Hey, everybody. Thank you. This is the home stretch. I really appreciate everybody attending this conference. This is, I guess, the second half of day two. We have had a lot of really interesting insight the first couple of days here. You guys, I am excited for this conversation with F5. I think there is just a lot of cross currents going on, whether it is AI, cyber, acceleration. You guys seemingly are at the crux of a lot of what is going on in kind of the modern tech landscape. Thank you for being here. Cooper Werner, CFO of F5, and François Locoh-Donou, the CEO. Thank you again from all of us at RBC. I guess from an evolutionary perspective, F5 has gone through quite a journey with the move from kind of hardware to now a multi-product software-oriented company.
I guess before we get into some of the specifics, when you sit back and you reflect on the journey and where you are today, the question I often get from F5 is, what is the market missing? Is there still a misconception out there about F5? Because you've gone through a paradigm shift. I think a lot of people think of you as a hardware company, and you're not. Is there anything that you would offer that you're like, no, the market still doesn't get this element of the F5 story?
I will in a moment. Let me just read my safe harbor.
Oh, OK.
Before I respond, I need to get our safe harbor on record. 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.
Excellent. I'm glad you covered that.
Really important.
Yeah.
OK. You were asking me about what may be misconceived about F5 and what has the journey been.
Correct.
I think the first thing is the journey of F5 has not been one of going from one consumption model to another. It has been one of expansion. Yes, originally, our technology has always been consumed in hardware.
Yep.
Over the last several years, we have evolved to offer our technology in hardware form factor, in software form factors, and in software as a service.
Yep.
That is really a response to the big market trends that we have been seeing. If you go back 15 years ago, the predominant mode of consumption of F5, so F5 supports applications. Fundamentally, what we are focused on and what we specialize on is delivering and securing applications and APIs. 15 years ago, that technology was consumed in hardware and private data centers. In the decades, in sort of 2015, when the cloud went mainstream, customers wanted to consume this technology in software, in public clouds, in private clouds, in private cloud environment, and increasingly also as a service. F5 evolved, and we made our technology available in software and in software as a service. Many of our customers now consume F5 in software and in software as a service.
In the last few years, what we have seen is customers have wanted to deploy their applications in multiple cloud environments and in hybrid cloud environment. This hybrid and multi-cloud environment gives customers a lot of flexibility to choose the best environment for each application. It also creates enormous complexity for customers. F5 has responded to that by bringing together our hardware, software, and SaaS into a single platform. We call it the Application Delivery and Security Platform. I think that is the thing that is misunderstood about F5. It is really that we are today the only player that has the opportunity to secure and deliver every app or API anywhere, in any environment, or any form factor.
That is a significant source of differentiation relative to any other player in the industry, be it a hyperscaler or a point security vendor or a point application delivery vendor. That position with our customers, being able to be the partner that they go to that can secure and deliver all their apps and APIs across environments, is, I think, the piece that is not yet fully understood by the market.
You sort of read my mind for where I was going with the next question. There is just incredible complexity with that. I think we went from this paradigm of everything is going to be in the cloud. Then, maybe not everything. I think customers struggle with, you know better than I do, how customers struggle with what is the right underlying infrastructure map of an organization. I guess as you talk to customers and you think about that paradigm, for most large organizations, what is that mix of apps? Do you get a sense for, no, the mix is, I do not know if it is probably hard to generalize. What is that right mix of where apps run in the future from your perspective?
You know, as you said, it's hard to generalize because you'll have some customers who will have 100% of their apps on-prem. You'll have some customers that will have 100% of their apps in the cloud. You will have all kinds of variations based on their business models, their needs, regulation, compliance, security challenge, cost optimization. All of these variables come into play. Now, if you fast forward over a long period of time, because so many new apps continue to be born in the cloud, I think over time, there'll be a general tendency that there will be more apps that are in public cloud environment. The mix will vary from one company to the other.
Certainly, what we are seeing now, and it's frankly an inflection that has happened in the last year or two, is we're seeing a lot more enterprises reinvest actively in private data center capacity. You look at private data center capacity between 2015 and 2023, it was pretty much flat. There is an inflection now where it started growing again. That is driven by a number of factors. It's companies doing cost optimization, repatriating applications more aggressively from public cloud. It's for a number of sectors. It's regulation. Companies really have to build resilience and have maybe an environment on-prem and be able to fail over in public clouds. It's geopatriation for some customers internationally that don't want to have all their apps in hyperscalers. Increasingly, it's AI. The issue with AI is the data gravity of AI.
