Conference Transcript - FROG 2026-05-27

Andrew Sherman, Analyst — TD

We are live. Thank you, Jeff, for being here. I'm with Jeff Schreiner, head of IR at JFrog. I'm Andrew Sherman, analyst covering it as I get my stuff together. You guys have had an amazing past year or so, certainly for the stock and your numbers. Talk about what has driven that. The AI tailwinds you've seen from greater usage, how sustainable are those? What exactly is driving it? All that stuff.

Jeff Schreiner, Head of Investor Relations

Yeah, thank you, Andrew. So, again, thank you for having us. We really appreciate attending the conference. I think as it relates to cloud, what we said kind of started last year and has continued on into this year is that last year we started to see the spark. And it's a spark that we think will eventually lead to a fire. and that spark was the introduction of experimentation and the use of coding agents into the enterprise and we started to see that really accelerate here in the first quarter of this year as the adoption and usage as has continued to increase year over year since last year at this time and I think that's been one of the core drivers as teams within the enterprise are working just as JFrog's team is doing internally at training the frontier models you might use one you might use two as in our case we use multiple models training those models creating those agents all of this is creating activity it's not yet necessarily led to the fire that we would describe to where we eventually lead to a higher level of those that activity into production binaries but What I think it has done is it has created a lot of activity that has generated a lot of cloud usage. And I think more so what we've seen is that in the era of AI, what has shifted for JFrog is that the era of human interaction is starting to be phased out. And in the era of human interaction, that was where people viewed the most importance historically, i.e. source code. And I think it's starting to be demonstrated that the real asset starting to emerge in a world of AI is, in fact, the binary.

Andrew Sherman, Analyst — TD

Yeah, yeah, definitely. Q1, cloud group, 50%. You've kind of been in this 45%, 50% range. Some of that is from overages, not commitments. And I know you're only guiding to commitments. But if we think about the sustainability of that cloud growth and the building blocks of it, what is the path forward there? What's like a normal growth rate number? I know it's probably hard to say.

Jeff Schreiner, Head of Investor Relations

No, I think it's a fair point, and I think it's what we're trying to achieve in the way that you asked your question, Andrew, is how durable is this growth? and I certainly think we're back into a time at least for now much similar to what I described as the Wild West the years of 20 through 22 where the word budget was somewhat scratched out of the financial software dictionary I think we're we're back at that time right now because everyone is pushing towards this you know build my AI models build my autonomous agents and and we're all working towards that and just like that period I think that eventually we'll have to get back to a period of more normalized, whatever that may look like in an AI world. And I think that what we're trying to explain to you guys as durable is how we guide, as you noted. What is committed? What is the customer signed a commitment to JFrog for? For which in fact, we have a legal recourse in which they must use that amount and pay for that amount. And I think that you've seen the floor in our revenue. So you want to know the durable growth. well that would be the growth without the variability that we've benefited from from this high degree of usage above those commitments and that was 116 two quarters ago 117 last quarter 118 this quarter knock on wood that during this euphoric stage we're able to continue to capture some of this into a higher commitment get a two handle on that at some point hopefully and then That would represent, at some point of correction or rationalization in budgets, a good go-forward growth rate for the company from which to build off of.

Andrew Sherman, Analyst — TD

So it sounds like you have some line of sight to a 120 NRR and it's not guidance.

Jeff Schreiner, Head of Investor Relations

We're hopeful. We're at 118 now without the benefit of the overconsumption, and it's on us to continue to try to convert that consumption into a commitment.

Andrew Sherman, Analyst — TD

Let's talk about that. Last year, you did have some over-consumers come in into higher deals. Is there, and I know you can't, you have thousands of cloud customers, you can't chase them all, and some of them just don't want to, they don't mind paying the overages. But eventually, there's no reason why they won't convert. They're a happy, high-using customer.

Jeff Schreiner, Head of Investor Relations

Well, I think what you're describing is, and I described this to our CFO, Ed Grabscheid, that I thought that we would have some people that might get a blank stare when we talked about this, and I can understand it being a finance guy, because the world in which you described, Andrew, is the world that we had worked in and were familiar with, where the customer went over, was very apt to then come back, work with us, work with the rep, and find a higher commitment level to get them off that overusage or that penalty rate if they had, in fact, gotten to that level at that point. The reverse is what's happening today, which I think is a little hard for people to understand is that I am not wanting to take the economic value that I may capture and sign in a higher agreement because I'm unsure if that's really going to be enough for me or is it going to be too much years from now if I'm in an era of experimentation. So we've ironically found that there has been pushback from customers to signing a commitment and much more willingness in an era of again, they're not receiving pushback from their executives so in an era where I'm being forced to continue to buy that next token we're continuing to see the customer want to spend and is willing to spend and pay the higher penalty rate in this era that we're in today now as we just discussed we think that eventually will change over time but that's how we then capture the commitment and also in the conservative and you know prudent philosophy we use in providing you guys guidance and only guiding to what the customer has in fact committed to us.

