Awesome. Great. Well, my name's Lucky Schreiner, software analyst here at D.A. Davidson. Really happy to have Jeffrey Schreiner with me, a long-lost cousin, I guess. Something of the sort. Really happy. JFrog's been a great story. A lot of momentum in the business. Security's an interesting kicker too as well. Maybe just to start though, for any new investors, how would you describe JFrog's core role in the software supply chain, and why is it becoming increasingly mission-critical today? Well, first and foremost, thanks for having us, Lucky. We really appreciate the support and appreciate attending the conference today. I think in a general sense, when you think about JFrog, you should think about the word automation. That's what we've always done, is it for the software supply chain, we've been the aspect that you are trying to deliver speed and security at the same time, and that's been the crux for many who have tried to do what we do, is the ability to keep things in line, both speed and security, as you scale and add more support and more open-source languages into the architecture. What we are akin to is that in an organization, you are building code that has typically been done in the past with human developers generating source code. That source code today has been written primarily in a language called Git. You might be familiar with the likes of a GitHub or a GitLab who are very successful in the source code business. They generate that source code, and then that source code is compiled, and maybe 10% or so of that source code is combined with packages or open-source software from outside the organization, and that creates a package or i.e., the binary, and that binary is then transported through governance gates and is then eventually handed off and deployed by the organization into production, where it's then managed by the observability and runtime players. We sit in between kind of the source code and the runtime and provide governance to source code and provide data and visibility to the runtime organizations who may need to access the metadata that we have that is stored and held within binaries. Awesome. I think we got to start with last quarter two. Very strong results. Everyone's asking about why cloud growth is accelerating. A lot of that's driven by overages. What are some of the dynamics there, and how should investors think about the sustainability of cloud growth moving forward? Yeah. Cloud's been strong. I think it was strong last year, and it was strong in the sense that I've been here almost five years at JFrog, and for the majority of that time, the monetization strategy within JFrog, inside of cloud, is that you can either migrate from self-hosted, where you're managing that type of deployment yourself, and then shifting it to the cloud. A lot of our growth, when I first started, that was the way in which we grew. Last year was unique in the way that it was the first year that we saw consumption the other way. I buy a data package from JFrog, and each time I access a binary, I'm utilizing some form of that consumption that I buy from JFrog. We saw the use of consumption really be the driver of what was driving cloud growth last year, and very much so in Q1. In Q1, we talked about the broad use of all packages, and thus created a very high level of cloud usage amongst our customers in the world of tokenization. What we saw is that we were able to outperform the expectations, the guidance level, by roughly around $7 million. We carried forward of the combined business only four and a half into the full year. What we've tried to say to you is assume that that's the committed portion, security, other things that we've won and signed commitments to, versus the two and a half delta more or less being attributable to that penalty rate, that usage above the commitment rate that the customer has been running while they're running hot right now. I think that that is something that is happening because we still remain very much so in the world of experimentation and training the models. I'm telling the model to build this process. I don't like it, do it again. I don't like it, do it again. That's creating this activity. Eventually, I think we'll get to a portion when you talk about our cloud growth, where that activity will become more specialized, more production volume type levels. When the model starts determining that I don't need to use source code, I want to make binaries because that's what I'm starting to produce, these production-level binaries, whether it be from source code, human, agent, what have you, will be inherently more valuable than the binary activity we're seeing today because those will be assets deployed valuable to the company. They will need to be updated, which causes builds, secured, which causes the buys for security, stored, which is part of your consumption, and maintained. I think there's still a lot of activity as it relates to production binaries that we haven't seen the full effect of yet. Yeah. On that note, how do you see the model usage evolving with customers today? Do you think the models they use will keep getting bigger and bigger, or will they become more specialized, and how might that impact JFrog? That's a very timely and good question, Lucky, because as it relates to JFrog, I'm hearing it from others in my conversations when they speak to other