We have a lot more companies that are collecting a lot more data about their business than they used to because they know they're going to be able to monetize this data with AI. Increasingly, we see a lot of companies choosing to keep that data on-premise because they've learned that it's quite expensive to have that data in the cloud or to move data in and out of the cloud. They want to have their data on-premise, but their AI models could be on-premise or in public clouds. It creates a need to manage all that traffic. For all these reasons, we're seeing a lot of companies actively reinvest in private data center capacity. The question isn't just where applications are going to be. It is where are applications going to be and where is the data going to be.
Ultimately, for us, we think.
It doesn't matter.
It doesn't matter. We think it's going to be what matters to us is it's going to be distributed across multiple environments. The complexity and scale challenge and security challenge that that creates is where F5 comes in. F5 delivers things that no other company can bring to that equation today.
One more question for you, François. Then, Cooper, I'll get you with a question. Kind of continuing on the thought process of AI, two years ago, AI was sort of like the last question that I was asked, sort of, oh, just what do you think about AI? Now it's sort of question one, two, three, and four. When you think about the demands that AI puts on customers in terms of, like you said, data privacy, speed, latency, what are the things that we need to consider from an AI lens that F5 is uniquely positioned to help customers deal with?
I think the first thing to really, I think, internalize, which I don't think the market really understands yet, is AI requires more fine-grained processing of traffic and more fine-grained securing of traffic. Specifically, you have to secure every token, and you have to process every token. That fundamentally requires layer seven capabilities, which is what F5 has been building for the last 20 years. We are very excited about the challenges that AI brings because we are very well equipped to solve these challenges. Where that manifests itself today, we're in the very early innings of that. We are already seeing the use of F5 in places where we were not used before. Take data delivery, for example.
Increasingly, companies have their data stores on-premise that are being accessed by AI models that need that data, either for training or for inferencing and making a lot of read requests to these object stores. That requires high-performance traffic management to make that work. F5 is now being put in front of these data stores to make those use cases work. We were never in front of data stores before or databases. That is a net new use of F5 because you need that layer seven high-performance traffic management. Another example, which is nascent, is AI security. As I said earlier, you're going to need to secure every token. That requires layer seven capabilities. F5 has been fronting applications, securing and delivering apps and APIs for many years. Now we need to do that specifically for AI applications. There are specific threats on AI applications.
We made the acquisition of a company called Calypso AI a couple of months back, in part to accelerate that. We are starting to see demand for these AI firewalls, red teaming of models, penetration testing of models, and then being able to provide guardrails against AI applications or AI models. That is a net new opportunity for F5. A third example of that is inside of these AI factories, these large GPU clusters where some of our customers are investing large amounts of money. The traffic in there is not optimized. Sometimes it creates lack of utilization of these GPUs. You bring F5 traffic management in these environments, and you can make those environments way more efficient. There is a lot of dollars at stake for companies.
We have partnered with NVIDIA to create an integration between our software and their GPUs that enables this traffic management to be way more efficient and increase utilization of GPUs, allow folks who have a large amount of GPUs for a fixed investment of GPUs to generate way more tokens because they use F5 software in addition. That is another environment where you bring F5's layer seven expertise, and the AI scaling challenge gets solved way better. That is why we're excited with these opportunities in AI. I think it's just the start of it.
I lied, Cooper. I just wanted to double-click on that AI guy. One thing that we're picking up in some of our conversations with the executives, I think we saw these digital natives pushing the envelope on AI and the obvious frontier models and whatnot. You have such a diverse customer base. Are you starting to see evidence of sort of the non-AI native cohort of customers increasing their AI journey? We may not be at an inflection point yet. Any observation on customer behavior in terms of their AI adoption of the non-AI native cohort?
Definitely. I mean, you start with at the start, I would say it was AI native companies. I think, of course, as an example, there was a company in that cohort. I think you're now moving to see what I would call not AI native company, but AI leaning companies. So it's typically.
The subtlety to that.
