as they increasingly see the need for JFrog to function as their infrastructure for the software that both machines and humans build and consume. We see their usage of AI-specific packages and their dependencies expanding on top of their traditional DevOps and DevSecOps workload goals. In Q2, we were excited to announce partnerships and integrations that are solidifying JFrog as the enterprise standard for AI-powered software supply chain infrastructure. These investments included the tight integration with Anthropic that brings JFrog security and governance solutions to the millions of cloud-core developers. We also announced the partnership with Cursor, which powers over a million daily users that now have access to development and governance workflows directly in their development environment. We also continue to expand our footprint with the world's leading AI native companies, welcoming a new logo, displacing a competitive solution that failed to scale with its growth. This customer migrated to JFrog Platform as its software supply chain system of record and binary distribution engine across a multi-region hybrid deployment. As AI leaders increasingly build software with AI for AI, this win further validates our strategy and reinforces JFrog's position as the trusted software infrastructure for the next generation of AI native, autonomous, and multi-agent software development. The accelerating adoption of AI development practices and coding agents across our customer base, combined with our deep partnerships with the world's leading AI companies, is also helping us navigate the enterprise shift toward token economy optimization. As organizations increasingly govern and cap token consumption, the economics of software development are changing. JFrog's value proposition is fundamentally aligned with this transition because we are focused on the compiled output of software, the binary, not on the number of prompts or lines of code generated. Whether software is written by a human developer, an AI agent, or both, it ultimately results in more trusted binaries that must be secured, managed, governed, and distributed. As AI reshapes how software is created, we remain focused not only on what is growing, but also on why it's growing and what matters most, the trusted binaries that power production. And with that, I will hand it over to Ed for a detailed review of our second quarter financials and our updated outlook for the third quarter and full year of 2026.

Ed. Thank you, Shlomi, and good afternoon, everyone. We are incredibly energized by our second quarter results, which, as Shlomi noted, exceeded the top end of our guidance range across every key metric. These results once again showcase our consistent strategic execution and operational discipline. During the second quarter, total revenues equaled $163.8 million, up 29% year-over-year. These results demonstrate the continued execution of our go-to-market strategy, fueled by our cloud revenues, growing demand for our security core products, and expansion in our Enterprise Plus portfolio. Cloud revenues in the second quarter accelerated to $87.5 million, up 53% year-over-year, now representing 53% of total revenues versus 45% in the prior year. Our outperformance in the cloud was driven by robust usage across our customer portfolio, which continues to exceed contractual minimum commitments and increased adoption of our security core products. We strategically work towards converting this usage into higher annual commitments. During the second quarter, our self-managed or on-prem revenues were $76.3 million, up 9% year-over-year. We continue to proactively engage our on-prem customers to migrate workloads to our cloud or hybrid offerings as they explore solutions better aligned with the rapidly changing security environment. In Q2, 59% of total revenues came from Enterprise Plus subscriptions, up from 55% in the prior year. Driven by the ongoing execution of our enterprise go-to-market strategy and broader customer adoption of the JFrog platform, revenue contribution from Enterprise Plus subscriptions grew 39% year-over-year in Q2 2026. Net dollar retention for the four trailing quarters was 121%, representing a year-over-year increase of three percentage points and a one percentage point improvement sequentially. These results continue to highlight the strong adoption of our security core products, increased cloud usage, and conversion of customers into higher annual contracts. We continue to demonstrate the strategic value of JFrog as a mission-critical, trusted system of record for our customers' software supply chain with gross retention of 97% as of the second quarter, 2026. Now, I'll review the income statement in more detail. Gross profit in the quarter was $136.2 million, representing a gross margin of 83.2% versus 83.1% in the year-ago period. We remain focused on cloud hosting cost optimization as we anticipate a larger share of our revenues being generated from the cloud. Given our expected increase in cloud revenue contribution to total revenue, we reiterate annual gross margins to be in the range of 82% to 83% in 2026. Operating expenses in the second quarter were $103.6 million, equaling 63% of revenues. This is compared to $86.4 million, or 68% of revenues in the year-ago period. Our operating profit in Q2 was $32.6 million, or an operating margin of 19.9%, compared to 15.2% operating margin in the second quarter of 2025. The continued balance between strategic investment and operational efficiency demonstrates our ongoing commitment to profitable growth. Cash flow from operations equaled $57.1 million in the second quarter. After taking into consideration CapEx requirements, our free cash flow reached a record $53.8 million, or 33% margin, compared to $35.5 million, or 28% margin, in the year-ago period. Now, turning to the balance sheet, we ended the second quarter with $824.5 million in cash and short-term investments, compared to $704.4 million at the end of 2025. As of June 30, 2026, our RPO totaled $659 million, a 38% increase year-over-year, once again highlighting the successful execution of our go-to-market strategy as customers continue to make larger commitments to the JFrog platform. As a reminder, RPO excludes any benefit from the customer's usage over contractual minimum commitments. And now, let's turn to the outlook and guidance for the third quarter and full year of 2026. As we enter the third quarter of 2026, we remain optimistic by the strength in our pipeline and the tailwinds of emerging AI workload trends driving increased cloud usage and security core product adoption. Even as usage trends accelerated through the first half of 2026, our guidance philosophy will remain unchanged as we continue to de-risk large deals due to timing uncertainties and any benefit from cloud usage above contractual commitments. Looking ahead, our outlook remains anchored by three key drivers, growing contributions from security core products, ongoing adoption of our full platform, and cloud growth driven from higher annual customer commitments. We are raising our estimated full-year 2026 baseline cloud growth to be in the range of 41% to 43%. Given the anticipated contribution from our security core products and increased baseline cloud growth assumptions, we now expect our net dollar retention floor be 120% for 2026. Turning to our operating expenses, we continue to focus investments on innovation across our entire platform. We remain committed to a disciplined spending philosophy and confident in our ability to drive ongoing operational efficiency in line with prior execution. For Q3, we anticipate revenues to be in the range of $164 million and $166 million, with non-GAAP operating profit anticipated to be between $27 million and $29 million, and non-GAAP earnings per diluted share of $0.22 to $0.24, assuming a share count of approximately 130 million shares. For the full year of 2026, we are again raising our revenue guidance, now anticipating a range of $648 million to $652 million, representing 22% year-over-year growth at the midpoint. Non-GAAP operating income is expected to be between $116 million and $120 million, and non-GAAP diluted earnings per share of $0.96 to $1, assuming a share count of approximately 129 million shares. Now I'll turn the call back to Shlomi for some closing remarks before we take your questions.