organizations, one way that we're looking to manage token spend in this world of token maxing, as you guys have defined it, is to utilize very specialized models. We use Opus a lot. That's probably the highest usage model that we have today. We use Sonnet, we use other smaller models that we are finding can do certain tasks very well, and I think that that is how other organizations are doing that. That brings up a unique aspect that then brings favor back to us at JFrog, in the fact that if I'm using four disparate models to complete one build, I'm going to have to inevitably bring back the four models to one point. Where is that point? That point is the system of record, i.e. JFrog. Yeah. I wanted to touch on, too, the rise of different coding languages as agents use all kinds of coding languages. How does that impact your differentiation, especially today? I t's magnified something that was always there, Lucky. You know this in talking to me that I've talked about scalability as inherent value, that why we win. I think I define things kind of now in the software world, two years ago, and the AI world now, right? In the software world, the scalability was always there and inherent. We had a competitor who started essentially as the same year as JFrog did, but has been stuck at being able to reach a certain level of only about 15 languages that they can support during this entire path. We've scaled up to 36 languages, that scalability is now very evident in the world of AI because what we have learned in 2026 is that it is no longer that you can feel like you did in the software world, that I program in these four primary languages. They're widely used, so I can use something that isn't all that JFrog offers. It's very sophisticated. I can use this good enough solution. Unfortunately, we found in the world of AI, that will lead to you being penetrated. Because while you may not say you program in a certain language, the model inherently is looking to be efficient. It is going to say, I am building X. The most efficient package for X is Y. Just because you don't program in it, doesn't mean it won't go get Y, and now you've been penetrated. I n the world of AI, our ability to scale and scale with AI is becoming a big differentiator for us, as binaries are now becoming much more of the asset of focus. Yeah. I feel like that naturally brings up security. You guys had in your State of the Union report, malicious NPM packages grew by over 400% in 2025. Feels like that trend is continuing in 2026. That tends to create a catalyst for Curation adoption with you guys, and would you expect that to remain a long-term secular tailwind for you? Very good question. I think it's changed relative to maybe if you'd asked this question four years ago, I think you would be going down the right path, that it used to be that we would have an event like Log4j. That was the last major software supply chain attack. Enterprises would rush and spend to patch and fix, and then there would be a lull, and two years, three years would pass, and then maybe another software supply chain. Now it's happening daily, if not multiple times a day. They're all happening at the binary level. Thus, I think that people are starting to realize that a product like Curation, which is essentially a firewall for your software development organization. It sits outside the developer IDE station, and it curates various repositories and packages that you, as the central CISO or policymaker, have designated okay to utilize in your organization. I think the pipeline has remained very strong because the attacks continue to be in your face as a CISO, and it's something that you are now having to take note of. It's also becoming more important because there was an alternative you might've been looking at or looking to evaluate for Curation that had told you that if you only utilize a few packages, utilize this, you're better off. You now don't have that as an alternative, and people realize that, and I think it's just continued to fuel the demand and keep the pipeline strong for our sales team. Yeah. As the attack surface shifts upstream, how would you say your security conversations are trending from a budget perspective, and maybe compare and contrast the demand between JFrog Advanced Security and Curation? Yeah. A very good question. I think that, let's start with the second part first. Curation and JFrog Advanced Security. We've been quite open in the fact that we said, before the first Log4Shell event in Q3 of last year, the pipeline and the deployment was probably roughly 50/50, both Curation and JFrog Advanced Security. Since then, it's obviously skewed very heavy to Curation because industries are obviously seeing the need to have somewhat of a firewall type security out there protecting you, and those that had it in Log4Shell can tell you what it did for them in protecting their organization when the first Log4Shell attack happened, and the packages that were being blocked that you referenced in our State of the Union Artifactory report went up exponentially, right? That protected those organizations. Other organizations, think of it if you're a CISO who has now been given software supply chain, and that's now in your wheelhouse, but you've never dealt with it before. You've dealt with networking firewall security. You understand the concept there. A firewall-like