Typically, you take the, in any industry vertical, you take whoever was the digital champion of that industry vertical, so had moved faster than others in that vertical to digital transformation, had transformed their data faster than others, had invested in their data faster than others. Those are the AI leaning companies. It could be in a traditional industry, but they are deploying AI models. They are training models. They start to need to inference at scale. They start to need to secure these models and these AI applications at scale. It is a minority of companies because they were kind of the single champions of their industry vertical. I think past those, you'll start to see a lot more companies in these verticals then adopt AI at scale. It has already shifted from AI native companies to, I would say, more traditional AI-led companies.
Yeah. I'll add to that. I think that it's shifting some of the consumption patterns with customers as they're kind of planning for their use of AI. They're seeing kind of pressures around workloads driven by AI. Data gravity is becoming more of the forefront of attention. Even as they're doing their tech refreshes, what we're seeing is customers starting to migrate up to the higher performance end of our appliances. We're seeing more kind of advanced data center capacity expansion coming through in some of the hardware business. We think sometimes we call it shadow AI, but it's kind of that invisible hand of AI at the forefront of their planning as they're thinking about the next few years of what their data centers look like and then making sure that they're ready for some of the changes that are coming driven by AI.
That's actually where I wanted to go next, Cooper. You guys are coming off of a really strong fourth quarter. I want to touch on the security incident in a second here. Excluding that, could you talk, reflecting back on the year, which was really strong results all year long, as you sit in your seat today relative to a year ago, and you're looking at your model and you're looking at your forecast, what surprised you the most about the performance that you saw in fiscal 2025 versus when you kind of set that initial forecast? It was revised up really throughout the entire year.
Yeah. I mean, clearly, we had very strong growth in our systems business. We closed the year at just under 10% overall revenue growth. We started the year with a 4%-5% revenue guide. A lot of that strength is really coming through in the systems business. I kind of touched on it. Maybe what surprised us was the growth that we were seeing outside of tech refresh. We knew that tech refresh was going to be a big opportunity. We expected customers would really start to refresh some of their legacy equipment starting in fiscal year 2025 and that that would likely be a two- to three-year cycle. Outside of that, we're seeing more growth for new use cases. In some cases, customers are repatriating from public cloud to data center. Performance needs are clearly going up in a significant way.
We are seeing customers kind of ready their environments for the needs around performance. I talked about data gravity, having data sovereignty. Those are also driving what we call a data center modernization initiative across our customer base. There is strong growth, both from the tech refresh, but also the non-tech refresh component of our systems business. That is something that we think potentially is a durable kind of new dimension for our hardware business. Historically, we have said that that market is likely kind of a low single-digit declining market and that we have consistently taken market share. We likely would be closer to flattish long term. Now we are seeing growth in some new areas that we had not seen in prior cycles. That is something that we are going to be kind of tracking as we look ahead.
I have to ask about the security incident. It feels like even just yesterday, I had Cloudflare on stage. And there's an outage. It feels and CrowdStrike had one and Okta. I think your transparency was refreshing. I think you guys were as transparent as you could have been at the time. As you have had more time to pass since that, and I want to ask about kind of how you thought about it from a guidance perspective. François, have you learned anything more about the impact of that, the response, the customer response to your messaging?
Certainly, in terms of the customer response, yes. I mean, we went through a period. We announced this on October 15. We went through a period before October 15, of course, with several weeks of investigation into the intrusion to understand truly what had happened and be clear. There were several weeks of preparation to ensure that by the time we would announce on October 15, we could give our customers information that was actionable so they could decrease the risks for them. We were able to do that successfully. Specifically, we had a lot of work to do to build releases of BIG-IP.
This breach only affected BIG-IP, not the other products, but releases of BIG-IP that addressed all the vulnerabilities that we had, not that we had been working on but were not disclosed, that the threat actor may have accessed. We were able to make these releases available to customers. Their response has been incredible. By that, I mean, A, our customers generally have gone through that with other vendors or themselves before. First of all, there is an understanding of it and an empathy that was inspiring, frankly. Also, the speed with which they mobilized resources to patch their environments where necessary was incredible. The collaboration with F5 has been just incredible. F5, as a company, we really mobilized. As you said earlier, we had made a big choice. There are a lot of choices you make along the way of something like that.
