in the past, either in the cloud or on-prem. That's great. And, you know, it was always logical to me that an ADC company, or product that is responsible for load balancing traditionally, should have a right to participate in other areas, such as security given that, you know, you're looking at these packets and redirecting them anyway. Mm-hmm. I'm sure you could put that in a much more articulate way, but, like, why does a company that has historically been very good at ADCs naturally can extend into security? No, you're very articulate about it. The reason that we've extended into Security is because we are in line of application traffic, and so, you know, the traffic between users and applications or between APIs or between multiple components of an application, if you are in line of that traffic, and you are inspecting the traffic anyway, you can inspect the traffic to do load balancing, or you can do with that same inspection of traffic, you can say, "Actually, I'm gonna put a security policy here that blocks certain types of traffic, or detects certain types of attacks." And because we have built over many years the capabilities to inspect the traffic at a very granular level, it allowed us to move into, okay, we're getting traffic, but for security capabilities. And then beyond that, we said, okay, well, if we're doing it here, we can do it in other form factors. We can do it in clouds, we can do it in containerized environment, we could do it at the edge. But it's the same core competency of inspecting traffic at a very granular level. This also explains why we have not been focused on extending to all kinds of areas in security. We have been pretty focused on app and API security, because that is where we bring a significant competitive advantage, a fluency in applications and understanding in applications that no other company brings to the table. And so our security capabilities for apps and APIs are really deep, and that is where we continue to be focused, rather than attacking all areas of security. Right. And just to really kind of solidify this understanding with an example, like, when we talk about, like, apps and APIs, like, what are some like, obvious examples or like, categories of examples where, you know, F5's product has, you know, really helped to establish, like, Application Performance and Security Management? Is it like a mobile app? Is it a website? Is it like some sort of, you know, productivity software that an enterprise is using? So it's all of the above. Okay. It started, you know, you go back 20 years ago in the dot-com era, all we were supporting is load balancing websites. Sure. The dot-coms. Then in the early 2000s, enterprises started building applications that were facing the web. And so they now have this issue of, I have an app, I need to expose it to users that I will need to go through the internet to expose it to users. So I need to put in place infrastructure that load balances application, makes it secure, et cetera. Those were oftentimes productivity applications, whether it's, you know, Microsoft-type applications or Oracle-type applications. So large corporate applications have been part of, you know, what F5 has supported for many years. But it's also a lot of customer-facing applications. In the world of financial services, a lot of the apps that you use on your mobile to do your mobile banking or your, you know, buying your coffee or, you know, all of these apps, increasingly, are secured and delivered by F5. And so it's really all types of applications. And perhaps the most modern of applications today that we're starting to develop quickly are AI applications. And we're starting to see new and different use cases to secure and deliver those types of applications. Right. And when you talk about AI applications, are these like the applications, the inferencing that's being done, or? Are you kind of like a chatbot, or are you talking about something else? Yeah. So let's dive into AI apps a little bit. So when you're using an AI app, like a chatbot, the chatbot is what you interface with, but that chatbot is making API calls to an AI model, which would be, in this case, it could be GPT foundation model. That model itself has been trained or is being refined, leveraging some AI infrastructure, GPU clusters, and a lot of data, which for an enterprise, would sit in an enterprise data store, okay? When you're using that AI apps, there's a lot of requests between that AI models and the enterprise data store or these API clusters. Increasingly, these things need to be load balanced, and so we're finding use cases. Think of what F5 used to do for websites, the early days. We're starting to see opportunities where we're having to do that for a GPU cluster. Between an AI model and GPU clusters, load balancing in front of a GPU cluster makes the model way too capable of responding way faster and dealing with way more requests than otherwise it would. And these are opportunities for traditional ADCs, typically in hardware, actually, today, that are specific to the way that AI workloads are built and how they function. As that evolves, we think it will also create a lot of opportunities in API security, because there are a lot of API calls between the models and the clusters, the models and the enterprise data stores, the clusters and the data stores, all of that's API communications. And so making those things secure and capable of performing for large volume of requests is a whole domain that's emerging where we intend to have a role. Great. That's incredibly helpful. Mm-hmm. Thank you for sharing that. You know what? Why don't we have five minutes left. Why don't we go to some audience questions? Can you give us a sense of how that pipeline- Oh, would you just mind waiting for the mic? Hi. Thanks for taking my question. I was wondering if you'd give a sense of how that AI pipeline is building relative to prior, since you've been around for a long