All righty, we'll get started. Welcome, everybody, to day three of the UBS Global Tech and AI conference. I'm Roger Boyd. I cover cybersecurity infrastructure software here. My pleasure to introduce Todd Nightingale, CEO of Fastly, and Ron Kisling, CFO. So thanks for being here.
Thanks so much.
Awesome. Before we get started, if you have any questions you want to ask, please submit them through the app. There's a QR code on the table. I'm happy to weave them in, and we'll get started in a fireside chat. So Todd, maybe just a high-level summary of last year. I think there's been a number of changes since you joined the company. As you work towards, I guess, more business stability, kind of what's the progress there? And from a platform side, how do you think about the evolution of the Fastly platform?
Yeah, in large part, those two things are tied. The evolution of the platform has been really about broadening the functionality, unifying our solution into one cohesive, easy-to-use platform, and bringing a complete edge solution. And really, in this market, a complete edge solution includes content delivery, edge security, or web application and API protection, observability, and edge compute. And by focusing our efforts on building the most complete edge, we've done an enormous amount of good to our go-to-market. It allows our teams to land and expand through the customer base, and it allows us to acquire customers not just focused on user experience and performance, but also focused on reliability with the observability product line, and most importantly, and most successfully, users who are focused on security and some deals that actually originate in the CISO department.
And so by focusing on this unified and large platform, or broad platform, I should say, we're finding new ways to acquire customers and new ways to grow and cross-sell those customers, importantly. And as far as the kind of trajectory of the year, we did see a lot of volatility in our large media customers. And at the beginning of the year, we had very, very high concentration, revenue concentration in that customer set, 40%. And we've really focused on making this a year of transformation where we are diversifying our business and ensuring that we are growing outside of those top 10 large media customers at a very high rate and building a more reliable base, a more diversified base upon which to grow Fastly, double-digit revenue numbers in the future and in the out years.
And the key to that is a broad portfolio that is attractive to the non-media set of customers, which is really the larger part of the market, and to focus our go-to-market on the rest of the portfolio as well, CDN.
Yeah. Can you remind us of the growth dynamics between, or maybe this is a question for Ron, too, but the growth dynamics of the top 10 and then outside the top 10?
Yeah.
You talked about kind of chasing customers outside the media vertical. Kind of what are the verticals that are doing well right now?
Sure. So on the top 10, what we saw this year is outside the top 10, we saw in Q3 that segment grow 20% year over year, which was substantially an acceleration from what we've seen in the past. That reflects share gains that we've seen in that space, as well as continuing to sell. That's a space that tends to buy the whole platform. We see higher attach rates of various products.
Gotcha.
As far as the vertical goes, Fastly has always been really strong in e-commerce and high tech, for sure. But now we're seeing a lot of strength in transportation, traditional retail, healthcare, and hospitality. That's made a big difference, being able to, with this broader and more cohesive platform, being able to approach a much larger chunk of the market.
Got it. Okay. And then maybe just with the large customers, I think most are aware of kind of the headwinds there, and that's not a Fastly-specific problem that's being seen by your competitors as well. What's been going on there? I think the multi-CDN approach, I think, is also pretty well understood. But can you talk about kind of traffic optimization? That's been kind of an angle as well.
Yeah. I mean, we started to see a lot of volatility in the first half of the year in those accounts. And as a reaction to that, we built a new motion at Fastly, a very, very high-touch customer success motion for the largest media accounts. And that includes better telemetry for better revenue projections, better early warnings around traffic shifts or experimentation at those accounts, better executive engagement with the accounts, a more reactive organization to transition to new needs of theirs. But I think what we really saw, the one trend, there was some technology optimization, traffic routing optimization. The broader trend that we saw in that space was media accounts that were focused on profitability for the first time in a while, really focused on profitability. And they reacted to it differently. And what was important is that we reacted to their needs.
I think we're most of the way through that. In fact, I think the pendulum is starting to swing the other way right now, and I think this was a good lesson for the industry and for Fastly. Maintaining a very, very high-touch customer success motion around those multi-CDN customers is imperative, and it only helps us serve them better, and it also helps for us by keeping that ultra-high-touch motion separate. It helps us stay focused on this diversification of the revenue and growing, Ron mentioned, 20% outside the top 10, growing that cohort to build a healthier, more reliable revenue base moving forward.