The fundamental North Star for us was we have to disclose everything that would make it possible for customers to take action to protect themselves. That is why we are so transparent, keeping that North Star. Frankly, having learned from others that went through this before and in some cases, maybe regretted not having been as transparent upfront as they could have been. Having learned from others also drove this transparency. This North Star of the mission is to make sure that our customers can take action and to give them as much information as possible so they can take action. We have done that. I think just customers generally have been very complimentary of our response. Of course, there is work that continues because we and our customers, I think, have done most of what needed to be done immediately.
We want to make sure we continue to work on code integrity, continue to test our code, leverage AI tools to try and penetration testing on our code. We have entirely rebuilt new product development environments. We are also leveraging this to bring some innovation to the industry. We have done this partnership with CrowdStrike to make their EDR available on BIG-IP, which gives us an extra layer of observability and kind of belts and braces for our customers. Over time, there are good things that will come out of that. We are not letting go of the energy that exists in our organization to take this game to a whole new level.
Yeah. I think you said it best. In times of stress, I think it can make companies stronger. I think you said there's a lot of positives that come out of a situation like this. It feels like you guys are solely focused on that. I mean, I guess, Cooper, from your perspective, as CFO, how do you, because you weren't really seeing much of an impact, how did you kind of think about handicapping that from a full-year guidance perspective?
Yeah. I mean, it was a bit of a challenge just given the timing. I mean, the good news was we had a very, because the timing of the public announcement was after our fiscal year end, we had kind of a pristine view as to what the underlying health of the business was as we exited the year. And it's very strong. We're seeing good momentum across all dimensions of the business. The downside is that with the public announcement coming out in mid-October, we had eight business days before we're guiding. We didn't really have any new insights coming out of our pipeline in such a short period of time.
What we did was we spent some time kind of profiling our overall revenue base and looking at which areas of the business potentially could be more exposed to some of that near-term friction that we talked about in terms of customers going through their operational activities for remediation and that could put some projects on pause, additional approvals that may be required, but really kind of looking at what part of our revenue base might be exposed to that friction. If you think about our services business, our SaaS business, our renewal license business, that business has all either already been contracted or it is running in customers' environments, and it is just a renewal motion. It is more of a contract motion. That part of our business is not really exposed to any short-term friction.
Where we potentially would see more impact in the near term would be on new projects. Once we're able to kind of size that piece of our revenue base, we recognize that's a relatively small piece of the overall revenue mix. We just kind of layered some judgment as to what the kind of range of outcomes you could see in the near term. We balanced that against conversations we're having with our sales leaders. We may not have a lot of runtime with the pipeline, but we are having a lot of conversations with customers. We took that feedback from our sales force. That's kind of how we bracketed the range. The ultimate position that we came to as we headed into our earnings announcement was that the disruption should largely be short-term in nature.
We think it's more operational friction as opposed to customers looking to move on to alternatives in that our demand would stabilize second half of the year. We think the long-term impact should be relatively benign.
Yeah. I mean, I think history has shown that for others that that's largely been the case, especially with swift action and transparency. This question for both of you, sort of taking that piece aside, when you think about the 2026 outlook, and if we're sitting here a year from now and we reflect back, what gives you optimism about this year? You've alluded to some, François, in terms of some of the steps on AI and multi-cloud. What are some of these things that we should think about that could push you guys in a positive direction on this guidance framework?
I can start. For me, there's two trends that F5 can capitalize on that are not yet fully understood. The first one is hybrid multi-cloud. Hybrid multi-cloud is driven by a number of factors that are really fundamentally secular. When I talked about cost optimization, governance, regulation, geopatrioticization, and the gravity of data and AI, all these things are causing customers to adopt these hybrid multi-cloud architectures. That complexity does not go away. We essentially have built the only platform that can deal with that complexity, secure and deliver these apps across all environments, but create a single experience across all these environments, which reduces complexity for customers. That has me excited because as we continue to evolve this platform and engage customers with it, I think we're going to get more and more sticky, more and more embedded into their hybrid and multi-cloud environments.
