Conference Transcript - NTSK 2026-09-15

Rob Owens, Analyst — Piper Sandler

All right, everyone, let's get rolling. I'm Rob Owens with Piper, and I manage our cybersecurity and infrastructure software practice. Really pleased to be up here with Netscope and founder, Sanjay Berry, CEO, to talk through the story a little bit and understand what's going on in the SASE world, what's going on in the AI world, what's going on in the data security world. I think you've got a lot of opportunities, a lot of plays in front of you. So thank you very much for coming. Thanks for having me. Welcome to Nashville. So maybe just unpack Q2 for us. I think there was a lot of good trends from a net new ARR perspective, what you're seeing in AI. Just the business in general seemed to, Michelle's going to yell at me or throw something if I say the wrong words, but inflect to a degree relative to where it's been. And it seems like you're on a nice glide path to an upward trajectory. She doesn't like that.

Sanjay Beri, CEO

Yeah, I'll get it. So, look, if you look at our business, we've always had an amazing win rate, over 80% when we get to a POV. We have about 25-plus products. We released about six new ones already this year. So we're this growing, broad platform with now over one-third of the Fortune 100 as customers. And so we have strong – we continue to have strong upsell. You saw our NRR uptick to 114. We had some AI security wins. We continue to ramp our sales force. 50% of our sales team is being ramped. And so we have a lot of good things going for us. Massive TAM, broad platform, right in the heart of sort of securing AI. We have a right to win that because of where we're located in our architecture. And our customers love us. We have the one that highest GRR rates in the industry. And so, yeah, obviously we're happy with our Q2, and we're looking forward to the rest of the year.

Rob Owens, Analyst — Piper Sandler

So you guys have boasted these strong win rates for a while as part of this story pre-public. Why not more at-bats? What's the gating factor there?

Sanjay Beri, CEO

Yeah, so when you look at us, you're right. Like for us, our whole focus, if you're inside on the go-to-market side, is get more at-bats. And how do we do that? Well, one, why are we in the midst of a large sales expansion where half our reps are being ramped and are being hired in the last, say, nine, 12 months? Ultimately, that's because we want to increase our ground game with the number of reps. And so that's underway. We saw that. We started in EMEA, and you saw that in APJ. You saw some of our growth rates, and our North America next cohort of reps will ramp towards the end of the year. So that's one. Two is our partnerships, not just our partnerships with the ecosystem that we've expanded, whether that's Anthropic or OpenAI or NVIDIA or others, but our partnerships with MSPs to hit the mid-market and SPs and SIs. And so really expanding our channel and distribution and then expanding our ground game with number of reps. And then, of course, we're right in the heart sort of of the AI security and networking ecosystem.

Rob Owens, Analyst — Piper Sandler

Yeah, 80% win rates, very impressive in what I think the street perceives to be, these are not my words, a bit of a commodity. When we just look at the transport layer and connecting point A to point B, How do you differentiate in there? Because every networking vendor, every historical firewall vendor plays there. There's obviously some pure plays in the market. So what become the points of differentiation? What drives an 80% win rate for you before we get to the expansion opportunity, just on that true transport layer?

Sanjay Beri, CEO

So if you look at what we do, often part of what we do is categorize in the term SASE, Gartner's term, of course, beautiful name. And more and more gets thrown into that. And it really is this, hey, inline processing point. And the reality is that SASE is such a broad term that what you really need to do is look under the hood. And I call it SASE version one was move your stuff to the cloud, your security enforcement point and do web filtering. That's what we saw a lot, actually, with the existing vendors that we saw that we're replacing. That's not us. Right. If you want to do web filtering and you want to stick it in a cloud, you can pick any one of the three of us. But when you want to go beyond that, when you actually want to say, hey, I don't want a corporate, you know, I don't want a non-corporate version of cloud code with sensitive data. I only want that to be the corporate version. I don't want people sharing externally sensitive data from my OneDrive, right? I don't want this agent going out and retrieving sensitive data and then sharing it with a subagent. That's us, right? That's not the block AI or allow it. It's the world we live in, which is it's not about this binary yes or no. And so we're that sassy, right? And a lot of our patents were on, hey, how do you not be a web proxy? How do you be an high-speed, globally distributed API proxy? And that's a big differentiator versus PAN and versus eScaler. And that's why, for example, recently, one of the largest construction companies in the world, they moved because they got sick of bypassing 70% of their traffic, where all they could do was block or allow it. And so for us, we know the world is kind of coming to us. Cloud and AI is all a bunch of APIs. You need that granularity. So that's one. We're a layer-rate granular proxy, not a web proxy. Two, data protection. Over 90% of our customers use our data protection. We're well-known as sort of the best in data protection. And so whether it's your endpoint, whether it's inline, whether it's cloud, SaaS, AI, whether it's your Snowflakes data store, we will find your data and govern it. And so these two on the security side are big reasons. We use our own models. We have over 190 deep learning models that we use to solve these problems and many more. You marry that with the largest and most performant cloud network. If you measure us versus PAN, you'll see we're many times a magnitude difference in performance. We run our own infrastructure. If you look at our infrastructure versus Zscaler, a similar concept of, wait, you can guarantee sovereignty or better performance. And so we have this unique security and networking advantage, and we've got to increase the ground game to get that out to everybody.

