Cellebrite DI Ltd. ($CLBT)

Welcome, and thank you for standing by. I would like to inform all participants that this conference call as well as any Q&A may be recorded, where a company is presenting any recording may also be posted on their website. Views and opinions expressed by any external speakers on this call are those of the speakers and not of JPMorgan. Parts of this conference call may be reproduced in JPMorgan Research. Participants are prohibited from posting, sharing or distributing any part of this call or its content on social media platforms or any public forums without prior written consent from JPMorgan. If you have any objections, you may disconnect at this time. Members of JPMorgan Global Banking Department may be present on this call. Your host will now begin.

Great. Thank you very much. Good morning, everyone. My name is Brian Essex. I cover large cap, smid-cap software for JPMorgan. And we're excited to have the Cellebrite team with us for a tech talk. We have Chris Wade, the company's Chief Technology Officer; Shiv Ramji, the company's President of Product and Technology. Evyatar Ramot, Head of AI Innovation. And then over there on the slide, we have Andrew Kramer, VP of Investor Relations and Treasury, and we also have Dave Barter, the company's CFO. So thank you all from Cellebrite for joining us this morning. Before we kick it off, a couple of housekeeping items. I believe there is a box for Q&A. We'll save some time at the end if any questions pop up that are relevant. Please keep in mind, this is a tech talk, so technology-related questions would be appreciated. I would avoid sending me an IB chat, although you can try that and definitely don't send an e-mail because that's going to take too long. And then number two, housekeeping, I think Andrew has a disclaimer, he needs to read off and then we can get things kicked off.

Yes. Thank you very much, Brian, and thank you very much for hosting us today. I'd just like to remind everybody very briefly that today's discussion will contain forward-looking statements that may include, but are not limited to, the company's business operations, product road maps and financial performance. All forward-looking statements are subject to risks and uncertainties and other factors that could cause matters expressed or implied by those forward-looking statements not to occur. Actual forward-looking results could differ materially from historical results and/or from forecasts. Some of these forward-looking statements are discussed under the heading Risk Factors and elsewhere in the company's annual report on Form 20-F filed with the SEC on March 3, 2026. The company does not undertake to update any forward-looking statements to reflect future events or circumstances. And so with that said, Brian, I'll kick it back to you.

Great. Thanks, Andrew. And then I guess for those online, I think we have about 45 minutes, so we'll try and keep it to that to kind of like keep everyone mindful of the time.

Maybe to start with you, Shiv. Thank you for joining us. We certainly appreciate it. I'd love to get your thoughts on observations you have on the company's core competencies from a technology perspective. Certainly, before we go deep into AI and its impact on Cellebrite's business, I think it could be helpful to ground investors with regard to the way that Cellebrite has evolved from a leader for collecting a digital evidence from mobile devices into a broader platform. So we love your take on that as well as the core competencies of the company.

Absolutely. Good morning. Brian, thanks for hosting us. Good morning to everyone who's joined. We really appreciate it and looking forward to doing more of these with all of you. So look, I'm still relatively new to the company, but really, really excited about the amount of technology and products that we're building. So maybe from my perspective, when I still looking at this with a little bit of fresh eyes, historically, obviously, many investors have known Cellebrite as the leader in mobile device access and extraction. And that is a core competency. But when I look at all the assets we have in the company, where the company is today, I think the company today is much, much broader, broader. So like what I think differentiate Cellebrite is that we spent decades helping investigators collect, analyze and act on digital evidence across the entire investigation life cycle. And so the value is no longer just simply accessing a device. The value is really in helping customers turn volumes of digital data into evidence and actionable intelligence. And as many of you know, investigations increasingly are digital. Customers really need a platform that spans everything from the initial collection analysis, case management, collaboration. And now, which I'm really excited about, Evy will talk about this. But just this morning, we announced our Genesis product, which is generally available. So this is all about AI-powered insights. And so we'll talk a lot about this, but some of the early feedback we're getting from customers has been really, really amazing. And so I think that's where we see our biggest opportunity. We're looking to expand wallet share by solving more of the customer's workflow not just providing another point product. And so I think when I look at these decades of domain expertise, these very unique products around collection and then all the chaining you have to do of all of these this information and insights to produce evidence that can held to the scrutiny of the legal system. I think that's really, really unique. And the powering with AI, I think we will bring a ton of value to our customers. We're starting to see that with our launch this morning.