product then for Curation, to your point, seems very likely to have to basically stop the inflow of what I will allow from a central location, then tell the organization, I don't want to interrupt your productivity. Be very productive and creative, but just do so with the products you have today in Curation. Thus, it's become very popular, and our CTO snaps at Ed and I sometimes saying that, you speak a lot about Curation to the investment community as security, but it's actually governance, if they don't see it yet, they will. What I would say to you is that some mind you may be seeing from maybe some of the smarter organizations could in fact be those looking to purchase Curation not only for security, but for the governance aspect of it down the road. We equate Curation to more of a caffeine-like high, right? It sits outside your developer organization. You'll want to put as many contributing developers, those who interface with Artifactory, behind that right away. It's a seat sale, just like security is, so the number of developers. Quickly, you can get a high level of developer seats right away with Curation. Advanced Security, on the other hand, is for good actors gone bad inside the castle. With that, I'm taking a tool. I have seven disparate point solutions. I want to consolidate all of them into the JFrog Advanced Security platform. I do so in year one with one solution. I run it in parallel for 9 to 12 months. I replace it. I get to year two. I'm running one tool. Now I'm doing another one or two for replacement. We'll have to see over a multi-year timeframe if some of these customers even replace all seven or eight in their first three-year, multi-year agreement. A much longer tail aspect where it's just the early innings of this consolidation trend playing out for Advanced Security. Yeah. You brought up governance, I want to shift to that. You guys released at JFrog swampUP last year AppTrust and AI Catalog. Those were exciting and received well. How would you characterize that opportunity for investors and maybe any timeline for when that might start to become more significant? I would say that for all the investors listening out there, that's where you want to start shifting your focus. If you talk to our customers, they will tell you a lot of the same things that you'll hear coming out of my mouth. That is the fact that right now today, I'm very much concerned with make AI work. I'm not so worried about the cost, things of this nature. Once AI is working, I'm going to be very concerned with governing AI. You referenced AppTrust and AI Catalog. Those, along with Curation, along with things that we've now added, such as MCP Registry and Skills Registry, which is scalability on the revenue side, adding new revenue opportunities in Artifactory without any real investment, will be part of our governance bundle. The team is meeting in an offsite this week to start our initial thoughts for 2027, and certainly, I think governance is going to be a topic du jour about how we will bring this to bear and the monetization strategy that we will use. We've seen a high degree of interest in AppTrust, AI Catalog. I think AppTrust is still being worked with customers in developing that product and bringing that product to bear so it's ready for primetime when we really have it as part of the full governance platform. AI Catalog has been doing quite well and seen a lot of interest. I would even tell you this, Lucky, that the MCP Registry and Skills Registry, while not a driver today, we've been seeing successful POCs close down in several weeks for this product, with people saying, yes, it does what it does, and it does it well, and we'd like to move forward. I think it shows you that the governance Lego, as we like to call our Lego strategy, with Artifactory being kind of your base Lego. Now you add security on for more value. You add governance on for more value. This is what continues to make us very strategic within the enterprise. We think that governance is the key. Why is that? Because as I was going, in another answer, kind of alluding to, the machines are going to learn that they are only using source code because they have to speak with us humans, and they don't really need source code. What do they really want? They want to produce binaries because that's what's going to be inevitably put into production. Even in that scenario, when a model is building at the binary level, governance will be critical because the creators and the governors will always be separate in the organization. I want to tie this to your go-to-market and talk about net new logos. You guys have historically shifted to targeting the largest enterprises. On the flip side, I think what you're alluding to, a lot of enterprises are reaching a certain scale where they have to adopt JFrog moving forward. Maybe how do you think about net new logo growth moving forward with all this new product? It's a focus, and we've got a new CMO in Genefa Murphy, and I think she's definitely got that as one of her tasks in which she's trying to help drive logo count. I would always say that we can always drive logo count. Yeah. T hat it's quality logo count that you're after. I think what you're seeing from us is that the last two years we've been going through kind of this year's theme for JFrog, ironically, is a rebirth. It's rebirth in the sense