Gotcha. The other dynamic with the top 10, there's been some consolidation within the CDN industry, the delivery industry, most recently with the bankruptcy of one of your competitors. How do you see that changing the CDN share within those top customers? And how do you think about it as an opportunity for Fastly, as well as the impact of the pricing environment?
Yeah. Yeah, the Edgio bankruptcy, I think, certainly changes the landscape a little bit, and it's in those two areas. Number one is there's a lot of traffic that was carried on the Edgio networks, and as they get deprovisioned, that traffic's going to go somewhere. Some of it will flow through to Akamai as they bought some of those contracts, but certainly not all of it. There's a huge opportunity. We are tracking every one of those accounts right now. We've already had some early wins, and we're pushing hard to get every bite of that traffic. We plan on fighting for and winning as much of that traffic as we possibly can. I think we're going to see that traffic be transitioning over really over the next six to 12 months. It doesn't happen overnight. There are commits involved that are going to age out.
There is transformation cost and time. It's going to take maybe six, maybe 12 months for all of that to sort of shake out. But it's exciting because I think it'll help us bring our large media accounts back to growth faster, making that sector healthier faster, which is great. In the fullest of time, I think it also does a lot, especially in the out years in 2026, 2027, does a lot to just make the pricing environment a little more attractive to us and a little healthier for the organization, for the industry at large.
Gotcha. I think you mentioned kind of the headwinds broadly within those customers starting to abate a little bit, and now you have a potential tailwind from Edgio. Is it right to think about that top 10 could start to perform a little bit better over the next year? And I guess champagne problems, but how does that affect kind of the diversification goals?
Yeah. I mean, I think for sure, for us, there's some external causes you mentioned. There's the Edgio piece, which involves both the traffic and the pricing. There's the fact that I think this dynamic is sort of reversing already, which is great, or has sort of come to a shift to profitability. Those changes have already happened, so it's kind of shook out already. For us, I think there's also this new customer success motion, which is helping us win more of that traffic as it stands. And so for us, I think we absolutely have an opportunity to drive that large media cohort back to growth and maybe even double-digit growth much more quickly. We'll also have, especially in the back half of 2025, much easier kind of comparable numbers, which will certainly help us do that as well.
Gotcha. Maybe changing gears a little bit, the other big changes here was you brought in a new Chief Revenue Officer. Can you talk about his background, what he's been focused on with some of the diversification issues, and how you've segmented the go-to-market organization that way? And then his security background, I think, is pretty interesting, too. How does that kind of tie into where Fastly is headed?
Yeah. So we brought on Scott Lovett. He comes from Imperva. He has a background at Akamai as well as Cisco. He's been phenomenal so far. It's been remarkable to see how fast he's been able to enter the business and really start to already operationalize significant changes through the organization. It's amazing what a strong go-to-market leader can do. He specifically, he's got a very strong customer acquisition motion, Imperva especially, but also at Akamai and Cisco. He understands that hunting motion. And at Fastly, the name of the game is customer acquisition and gaining market share through customer wins. And he's been amazing at focusing the team on that and ensuring that we have all the pieces put together. And the second is a strong security background. For us, especially with the platform approach we mentioned at the beginning, a security focus is so, so important.
And having that in the go-to-market, we've had strong security expertise in the product team with Artur as the founder and engineering groups, et cetera. But now having that on the sales side is awesome. And we're really looking forward to a strong security focus in 2025. We've done a ton of investment on the R&D side here in 2024. 2025 is going to be the year we push that into the go-to-market.
You've mentioned this a couple of times, but as you think about going after the non-media cohort, can you just talk about how important it is to come to those customers with the full platform, with security and compute, and how that kind of changes your trajectory in that business, which is going pretty well?