And as you've developed a platform, just kind of curious or maybe this is a broader question for Evyatar or Chris. How has your user profile changed? I mean was it initially with extraction, was that primarily people in the field that we're utilizing it and using the platform. And I'm just kind of curious, particularly within customers as you think about the seniority level or the sophistication of the users on the platform. How is the development of the platform change that user -- the user base?

So I can talk to that. So yes, historically, we've been selling to the digital forensic labs, which is usually this dedicated, highly sophisticated unit within a public safety organization, which is relatively small, often, you have a few experts, the larger the organization, the more people you have, obviously. The more we expand, the more we reach and talk to and interact with other personas within the investigations -- within the agency sorry. One thing it does is it opens a whole new [ time ] for us because there are a lot more investigators, analysts, prosecutors than there are digital forensic experts. So that's one thing. The other thing that it does is it takes us higher up the chain, the command chain because it gets more visibility. We touch more units. We touch more people and I will say the value we deliver is broader, right? Digital forensics in itself is extremely lucrative, and we often talk about how rich the evidence that we help collect is -- but as you broaden that and of course, bring AI to the table, the impact then just multiplies and therefore, we can really interact with much senior people across those organizations.

Got it. That's helpful. And how do you think about your customers' technology investments as well as their investigator protocols and legal frameworks that your solutions have to support?

I can start there and then please add Chris and. So again, what I was really impressed by like how we solve our customers' problems because our customers are operating some of the most complex technology and legal environments in the world. So they need solutions that are trusted, auditable and adaptable to varying legal frameworks and data residency requirements. So what I have at least seen we have a very, very deep our foundation is really in this deep technical expertise in accessing and processing digital evidence. And then over the last few years, we've invested so much in building the software sits on top of all this. So everything from the analytics that we produce, the workflows that we have. Evy mentioned, we're now touching other personas and we have this collaborative capabilities across them. And then we're bringing all of that with the announcement we made this morning with Genesis, where now you're using AI to kind of stitch all of this together and really, in some cases, taking where some of those who work on our case would take days and weeks to get to information. We're now seeing that -- helping them get to insight within hours. So all of these things around investments that we've made in on-prem technologies, which obviously everybody is familiar with. We've invested in cloud, and increasingly, we're making investments with AI so that we can really take these mountains the heaps and heaps of data large area of volumes and give our customers the insight that they really need to make some critical decisions and investigation. So stepping back, I think we'll continue to operate in these hybrid environments. And another big update, which happened just when I joined is we have also worked on FedRAMP High certification. So again, that opens up another customer segment for us. And so again, we're investing where kind of our customers need our technology to be.

Great. And then, Shiv, how do we think about operating in a hybrid environment? How do you maintain -- I guess, what considerations do you need to think about particularly in a way that it might differentiate Cellebrite to maintain the integrity of the profile, particularly around things like chain of custody. When you have an on-prem environment where it may be a little bit easier, how do you do that with the cloud? And are you capable of doing that across all the different products that you have in your environment?

Yes. We definitely have to -- we maintain all legal privacy and restriction of jurisdictions in all the geos that we operate in. And so the integrity of the evidence collection and the storage and access of that information, obviously, is it's highly sensitive and very critical. So all of our systems, whether you are on-prem or if you're using a cloud product and now increasingly as we use -- that is our differentiator, actually, which is to make sure that at any given point, our systems are adhering to the jurisdiction and the standards that we're operating in.

Got it. And then how much of the platform is hardware-related versus software? And how should we think about where the foundation lies especially as your solutions evolve from on-prem to cloud?

Increasingly, obviously, we've been investing in the software world quite a bit on cloud. And so we expect that to continue to be a big part of our business. But there are certain segments of our customers where that is not a viable solution. So we have products that our customers use in the field where a cloud solution won't work. So it really depends on the mix of the customers that we're serving. So I think we will always have some hardware capabilities for very unique scenarios like using your products out in the field where you may not have access to the cloud and also where maybe investigations are more time-sensitive where it's just not feasible to bring evidence back to a lab, and you have to do that out in the field. So -- but I actually see that as a really big advantage because we're kind of meeting where our customers and the users are in providing value so that they can get to the answers that they want.

