F5, Inc. (FFIV) Earnings Call Transcript & Summary

Good morning, everyone. Thank you for coming to the conference. I'm Samik Chatterjee, and I cover hardware and networking stocks at JPMorgan. I have the pleasure of hosting the F5 team for the next session, Cooper Werner, who is the EVP and CFO of the company, Kunal Anand, Chief Innovation Officer. Thank you both for coming to the conference, and thank you to the audience as well.

Cooper, maybe this is for you, but Kunal, feel free to chime in if you have a view here. But what we're starting and primarily asking all of our companies to do is initially sort of give us their take on the macro at this point because I think if you look at the stock market, I mean, it seems like that's the biggest top-of-mind concern for investors at this point? And when you think about sort of the macro outlook, particularly going into the back half of the year, how likely or how likely do you see a significant slowdown or a recession. And particularly if you're starting to sort of plan around those scenarios and how you run your business?

Yes. Great. Thank you. And before I respond, I just want to get our safe harbor on record. 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. Okay. So to address the question of the day. Yes, so when we laid out our guidance earlier in April for the quarter and the back half of the year, effectively, we said that we are being somewhat prudent with our annual guide. We raised our revenue guidance by 50 basis points for the year, but that was really on the strength that we've been seeing in the business in the first half of the year. The conservatism was really just more recognizing that there's still some degree of uncertainty with the macro, but we have not actually seen any indicators in our own business of any slowdown. And so this is both for our most recent quarter and through April. The business velocity remains kind of at our expectations. The things that we track closely, just to get the earliest indicators, if there was any change in demand or things like our weekly commits, are we delivering what the sales team's expectations are on a weekly basis, linearity through the quarter. We've got pretty good patterns historically on what linearity would look like in a given quarter. And then beyond that, just general pipeline and close rates. And then, of course, meeting with our sales teams and our quarterly business reviews and just hearing kind of what they're seeing on the street level with their customers. And it was a very orderly quarter last quarter in terms of prosecuting the demand that we saw, and that's continued right up through April. And so again, the prudence is really just tied to uncertainty, but it hasn't translated into anything that we're seeing. So from where I sit, I just -- I don't have any visibility around any kind of real erosion to the demand environment.

I mean part of that uncertainty stems from tariffs? And how are you managing or navigating sort of the unpredictability tariffs as well as much as managing the levels that you see today versus yesterday? Just give us a sort of lay of the land, what you're doing on that front?

Yes, of course. So we've spent a lot of time kind of digesting all of the new cycles and kind of inspecting where we might have exposure from a cost perspective. Thankfully, the majority -- the tariffs, if there were tariffs that apply to our business, it would be on the hardware side of our business. The majority of our products are manufactured in Mexico and are USMCA compliant. So our cost exposure is pretty immaterial. It's mostly around peripherals that we might ship from tariff-impacted countries, but it's in the low single-digit millions is what we've kind of sized our exposure at. And we think we can manage that just through efficiencies that we've identified in our manufacturing process. So no real near-term impact to the business. From a demand perspective, we also haven't seen any indicators of an impact in terms of pull-in. I think that customers have a reasonable view as to what our pricing practices are. We haven't announced any changes to our pricing practices. And so the business that we've been seeing has been kind of within our expectations with our pipeline and our sales teams haven't identified any material cases of pull-in related tariffs.

Okay. So on the topic of pull-in, but maybe putting tariffs aside, just the VIPRION and the iSeries hardware refresh cycle where you do have sort of end of support on some of them coming up. What are you seeing in terms of the driver for the refresh cycle? And why shouldn't we interpret that as pulling forward some of the system-level demand?

