Matthew Prince: stay wherever your permissions are. So if you're a customer that really cares about, I need to make sure that all my data stays in Germany because I'm a German healthcare provider, then we can actually set that up now where your data can actually stay resident in Berlin, Frankfurt, Munich, the various data centers that we have. inside of Germany. And we can do that with a level of granularity that no hyperscaler can match. And so I think that will increasingly become an advantage of Cloudflare's network because of what we're doing. And then if you layer on top of that the Zero Trust and SASE tooling, that's going to allow us to make sure that agents can only get access to the things that they have permission to get access to. And I think that's actually going to be a bigger tailwind to that space. And, again, we won't be the only winners there, but I think across the board you're going to see a lot of these principles that sort of the most forward-leaning companies had adopted previously, that those will become much more standard with what people are doing with things like OpenClaw. I mean, it's amazing how the fact that we have a self-service version of our zero-trust products As people are experimenting with things like OpenClaw, they need to have that fine-grained data control, and we're basically the only game in town to deliver it.
Operator: Your next question comes from the line of Seket Kalia of Barclays. Your line is open.
Seket Kalia: Okay, great. Hey, guys. Thanks for taking my questions here. A nice start to the year. Thomas, maybe for you, I know the growth in different acts of course has different impacts on on gross margin and of course you know we spoke about sort of the proactive optimization here in the business as well as as we get through those changes here i think you said by the end of q3 how do you think about sort of the the the net impact to opx and how should we sort of think about gross margins as as those other acts continue to grow yeah
Thomas Seifert: You know, it's a topic we've been talking about for a while, that the margin structure is different, the cross-margin structure is different across the various apps, with the developer products being the cross-margin weakest. But we also said that despite this fact, all products are pretty much equal when we look beyond cross-margin and look at performance from a unit economic value. And I think this will be a transition that we will get you're ready for an investor day, that our operating margin becomes a better measure for competitiveness of products than cross-margin. Just keep in mind that one of the biggest moves in cross-margin the last quarter was free traffic moving to paid traffic and moving into cost of revenue. While that is decreasing the cross-margin, It's literally a wash from an overall P&L perspective. And I think you'll see more movements like that. It will be up to us to give you the right insight. But what is clear to us across all products with the opportunity in front of us, that the unique economic margin and value is going to increase over time. With the guidance that is in place already today, we are getting north of 46% from a rule of 40 perspective. And we think we have visibility to reaching, you know, north of 50% next year. So that all shows you how much potential is. And we just need, you know, to give better insight in how the various parts are coming together and provide force in that direction.
Matthew Prince: Yeah, the one thing I would add to what, oh, sorry, I was just going to say, the one thing I'd add to what Thomas said is is just to put a finer point on what I think is the most important part there, which is since our founding, Croppler's always had a free version of our service, and that provided a lot of different benefits. One of the largest was it gave us data to build the security models and do everything we did. And we haven't actually worked that hard to convert free customers into paying customers. And so the traffic that was associated with those free customers went into it as a marketing cost. What's been fascinating, though, is that that giant pool of free customers, turns out a lot of them are developers. And so as we've now built incredibly compelling developer tools, most of our developer tools aren't free. They require at least some payment. And so you're seeing a lot of that free traffic turning into paid traffic. And so it actually might be that the cost of customer acquisition that we have for those really high-growth products like our developer platform is are really fueled by what we've built over the beginning of CloudFlare and some of those Act 1 products. And so while that shows up somewhat funnily in gross margin, it actually is a sign that more and more people are adopting the paid products, including the developer platform products, which is the part of our business which I think I am certainly among the most excited about.
Seket Kalia: It's a great reminder and certainly a geography shift to your point. Matthew, maybe for you for my follow-up, I'd love to hit on Act 4 a little bit and CloudFlare's ability to sort of manage that relationship between AI tools, scraping type of tools, and content owners. Maybe the question is, what are some of the milestones that you need to see in order to see that business inflect? Is it big lighthouse accounts? Do you need industry consortiums? It's just such an uncharted territory. I'm curious how you think about it.