Rob Owens, Analyst — Piper Sandler

Talk about the advantages of owning your own network. And let's just start with, A, is the build where it needs to be, or is there still more work to do from a geographical location standpoint? B, two-part analyst question, of course. Just the functionality that it brings you, the fact you want it. And you mentioned data sovereignty, which I know comes up when we do customer conversations.

Sanjay Beri, CEO

Yeah, so if you look at our infrastructure, just to level set, we take space and service providers are colos. We run the hardware, the racks. We bring in the power, and we bring in the connectivity, and we run it, right? And we do that over 120 data centers in the world, 80-plus regions. Everything we do is in every location at the edge. It's about five milliseconds from anybody in the world. And then on the back, when you come to, for example, an AI destination, we are the fastest path for inference, or about five milliseconds from most AI destinations. And so when you look at this infrastructure, there's a couple of advantages. One, infrastructure operations, people look at it and go, wait, I can turn on your security and it's faster. That's a paradox that normally is never true, but you can do it with us. Two, it can guarantee your sovereignty. We run, like, for example, a large European government runs all their traffic through us. Not only because of data protection, but I can guarantee it's processed locally and their data is stored locally. Three, resilience. Because every data center has full compute and runs all our services, every single one is a backup of the other as well. And so that is sort of the product advantages. Remember, our software was designed to run in memory, in high speed, in all these locations. So it's also the design of how we built our platform and the marrying with the network. And then, of course, from a gross margin perspective, we expanded our infrastructure to cover most of the countries in the world. Eventually, you run out of countries, and so you're not expanding for regional reasons. You're just expanding for capacity reasons. And so the infrastructure in U-Edge also gives us a gross margin advantage as we drive towards our 80% of 77 right now.

Rob Owens, Analyst — Piper Sandler

Great. And I think one of the themes that came up at our security dinner last night, standing room only, was velocity. Velocity from an agentic standpoint, velocity from a threat standpoint. Now, you've built out your platform largely organically, And you mentioned six product releases thus far. Do you think you're hitting the speed requirements that customers are asking for right now? I mean, customers, I'm sure, are drinking from a fire hose and paralyzed. But ironically, this is not new for security. And the average tenure of a CISO has been 18 months to two years. Mine's in the back of the room. I won't make eye contact now. But for, you know, as long as we can remember. So talk about velocity relative to your story, whether you think you're achieving new product releases and milestones that you need to, or do you need to push that accelerator a little bit more?

Sanjay Beri, CEO

Yeah, so we always are the view of never being satisfied. I mean, that's not a good thing to say, but with the pace. And so we always want to push the envelope more. And I think that's part of our culture, just, you know, innovate or die. And I think all companies should kind of view that, especially in tech. And so one is, why can we move quick? One is we built our platform such that, one, everything is done once. There's one network, one client, one policy framework, one data protection system. And what that means is when we build a new product, everybody just gets to leverage that. It's just, to them, it's abstracted. It looks like a component and an API. So it's easier to innovate. and we get velocity. And so we used to spend 80% of our R&D was building this platform. Well, once that was built, 80% of our R&D now works on building new product on top of it. And so that's why we get the pace. Organically, that's been generally what we've done. You know, all our ARR, all our build is generally being organic. And it's simply because we built this true platform, which is not a bunch of Frankenstein things integrated. Many times it's just faster, better, and leads to a better outcome if we do it organically. It doesn't mean we won't pursue M&A for targeted technology or targeted teams. We do that. But generally, we haven't built a new product in organic means.

Rob Owens, Analyst — Piper Sandler

So to the extent that you lose one out of five deals or it goes somewhere else, taking the other side of the coin, what do you think are the core reasons? And I've often found through my 28 years of covering the space that a lot of times IT and CISOs are trying to solve a problem for now and then try to solve the next problem almost sequentially. And it feels like you built this mousetrap that, frankly, can do a lot of the things that are going to be necessary over the coming years. But to the extent that you lose, is it an 800-pound gorilla, easy button, speeds doesn't matter, functionality doesn't matter? Maybe you can go into some of those different scenarios.