Thank you, Ed. Our second quarter reflected the strength of our strategy and the dedication of our team. We delivered continued cloud expansion, security as a meaningful growth engine, strong enterprise execution, increasing customer platform consolidation, and solid free cash flow momentum, all while maintaining the discipline to grow responsibly and efficiently. To every frog around the world, thank you. These outstanding results are your achievements. Together, you turned Q2 into a major win for JFrog. Your passion, resilience, and focus on our customers didn't just deliver another great quarter, but laid the foundation for future growth. As we conclude today's call, we look forward to welcoming many of you to Swamp Up New York in just a few weeks. Together with customers and industry leaders from Anthropic, Cursor, Morgan Stanley, Microsoft, CoreWeb, NVIDIA, and many others, we'll demonstrate how the software supply chain is evolving for the AI era and how JFrog is helping enterprises control, secure, govern, and scale software creation and delivery in a world powered by developers and AI agents alike. May the frog be with you. Operator, we are ready for questions.

We will now begin the question and answer session. Please limit yourself to one question. If you would like to ask a question, please raise your hand now. If you have dialed in to today's call, please press star 9 to raise your hand and star 6 to unmute. Please stand by while we compile the Q&A roster. Your first question comes from the line of Mike Seacos with Needham. Your line is open. Please go ahead.

Hey, guys, this is Matt Kalitrion for Mike Seacos over at Needham. Thank you for taking our questions. We were hoping you could share some color on the conversations you've been having since the OpenAI models discovered the self-hosted artifactory zero-days vulnerability. It's worth noting from our perspective, it seemed like the whole ordeal was very well handled. But we were just curious if it's delayed pipeline conversions at all or if there's any plan to continue to leverage these AI models to search for other potential patches going forward. Just anything you could share on that whole experience would be helpful.