Matthew Prince: Yeah, I think the way that we think about Act 4 is that the business model of the Internet, which has historically been advertising and subscriptions, is about to change dramatically over the next five years. And exactly what it changes to, I think, is still an open question. And I think it might not be one thing. I think it might be several things. Because of how much of the Internet sits behind Cloudflare, we have a seat at the table of defining that. And so I think there are a number of different things that I'm watching. So one of them is I think that some part of this is going to be some kind of microtransactions for any request that agents are making. It might be fractions of, you know, fractions of fractions of pennies. But, you know, if you think about the, I don't know, it's about 500 billion requests that pass through CloudFlare, you know, in any given second. Some percentage of those, we think that there's going to be some ability to have some micropayment that is made for that. Because something has to do with the infrastructure. And if you look at the growth in energetic traffic, if you look at the growth in sort of non-human traffic on the Internet, somewhere in 2027, we think it's going to surpass human traffic, and it's not going to slow down. And so we've got to figure out something else to build it. And so I think one of the milestones we'll look for is how do we figure out what that sort of micropayment infrastructure looks like? And the challenge is, like, nobody can handle the volumes right now. And so we're looking around to partner with people. We're looking around for everything. But right now, the sort of transaction volumes that people are excited about, like a million transactions per second – we need something that's significantly larger than that. So I think that's one thing. I think on the other side, first of all, it's really important to be clear that the answer isn't that everyone wants to be blocked from AI or everyone wants them to be paid. For example, Cloudflare has a whole bunch of developer documents. We want those developer documents to be in every single LLM that's out there. And so we make it as easy as possible for LLMs to crawl our developer documents. On the other hand, if you're an ad-supported business, then your content being crawled is actually a threat. So I think we're trying to provide tools on both sides of that. The side that you focused on was the folks that want to block it, the ad-supported folks that are out there. And I would say that the first milestone that we've seen is that we went from being relatively low in terms of our penetration in the media space to today dominating that space. And so I think that's the first sign. And what I hear from media company execs is they are signing better deals with AI companies because we've given them the tools to be able to control who has their content. That's the early lighthouse signal. I think the question is, how can we then take that down to the long tail. So, sure, the Condé Nast or the Datas Meritas of the world can negotiate their own deals, but how do we make that available for everyone that's on the Internet? And I think that's going to be really some lighthouse sort of deals with some of the foundational model companies. And, you know, I don't know exactly when that will come, but I will say that when we listed what our top six priorities were, were for 2026. One of the six was making sure that we make real progress and see the first revenue that we can then pass back to that long tail of the internet in order to help make sure that we continue to create a healthy ecosystem for content creators. And I'm pretty confident we'll make that goal.
Operator: Your next question comes from the line of James Fish of Piper Sandler. Your line is open.
James Fish: Hey, guys. Thomas, you just mentioned Rule 50 there potentially. But given what you're seeing with the demand behind inferencing, I guess what's the team's willingness to go after more of this opportunity and really drive more megawatts behind the network to host more of those inferencing use cases like what you're seeing out of some of your edge peers?
Thomas Seifert: I can get started, but Matthew for sure has an opinion on this. I think you see us leaning in in this opportunity already with all the force that we have. This is the reason why you saw developer count go up by a million in the first quarter alone. We are continuing to optimize margin for these products, but we are not at all restricting growth. Just the opposite, I would say. There's no restriction on leaning into this opportunity.