Sanjay Beri, CEO

Yeah, so depending on where we are and what we're selling, what use case, you may see different competitors. But if you look at a quadrant, you generally see Apollotos there or so on. In the data protection side, you'll see some others, and then you have legacy. So there's a couple of reasons when you lose. One, a lot of these customers have legacy. They have Cisco VPNs. They still have Symantec or Trellix DLPs. they have. And there is the notion that, okay, you know what? I'm not going to do it right now. I'm not going to move. I'm just going to keep what I have. I'm going to renew it. I don't want to migrate and so on. So there's that. That's one. Two, I think to your point, there is the notion that in sales, there's also the political battle. And so product is one, but as we all know, there's enterprise sales and you've got to manage the politics and you've got to get high and you've to manage the low. And so for us as well, one of our focuses has been get to what I call the triangle, the CISO, the infrastructure ops, and the CIO. Because ultimately, all three, in many cases, for these wins, platform wins, you have to win all three. And so where you lose, you may not have done that.

Rob Owens, Analyst — Piper Sandler

Maybe delve a bit into the agentic opportunity relative to how you see it. And I've always thought you guys were in a very unique position. I think that in order to secure a Gentic, you have to solve for many problems. Not only guardrails around what these agents can do, what they can touch, every transaction. And I've always used the analogy that if I'm throwing a party, I can either walk around and hand different keys to every single person to get in different rooms of the house, or I can put a GoPro on the shoulder and hopefully monitor in line in real time what they're doing. And you have that vantage point. So maybe talk a little bit about understanding how you articulate the agentic vision, where customers are at at this point. I mean, as we said, everyone's drinking from a fire hose right now, and I think that the water pressure continues to increase.

Sanjay Beri, CEO

Yeah, I think so. I think like if you're in a customer's shoes, like a CISO or a CIO, I always tell people for them, AI is also new. It's not like they knew what agents or LLMs were two years ago. they're learning and they're ramping quickly. And so for many of them, they don't come in knowing what they should do, right? There's no textbook or blueprint. And so when we engage with them, in many cases, first of all, you are educating them on what's the right architecture. And the way I think about that is you have, you've always in the history of security had identity, endpoint, and you had this inline processing point, and you had a SIM or a SecOps platform. And that industry is transforming. And so I often tell them, look, with AI security, it's actually not that much different. You have these three points. You have to care about identity of the AI agents and users. You have to care about your endpoint, and you have to care about inline processing. And for us, obviously, like our right to win and easy button is, hey, cover the inline part, enable it, see all your agents, tool calls, protect DLP, watch every prompt and response, and do data threat and content moderation. A lot of our customers do that. And some of the wins we talked about in Q2, that's what they were doing. Now, we've chosen in this area to go horizontal, and so we will also enable that on our clients for coverage. We'll enable it on servers and out of band. But in the end, I think AI security will breed multiple winners. And I think eventually that's where people will get to, and it'll be partitioned in this normal endpoint identity and inline processing point ecosystem.

Rob Owens, Analyst — Piper Sandler

And I believe in the last quarter you talked favorably about the AI portfolio, a lot of the new solutions, what it does, pipeline build. Maybe you can share that again.

Sanjay Beri, CEO

Yeah, so from an AI perspective, we think about AI security as outside of just SASE. We think about it as this new pillar. We released our agentic broker and AI guardrails. Gentic broker is sort of understand agents, what they're doing, monitor the data that they're trying to transact with, catch rogue agents, show shadow agents. You can quickly do that when you enable this on our product. And then we have guardrails, which is sort of zero trust your LLM. I don't know what an LLM is going to say. I don't know what is going to come out of its mouth. How do I moderate what comes out of its mouth? And those two products are really the first products that people started POCing and adopting. We then followed that with releasing our AI gateway, which can live in your premise and do the same for East-West. And we just released our AI command center, which gives you a holistic view of all agentic use in your company, how people are using it. It visualizes it, shows you how data is traversing, and then lets you shape it and govern it. So this is sort of the progression. And really, if you look at our pipeline, it's one of the fastest builds of any product we've had in Netscope. About one-third of it is in POC, and we mentioned that enterprise sales cycles, procurement, getting budget, six to 12 months. And so we see some of that continuing to come in fruition in these coming quarters, but a lot of it towards the end of the year and next year.

Rob Owens, Analyst — Piper Sandler

So the events of the weekend and potentially putting some governors on the release of new models, What do you think the chances are that this throws friction into the process relative to security buying? Obviously, the conversations have been there, again, drinking from a fire hose, but you've got a ramping sales force, which you've mentioned. And so just how are you contemplating that potential risk here versus a groundswell of momentum?