Yes. So you're referencing the end of software support dates we have on our VIPRION and iSeries families, which VIPRION goes end of software support in April of 2026 and iSeries goes end of software support in January of 2027. Collectively, those 2 product families are well over half of our customers' installed base. So that's a sizable opportunity in terms of tech refresh, and we're seeing that come through in the numbers this year that's driving a lot of the growth that we're seeing. We don't look at it as pull-in. If anything, it's been deferred demand, where customers have been prolonging their existing installed base and sweating assets beyond kind of what their normal replacement cycles would be. And we're finally starting to see customers kind of get caught up on their capacity. And so that's where some of the strength that we're seeing on our year-over-year growth rate is coming from. But there's not really any pull-in that's associated with those refresh cycles. It's more of just kind of modernizing their data centers. But what we are seeing beyond just the normal tech refresh is we're seeing some increase in demand related to capacity expansion and preparing data centers for how applications are expected to grow over time. Things like AI are adding a new dimension of growth related to our systems business. It's still pretty early, and it's a little bit of an exercise trying to assess how much of the demand is coming related to preparation for AI. But clearly, we're seeing growth beyond just normal tech refresh. And we're also seeing kind of an uptick in our competitive takeout opportunity. We've got some competitors that have invested less in innovation and the quality of the support that we're hearing from our customers is also less than what our customers would expect. And so we're seeing a lot of interest from customers in coming to F5, and we think that represents a share gain opportunity over time.

Okay. Maybe just following up, I mean, how do you expect that curve of the fresh cycle to play out, particularly if iSeries has a January 2027 sort of end of software support, that's a fair amount of time between now and then when you've gone through these sort of similar dynamics in the past with some of your products, is it really sort of just ramping and probably you see a peak more closer to that January 2027?

Yes. So all customers kind of have their own time frames in terms of their planning. You've got some customers want to get well ahead of these end of software support dates. Other customers may choose to kind of run right up into those software support dates. And you actually have other customers that may go beyond those dates because we still do offer maintenance on those products. So that's one of the more difficult things is to predict is the shape of that demand curve. We expect that we're going to see strength at least through FY '26 and into FY '27. And it's really just a question of how fast customers move forward with some of those refreshes through next year will kind of give us a better indicator of whether that opportunity goes well into FY '27 or if it starts to kind of tail off.

Okay. I'm going to take it back to compared to some -- at one point in time, your big expectations for systems revenue was that you would be in sort of a steady decline in that business. And I know that hasn't been reiterated for a while. But when we certainly see a year like this year, where you're guiding to double-digit growth, I mean does that change overall your thinking in terms of what the long-term drivers for the system business are? Does that change sort of how should investors think about the system business and the sustainability of some of the growth rates versus this decline that you had at one point thought it would.

Yes. So you're right. The -- our messaging that we've had around our systems business is that we think the market is kind of a mid-single -- low to mid-single-digit decliner, but we think that we can do better than that just by continuing to take share. And if you look at kind of what's implied in our systems business right now back to kind of what a normalized level was back '18 -- FY '18, FY '19, kind of pre-COVID, it's right in that kind of low single-digit decliner on a CAGR basis. But clearly, we're seeing a lot stronger growth in this year's numbers. We've talked about some of the newer dimensions of potential inflection on that growth rate. It's a little bit early for us to say that this is kind of a new expectation, especially around AI-driven demand. It's very early. So it's something we're watching closely. But between the potential for AI demand and some of the competitive takeout opportunities where we may be able to take a little bit greater share than we had in earlier years, there is potential that, that market may be a little bit healthier for us than we had previously signaled.

Okay. Got it. Since we are on the topic of AI. Maybe just outline how do you think about how F5 is associated with the enterprise adoption of AI or sort of AI adoption by the enterprises and how F5 can help on that?

So I think enterprises are still in more of a foundational phase when it comes to AI today. We roughly see about 3 different use cases related to AI. One is sitting in front of data stores that organizations have. For those that don't quite know in order to get the most utility out of these large language models that are out there, enterprises have found success infusing it with their own enterprise data. And what customers have begun to do is put F5 technology in front of those data stores to help with scaling and to help with the information retrieval that eventually going into these large language models. The second area is securing access and securing the information that's going to and from those large language models, whether they're locally hosted or in the cloud. And for that, that's where we introduced the F5 AI gateway effectively sit between these applications and APIs as well as these large language models, whether they're deployed somewhere else or locally. And then last but not least, is around load balancing specifically load balancing these AI clusters. How does information get to these clusters and then intra-load balancing as well. So we recently announced the GA of what we're calling BIG-IP next for Kubernetes that runs on NVIDIA's BlueField-3 DPUs, which now gives organizations the ability to do delivery and security within an AI factory or inside of an AI super pod. So we really think that those 3 opportunities are very meaningful.