Yeah. So in the large media customer, large media tends to run a multi-CDN architecture. Their spend tends to be 90% or more, 95% even, content delivery. That is their business, is delivering content. They have a huge spend in that area. And they will spend a little bit on observability and security, but it's like a diminutive. But that's not true of a retailer, hotel chain, hospitality group, healthcare organization. Number one, they're not bringing on multiple vendors. It's not worth it to them. They want one technology partner that's going to provide a complete edge solution. The second is that they have a much more balanced spend here. They need DDoS and bot mitigation, and especially best-in-class WAF, web application firewall. Of course, they want best-in-class performance, and they're happy to use either content delivery networks or edge compute in order to deliver it.
Their business, their security teams, their executives all demand very, very high reliability. They're looking to orchestrate an observability solution that feeds into their full-stack observability offering. Their spend is a lot more balanced, and they're only looking for one partner. Having a broad portfolio is really table stakes. Fastly's core differentiation for years has always been ultra-high performance and incredible flexibility and programmability by developers. It's always been possible to build these things on the Fastly platform, build best-in-class DDoS protection or best-in-class observability. Now with turnkey products, we make it so much easier for these customers to come and really partner with Fastly and use that full product set. I think that is in large part what's driving this growth outside of the media vertical, outside of the top 10.
It's driving a customer acquisition motion that I think is going to serve us well and fuel our growth rate for years to come.
Yeah. Makes sense. Can you just talk about the maturity of the security offering? I think when you joined the company, it was maybe a weaker point of the platform and sort of in a transition period. But over the past year, year and a half, it's been a pretty steady flow of innovation. You developed, I think, a DDoS and bot mitigation product internally. You have a private relay product. Web App Firewalls has been kind of iterated on. Just talk about the maturity of that and your right to win and maybe security focus in these cases.
Yeah. When I joined the company, Fastly was already shipping really best-in-class Web Application Firewall. They'd acquired that technology through the Signal Sciences acquisition. And it was and still is. There's been a lot of investment along the way, but still is a highly differentiated part of the security portfolio. The gaps, I think, that we saw were really twofold. One is the breadth of the security portfolio. And the second was the product, the portfolio cohesion. It was still being run as kind of a completely separate product or really a separate platform. And that limits our ability to cross-sell content delivery customers into security and vice versa. It also just makes it harder for the go-to-market to run those kinds of portfolio-wide sales engagements. And that means it's more expensive to run them. It's less efficient.
You're not moving on to the next customer in the next quite so fast. So there's a lot of efficiency to be gained. So what we did is we pushed very, very hard, and we did quite a lot of investment on platform unification. First step was to deploy the WAF at Fastly's edge. So every single server in the Fastly network that's running our content delivery service is running our full WAF solution. It's one completely cohesive, fully unified edge. And then the second piece was to take the management control plane, unify that as well. So the developer and the user experience is fully unified. And that's what can drive that very simple land and expand sales motion. It drives a better customer experience, better developer experience.
And in some ways, it gets back to our core, one of those core Fastly value propositions of the most programmable, the most flexible solution. By being on one platform, we really deliver on that promise. And then the last piece was filling out the rest of the portfolio. The infrastructure has been so programmable and underlying tech so good that we've been able to drive a really solid feature velocity here. The bot mitigation stuff, with all the WAF data coming to our edge already, we had a really significant competitive advantage there to build some really best-in-class technology. DDoS with the breadth of our edge and the performance of our edge. Again, we were already running and have been for years really best-in-class DDoS protection. This is a truly turnkey DDoS solution, which is great.
And it just gives our sales teams, it gives our customers so many more options. So we're really happy with where we are on the portfolio side right now. And with Scott and the team, we're very excited about the kind of go-to-market push for next year.
Yeah. Maybe just to flesh that out a little bit. I mean, I think coming to those customers with a full platform is in itself a benefit. But what has Scott been doing in terms of adjustments to quotas and comp plans, et cetera, to better push some of the security portfolio?
The new comp plan comes out in three weeks, so I'm not going to steal his thunder. What he's been, I think, focused on in addition to comp planning, which of course will have a strategic component, so a component that incentivizes our sales teams to capture not just the security part of the house, but the compute and observability as well. We will be incentivizing that. I think it's more about ensuring that our teams can engage with the CISO organization at the highest possible level. That means bringing in top security executive talent, not just himself, but security executive go-to-market leaders who are on his staff. He's done an amazing job there. He's brought on leaders even from some of our direct competitors.