that some of the logos that had fallen off were logos that were in our lower price subscription tiers, had never really grown, liked JFrog, but were never really intent on upscaling their subscription or utilizing more JFrog. Were sitting at subscription layers that just weren't going to lead to that land and expand type of customer we've shown you in the slide deck. I think that we've seen a roll-off of a customer that came to JFrog four or five years ago when I joined, when we were selling you Artifactory as a product, hoping that Jeff's organization takes Artifactory. He's successful, he's productive. Lucky sees it. He wants to bring it into his organization. It was more product driven, bottoms up from the developer. Circa today, we're being sold top-down. It's strategically placed inside the enterprise that the developer's told what they will use, JFrog has become a strategic partner. I think that's what you've seen in terms of logo as it relates to count. I would tell you that the ASP is up materially in the five years that I've been here from a new cloud land to what we see today. That's just because I have customers now landing in the platform, and if you land on the basic version of Enterprise+ in the cloud, you're going to already be in that 100,000 cohort just with that. T hat's the difference that is what's happening, is that the logos are not maybe accelerating in terms of a high number, the contribution per logo is certainly growing, and certainly for new logos, it's growing. Yeah. I want to talk about some of those high-quality lands, specifically the AI natives. You guys have three out of the five. Maybe first, level set, how do you define an AI native? There's a lot of definitions out there. Two, I think most investors would assume they would build a solution like this internally. What was the catalyst on their end for adopting you guys? Yeah. Look, I say there's probably seven, you can interchange the last three for every person, depending on what they view as the models. When we talk about the frontier models, I think you can be safe to assume that we're talking about many of the models that we are all using or all hearing about in the news today, some of the major model companies out there. While there have been talk by many, and I think that that was a concern earlier this year, that perhaps inevitably these things are so large that they will encompass and take over so much in software and even binaries. What we learned is that there was one that tried to do just that, and that was one of our first frontier model customers who had tried to build JFrog on their own without benevolence to JFrog, and failed at doing so for that very key word I have said many times today, scalability, thus came to JFrog and began working in tandem with us. Now they are also one that has us on a moonshot-like project, which would lead to the management in a model as a package type of business architect that you would have at almost line rate speed, thousands of large language models being trained, secured, stored, transferred, all of this information being transferred to and maintained from Artifactory. The unique aspect of the way AI looks at Artifactory and the system of record has changed dramatically the way the intelligent people inside JFrog look at Artifactory versus maybe the restrictions of the software world before. I think if anything, what investors should take away is that there has been a significant intangible benefit to working with these guys and becoming very close with the frontier models and where they want to go, and solving the pain points on the binary aspect of getting there. Yeah. On that front, how do you target these companies moving forward? Does the go-to-market look different? Maybe in terms of the deployment, any nuances between cloud versus self-managed? Well, let us start with the second first, because the three we have today, as of the March quarter, have all chosen to go self-hosted initially. I do not think that precludes them from becoming hybrid at some point and moving to private or public cloud, what have you. A fourth model that we hope to sign during the second quarter, nothing is ever official until it happens, but to the extent that it does happen, the uniqueness that we have been trying to communicate to you guys, because we will not be issuing a press release, we cannot talk about these types of companies, but the uniqueness is that this particular one would like to engage with JFrog on a hybrid basis, which would bring cloud into the frontier model contribution, not just self-hosted in terms of how they are doing it. That is how the major models today have chosen. You had a group that says self-hosted is the right route, and what gets back to management myself from the sales team is as simple as, I have something special I don't want to share with the public cloud yet. Okay, understood. You can understand from one, the one that's working on the moonshot, why they want to get to building the data center, because that's inevitably where they're trying to go. Right? I think that each one of them, and I do it akin to the buy side when I have these events at conferences, right, that they all are asking the same question. They're asking it a different way, all trying to get to the same answer. Right? I t's just a matter of preference for the model guys of trying to get to the