Okay. And do you mind just giving us maybe one more layer down when you think of these 3 use cases where you're seeing sort of the most appetite from customers today and which would be sort of more, I would, I guess, more sort of staggered in terms of ramping up maybe a bit later with the enterprises?

Yes. So it's interesting when we look at it because I think each of these fit a different persona and a different use case. Like, for instance, the first 1 around object retrieval and information retrieval, you're squarely talking to a CIO who's worried about scale, who's worried about uptime and availability of core systems and core information in their networks. And we typically see that across enterprises today. It's one of the more surprising things for us in the conversations that we've had around organizations who have collectively put BIG-IP and our systems in front of those data stores. It's what informed our partnership with NetApp that we announced. Then the second use case around securing access is speaking more to a CISO, and they're truly worried about governance. They're truly worried about information security. And that is something that we see cutting across all enterprises. We announced the AI gateway, and we're seeing early wins right now, which is very promising, but at the same time, I think it's very telling of the moment that we're in, which is a lot of companies are experimenting with these large language models, but they don't have the security and the governance that they need to properly lock that down. And then the third domain, which is around this work around load balancing. I think there are 2 things that are just fundamentally important, which is performance related to these AI factories and these AI workloads. And then the other is energy efficiency. And what we're finding is through early customer validation, we're seeing that we're able to deliver that for organizations, which is improving the time to first token, which is essential in these AI workloads, how fast do these models produce output. And then also the energy consumption, the fact that energy consumption over time can go down because a lot of the networking work that would typically happen on these GPUs is now getting done before the data even gets to the GPU, which increases the lifespan of these GPUs and allow them to take on more work.

Okay. Okay. Fair. But maybe just asking it going about it in another angle, which is now when you think about these use cases, translating it back to your product portfolio, how does it translate back to hardware versus your software demand?

I mean, look, I think for these AI workloads systems is definitely something that organizations are looking for. They want high performance, and that's something that they get with the coupling of hardware and software.

Okay. Got it. On the last call, you talked about AI for ADCs and using AI to enhance the product offering and customer experience. Can you talk about the significance of those investments? And what's the early feedback from customers?

We announced a bunch of capabilities over the last several months. And going back more than 6 months now, we announced our F5 AI Center of Excellence, which is effectively our own organization within F5 that is doing the things around AI, like building, training, fine-tuning models which allowed us to put out capabilities in the market that our customers are absolutely loving. We released something called our AI assistant for our distributed cloud customers and the attachment has been phenomenal. More than 50% of our customers are using our AI assistant today and more over 50% of our -- sorry, to scope it specifically, our distributed cloud customers are using that AI assistant today. And that's been extremely interesting to watch and observe and see how they're leveraging that solution. But it speaks to, I think, the broader challenge that I think that every enterprise has, which is the complexity of the environments and needing to get help when it comes to building policies, rules, understanding what's going on. And then for our BIG-IP customers, we announced that we're bringing that AI assistant to BIG-IP later this year. And we also announced at the beginning of the year work around iRule, which is a programmability construct that we have for BIG-IP. Basically, we're allowing customers to generate these iRules using natural language and AI. So this is where we're leveraging AI for ADC of how do we make our life easier or the life easier for our customers.

Okay. You talked about the gateway product and how it sort of helps in these security use case, particularly around AI. Distributed cloud product, I mean, can you just talk about whether there's a relevance here for enterprises in relation to AI with the distributed cloud offering that you have?

Yes, absolutely. I mean when you think about it right now, every organization is on some AI journey or AI curve. But we talked about it earlier around how the world is pretty distributed. Data is everywhere and data is needed to get the most value and the utility out of these models. And a lot of organizations are not able to move their data for compliance reasons or security reasons, which means that data on-prem is likely going to remain on-prem. That data is not going to get transmitted somewhere else. So if you think about it fundamentally, we're not bringing data to the AI model. We're going to have to bring these AI models to the data or we're going to need to find a newer way to bridge that level of connectivity. So what we built with distributed cloud, we have a capability in there for multi-cloud networking. And what we're finding right now is organizations are leveraging multi-cloud networking, and they're building this interesting fabric across their application layer. It doesn't matter if it's on-prem, in the cloud, across multiple environments, and they're able to now connect these applications, APIs and these AI workloads together. So we think that there is absolutely a play around what we can provide there.