Yes. Well, thank you for the question, obviously. One that we were dealing with in the past two weeks in great partnership with OpenAI. So as you mentioned, OpenAI have a self-hosted artifactory, and they ran a model that worked on a sandbox with limited guardrails. AI models in today's world should not be treated as free. They should be treated with zero trust, with the security practices that are required around that. Once this AI model found a vulnerability within our factory, they contacted the JFrog team immediately. We remediated fast, worked in great partnership with the security researchers of OpenAI, and throughout the last week, kept improving this communication between us. Obviously, this is also a great opportunity to discuss the cloud solution, of the SaaS solution, to remind everyone was not breached, and also to discuss the security solution that JFO can provide on top of Artifactory. So great relationship, we'll build a better product. More and more vulnerabilities will be found as models are getting into the pipelines. And I think that what counts is how fast vendors are remediating. We are very pleased, very honored with the relationship we built with OpenAI. before and during the incident.

That's great. Thanks so much.

Your next question comes from the line of Miller Jump with Truist Securities. Your line is open. Please go ahead.

Hey, great. Thank you for taking my question. And congrats on the continued really strong momentum here. I want to stay on security. You know, it was really great to hear about the momentum you saw in the quarter. there were a number of pretty significant open source vulnerabilities that came to light at the end of Q1. I'm wondering, did that have an impact on the Q2 security contribution in your view? And was there any of the second half pipeline of security that actually got pulled forward into the first half as a result of those vulnerabilities? Thanks.

Yes, Milo, that's a great question that represents everything we've seen in the past few quarters. Software supply chain attacks are becoming a daily thing. Just two days ago, another massive attack over millions of open source packages coming from NPM, yet another one. So obviously, we start to see that every CISO asks herself or himself, what's the right firewall we should put from the get-go? than what the right scanner we should put on top of our system of record. Obviously, this generates a lot of traction around JFrog because JFrog is one of the unique security solutions that is not just providing a security solution, but also the system of record that needs to be protected. So, yes, the pipeline is impacted by it. We were very pleased to see the results in Q2 that are showing yet another growth after the results of Q1. It supports not only our penetration into the DevSecOps world, but also the growth in the cloud. And looking forward, we are very optimistic, as mentioned on the call, that security will keep being a very strong growth engine for the company.

Your next question comes from the line of Mark Tash with Raymond James. Your line is open. Please go ahead.

Thanks. Yeah, Shlomi, if I could go back to the OpenAI incident. Look, absolutely novel. Your team was great and transparent with disclosing what happened. Had solutions in place to take care of customers. And I completely understand this is a strong argument for adopting cloud and for security. So I did want to ask, though, considering 47% of your business is self-hosted, how have customers reacted? What have you done to ensure customers are patched and the risk wouldn't spread? And then could you actually turn this to a positive for demand, as we've seen with some other security companies that use incidents to actually get closer to customers and then drive broader platform adoption? And that's it for me.

Thank you, Mark. A great point. So I'm thinking about how fast you remediate and how fast you take it to the market. But there is also a very responsible way to treat your customers, and that's transparency. The moment it happened, the first thing that our team did was protecting our cloud customers and releasing a patch to the self-hosted customers. Obviously, this is not in our control, so it can become a tailwind of customers that see the SaaS as a more secured environment. But we immediately released the patched version. It was confirmed by OpenAI as resolving the vulnerability. We were very happy to see that they keep running their models to check if Artifactory is secured and bulletproof, and the answer was yes. And just yesterday, on a black hat stage, they shared with more transparency what happened there. And obviously, some of it has to do with how you configure your environment and what guardrails you put around models. I think that the entire industry is learning about it. What I see inside JFrog is how fast we remediated, the level of transparency with the market, no hidden stuff, like going straight and bold to our customers and demand that they will protect the software supply chain. But I also see an opportunity here to say, hey, you can be much more secure in the cloud. You can even be more secure with security around modern behavior. Now, what we will see next, and have no doubt about that, we will see more models getting more sophisticated, finding more vulnerabilities. It's not a matter of being the scanner anymore. It's a matter of how fast and how efficient you are in remediating and communicating to your customers.

Your next question comes from the line of Howard Ma with Guggenheim. Your line is open. Please go ahead.

Hey, guys. Thanks and congratulations on a really strong quarter and the full year guidance raise. One for Ed, if you look at the Q2 outperformance, how would you compare the mix of higher commitments and overages relative to Q1? And if you could comment on if there was any contribution from the fourth Frontier Lab customer that you added in the quarter and for the full year as well. So, you know, kind of two questions there. Thank you.