Matthew Prince: I think that, Jim, at risk of being critical, I think that people don't understand the difference in our business model versus the hyperscalers. The hyperscalers business is to buy a server and then to lease that server back, you know, ideally for five times or more of what they paid for it. And so if they don't have servers to lease, then they can't grow their revenue. And so their CapEx has to invest ahead of whatever that demand is that's out there. We focus on very different things. So the thing to watch for us is when you see us publish a blog post about how we figured out how to get more utilization across our fleet of GPUs or how to get more models loaded quickly across GPUs. That's real IP that we're inventing internally. And the metaphor to think about is, once upon a time when I was in college, I remember this new thing called the web was starting, and so we needed to have a web server. And so we literally, from Gateway, I remember ordering a box that came with cow prints on the outside of it. We bought a gateway server and we plugged it in because there was no idea of virtualization. And then VMware came along and after that you had Docker and containers and that was sort of the journey that everyone went on. We're still at the stage with GPUs of buying the physical server and needing to use that for most of the industry. And so across most of the hyperscalers, you're seeing utilization rates of their GPUs that are in the single digits. Whereas we're... slowly getting our GPU utilization to approach what our CPU utilization is, which is up in the 70% to 80% range. And so as we do that, we can actually continue to, with the fleet we have, service the requests that come in and invest behind demand as opposed to investing ahead of demand. And so because our business model is different, that's allowing us to continue to keep up with the inference demand and and also do a lot of experiments and trials and things to capture developer mindshare that is very different than if, again, what your business is, is essentially leasing a physical server that you bought and trying to get at least five turns out of that lease.
James Fish: Yep, makes sense, Matthew and Thomas. Maybe just following up on the security side, look, you guys have been aggressive about displacing customers The legacy hardware guys across security, firewall, VPN, and so forth. Are you seeing any compression in enterprise sales cycles for large-scale zero-trust deployments or macro approval still kind of elongated? And really the crux of it is the supply chain issues, the component issues, causing more customers on the enterprise side to start to evaluate more of the cloud opportunities out there for protecting their environments. Thanks, guys.
Matthew Prince: I mean, the hardware companies seem to have nine lives, and I don't know how many more they have had to use. But I think that as you see that there are vulnerabilities in hardware, which some of you, Palo Alto, had issues with. As you see, again, supply chain shortages, especially around memory right now. I do think all of those things are pushing more and more people to evaluate that they need at least some part of the cloud as part of their infrastructure. But again, I've been... I have been impressed by how long the hardware players have continued to operate and hold out. So I'm not willing to call that this is the time that there's going to be a complete change. But I do think that there's – I do think that the cloud continues to show its resilience and that all of these things are tailwinds behind our business and other cloud-native businesses that are out there.
Operator: Your next question comes from the line of Gabriela Borges of Goldman Sachs. Your line is open.
Gabriela Borges: Hey, good afternoon. Thank you. Matthew and Thomas, I wanted to get your observations on how the fleet mix may be changing between GPUs and CPUs. And specifically, Matthew, you were just talking about how GPU utilization is approaching CPU utilization. Are you also finding that there are some AI inference that you can now route to CPUs? And Thomas, I imagine that has an implication on the unit economics of how you're serving up the AI influence market?
Matthew Prince: Yeah, I mean, so we want to make sure that for customers, we're just abstracting what the most optimum silicon is behind the scenes. And so for some models, CPUs work great. For others, we want to make sure that GPUs are available. As we deploy... You know, a server, those servers today have, you know, they sort of come with all of the various parts, and so it's not you deploy a GPU server, you deploy a CPU server. When we deploy a server now, it's got a CPU, it's got a GPU, it's got a certain amount of memory, it's got a certain amount of storage, it's got a certain amount of capacity in the network. And those are all resources that we're constantly trying to both balance and then create opportunities around. And so I think what's different about us is we're not renting an H100, but we'll have H100s across our network. And then we're trying to match the workload that makes the most sense with the silicon that's behind that and with what a customer is paying for. So if you're paying us more, then we're going to give you a faster and better experience than if you're paying us less. And so I think that that has always, like, at some level, Cochlear has always been just a giant scheduler. And so what we're effectively doing is dispatching those jobs across this and then being able to drive it. The question earlier about, is our cost going up? because our developers are using more AI tools, the answer is yes at the margins, but much less than we see from our peers because we can use that giant scheduler to essentially route those tasks to anywhere we have excess capacity across our network, and we can prioritize those based on what is more important or less important.
Gabriela Borges: Matt? there is the data point intra-quarter on entropic announcing managed agents. I would just love to pick your brain on where you think that type of infrastructure intersecting with the LLM creates opportunity and or risk for the collaborative business model.