same, solve the same problem, but we may do it different ways. Yeah. I think shifting gears just a touch, going back to the overages in the cloud, how has customers' willingness to commit to multi-year deals maybe changed today? What are customers telling you, and how are they approaching it? It's changed quite dramatically, and I would say that's just over 12 months. A year ago, we were going through the same thing in essence, in that we had strong usage in Q1. Agents were just starting to be introduced to the enterprise. We saw this tick up in usage in Q1 last year, and it was still very much the behavior of the historical trend of that was more of an oopsie, and let me come back now and capture economic value by signing a higher commit, a new commit with you that has a higher commitment of consumption, and thus I'm no longer utilizing that overage. That's captured in the new commitment I've made with JFrog. What's different today is that our customers are saying, I don't know what my cloud usage is two days from now, two weeks from now, or two months from now. Signing an agreement for two years, I don't know if that's right because I don't know whether my cloud usage may change, the way I use cloud could change. If I move from experimentation to productivity, does that change things? I try to equate it for you guys on Wall Street is to think about, is your PM asking you for a table-pounding idea? If he's not and you walk in there with one and it blows up, did you do yourself any service? Probably not, because procurement office right now is not feeling the pressure from the office of the CFO to go out and rein in costs or be my cloud expert and buy me the exact amount of cloud and don't get it wrong. Think about it. If you're that procurement guy and you're Joe Procurement and you think you're the best at buying cloud ever, you go in, do a two-year commitment, and in four months you've used it up and you're back hat in hand to the CFO, you didn't do yourself any favor in the environment you're in. I think right now, customers feel that running hot is the better choice. Inevitably at renewal, they will have to make a choice, and we'll see what that choice may in fact be. There has been a shift and change where typically an overusage would lead quickly to a new commitment, and we aren't seeing that just yet. Yeah. I think historically, you guys spoke to seasonality being more back half weighted with the large deals in the pipeline. It feels like with the 24/7 nature of coding agents, seasonality is changing a little bit. I guess, how would you characterize seasonality moving forward? Not to be flippant, I'd probably crush it up and throw it out that window. The reason I say that, right, Lucky, is because to your point, when I first joined here a few years back and cloud migration was the monetization mechanism for cloud, it was that I spent the first half really planning that deployment. The second half executing that deployment, and thus the second half more waiting typically in terms of our business. Last year, being consumption-driven, it was more steady Q1, Q3. The dynamics of Q4, which will always have this dynamic, at least for the near term, is that we all have holidays to celebrate and we work less. There's less development work, so there's typically rarely if ever any usage above commitment in Q4. T hat seasonal trend will stay, but that could even go out the door eventually someday. If Agent 101 is working while we are celebrating the holidays, that could change. That could change the dynamics of seasonality, but not this year, not anytime soon. I think as it relates to the consumption trends, it certainly seems like it is a steady as she goes for the first three, and then we have a seasonal trend of more of a slowdown in work, slowdown in development in Q4. Yeah. Anything to call out in terms of sales compensation with overages being a bigger piece of this quarter? How are you thinking about that moving forward? I think that our salespeople are kicking dirt right now. That is the way that works for us, and I think it is the right way. It is akin to paying you, the PM, for a high watermark, right, for something that could evaporate. They will not get compensated, our reps, until they are able to sign that higher usage into, in fact, a commitment. That commitment goes into RPO. Usage is captured today via revenue, and we can capture that, but inevitably, we would like to capture that into a longer-term agreement to have better visibility for ourselves and the customer, and the rep would obviously like to do it because that would help them in their comp structure. Yeah. Maybe taking a step back, how do you think investors should think about your long-term growth, the algorithm between cloud versus self-managed security and some of the newer AI products? I'll take security out right away because security's always going to fall, whether you're self-hosted or cloud, in the cloud. We maintain that database now, so that it is always updated, and you get five updates over the cloud through the day. That's the delivery mechanism, we recognize all security, whether self-hosted or cloud, in the cloud. Certainly a large driver for us. At only 7% of revenue, 10% of ARR as of last year, it's the new growth driver for us on top of cloud growth that we're seeing. I think longer term, you're going to add governance