Okay. I mean is the distributed cloud opportunity sort of, I guess, in the 3 pillars that you outlined already, do you see as a fourth pillar? Or is the magnitude a bit sort of lower in terms of near-term contribution than the other on -- particularly on the systems side?

I think it's just obviously different compared to the high-performance systems. This is speaking more to connectivity. So it's not going to be around sitting in front of data stores directly. It's going to be about bridging the different environment. So I view it more as a broader connectivity play rather than just specific AI play.

Okay. Yes. Okay. Maybe talk about the competitive landscape, the way you see it today, particularly related to traditional competitors that you've had in the space, Citrix and Radware, and then we can go into sort of the more sort of public cloud competition, but maybe let's talk about the traditional competitors that you've had in this space?

Yes. I mean what we've been focusing on is obviously a bunch of innovation around, I would say, modernizing what would be a traditional ADC. When we look at our competitors, Citrix, Radware, et cetera, they spent a lot of time in ADC, but relatively one-dimensional in their capabilities. What we've been focusing on is bridging both delivery and security together. I think for the longest time, ADC has been synonymous with things like load balancing, traffic management. What we've been able to do with a lot of investment and obviously, a lot of work is bridging a lot of the security use cases along with the load balancing and traffic management and that includes application security, API security, bot protection. We really believe that the infusion of security and delivery is fundamentally important. So for us, we think that, that investment in building out that platform play is a core differentiator as it relates to our traditional competitors. Then there's the sort of modern players. Let's take the hyperscalers as an example. Many of them have things like load balancing, some of them have Web Application Firewall. And for those organizations that use those technologies, those are limited to just that cloud. Now 90% of the enterprises that we cater to are sitting across 3 or more different enterprise environments. That could be multiple clouds that could be on-prem. And so for them, when they try to leverage something like a hyperscalers load balancer or hyperscalers WAF, that is limited to just that environment. They're not able to extend that to other environments. And so for us, the competition there is for us, playing in a hybrid multi-cloud space, which when we go in and have a conversation with a CIO or a CISO is more appealing rather than them adopting lots of different point solutions and trying to figure out how to make that work.

Okay. Got it. Coming back to some of the recent trends, and you did have a softer-than-anticipated software performance in the last quarter, at least from our standpoint, native to what we model versus sort of what was very clearly a strong quarter for hardware. And one of the things that always comes up with investors is how can you improve predictability in terms of quarterly software and hardware revenues for investors that follow F5 closely. I know you probably have much more insight, but the productivity on the investor side seems to be a bit lower, just given how new numbers move around quarter-to-quarter.

Right. Yes. And there's -- that's just part of our business model. We're going to have variability in the growth rates on a quarterly basis, and it's really kind of tied to the majority of our software revenue is coming through in our renewals motion on these multiyear software contracts that we book with customers. And so what we've really tried to do is focus investors more on the kind of annualized growth rates because that quarter-to-quarter variability will kind of normalize over a longer-term time frame. But also, I want to just be crystal clear that the lack of a smooth growth trajectory does not equate to a lack of visibility on our perspective. So we have good insights as to what the timing and shape of these renewals looks like. And we try to give some kind of indicators. Going back to last July, we had said that we thought first half of the year would be closer to flat growth for software just based on the timing of these renewals with stronger growth in the second half. And that's pretty much playing out to our expectations with the exception of in Q1, we had an exceptionally strong growth quarter driven by some really healthy expansion across a couple of our larger renewal opportunities. But we had said we're not -- that was acute to that quarter's renewal base, and we weren't flowing through that expansion, changing our expectations for the rest of the year. So from our perspective, Q2 was pretty much in line with what our expectations were. And we expect to have a very strong growth quarter in Q3. So what we could do to make it more smooth would be to reduce our flexibility that we give customers, give them less choice, have every customer go on an annual subscription agreement, and that would smooth the numbers out, but they would be smaller. And we think that the better opportunity for us is to give customers choice that leads to them growing faster with F5, and we'll take the kind of ups and downs on the quarterly growth rates that come along with that because longer-term, it's a better software outcome for F5.