time, but you have been on the internet, and all these corporate workloads you're talking about facing consumer internet applications, how is this pipeline built relative to the prior two mega cycles? I think it's difficult to say today because it's very early days, and so the companies that are building, you know, these AI factories with large GPU clusters are very few and far between because it takes a lot of money to do it. Those who are doing it are spending tens of millions of dollars doing that, but it's typically gonna be a very small number of very, very, very large enterprises or service providers that are building infrastructure more for Sovereign AI environments working with a government. While the individual spend of a company could be large, the number of people doing it is very small today. I think it will take a while, and we've said probably 12-24 months, to see that start to go mainstream with large enterprises. Doing inferencing, having their AI infrastructure to refine models, having built their enterprise data stores, I think it's, you know, 12-24 months before we see that. So the pipeline today is very narrow because it's a, you know, small number of individual companies doing that, but the early use cases are really interesting. Yeah, and then I would just add to that, that the pipeline that we are seeing today tied to supporting these large AI factories is more aligned to our hardware offerings, where performance is really critical to doing that load balancing work. I think longer term, as we see these opportunities go more mainstream across a broader set of enterprises, that's likely gonna play out in software. And then just a quick follow-up. Is from a software renewal perspective, do you expect the upsell and cross-sell to be similar to the upsell and cross-sell that you've seen in terms of expansion, as last year? Or are there greater opportunities going forward? Can we see an uptick in those metrics? Yeah, I mean, what we've seen is continued momentum. So let's take a step back. Very early days, we saw large expansion rates, and we called out that we felt like some of this expansion is more tied to how we initially size these opportunities, because we were introducing this flexible model to the market for the first time. And so likely we were undersizing in some cases, and so some of that early days, really high expansion was more tied to aligning these opportunities to the right level of usage from customers. B ut since then, we've continued to see good expansion with customers that are really expanding the role of F5, but also just the growth in the workloads, and so we're continuing to get better and better at that. I think that we'll continue to see the growth in the workloads, and I think with a lot of the work we've done across the portfolio, that's gonna really start to feed a bigger expansion opportunity with the cross-sell. And so that's been a big focus of ours, is working with customers to expand across the full portfolio, and I think that trend is really still early days, and there's a big opportunity in front of us. Great. I just wanted to follow up with the Sovereign AI commentary. So if you were like a Microsoft, Google, Amazon, wherever, you kinda build your own. But for someone like an Omnissa, they would go to you guys, is that the opportunity there for the hardware side? And then on all these API calls from the enterprise, is that a net new service for you? Because I assume that when you deploy these chatbots or whatever tools they have, there's gonna be tons of API calls. I mean, that's how, actually, I mean, that's how these guys are gonna make their money, you know, locking up people into these APIs. And if I'm an enterprise and I'm gonna do a lot of API calls for like, you know, call it, OpenAI or whatever, does that, is that the opportunity for you? I just wanna make sure I kinda understand. Those two opportunities. Thanks. Yeah. So let me start on the second one, come back to the first. The API calls, yes, it is a net new opportunity for us, largely because, you know, APIs have grown in organizations and have not been a category of assets that's been super well managed. A lot of companies don't know how many APIs they have. They have zombie APIs, shadow APIs. So, first of all, discovering all of those APIs, being able to catalog them, understand the vulnerabilities that exist with these APIs, and then being able to mitigate the vulnerabilities in line of traffic, is a whole discipline that is emerging. And we've built an end-to-end solution to be able to do that, API discovery all the way to API protection. We've made a couple of acquisitions to do that, and we are now unique in the market with an end-to-end solution for API security, which is emerging as a more required service. So that's why it's a net new opportunity for us. It's not a business that we've been in. I shared on the last earnings call that the number of API security customers we have has tripled in the last year alone. Back to your first question, no, the Sovereign AI opportunity is more where, you know, a government or a service provider is building infrastructure to, say, build and refine a large language model in their own language c ause they want to have, you know, say, in Europe, a large language model that doesn't depend on, you know, the U.S. necessarily, but is in their own language. So that requires a build-out of new infrastructure, and then training of a model in that language. And as you build that infrastructure, you build the GPU clusters, you have to have the data that the model is gonna be trained on. That requires load balancing in front of these clusters, and it requires security in that. And that's a net new opportunity for us. It's supporting the build-out of that infrastructure. We're over time, so we'll just cap it there. But François, Cooper, it's been a privilege to have you on stage. Thank you so much. Thank you. Thank you so much. Thank you.