on to that growth mechanism as that next Lego of growth that will be additive to both the cloud and the security aspect of what we do. O ver a longer term, certainly we've now achieved 51% as of Q1, the business in terms of cloud mix, so cloud has now finally become the larger portion, and I think that that will continue. However, there will be a cohort that will remain in self-hosted, self-managed, either through regulation or through a choice to do hybrid architecture. We will continue to support those customers, and I think you're going to see a lot of federal type of penetration. We had some success in federal last year. I think that's a new opportunity in self-hosted for us. A lot of work is done in those regulated industries within the government. To the extent that others that don't move over time in self-hosted, I think that those are the areas where we continue to add a lot of value, and we'll have to see how we can maybe extrapolate additional value, whether it be pricing or other things, into self-hosted, as that becomes more of the niche business and cloud becomes the real growth driver. Yeah. You talked about governance, I want to bring up again the report that some enterprises are maybe a little overconfident with their AI usage and not as well prepared on governance. Is it going to take something bad happening, or what maybe is the catalyst for customers to recognize this? T hey recognize it. I just don't think we're at the point where they're trying to What are they going to govern? The agents aren't there yet to govern, so to speak. We haven't trained the agent, or we haven't turned it loose. To your point, could there be some that don't see the importance? Sure. I think that what I'm hearing back from my own team, and then in conversations with the likes of investors and others, is that they're very much hearing back the governance conversation from other companies, and that this is very much a concern for them. I think more so you've seen the behavior you're describing happening somewhat on the security side with the, I'll try it good enough, I get penetrated, or, I haven't worried about binaries before. Why do I need to worry about securing them now? Something happens. Okay, now I'm wanting to step up to the table and talk to you more about Curation and the JFrog security products. Yeah. Maybe one more question and I'll open it up to investors if they're interested in asking a question. In terms of the share buyback that you announced recently, that was nice to see. How do you guys think about capital allocation moving forward and maybe the build versus buy equation, especially given your successful past acquisitions? Yeah. We've had success, as you noted, thank you, with two large acquisitions in Vdoo. That was the security that you're seeing today. In Qwak in 2024, which has emerged into really just AI and MLOps together, and those guys are the core of AI Catalog and AppTrust, delivering a lot of value to us today. I think that when we look overall in terms of where we're going with the business and how we see each one of these playing out, is that we're really looking at the business as taking security and driving it on top of Artifactory. They're going to be sold together at all times. The go-to-market there is never that we needed to go and hire a security staff to sell security. We were lucky enough that the industry started to shape a security sale to a panel, and we would bring in an overlay. We treat the customer has one main rep that you interface with, but you bring in overlays for security. We'll bring in overlays for governance because now at the panel level, you're selling to the CISO, the CIO, the chief risk officer. I think on a go-to-market basis, that's one of the changes that we've been seeing in terms of how things have been getting done and how we're approaching things and how the industry is looking at things from a top-down aspect. Yeah. Any questions from the audience? Hi, Matt. I'm curious, so what do you view as your biggest competitive risk today? Maybe differently asked, three years from now, what would be like, hey, we've got to monitor this. If this were to happen, we'd have to Yeah. Thank you for the question. The question was what is our biggest concerns today in terms of maybe competitive concerns, and what might it be for three years from now? I would say to that question, sir, that it may sound simple, but it is very much true. This is an opportunity that this management team and these founders have focused on in the binary diametrically for 18 years. One asset that they have told you was going to be of utter importance, and it never really seemed that way for some time because it was not where humans interfaced with software. However, in the world of AI, it has become very inherent that the binary is, in fact, that asset. I think that execution is our biggest risk in the near term, that we have a very large opportunity that has come to us. The market has come to us, the asset that we have focused on, and I think that it is execution. I think three years from now, it is does someone else try to work the binary in a way that we did not conceivably think. I think that is the biggest risk to us, not others that have tried the old way of I will support a certain amount of languages, and that is good enough to compete against JFrog. Maybe is there somebody that