Right? That makes sense. I know you guys have talked at all about moving the use case to the edge. So that's -- that makes sense. Maybe if we can start like moving into AI a little bit. I want to approach is for make two different angles. One of the questions that I've gotten from investors is just to kind of like get a grasp of the core extraction technology, particularly as it relates to mobile device OEMs having access to AI coding models? And how should we think about, one, I guess, maybe can you walk through how it currently works when you unlock a device to extract decrypt and decode data from that device? How much of it is hardware versus finding holes in the software to -- I don't know if exploit is the right word, but maybe take advantage of so that you can unlock a phone.

It depends on the situation in a specific device. It's a combination, obviously. The software utilizes various hardware components inside our value fits to export a device. And it depends on, like it's very vendor specific. But generally, there is a significant combination of the hardware, and we are dependent on the hardware in those situations. So it's not purely software. And some of the hardware is for protections of the software to ensure that our IP stays safe.

Got it. But I guess one question is, so when -- as OEMs have access to some advanced foundation models like a ethos to uncover where there might be vulnerabilities that could be exploited and potentially improve the quality of their software, so they don't have any as many vulnerable dependencies or holes in the code, how do you think about the way that the higher-quality software code that's embedded in these mobile devices could affect your ability to extract the data in the future.

Brian, I'm having a bit of a technical issue. I lost the first part of the question. Could you just repeat that please.

Sure. Yes. I guess -- so just maybe like Chris, to follow up on the question I asked before. So as mobile device OEMs have access to advanced foundation models, potentially have the ability to improve the quality of the code, so there aren't as many vulnerable dependencies or holes of software, basically improving the software quality of the device on the software quality of the software on the device. How might that affect your ability to I guess, unlock a phone in the future or unlock a mobile device in the future to kind of extract and collect the data.

Yes. I mean it definitely makes it harder but we're using AI on outside as well to find vulnerabilities. So...

I think we froze Chris.

I think we might have lost him. This connection doesn't seem...

I don't know if anyone else wants to pick that one up?

Asymmetric cycle that we've been...

Chris, if you could just -- we lost the first part of your answer. So if you can just sort of pick up at the outset of how you were framing the answer.

He froze again.

Okay? You can put on your CTO hat.

Yes, we can. Well, we'll get to Chris, but I can maybe share just.

I think we got -- I think Shiv, I think we have Chris back. I see...

Takeaway it away.

[indiscernible] under right now is not agreeing with me. Yes, I think we see it as kind of a faster treadmill. Like the half-life of vulnerabilities is definitely shorter with AI in the mix. But...

So Chris, maybe we should take you off of video just to try to conserve some bandwidth and get your audio answer on.

Yes. Sorry about this. Yes. So as the OEMs like Apple and Google patch vulnerabilities faster using AI, finding and patching -- of vulnerabilities that we've been in since Cellebrite began that shortens, but the attack surface really grows. They are adding new features with every new chip. They're constantly churning the code as they change and fix vulnerabilities. Essentially, they have to patch all of these vulnerabilities, and we only need one vulnerability. So it makes it a shorter half-life, but it puts us in the prime position to maintain that edge because we have this decade-plus long knowledge of these devices that we've been playing this cat and mouse game with Apple and Google. So I see this...

Shiv, do you want to pick up where Chris left off?

Yes. I think the TLR is -- so a couple of things on this model stuff. Just maybe a little bit of might take to from just still new to the company. In general, these powerful models are good for the entire ecosystem we live in. We obviously all want secure software. So I think the entire ecosystem is going to benefit. That's a good thing. As it relates to our business, I think what Chris was trying to highlight is, look, we've been at this for a very long time. So obviously, vulnerabilities, he was talking about the we expect the span of these vulnerabilities to be much, much lower than in the past. So will this make our job harder 100%, it is. but this is also our area of expertise. And the second thing, what I think I was trying to highlight was, look, anybody who's experienced using AI models you will notice as fast as it's generating code to help you build new features. It is just -- it is introducing new bugs and issues just as fast. So this is the nature of being in software, by the way, you're constantly patching things but as you're patching and you're building new features, you're always introducing new bugs or vulnerabilities that will need to be so -- that will need to be addressed later. So I think the -- what he was trying to highlight is it's will be harder for us, but also the surface area of the products are growing. I'm sure features and experiences will also grow -- and it's just the nature of being in software. You're always going to have bugs and so -- but we still believe we have the expertise to continue to keep up and serve our customers so that we can help them do good in the world.