Got it. But 1 of the questions that stems from what Kunal was saying in terms of high performance systems and demand related to AI is when we now think about a lot of the investments for enterprise going forward, incremental investments will be towards AI. And that pulls in a lot of high-performance systems. Does that automatically cannibalize some of the appetite to spend on the software offerings from F5? Do you see any cannibalization happening on that front?

We have not seen any signs of cannibalization to date. Most of the increased demand that we're seeing on the -- related to AI is on the hardware side, but it's mostly new use cases. It's either for AI -- direct AI capabilities that customers are rolling out or growth in their apps that are driven by AI that is leading to additional capacity needs in the data center. But we're not really seeing a shift in customers' architectures for how they support their hybrid multi-cloud environments. So no indicators that we can see to date is as to where software is shifting back to hardware related directly to AI.

Okay. Staying on software, the revenue growth this year in fiscal '25 is expected to be low double digits following a similar year in fiscal '24 and then you had 0% growth in fiscal '23 in the software revenue. As we look through a cycle and particularly think about the software revenues through a cycle, like how do you reassure investors that there's enough there to drive double-digit growth consistently in software revenues?

Right. So you're referencing what we call kind of a math headwind related to that '23 software, where growth was flat because we saw pressure on new projects related to more of a macro downturn in customer budgets. So that does flow through on that renewal cycle for next year. Now what offsets that are the increased rates of expansion we've been seeing with customers we're seeing more and more customers that are expanding into new use cases and new functionality across our portfolio at the time of renewal. And so that's where we have an opportunity to kind of drive growth in spite of that math headwind. And then also point out that, that same math headwind becomes a tailwind in FY '27 where you see a growth cycle on our FY '24 numbers that comes up for renewal in FY '27. Long-term, we believe that the software revenue opportunity is a double-digit grower on a CAGR basis. There -- in any given year, you may see a little bit of a fluctuation up or down just based on some of those cycles. But broadly, we think the opportunity is only increasing based on the expansion that we're seeing across the portfolio.

Okay. I mean, clearly, on the software side, there's a lot of reliance on renewals and customers coming back and sort of raising capacity overall. That does tend to sort of lead to concerns about your TAM growth being limited and it's being driven by a set of customers that you're renewing Yes, you're expanding with them, but the pool of new customers coming in seems to be relatively more modest relative to the renewals that you're addressing. Like overall, how do you think about the TAM growth on this front? And what can you do to sort of accelerate TAM growth?

Yes. So the TAM, we actually think is has a very healthy growth rate. It's just that a lot of the new projects come through with customers that we've already contracted with. So we don't want to conflate the renewal opportunity as being just a cyclical opportunity. The growth in a lot of the new workloads that customers are seeing in their states. They're very often building that into capacity expansion or new use cases at that time of renewal. So -- we are seeing a lot of new business opportunity come in either as net new projects or new customers or new functionality with existing customers. So you want to be careful not to just separate them as either business that we're winning or business that we're recontracting. It's -- the growth is really coming on both dimensions.

Okay. Okay. Maybe just going into smaller pieces overall. One, you talk about competitive displacements, but give us sort of what are you thinking in relation to how much more there is in terms of headroom on the competitive displacements because that is coming up a lot more on your commentary as well in terms of where the share gains are coming from?

Yes. So we've always had competitive displacements. That's part of our revenue opportunity going back several years. I think what we've seen is an increased appetite by some of our competitors' customers to move towards F5. And that really starts showing up in the numbers more in this past year -- past kind of 4 quarters. But we think it's pretty early. We're just kind of scratching the surface in terms of not just winning some of those accounts, but then the follow-on opportunity. Most accounts are going to replace their entire state on the initial buy, right? So they're going to replace part of their environment. They do the work to standardize on F5. And then the -- in a more cadence fashion, replace the rest of their state. And so we think that this kind of increased rate of competitive takeout opportunity probably a multiyear cycle, and we feel like we're still kind of in the early innings of that opportunity.

And would you care to sort of outline between Citrix and Radware, like where are the bigger share gains coming from?

Yes. So it's Citrix is definitely the bigger opportunity. Just they've got a larger installed base in general. And so I would put that opportunity kind of at the top of the list.

Okay. And then again, maybe following up, fair to assume that most of that share gain opportunity comes through on the system side first and then sort of expands into the software domain?