Andrew Sherman, Analyst — TD

Wow. Okay. Great. So, is it fair to think about the benefit in usage you've seen? A large part is still experimentation, not live AI apps in production, but you've already seen a usage benefit, so is it fair to think about that as a leading indicator and you'd see even more binaries created once these actually go into production?

Jeff Schreiner, Head of Investor Relations

Well, I think that what you're highlighting is I'll point this out. So we just released our State of the Union report and it kind of tells you about artifacturing, what's been going on inside artifactory over the last year. Over the last year that we released this report from last summer to this, binary growth inside artifactory is 136% in the amount of binaries. So you're seeing a strong amount of binaries that are in fact being created in the era of AI. And I think that this overall consumption of binaries will continue to accelerate because what we've started to see now in the world of AI is that it is no longer good enough. There was always an alternative to JFrog previously that maybe I was good enough. I supported enough languages for your organization that you didn't need everything that JFrog offered. Well, those customers are now finding that they are, in fact, exposed even in languages they, in fact, may not feel that they program heavily in. I think that has been a change for JFrog and that will then shift how customers are looking to interact with JFrog and the type of commitments over the longer term that they are willing to then engage in and enact with us.

Andrew Sherman, Analyst — TD

Wow, that's great. Okay. You have some of the AI native companies as customers already, I think three of the top five. What are some of the use cases, or what do they have in common? Are they cloud versus self-managed, some of them? Is there a pipeline of more out there, anything like that you could?

Jeff Schreiner, Head of Investor Relations

Sure. So as of March 1st, we've described having kind of three of the five. I think the five can be somewhat interchangeable depending on your views of different companies and their models. We're not really able to talk openly about who they are. I would say the three that we've had as of the March quarter, the uniqueness there was that they're all using us with artifactory primarily, similar to our enterprise. No one has yet taken security, and all of those three are self-hosted. Now, nothing to report finality yet, but we hope to have something to discuss in a general sense as Q2 earnings come to pass and talk about possibly a fourth lab that has worked with JFrog that would be unique, and we won't be able to issue press releases or anything of this nature, but in talking about it, I think the uniqueness that we would bring to you is that they have approached JFrog to work with us in a hybrid nature. which would be the first one of the AI frontier labs to do that. I don't think that the others aren't precluded at some point in the future of not being hybrid, they've just chosen today that they see that as the differentiator for themselves.

Andrew Sherman, Analyst — TD

Wow, that's exciting. I look forward to that call. The one thing about this market is you don't have that much competition. I mean, there's a couple, but you kind of own this market. And I do, in the past, have gotten some questions on, okay, this doesn't sound that hard, but you realize it is after covering it for a while. Well, what would you kind of highlight as the top three to four competitive differentiators, reasons why this cannot be replicated?

Jeff Schreiner, Head of Investor Relations

Well, I think it can be captured in one word. It's a word I'm using a lot this quarter. And it's the word I hope everyone starts to realize and associate with JFrog. and that's scalability and the scalability of focusing on that one asset you know i what shlomi has done and the team has done at jfrog for for the better part of you know almost two decades is say look i'm that friend of yours that was you know focused on this one thing and we all thought he was crazy told him hey come out come have fun do this and then that that crazy idea turned into something huge and we're all like wow i wish we had done that and i think for shlomi and the team at JFrog what they did is they stayed focused solely on the certain asset and never deviated from working with that asset and seeing the importance in their eyes of that asset and now that asset has evolved to where the software industry also recognizes that as a very important asset and so I think that that's one of the big changes that we've seen as that in the world of AI we've shifted from the human interaction to the machine becoming start much more important And I think the scalability of supporting all these languages is starting to become something that's just recent, I would say, Andrew, is things that we're hearing back from customers and talking to some of the key engineering teams inside of JFrog is a unique aspect of the fact that having four languages, as I said before, that I work with primarily. In the past, I might have gone to an open source alternative, a hyperscaler container registry, but I can no longer feel safe that because I program in only those four, those are the only four I need to secure, protect, manage. And we've had already incidents in which customers have looked at alternatives and then found to be hacked because they were hacked by a language in which they weren't even programming in at the time. And so I think that that scalability of JFrog and building that architecture up to now support 35 plus languages and continue to add because I think that the next programming language that's widely used may not even be built as we discussed. And I think that that's been the real differentiator that I think people have missed, is that no per se to get up to 10 or 12 languages may not be that hard, but it's that incremental, the ability to surpass that and get to the level that we're at and add functionality like security, add additional languages, and keep everything in terms of speed and security in line in that platform.

Andrew Sherman, Analyst — TD

Yeah, wow. Great answer. Sure. Security. Speaking of security, it's been a huge growth driver. It's 10% of ARR. Curation has seen great demand lately. There's a new software supply chain attack seemingly every week now. How is business there? How's the pipeline there? What kind of adoption of the customer base do you think? Like, is every large customer a potential curation customer? and I know it's a pretty big uplift when they do that. I know you're not sizing it, but how's your ability to get budget for that? Has there been an unlock because of what's happened?