Got it. Super helpful. This might be -- I don't know if Chris is back on, but this might be more of a Chris question than I had next. That's basically what does Corellium fit into vulnerability research. And if we think about Corellium Falcon, one thing kind of back to that point of vulnerabilities foundation models are very good at finding vulnerabilities in code as well as Logic. So how do we think about Corellium's future functionality and how that matches up against those are the foundation models?

Well, if you think about it, it's -- they're complements because...

Brian, maybe we can switch to a different question. We'll come back to Chris as his Wi-Fi...

Okay. Yes. We'll flip to a different question. I guess maybe for Shiv or Evyatar. Cellebrite had AI capabilities within its platform for a decade or so. A lot of that was machine learning, proprietary machine learning focus. But now we're seeing something that's very different with the emergence of the foundation model. I guess how do you assess what the models do well? And how durable is your core platform in the face of these capabilities. And I think you -- obviously, you mentioned you use AI as well. So maybe you can talk about some puts and takes there.

Well, maybe Evy, why don't you start talking about how we're using AI specifically solve the customer challenges, especially the stuff you have built. And then I'll talk a little bit about our history and kind of where I see continue to fall.

Yes, Shiv. Sure. So we believe -- and I would say we are more than convinced that AI has a significant potential in our space. It is almost like the perfect match for us. When we look at the customer problems, which is having so much data from so many different sources, lack of resources, this is like the perfect storm for AI to really be leveraged in a significant way. And so when we started experimenting with Gen AI, the frontier models and how they can help, of course, we were blown away from I'll say, the opportunity, but then actually seeing it work in real life and customer environment has really convinced us that this is something we need to go all in with, to be honest, something that we are investing a lot and in -- sorry, and Shiv just mentioned, we announced the general availability of Genesis today, which is catered towards that use case of how can we help investigators be not just more efficient, efficiency is probably the biggest thing, right, doing in minutes what would have taken maybe weeks, but also just bring justice with -- now I can find things that I had no idea whether AI was not able to do it because human brain is limited, right? When it comes to that scale. And so from the perspective of solving the customer problems, the potential is huge, but it does require a lot of investment in -- especially around the accuracy and catering it to that environment, which you've talked about before, right? Ethics, trust, compliance, all those things require a lot of action and investment which is what we are constantly doing and continuously improving what our products can do with those models. But we are already seeing a significant impact on those customers who are willing to adopt those solutions in a very meaningful way. And you're right, we had AI for probably a decade, if not more. This is something different. It requires a different approach, different talent, different structure, different way of doing things, which is exactly what we're doing with Genesis today.

And I guess from your -- from a user perspective, how open are your users to or your customers to utilizing the AI that's embedded in your platform? And I just think about -- is it one old tariffs that don't have a lot of technical sophistication and getting them to use technology to begin with is really challenging or is there maybe a younger or more enthusiastic profile of users that's accelerating your TAM because they're very open to using technology in what AI has to offer?

Yes. So all of the above and probably a few more examples of people who are actually looking at this technology and immediately realized how significant it can be in terms of the impact on the line. The one thing I will start with is, I think, given us were surprised at the pace of adoption -- we've been investing, let's say, in cloud technology in this somewhat traditional space for a while, and we know change requires time. I think here we're seeing something different in terms of the impact in the value is just so obvious that people are much more open to it than we have expected, which is really encouraging. In terms of the profile, so it is actually quite wide in terms of -- yes, you're right, you have the older, less technical detectives who see this as a really easy way to leverage technology because everything they had until today was complex. It required training. It required a lot of change management. And here you have something that as long as you can type and ask questions, you can basically use it. There's the younger generation, you're right, who are expecting to have that kind of experience because this is what they have in every other avenue of life, if you like. But what I'm most surprised about is that even the technical people, people who we have been engaged with for a long time, we know their technical, we know they're very forensic and they're very traditional in how they run their operations are looking at this and realize this is something that is going to be meaningful either to them or to their end customers who are the investigators, so right, and sometimes both. And so I'm really encouraged. I've been here, by the way, 6 years. I'm really encouraged by what I'm seeing from the forensic community. It comes with a very high, I'll say, standard or expectation in terms of what we can do. But given that's natural in our space. But the majority of the people today are in a position of, okay, I'm willing to adopt it as long as you do it right, which is really where we want them to be.