Yeah, thanks for the question, Howard. We don't necessarily split out in terms of providing guidance or updates of what was overcommitted revenue versus commitment revenue. But what I can tell you is in Q2, we saw something very similar to what we saw in Q1, which was strong usage across the customer, a diverse group of customers and our install base, continuation of packages going through the software development lifecycle and Artifactory. And we didn't see a decline by any means in the usage over that minimum commitment. And we're very pleased with the end result. In addition to that, we also had, as Shlomi talked about in the prepared statements, we extended our foundational AI labs. We have four of those customers, and Shlomi can share a little bit more about that customer that we landed during the quarter.

Your next question comes from the line of George McCrean with Bank of America. Your line is open. Please go ahead.

Hi. This is George McCrean. on for Koji Akita at Bank of America. Thank you for taking our question. I wanted to ask about kind of the contribution you guys are seeing from your suite of security products, you know, between the three, advanced security, runtime, and curation, kind of how do you stack rank the contribution from those?

Yeah, I'll take it, George, and thank you for the question. And we spoke a moment ago about the amount of software supply chain attack and the kind of, I don't want to say panic, but alerted response from customers and prospects. So obviously, the first thing that they are applying is a firewall between the software supply chain, the organization, and the open source apps. This is J4 Curation, and we saw J4 Curation being adopted rapidly and also referred on the call. J4 Curation by itself is an amazing firewall, but when it comes with the artifactory integration, that's a bulletproof solution that prevents any malicious package, any vulnerability known, any type of unrequired packages to come into your organization. So because it's so simple and because it's implemented in very high integration with Artifactory, obviously our customers and prospects are betting on curation. And as mentioned on the script, none of our customers got affected by this massive attack of software supply chain out there. The second thing is that what happens once you start to run your pipeline, This is where JFog X-Ray, JFog Advanced Security are providing a comprehensive holistic solution on top of your source code or sorry, on top of your system of record to make sure that all the binaries, all the software packages and everything that you will distribute will not only be secured, but also known, traceable and monitored. So later on, when you need to govern it, when you need to audit it, you have all the information. So, obviously, because of the amount of software supply chain attack, curation gets the spotlights now, but our customers are requiring more and more security solutions from JFrog, mainly because of the advantage that we also manage all the binaries for them, and we also play as the single source of truth.

That makes a lot of sense. And if I could follow up, you know, with the second question here, you know, last quarter in cloud consumption, you know, there was usage above commitment, you know, this quarter you guys noted as well. But I'm kind of wondering if there's any change in kind of customer behavior in terms of our customers kind of getting a better sense of maybe how much they're going to be consuming on JFrog over, you know, in the future and kind of getting more comfortable, you know, committing at higher levels of usage, or is that maybe not the case? Thank you.

Well, George, this is a wonderful question. question. What we see is what I'm going to share. A, we see more AI tools being part of the software supply chain that drive scale. B, we see more AI software packages, unique software packages, and we call some specifics like MCP and skills. Just half a year ago, you wouldn't hear those terms. There are new assets. These are all binaries. So, new assets with new software packages. And the third thing is the uncertainty. JFork provides this flexibility that is amazing, not only for the CIO, but also for the CFO. They need to settle on what is the right estimation a moment before they decide what would be the budget of 2027. Now, with our philosophy of guiding you guys with the commitment only. If you combine that with the number of customers over a million dollars, a number of customers over $100,000, the growth in the cloud, you probably understand that our team is doing great work, converting those over usage to commitment. But we provide you with the full predictability and the certainty around our model, and this is why we follow commitments and not usage. usage. Usage over commitment is growing, and we will still be focused on the commitment and the cloud migration of our self-hosted customers to the cloud.

Your next question comes from a line of Roddy Sultan with UBS. Your line is open. Please go ahead.

Awesome. Yeah, thanks for taking the question. Just one from me, Shlomi. I wanted to ask on the shift we're seeing towards increased adoption of open source and open weight models. Like, can you just walk through how you see that trend impacting demand and usage? I'd imagine maybe it's increasing pull through on the security side, but I'd be curious if this could increase the need on Artifactory as well.