thinks about doing binaries in a way that we did not, and does that create disruption? I would say that there has also been disruption in the five years I have been here. Right before I joined, they bought Vdoo and Security, and that was something that could have disrupted the business. Buying Qwak and MLOps, that could have been something that would have taken over binaries and possibly disrupted JFrog. T hat is the other thing that we try to do as well, is run fast and see these holes that could be there in three to five years and use to your capital allocation conversation. Sorry, I got lost a little bit on that answer there. On the capital allocation, why we have used it to be inorganic when we need to, now we are using our strong cash balance as a way to help us maintain the merit increases within the corporation. We have shifted last year to higher merit increases from cash since we have such a strong cash-generating business and less in equity since we are a little bit more of a We have become a larger organization and now public for five years, and thus, we are trying to start to manage the burn rate. To come back, I guess I woke back up. The capital allocation is the cash will always be favorable to us. We get a very good return on that cash. We can use cash to be very quick and nimble in an inorganic fashion, as you noted, and I just noted with the two acquisitions we have made. It is not that something that it will burn a hole in our pocket. It is something that we see as valuable, and we will utilize it as best we can. The $300 million buyback that we put in place was put in place to be a buyer of last resort, as I think we need to have when we saw the disruption from what we call on February 20th, Anthropic Day. That will be that if we feel the market has disconnected in some way, we may be active in purchasing the stock, and we will certainly communicate what we were able to do with that initial authorization, after Q2 and with the Q2 results. That's how we're going to be looking at that, but that can also be a tool, not that that was created for, but the fact that if we're able to buy some treasury shares and use those in the future for employee issuance, well, that takes off the plate then of coming hat in hand to you guys asking for more shares, asking for dilution to continue to enhance and merit those employees that are helping drive the business higher. Yeah. Any other questions from the audience? Great. Well, maybe, on the profitability side of things, you guys have always had strong profitability. Where do you see additional leverage in the business, and how do you manage the push and pull between investments and maintaining your market leadership position? Yes. We very much see margins and the rule of, recently it's been a rule of 50, it's been nice. It's been a rule of 40, as how we kind of manage the finance organization is managing and running the business. I think that there's leverage still inherent because what we've done today with our guidance philosophy is that we provide you with a conservative, in our view, more of what has been contractually obligated to the customer. That's what we guide you on revenue, but we're not doing that with OpEx. We're giving you what we believe will be our essentially full OpEx. To the extent that we can continue to outperform on the top line, a good portion of that should flow through down to the margin line, and we should continue to expand margins. Certainly, we've had a long-term target out there, issued that target in May of 2023 when we were coming off a year of half a percent of EBIT profitability. Not a lot of profitability to speak of at the time, but told you that we saw a model that could grow to 21%-23% EBITDA margin on a non-GAAP basis. Well, in the first quarter of this year, we reported over 21% EBIT margin. We're along that target, and because the EBIT margin has been expanding, and also due to the fact of things like the cloud marketplace and customers paying multi-year deals up front with strong collections. We've also had strong cash generation and thus strong free cash flow margins due to the fact that profitability has been enhanced over the last few years as well. Yeah. Maybe final question, the NVIDIA Skills product, you guys were able to create that fairly quickly, and NVIDIA asked you to. How do you feel like that helps from a marketing sense and some of these strong partnerships and endorsements that you have in the ecosystem? Yeah. Well, I think that that's a key, right? Is I think that any one partnership shouldn't be looked at as the partnership. We've had partnerships with many companies, and I think if you look at the history of the company, if we had tied ourself to any one of those horses at one time, it might have looked like the right decision then, but down the road it might have not been. I think this universality is what really separates us and why our customers love us because we sit at that center that you want to choose what you plug in. You don't want to be told what you use and how you build your organization. The holistic nature of JFrog has been a competitive advantage for us in how we sit and can see holistically the entire software build and bring more value to you or as the organization from where we sit. Awesome. I think that does it. We're out of time, Jeff. Thanks very much for coming. Thank you, sir.