Sanjay Beri, CEO

Yeah, so I think there's been a lot of discussion on pacing frontier models. There's been discussion on mythos and Fable and so on and what part of security does this help, and what part does it hurt? So I think there's two questions there. One, look, in security, I view it as a spectrum from the left to the right. On the left is code scanning, vulnerability scanning, CICD. You get to the middle, it's sort of your data protection and network security. We live in the middle to the right. That's where Netscope lives. Those parts will be accelerated by AI because you will need more of those, and they'll be more important because of agentic use. whereas on the left, it's true. There's companies that are doing code scanning, vulnerability scanning, which will be disrupted because of mythos and fable. I kind of view it as this, you know, if you're on the left, yeah, maybe you're going to be hurt. If you're on middle to right, it'll be actually a big boost. So we live in the middle to the right. The second part is the pacing of frontier models. Pacing of frontier models doesn't actually, whether they pace or not, doesn't actually help or hurt enterprises trying to govern AI security and AI. And the reason is, look, first of all, pacing frontier models, what does it even mean? Well, if you read it, you're trying to put regulators into frontier companies to make sure they're using these models in the appropriate way. Well, you're not going to do that in China. You're not going to do that in open-weight models. Attackers, if they don't have access to a model, will use an open-weight model, and it's extremely powerful. Second, companies aren't worried about pacing frontier models. They're worried about the AI in their companies today. They already have agents. Some of these people have hundreds and thousands of agents that are unsanctioned. They're already using LLMs. They're already using MCP servers. Most of them have no security, no visibility on these. And so I kind of think pacing the frontier is somewhat of an orthogonal and irrelevant topic to enterprises. And what do they do to secure AI? Whether you pace it or not, you know, they're going to have to deal with that as well. And then more largely, I think pacing, there's been a lot of obviously things written about how is this altruistic or is this more to help volunteer models regulate, you know, a duopoly. We stay out of that. I wrote a marketing hype. There's a marketing hype. But if you want more, I wrote a good blog. Michelle's laughing, but I'm not pointing to it. But I wrote a blog on it yesterday, and it's on our site, which goes into more a worldview. you? How does the EU think about it? How would you, if you were in Japan, how would this play with you? And so for a lot more detail, I wrote it up there.

Rob Owens, Analyst — Piper Sandler

So the defenders have always had the challenge of having to defend every single element, the attackers of just exploiting one. Has the balance of power shifted with AI? And do you feel that we're in a less secure place than we were before? And the defenders need to do some catching up? Or is that balance still the same as it ever was in your view?

Sanjay Beri, CEO

No, I actually think it has tilted. You gave the attackers one of the greatest weapons they've ever had. And they have access to it. They don't have governance committees. They don't have to go for approval for spend, right, and so on. And so you've given them one of the greatest weapons. Yet on the other side, enterprises are a little bit on their heels because they have normal process and bureaucracy that they have to go through to enable protection of it. And so I think CISOs have a tough role, but I also think the ones I talk to are actually energized a little bit and excited about, you know what, we can protect this new world. And I think, you know, they're moving in the right direction. We've got to move quicker. And I think in the history of time, most great technologies have come in not through IT. You look at cloud and SaaS, it came in through end users. AI came in through end users. And so I think it's sort of, that's the history of how tech comes in. And I think enterprises, unfortunately, no one's coming to rescue them. You can't rely on a government. You can't rely on a frontier model or lab. You have to protect yourself. So I think people are starting to realize that and, as a result, start to move with pace.

Rob Owens, Analyst — Piper Sandler

I've got time for a couple questions if there are some.

Sanjay Beri, CEO

Go ahead, E. Yeah, so like today, we have a site. You can go to ai-index.netscope.com. It's public. And it actually shows you worldwide AI transactions across enterprises. And we did it for our customers because they wanted to know, hey, what's happening in the EU in pharmacy or in health care? And we show you what AI usage looks like. And so for us, we take trillions of AI transactions on our infrastructure today. And we see that growth is quite rapid. And so the opportunity for us is we know we're the fastest path for inference. We released a study last quarter which shows we cut inference time, the network part of inference time, by 90%. And so we think that's a big advantage. The second is because we see the traffic, it's very easy for us to inspect it because it's already coming to us and protect it. And so, yes, we definitely see both the network part being a big advantage because agents actually communicate and work a lot faster than users. So performance is going to be even more important. And then from a security vantage point, if people want that traffic on your network, then you're naturally going to have a right to win the ability to govern it.

Rob Owens, Analyst — Piper Sandler

All right, I guess one final one, CFO search update.

Sanjay Beri, CEO

Yeah, so one, as you know, Drew, our CFO, is retiring, and he'll be here until we find a CFO. He'll be here to help transition them and so on. And so, you know, we're not in any per se rush, but we did kick off our retained search. We're just kind of a little earlier, but we're got a lot of candidates coming in and we'll keep people apprised. But like I said, Drew's with us till we find one and beyond. All right. Well, thank you very much. Thank you. Thank you. All right.