I don't know that I'd say it's mostly systems. It's really spread across both systems and software. I would say deployable software. So BIG-IP software, in some cases, NGINX -- and then over time, one of the differentiators we have is that we have a true multi-cloud set of solutions. So there's opportunity to expand across the rest of the portfolio where some of the traditional competitors really were more kind of data center and cloud-based offerings, but didn't really have an offering for SaaS-based environments.

Okay. Let me take a quick pause and see if anyone in the audience has a question. Any questions in the audience?

Very insightful. How are you thinking about your equity compensation or incentives for executives and employees? Any change or things you're thinking about differently with the market environment?

Not at this time. It's -- we've had a pretty steady equity compensation in terms of percentage of revenue. It's slightly declined over the last couple of years. I will say that as we look ahead, things like AI and the competitive nature of talent acquisition. That's something we will have to consider as to what's the best way to attract and retain top talent. We see really exciting opportunity with AI. Kunal outlined several dimensions where we think that we can drive an increased revenue growth opportunity related to AI, but there is a cost to acquiring that talent. So it's something that we're spending a lot of time on in terms of how to attract the best talent.

Maybe last 3 questions here. One, federal spending, a lot of discussions and noise around pullback there. How are you sort of maybe quantifying the risk around it or embedding any risk around it? And what is your exposure to the federal spending specifically?

Yes. So federal is typically in the kind of 5% to 10% of revenue. That's the mix that we get from federal government depending on the quarter. Our prioritization is pretty high within federal budget. So -- we understand that there is, of course, some uncertainty in terms of various budgets within federal government. In terms of the appetite to reduce waste, our products we don't really feel are exposed to that dynamic just because we're not a seat-based model. Most of our solutions are tied to kind of mission-critical engagements. And so we're tied to security as an example. We haven't really seen any notable change in demand from federal government. We haven't seen any pull-in either. That's always the question is with uncertain budgets, do you see some pull-in while you have the budget, but we haven't really seen that to date.

Okay. You've been building cash on the balance sheet and like now you generate $700 million $800 million, $700 million to $800 million free cash flow. You're buying back $500 million. There is some cash that keeps building every year on your balance sheet? I mean, how are you thinking about allocation there, allocation priorities on that front?

Yes. So we're still in kind of the same commitment of repurchasing shares greater than 50% of our free cash flow that's our model that we'll continue to operate with. We want to make sure that we've got flexibility strategically in terms of our balance sheet. I'll also note that about 60% of our cash is held offshore. And so there are there are opportunities to bring some of that cash back across, but we want to be mindful in terms of any tax implications that, that could have.

And so when you say you want to keep balance sheet flexibility for strategic actions, M&A being a part of it. I mean, if I did do a bunch of acquisitions, but it's been a while since you've been the...

Yes. And we've been doing more kind of tuck-in type acquisitions, and those have been incredibly successful. We've had what we feel are really good returns in terms of the capabilities that we're able to bring to market quickly with some of these smaller acquisitions. We're constantly kind of surveying the landscape and seeing where there's an opportunity to accelerate our road map in terms of M&A. So it's not to say there might not be larger opportunities that come our way, but those tend to be a little bit fewer and further between. But M&A will always be a part of our strategic.

Last 1 for you. You are 1 of the few companies I cover that's hitting the Rule of 40 at this point. So while that's always good to see. How do you think about sort of upside on the operating margin from here on? Is it purely a case of continuing to drive operating leverage? Or is there a ceiling? I mean you're at 35% operating margin, is there a ceiling for a company of the sort of P&L structure that F5 has?

Yes. I don't know that we'd characterize it as a ceiling. We do think there's an opportunity to continue to drive operating leverage over time. I think we're always going to be balancing that with what we think the revenue opportunity is. We try to plan relatively conservatively in terms of our assumption on top line and how we then resource the rest of the model to make sure that we're able to continue to drive double-digit earnings growth sustainably over time. We've outlined several vectors potential inflection in revenue growth over time. So we will kind of make that balance in terms of how we resource the business to go after those opportunities. But we do think that there is more operating leverage in the model.

Great. I'll wrap it up there, but thank you for coming to the conference, and thank you to the audience as well.

Great. Thank you.