Well, yeah, if you refer to the security, you are very much right. That's a great driver that fuels the growth and the adoption of our security solution. But it's not only because of the open source that is brought to the organization. It's also because of the new practices that requires new packages to be managed and secure. Now, every vendor will provide an MCP software package. You need to govern that. You need to have an MCP registry, and you need to make sure that it's secure. JFOG security provides all of these assets on top of JFOG Artifactory, which makes the solution far more comprehensive and holistic. I'm not talking now about who has a better scanner. I'm talking about the outcome, what's the real value that we bring. So, of course, we see growth there, Roddy. And the second thing, it's also the speed and how fast things are happening. And we also mentioned that AI starts to be a wallpaper. It starts to be ambient infrastructure. People just use AI every day for everything, and they expect the software supply chain to be secured, to expect the pipeline to be secured. Now, with the amount of attacks that they see out there, our customers were not affected by it. Just think about the amount of time that they save for not having the need to remediate and recover. So I think that more and more customers put their trust in JFrog security and the holistic solution. And coming up next, also governance, because you also need to trust those software packages that you mentioned before you ship them. Awesome. Thank you.

Your next question comes from the line of Brian Essex with J.P. Morgan. Your line is open. Please go ahead.

Hi, good afternoon, and thank you for taking the question. Congrats on the results. I just want to follow up to a previous question that was asked about the way that customers may or may not be managing their overages. It seems like demand is pretty healthy. Shalomi, are you privy to any conversations with your customers in terms of how they might be addressing more efficient spend just in general, but also on your platform? That's question number one. And then as we approach Swamp Up, just wanted to get a sense of, you know, sometimes it's difficult to time product releases with a specific conference. You know, sometimes companies release products when they're ready. You know, how should, like, what should we expect as we head into Swamp Up, being kind of relatively new to the story here?

Brian, thank you for bringing Swamp Up up. It's going to be an amazing conference. Again, we already see the speakers line up. I'll start with that. I'll just say that obviously the main thing that we see at Swamp Up is that the world's biggest organizations are using JFrog and willing to share their best practices. And this is gold for our users. Alongside that, there are the JFrog announcements, which are exciting, very much aligned with the future roadmap. But in the world of AI, if I will wait for SwampUp to release our product, our company will go backward and not leap forward. So expect a lot of excitement also on stage, but also with the customers' conversation. regarding the efficiency and management of budget. So what is it that we see, right? We see more token being spent that generate more software because AI agents are being fueled, and that generates probably a higher cloud consumption. But people are asking, will that be forever like that? The answer, for sure not, because now we are transforming forming from the CIO FOMO, which was around technology, to the CFO FOMO, which is around budget control. And then we have to ask ourselves, what is the desired outcome of software pipeline, of software supply chain, of the AI world? What's the desired outcome? The desired outcome is that you will have more software being shipped in a higher quality. What I just said equals binary and this is what jfrog is monetizing on we are monetizing on binary traffic so we think we suspect that cfos will be smart enough and disciplined enough not to block innovation inside jfrog we keep saying that you can also optimize how much you want to pay electricity you consume at home, but you still don't sit in the dark, right? You just train yourself to close the light before you leave. That's the difference between source code and binaries. The desired outcome of a better AI is better binaries in high quality and higher consumption. Got it. Thank you very much. I appreciate it.

Your next question comes from the line of Andrew Sherman with TD Cowan. Your line is open. please go ahead.

Oh, great. Thanks, guys. And congrats on another quarter of acceleration here. Ed, the billings at RPO were extremely strong. RPO added a record $84 million. Could you talk about the breadth and nature of the big deals that went into that? And was there any pull forward from the second half pipeline? And how is the second half pipeline looking? Thanks.

Yeah, thanks for recognizing that in the RPO. We're very proud of that. And that actually comes from the efforts that we have around our security products. That's a driving increase in our ASPs. And you saw that in the million-dollar number. 80% of the million-dollar customers that we added this quarter had security attached to it. And that, along with even the new customer lands, 40% of those customers with security are driving much of the RPO. In addition to that, customers that land with security typically take a multi-year agreement as well. So the construct is a larger ASP, longer in duration, and that results in strong RPO. And that's what we're seeing. We did not pull really anything in from Q3. We just had strong build in the quarter and execution from the team. And that pipeline that we built really came from the Shai Haloud events that started in Q3 of last year and it continued to build. The sales organization executed on those deals, and it's reflected in our results.

I might just add to it that I completely agree with everything that I said. Some of those macro impact of open source attack is out of our control. Obviously, these are awful things that are happening, but AI also make the hacker more sophisticated. And this software supply chain rapid attack might expedite some of the opportunities we have in the pipeline.

Great. Thanks, guys.

Your next question comes from the line of Jason Salino with KeyBank. Your line is open. Please go ahead.

Great. Thank you. Shlomi, I don't think you've had the opportunity to talk about your important fourth AI customer. It seems like the hybrid deployment model is unique and interesting. Maybe can you speak to why this is maybe different from your other, you know, AI native customer deals? And then secondly, you know, we on the street always want more, right?