Yes. I'll just say, I think Evy is being modest here. Look, we've taken dads of experience in this world and brought it to life in this product. I mean when you get the demo of this product, you can just tell how incredibly powerful it is, meaning it to insights within minutes, what would typically historically take hours days, maybe even weeks. And so it's incredibly powerful. And whether you're a technical matter, the reason why this is powerful and we're able to get to insight is that the team has built this product with that idea of like you always have to earn trust. You have to make sure you have accuracy in their product. They also have an investigative lens to the experience. And so it's not just like the easy conversational experience which you would have, it's all the underlying work that the team has done to make sure that when we produce the answers, and the insights or even the visuals or materials that can aid in an investigation is really purpose-built. And look, you don't get this stuff from just any regular language model. There is a lot of work that the team has done. And I think that's really important to highlight because this notion of like you can use any large language model and just do this yourself. I don't think that's true because I think we have invested so much knowledge and proprietary ways of getting at these accurate answers. And so I think that is kind of invisible to the user, but honestly, I think that's powering kind of the time to value that we're seeing for our users.

And how do we think about your -- I mean, I don't know if maybe you have a good example to share of your typical user that's using your platform to extract data, maybe they're using some other products to store the data or collaborate on the data, how do they get introduced to the different -- I mean even before Guardian went [ GA ], how do they get introduced to the different AI-supported products on your platform and realize the value that you have?

Yes. So with Genesis, what we did, and again, I said it requires a different approach is, one, we interacted with customers, early design partners from the very beginning to help us design this in the proper way that will actually be usable to them. But then what we also did, I think it was exactly -- almost exactly 3 months ago is we announced early access for Genesis. So we wanted to flood the market with this. We understand that in order for us -- the most important thing for us with this is to get the adoption in the market. And so we've announced early access for Genesis, and we had hundreds of requests hundreds of people coming to us and saying, raising their hands saying, I want to use this, and we've actually had very, very large number around 800 today of people using it in the Early Access phase. Now that we are transitioning to GA, the first priority is to convert those people to be paying customers. But at the same time, we're maintaining a free tier or free trial so that we can really be everywhere with this product. We have our sales team all behind this pushing this and introducing this to customers. So we're doing this going in a parallel path where we have sales led. We have almost like a PLG movement going on already today. And so we're expecting this to be exposed to a very large proportion of our customer base.

Got it. Got it. Super helpful. .

Brian, I think you might have Chris in your waiting room. He's pending to be readmitted. If you could...

Yes.

So if the OpenExchange team can let him in, that would be great.

There curious. Great.

Thank you very much.

Chris, should we try your broadband again?

Yes. Let's do that. Do you want to swing back to the Corellium question? Do you want to ask that one again?

Well, actually, maybe just given all the audio distortion, Chris, to meet those questions, maybe we could just take that from the top.

Sure for which question you...

To Mythos question.

Yes. Okay Yes, let's hit that one. Yes. So back to the Mythos question, sorry. I honestly think Shiv did a great job. I heard Shiv kind of answer on this. I actually think you did a good job on this. But -- and this kind of ties into Corellium, I want to get into -- it kind of covers a little bit of what you asked about Falcon as well. One of our biggest advantages is that we have the ability to test the different vulnerabilities that AI finds on our side in Corellium. So if you think about vulnerability detection like AI is finding vulnerabilities in source code. It's all static. We're able to take that and then validate that in Corellium as like a dynamic layer. We're also able to do the same at scale with Corellium in terms of looking for vulnerabilities in running code, which is a completely different set of requirements to just analyzing source code. And given that we're one of the only companies would be outside of Apple and Google, who can do this at large scale because we can -- we're not using farms of devices, we're virtualizing these devices and spinning them up on servers it does allow us to iterate on vulnerabilities and then the code required -- the exploit code required for our product to unlock these devices. So that's a very distinct advantage to Cellebrite. Our competitors don't have anything like that, like Corellium, like Falcon. But moving back to the Mythos thing question surrounding vendors patching vulnerabilities at a higher rate, essentially having the life of vulnerabilities. And I think Shiv cover this a little bit, but it does create a lot of churn in the code. They're patching vulnerabilities much faster which means they're modifying a lot more code across a large area inside product, which generally introduces other bugs, other problems. And we only need one vulnerability to get in, right? They have to patch all vulnerabilities. So they're patching hundreds of vulnerabilities. And most of these vulnerabilities you see coming out of ethos are kind of like the low-hanging fruit, the much more complex vulnerabilities that require like logic conditions, race conditions, like I mentioned before, where you need dynamic analysis to find these vulnerabilities, which Mythos is not capable of. Those are the vulnerabilities we rely on, and we've yet to see methods come close to finding these kind of vulnerabilities.