That's trend number one. Trend number two, of course, is AI. And it's really what we're starting to see, data delivery as an example, as a use case, the need to use F5 in new and different environments, whether it's our hardware in data delivery, or it's our software in AI security, or it's new software on DPUs in AI factory load balancing. But these use cases in AI, I think, are more and more going to come, and more enterprises are going to adopt them. And it's going to create a net new opportunity for us.
Yeah. It sounds like the foundation has been laid for this journey. It feels like a lot of what you saw last year just continuing that evolution. I mean, Cooper, anything to add from your perspective?
Yeah. I mean, the other thing that we've talked about is with the software business, there's some near-term math headwinds that we referenced for the current fiscal year just from that kind of flattish software growth that we saw in fiscal year 2023 that's part of the cohort that's up for renewal this year. As I look ahead beyond 2026, we think there's a real opportunity to accelerate our software growth. Some of it will be the math tailwind that comes in the following year. The other thing that has me excited is where we're at with the momentum on our distributed cloud services business. If you look at the headline ARR, it was relatively flat this year. That was some of the transition we've been doing with some of these legacy offerings, which were largely through those transitions.
The underlying growth we're seeing from distributed cloud is very strong. It's going to be the majority of that or more than half of that ARR when we exit this year. That's going to be a real foundation for a growth contribution from that part of the business. Also, as François talked about, just the complexity of hybrid multi-cloud is driving customers to adopt across the portfolio. We think that that can be an accelerant to that opportunity as well.
As I reflect on the DNA of F5, I wonder, you guys are such an interesting part of the network topology. I think about your security business in particular. I mean, are there any aspirations to go even deeper into the security stack? It could seem very logical for you guys to start thinking about SASE or data security, data privacy, kind of like DSPM. Any thoughts on kind of how far you could push the security envelope of the portfolio?
It's a great question. Look, before I answer that specific, I want to say one of the trends that also has us excited is this convergence of delivery and security. I think AI will accelerate that convergence. We are not just a security player or just a traffic management player. We are both. Increasingly, in these hybrid multi-cloud environments, customers need both. Our ability to put both in a single platform is a huge differentiator relative to players that are security only, players that are delivery only. Convergence of delivery and security is really our play. In that, the number of security services that apps and APIs need is only increasing.
Our focus is not necessarily to go into other areas of security, meaning leave application and API security, but it's to continuously add the security capabilities to be the player you come to that has all of the security capabilities for apps and APIs. AI security is a great example of that. We're going to continue to do that. We're going to continue to build a portfolio of delivery and security services in a single platform that does all apps and API security. Beyond that, going outside of security for apps and APIs is not a priority today, but there are adjacencies that we could consider. Data security is a good example of that. I think AI security may lead us into that direction.
I would say if you look at it today, it's continuing to build and grow the platform with more and more security services that create cross-sell and upsell of our portfolio and create the expansions that Cooper talked about on our multi-year subscription agreements. It's because we're adding security services and more customers are embracing them because there's no friction in embracing a new security service from a platform that you already have. As we do that, we're adding new. We've added data loss protection or data security for data in transit recently. That was from a small acquisition. That's a new service that we're having customers adopt: AI security, AI guardrails, a new service that we're going to have customers adopt. That's more our playbook: build a platform, add security modules. As we do that, we're going into adjacent times in security.
Yeah. It strikes me that you guys are, in an increasingly complex world, you can take a lot of the complexity out of the equation for folks on many aspects of their network topology. Maybe just, I mean, that was a great conversation. We have 30 seconds left. Maybe put a bow on the conversation. In some ways, it's sort of like if you were to leave us with the most important point about F5 as we think about the future, we talked about a lot of exciting things today. I mean, how would you sort of wrap the conversation up?
You know, I would say the most exciting thing or important thing about F5 is we're building the platform, the strategic platform that can secure and deliver every app and API anywhere. We're the only ones that have the technology assets to be able to do that. Large enterprise customers around the world need that kind of platform to reduce their complexity. I think hybrid multi-cloud and AI will accelerate the adoption of this platform.
Yeah. Yeah. It's an exciting time to be in technology, and especially if you're the trusted.
Trusted with AI and applications.
Exactly. Exactly. We are out of time, unfortunately. We could use a lot more time than that. From all of us at RBC, thank you both. Thank you for coming. Really appreciate your time and attendance. It was great.
Thank you for having us.
Thank you.
Thanks.