And we've also seen some really deep thought about how security specialization will work through the organizations, how security sales teams can be targeted and comp directly on the success of the security business, and how we ensure that we have the security expertise deployed regionally throughout the teams. I've been very impressed with Skysix. It's a very strategic approach, a long-term approach. And yeah, I guess that's giving me confidence in the kind of go-to-market push for next year.
Got it. Okay. I want to come back to compute in a second, but maybe we'll get Ron back in the mix. The big news on your side, the conversion that you just passed through, can you talk through kind of the strategic rationale there? I think the balance sheet looks like it's in pretty good shape.
Yeah. I mean, I think we had set aside the plan that we were going to address the upcoming current of the 2026 debt. Over the last two years, we've paid down that roughly just under $1 billion or about $1 billion of debt down to about $350 million of debt. Earlier this week, we issued a new convertible, 2028 maturity for $150 million. We were able to repurchase about $157 million, almost $160 million of that debt, leaving a balance of about $188 million. That with our cash gives us kind of net cash of that 2024 maturity of over $100 million. Given our cash flow trajectory, we believe that we have a healthy balance sheet and that we've taken kind of the action that we said we would do.
When we looked at kind of the opportunity, we were really focused on, given where valuation has been, how do we optimize really the dilution that we would take on this? I think we were able to time it to some favorable market conditions. We struck a balance to take a higher conversion premium. We got almost a $20 conversion premium on the equity for a slightly higher interest rate.
Gotcha. Very clear. Maybe just to touch on margins in general, it's been a pretty strong push towards operating margin, better work there. You did have the restructuring earlier this year. Kind of where's that? Where are you on that? And how are you thinking about the trajectory on operating margins?
Yeah. I mean, I think the restructuring was important for two reasons. First and foremost, it allowed us to restructure our organization, to be able to align our organizations for the scale and strategy that we have today, to drive organizational efficiency. It also reduced our exit rate coming out of 2024, which gives us the headroom to make some of the incremental investments in go-to-market that Todd spoke about while staying on our operating profit model of driving 35%-40% incremental operating margin on incremental revenue.
Gotcha. And then on the gross margin side, I think you're getting pretty close to the long-term target. How do you think about balancing gross margins, especially as you're maybe going to customers with a broader portfolio, potentially some benefit there?
Yeah. I think when we look at, we've sort of been targeting a 75%-80% incremental gross margin. It's going to fluctuate from quarter to quarter, but we think over a 12-month period, we're going to see incremental gross margins in that level. The biggest drivers to being able to continue to deliver that are going to be scale. That's a big opportunity, as well as, as you touched on, continuing to increase the revenue from security and compute. Those are both incremental adders to gross margin and give us the opportunity to continue to drive those incremental gross margins even as our gross margins move into the 60s.
Gotcha. Okay. And can you just talk about the high-level operating margin for cash flow margin targets? I mean, it sounds like there's going to be additional push to accelerate kind of sales of the broader platform, kind of how that's considered in your long-term margin guides. And then with the convert, I think taking on a little more interest expense, kind of how that could potentially impact that.
Yeah. So a couple of things. Coming back to kind of the restructuring, that's certainly unlocked an opportunity to make some of those investments in go-to-market to drive some of those incremental revenues that we spoke about. We continue to drive kind of the financial model around the incremental gross margins that I spoke about, driving incremental spending in that 35%-40% of what that incremental revenue is, which drives that 40% incremental margin. Coming off of three years of improving our operating margin loss, we expect to continue to do that into 2025, and as we've said, we expect to be free cash flow break-even. In that number, while we are taking on this interest, we had planned to do a refinancing. We did index on a slightly higher interest rate.
That's maybe a $3million-$4 million headroom that we can accommodate in our plans for 2025 to still achieve free cash flow break-even.
Very clear. Todd, I wanted to finish just on the compute opportunity because I think it's pretty interesting. But can you just talk about what Fastly has from an architecture perspective? You talked a lot about kind of developer flexibility. How important is that? And then as you think about the compute opportunity going forward, some of your competitors are investing a little more heavily in the ability to run containers and VMs at edge. Just how important do you think that is for kind of the edge compute opportunity? And where's Fastly kind of on that path?