Matthew Prince: I mean, I think that we're, so without talking about specific customers, I'll say that what we're is, first of all, that we're partnered with all of them and get access to all of their labs, and that they see our infrastructure as critical to being able to deliver this. And so they are actually looking for partners to be able to have that infrastructure run on. And so, like, the examples of, you know, in the case of we launched something called Dynamic Workers, which allows you to very quickly stand up something which is significantly more efficient than a container. Containers are too slow and too heavy to actually be able to respond to these incredibly fast agentic workloads. And so what AI studios are doing is they're looking at this and they're seeing the opportunity. And so to give you a sense of what I'm naming them, One of the, you know, large AI studios, in just the last 15 days, went from essentially zero dynamic workers to over a million dynamic workers running across the platform. And we're seeing almost everyone excited about the underlying tools and technologies that we built. So I see what Anthropic is doing now. as being very positive to the infrastructure that we see. We see that as an opportunity to deliver incredible value across that. And we see ourselves as very differentiated in the space and able to provide significantly better performance at a significantly lower cost than anyone else who's playing in space today.
Operator: Your next question comes from Fatima Boulani of Citi. Your line is open.
Fatima Boulani: Good afternoon. Thank you so much for taking my questions. Matthew, along those lines, with respect to the agentic edge and agentic AI systems moving into production, I wanted to harken back to a lot of the announcements that you made at Agents Week. It's very clear that there is a fundamental rewiring that you're working on to natively power agents to basically do their best work, to put it simplistically. And a lot of this does seem like it is a frontier monetization opportunity for you, but I'm wondering if you can speak to what sort of halo the rise of production agentic systems could have on the more flagship established parts of your business, i.e. the revenue and monetization halo to Act I, Act II, and certainly Act III products. And then I have a follow-up for Thomas, if I may.
Matthew Prince: Yeah, I think for the Act 3 products, the developer platform products, I think it's the most obvious, which is that we're just seeing more code being created. It needs an efficient place to run. It needs to have a set of of primitives that can act as fast and as ephemerally as agents need to act. So if you imagine that you're a company that's building something that's going to plan a vacation for someone, what you really need is the ability to essentially spin up an extremely lightweight sandbox that writes code that assembles kind of all of the different parts of what that vacation may entail, puts it together, pulls it back, and then blows the sandbox up very, very quickly. To do that with any other platform is extremely expensive and slow, and we've made it with some of the things that we've announced during Agents Week and otherwise, we've made that just drop-dead simple. And so that's driving a ton of use. And again, I think that the way to see that is the fact that we added a million developers to our platform just the last quarter, almost as many as we did in all of last year, is extraordinary. And if you look at growth across the workers' platform, more than three-quarters of that is from new customers. And the growth rates are pretty extraordinary across that. That's also, again, to some extent putting pressure on gross margin because those are less optimized for gross margin products. But be confident that it can continue to drive GPU utilization and all those things that we'll be able to get more out of that. I think to the other part of your question, just as more – every time an agent does something – If you think about it, if you type something into ChatGPT or any of those things, to search the number of sites that get searched, the amount of traffic that gets generated. If I'm looking for a digital camera as a human, I might visit five websites if I really care about it. My agent is going to visit 5,000. And so that's going to just drive significantly more usage there. which is the biggest driver of our Act 1 revenue. And again, I think unlike some of the pure play CDNs that are out there, the agents aren't going to go watch reruns of the Super Bowl. They're going to drive things that actually drive real traffic to real e-commerce sites. And that's where CloudFlare and the huge block of the internet that sits behind us is really valuable. And then for Act 2, Again, as we talked about already, I think being able to very narrowly define what data an agent has access to and what data they don't, we're just seeing more and more of that usage, especially in the self-service category, which there really isn't another sort of sassy, zero-trust, self-serve competitor out there with any sort of scale. And so that's with things like OpenClaw driving a lot of usage there. And what we've found time and time again is as hobbyists or individuals adopt technology, they inevitably start to bring that technology more and more to work. And that's what we're seeing as we win more of the enterprise accounts across that too.