Jeff Schreiner, Head of Investor Relations

I think that there was certainly, we had the incident budget things occur in Q3 and Q4 of last year where there was an incident and then there was a budget right away to meet that incident budget. I think that would be more attributable to what some might classify as fear buying, that there were procedural steps skipped by many customers during Q3 and Q4 to implement curation quickly given the environment you described in which we're now seeing these attacks happen almost hourly, daily if not, right? And so I think what we have tried to do in that sense is continue to maintain the ability to support and secure all of those customers and also benefit from the fact that it's becoming widely kind of accepted or at least all of us in the investment community, I think look at the fact that a firewall seems to be very logical in the overall way that you structure and protect your software supply chain. And so curation has seen an outsized benefit. I would say that to give you that idea of such was maybe deployment and pipeline before Shai Halud in Q3 was 50-50 between our advanced security product, which is kind of the inside the castle security, and the firewall and curation. And since then, it's been very much more weighted to curation. I don't think that this is an incident deterministic type of buying pattern anymore. It's no longer that an incident like Log4j happens. We solve it, kind of subsides. A few years later, another incident happens. That creates some buying motion. I think this is constant and in front of our customers' faces. And I think that they're looking at how and when they'll be deploying curation. And I think our pipeline continues to show strength in that nature.

Andrew Sherman, Analyst — TD

Wow, that's great. and there's no competition for curation you guys created it this was a customer customers asked for this basically right well there has been no

Jeff Schreiner, Head of Investor Relations

alternative no I think that others have quickly tried to come up with some alternatives seeing the value there but that brings you back to the fact that one knock on wood that today curation has not been penetrated penetrated and I think that some of these other alternatives have come out claiming to be that good enough solution and unfortunately for some people they were then penetrated and had to learn the hard way as it relates to that and so I think that we've been able to continue to demonstrate the value of what we created with curation and that it is very beneficial when you create curation to be tied in with something as a system of record is artifactory.

Andrew Sherman, Analyst — TD

Moving to some of the newer products you talked about at Swamp Up last year AppTrust with the DevGovOps, more recently of Skills Registry with NVIDIA, there was also Fly, the agentic repository from last year. How would you stack rank those as far as what we should be paying attention to near term?

Jeff Schreiner, Head of Investor Relations

I think governance is the next big item for JFrog. That becomes much more critical in the enterprise as we move from training Agent 101 to governing agent 101 as he's now working more autonomously and without human interaction and how we equate that and how I think people kind of caught on or the light switched is that you know if you think about a courtroom Andrew you have the lawyers in a courtroom and the lawyers are creators and they're doing their brilliance but I don't think any of us would want to go into a court situation with something hanging over our head in which only the lawyers were present and they were trying to deny each other's motions, and there's constant chaos and chaoticness. I think where JFrog sits and what differentiates us is that we're not part of the creator class. We're part of the governor judge class and enforcing the policies in which we're trying to say, okay, Mr. Lawyer, be brilliant. Do what you do, but here are the rules for my enterprise for which you can do what you do. I think that's what separates JFrog and will separate JFrog even when the machines no longer need us and are no longer really interacting in our language and they begin to create what they really would like to create, which is the binary, which goes into production, you're still going to separate the creator and the

Andrew Sherman, Analyst — TD

governor. Yeah, love that analogy. There's no update on like number of customers using any of these new products or anything, right? No, and thank you. I think

Jeff Schreiner, Head of Investor Relations

that, again, governance is on the come and is very critical. It is not here yet today. It is not in our 26 guidance in any major way. We'll see what unfolds as the world moves quickly in 27. But again, I think this is something that we're building and compiling. You added the skills registry. That's the other side of scalability. That's the ability to scale new revenue opportunities inside Artifactory in the world of AI. You think about it. I create Agent 101, and it has that specific skill and what it needs to do. Artifactory is now being able to then track that agent and say what is it doing, what did it last do, is it hallucinating and is it acting differently and we found that customers are willing and ready to happily accept what we're offering as an MCP registry and skills registry. We're seeing some great POC success initially with some of those but that's going to be the next leg I think for JFrog and where you know people will want to start looking as to the next growth opportunity for JFrog.

Andrew Sherman, Analyst — TD

Yeah, that's great. And the agentic or artifactory light, is that still with Fly?

Jeff Schreiner, Head of Investor Relations

Well, Fly was built really, Andrew, to help us in a world of these smaller teams who are starting off pure agentic. And they can easily plug and play with artifactory and make it a learning experience to how agentic problems are solved, maybe at a smaller scale but then take that know-how and so when our large enterprise customers come to us and say I want to solve this problem we're able to go back to them from our interactions with Fly and say well you know what we solved that maybe on a lesser scale but here's how it was implemented in an agentic world and I think that's the value that we're looking to extract from Fly today.

Andrew Sherman, Analyst — TD

Awesome. We're out of time. Thank you Jeff. Thanks everybody.

Jeff Schreiner, Head of Investor Relations

Thank you Andrew. Appreciate it.