Yeah. So from an architecture point of view, I always feel like we have a pretty strong competitive advantage. We have a purely software-defined infrastructure. So our POPs are basically just compute and storage behind data center switching, all of our infrastructures and software. It means it can be upgraded. It can be made more performant, more efficient. We can drive up our margins through R&D, which has been enormously helpful. Along the way, it's helped us drive up gross margins over the last couple of years. The software architecture on top of that is also, it's pretty cutting-edge. It's a pretty modern architecture. So it allows us to be nimble. It allows us to demonstrate some of this feature velocity, like you've seen with the security feature set this past year. And that's only getting stronger with a unified platform.
The more unified our platform is, the faster our feature velocity can be because we're always leveraging the entirety of what's already been built for each additional feature and product line, and that brings us to compute. Fastly's core value propositions for years have been sort of performance and this programmability, flexibility, developer experience, and compute in a lot of ways is the next step in developer experience. We focused on where developers were most excited to bring best-in-class performance at the edge, which is in serverless, and that's what our current compute architecture is built on. Our compute technology is built on. It's on a technology called WASM, WebAssembly. And that gives developers the best-performing, easiest-to-target edge compute platform. As far as the competition looking at VM solutions and container solutions, I do think there's some merit there, to be honest.
I think there's a lot of workloads that really, for performance reasons, even for security reasons, should be deployed at the edge, but they're currently written probably in containers, and they should be targeting to a serverless solution, but that's a more difficult port, so in order to ease that transition, I think a container solution has merit. We're looking at that right now. Maybe VM might be a bridge too far. Like how deep into the legacy should you explore? I'm not so sure about that, but I think from a container point of view, I think there's merit in terms of accelerating adoption. I think the future is going to be serverless at the edge, no question about it, but as people look at how they port it over, yeah, I think there's no reason to believe that containers wouldn't help there. We're already exploring that option.
Super interesting. Okay. The other, I think, interesting point in the compute business is just the opportunity to help deliver some of the generative AI inference. And you launched a product, AI Accelerator. I think it's still in beta. Just kind of, what's been the reception from customers? What have been the interesting use cases you've seen? And how do you think about that becoming more material?
Yeah, I'm excited. So we're planning to launch this month. It's going to be very exciting. There's been amazing response from customers. In some interesting ways, I wasn't expecting. The AI Accelerator is largely driven towards an LLM use case. So for customers who have user-facing chatbot engagement, tech support chatbots, sales representative chatbots, we see this on websites all around the world. We have one on fastly.com. You can check him out and ask all about our products. And so for people who are running those chatbots using an LLM, the AI Accelerator can deliver a much better user experience, a much more human-like chat experience by lowering latency of responses. Like for example, on the OpenAI platform, it might be five to six seconds. We can make all of those responses sub-second. And we can radically reduce the cost that people are spending on those central AI clouds.
So there's a lot of benefits to the AI Accelerator. And it can really be built. The developer experience is amazing. You can integrate it with a single line of code. So that's a dream. I wish all of our compute products operated that way. The feedback through the beta, we got a decent amount of beta adoption, more than I expected. We got some interesting feedback. Like for one, we launched the beta with only OpenAI support. And we got a lot of feedback on different models people wanted to use, different systems people wanted to use. We added Gemini to the beta. We'll continue to look at new models and languages, sorry, new central AI clouds that we should be integrating with here, even private clouds, which we got a lot of requests for.
We got a lot of interest on the financial side as much as the performance side, which usually isn't the case for a beta. So that was, I think, really interesting feedback. And we got people really pushing and testing the efficacy. How intelligent are the responses? How good is the semantic match? The actual AI running in our cloud does the semantic match in order to pull the correct responses. And making sure we get that right is so key. So super interesting feedback. I am very excited to launch. And I guess I hope to have a more comprehensive update for you next quarter.
Amazing. All right. I think we'll finish up there, but super interesting conversation. So thank you all for joining. And thank you for being here.
Thank you.
Thank you so much. Appreciate it. Thank you all.
Thank you.