I mean, is there a pipeline or opportunity to land other, you know, AI natives of this magnitude? yeah well uh listen we we are so honored and so excited to add more and more companies that are building the the world of ai and this one that we just uh won this quarter was a great win and it was sweeter also because of the fact that we displaced a competitor that couldn't scale the the amount of binaries and the traffic that needed to be uh supported was not something that they could do, this AI factory moved to JFrog, and they moved to JFrog in a very interesting way. They took JFrog platform as their mothership in the cloud and with some self-hosted artifactory servers in their data center, so they will not only have the power of the JFrog platform supported by our services, but also to have a super robust distribution mechanism from this mothership to all data centers, from artifactory to artifactory. You know, working with these companies that build with AI for AI makes JFrog better. We spoke earlier about the open AI incident. This is how JFrog become a thousand times stronger, because these guys, they take you to the limit with security scanning, with scaling, with different deployment environment, and we are very pleased and very honored to have them on board. Obviously, they look at JFrog as the planners of the pipelines, the providers of the infrastructure to the software supply chain.

Thank you.

Your next question comes from the line of Sanjeev Singh with Morgan Stanley. Your line is open. Please go ahead.

Yeah, thank you for taking the questions and congrats on an awesome quarter. It was great to see. Shalami, I think you and I discussed before in terms of some of the evolution that JFog is going through, you guys were one of the key destinations for container registries and those kind of traditional software artifacts. You mentioned the initiative to be the home of models and some of the newer kind of AI native artifacts. And just to get a sense of how that trend is evolving in terms of winning those newer artifacts.

Yes, Sanjit, great to see you on the call. You are right. And you watched the JFrog transformation before AI and hopefully with AI. So what do we see happening? We see that the world of software supply chain management move from just managing the pipelines to also secure them and govern them. And by being focused on the right asset, what we keep saying from the foundation days of the company, binary is the primary asset. By being focused on that, we are not only providing our customers with the smart storage that can scale in the cloud and on-prem, but we also provide them with a comprehensive security solution. And now, as you probably heard on the call, we just started to win our first DevGov Ops customers and deals. People understand, and we saw it again this week with OpenAI, models need to be governed. So to your point, Sanjit, we became the biggest container registry after Docker boom, and now we are becoming the biggest model registry for our customers, mainly because of their following. They can scale with us, they can trust, and they can secure, and they can govern. So if we will provide them this fundamental, and it's on us to prove, then I think that you will see the next leap in our evolution. What also matters is that the AI companies, the AI native and the AI labs companies, trust JFrog to do it for them. So we are learning a lot. We are improving a lot.

Just to stay humble, we are in the beginning of the journey. your final question comes from the line of kingsley crane with canaccord your line is open please go ahead great thanks for taking the question fit me in uh just to build off what we've been talking about earlier as novel agentic attacks they become more apparent how do you manage an intelligence that we don't necessarily fully understand that's not capable enough to be dangerous so you mentioned earlier that ai models should not be treated as free Is it consensus that customers are treating coding agents with that same scrutiny they do with third-party packages, or are they still adjusting to that? Thanks.

Well, Kingsley, I'll be honest with you. With over 6,000 customers, you see everything. You see customers that are just kind of being attracted by the technology and adopt AI with no responsibility. and you see customers that are terrified and take it one step at a time. I think that what is special for the experience that we get from companies like OpenAI and other leading AI labs is that they are showing us not only what needs to be done, but also what happens if you don't do it right. And if you don't do it right, models are becoming sophisticated and even smarter than our most senior developers and engineers. So I believe that the world will become mature and the regulation and guardrails and security around models will become a bit more powerful. What we are discussing with our customers is how important it is not only to host the models and to host the binaries and the outcomes for them, but also how we secure the entire pipeline. And there is no better school than this AI labs to learn it from them. It's a great privilege to have all of this as our customers.

This concludes the question and answer session. I will now turn the call back to Shlomi for closing remarks.

Thank you, everyone, for joining our call. Obviously, Q2 was yet an amazing quarter for us. We are focusing on delivering what we committed in 2026 and looking forward to seeing you at Swamp Up, where we will host an investor meetup to keep answering your question in full transparency and with the right excitement. And may the frog be with you.

This concludes today's call. Thank you for attending. You may now disconnect.