Fatima Boulani: Thank you. Thomas, just on the pool of funds, you're sort of in year three in earnest of having this motion with your customers mature. I was wondering if you have any comments or observations on a pending renewal cycle from maybe your earliest vintage of pool of fund adopters and maybe what trends you are seeing from a renewal and expansion and expansion of usage vectors as you reengage with some of these customers that are coming back to the well, so to speak. Thank you.
Thomas Seifert: Yeah. Well, as we are now in our I would say six quarter of pool of funds. It becomes a much more standard tool in our go-to-market motion. Folks are familiar in how to deploy the tool, when to deploy it, when it makes sense. So I think we get efficiency in the process. From a renewal perspective, you heard this in Matthew's prepared remarks, we had our highest ever renewal rate in the last quarter, and that includes all their pool of funds deals that were up for renewal. So I would say the hypothesis that this would be a tool that not only allows us to work expansion really well, but also becomes a very sticky tool from a customer engagement perspective has turned out to be true.
Operator: And your last question comes from the line of Shaul Yal of TD Cowan. Your line is open.
Shaul Yal: Thank you. Good afternoon. Thank you for squeezing me in. Thomas, you mentioned quota carrying sales capacity continues to accelerate. Could you provide some more color on your expectations to continue to grow capacity relative to productivity? And I have another follow-up for Matthew.
Thomas Seifert: Yeah. So when I said we are not touching quota carrying AE sales capacity, What goes along with that is where we see significant productivity is in the support ratios for these AEs. So the ratio is going to change significantly, which means we are freeing up dollars, and within the same spent envelope now of dollars, you can deploy more quote-occurring AE capacity towards our market opportunity. And this, of course, then will allow us to continue to drive productivity from a go-to-market perspective.
Shaul Yal: Got it. Matthew or Thomas, partners increased to 30% of revenue this quarter. What's driving this continued increase and how much more channel mix would you expect going forward?
Matthew Prince: Yeah, I'll start, and Thomas might have more to add. I think that, you know, this really started with Mark Anderson laying out a story two years ago that we were going to have a motion that really included partners and made sure that they were able to deliver on that. I think that it has been an incredibly successful way for us to sell, especially our Act 2 products, which require a lot more of a consultative sale and a lot more of work making sure that the integration is done extremely, extremely well. That I think will continue going forward. I think that the big question is going to be what partners are really able to leverage this new world of agentic AI in order to just get additional value and scale and velocity. And I think that's what we're evaluating across the partnership world. I think there's going to be a lot of change in that space. But I think that partners will continue to be an extremely important part of our strategy. And just like, again, I think a lot of our businesses is changing. A lot of our partners' businesses are changing. And we're seeing that the ones that are delivering the most value to customers and are the best at getting success at selling, our tools are the ones that are embracing new ways of selling and servicing and making sure that the customers are successful with our tools.
Operator: Thank you. That concludes your Q&A session. I'm going to turn the conference back over to Matthew Prince for closing remarks.
Matthew Prince: So, again, just wanted to emphasize this has been a hard day. We've never done something like this in Cloudflare's history, and we take it extremely seriously, and we know how much it has affected people who have been friends and colleagues. I am confident that those people who are leaving us today are going to go on to take what they learned at CloudFlare and help build many more great companies. In fact, it's amazing to see how many people are already writing in saying, anyone who got traded to CloudFlare, we would be happy to interview. We're going to make sure that we take care of those people, but we also want to make sure that we are hiring for the right roles. This isn't about us downsizing. This isn't about us saving costs. This is about making sure we have the right people in the right roles to build the future. Our mission is to help build a better internet. That's an important mission. It's never been more important as the internet goes through all of the transitions with AI and agents, and CloudFlare is going to lead the way. I'm proud of everything that we're doing. I'm sorry that we had to take the action that we did today, but I believe it's going to make CloudFlare better for the future. Thank you. We'll see you back here next quarter.
Operator: This concludes today's conference call you may now disconnect.