Got it. And then on the static side, are you able to leverage those models as well, models like Mythos or even maybe ones that are a little bit less sophisticated and less exclusive for your own vulnerability research.

Definitely. Not Mythos. We would love to, if Anthropic is listening and they want to give us access. We're happy to take a look at -- it is. It is. But yes, so we use internally. We have our own AI that we use for vulnerability analysis. And we've been quite successful with that. We've seen this kind of [indiscernible]. There's a couple of terms out there for this [indiscernible] all these vulnerabilities that found before. We saw this like maybe 5 or 10 years ago with the fuzes from Google and Apple, they were fuzzing heavily, and there was tons of bugs being found. And it didn't really change anything for us. We've been playing this game with Apple and Google for a very long time. So we're used to their patch cycles, and we know what to look for. to find vulnerabilities in their patches. So yes, their patches for vulnerabilities contain vulnerability sometimes. So we haven't seen any kind of slowdown in the number of vulnerabilities in their products.

That's helpful context. And then I want to -- I think we've only got a few minutes left, so I want to make sure we get into a little bit more product focused. But obviously, you've got a new AI product -- AI-powered products like Guardian Investigate and Genesis. Could you just maybe frame out for us, obviously, Genesis going GA today, but how does Guardian Investigate differ from Genesis, maybe just kind of like level set for those that maybe aren't too familiar with the platform and then we can go into a little bit more detail.

Yes. So I'll take that. Yes, so with Guardian Investigate, so Guardian is our SaaS platform. We launched it few years ago, Guardian Investigate essentially an extension of that to cater for the broader investigative use case that we discussed before. Now Guardian Investigate is a [indiscernible] product is part of our platform. This is the system of record, if you like, for investigations, where you manage your evidence, you manage your cases, your tasks. You also run analytics on top of it and then you're getting a lot of the AI capabilities that are similar to what you have in Genesis within the context of that platform where agencies actually go and transform the way they run their entire investigative life cycle. With Genesis, what you have is a product that is not still on the platform, but actually something you can adopt really quickly, and that was intentional from us where we're saying our customers, we know on this maturity curve when it comes to transforming their entire operations. But they all have that same pain, same problems that we are analyzing the evidence. And so we are providing them with a quick entry to that new age of investigations as we define it, that will then help them maturing and evolve, and we believe that at some point, yes, we'll adopt the entire platform solution. But if they're not there yet for whatever reason, then they have a quick entry point to something that is really powerful. But in terms of the underlying capabilities and I'll say the use case, they're pretty similar in that respect.

Any good examples of how much more effective one of the other platform is versus the way that investigators may previously approach their workflow?

So just to make sure I understand the question. So compared to what they're doing today is what you're asking?

Yes. So if you have investigator Genesis, a customer that adopt one of those platforms. Is it for everything they do and then how much more efficient can they be on that platform versus maybe what they've done before? .

Yes. So one thing to highlight about both actually is that this goes beyond mobile forensics, right? So with Genesis and Guardian Investigate, you're actually able to create cases that include more than our mobile extractions, but mobile computer, CDRs and many other file types, including media all types. To create one case that includes all those data sources in one place. And so the ability to do that is really where you're seeing the multiples come into play, right? Because you're not only saving the time of analyzing a mobile device, which can take a long time. But now you're looking at multiple mobile devices, but also corroborating that with police report, with [indiscernible] records, with body-worn cameras, with additional data sources that are part of an investigation. And so what we're seeing is we have some really extreme examples of people telling us. I've been investigating this case. I had 3 investigators doing manual work for 2 or 3 months. And within 5 minutes, I was able to find what I needed to actually go and prosecute that case, which is incredible. A couple of other examples we had one case that involved sexual abuse of miners, where we had 3 devices now to put that into context. Analyzing 1 device thoroughly can take days. Now having 3 of them as part of the case, the investigators came in, the new of 1 victim in that case. And within 15 minutes, they identified 15 more victims they had no idea about. Now they said transparently, manually reviewing those devices would probably result in a similar result in terms of identifying more victims, maybe not all 15 additional ones, but most of them. But the time there was a really big factor because within 15 minutes, they were able to do something that they estimate would have taken 2 weeks. And we're talking about 15 more victims here that would have been suffering during that time. And so that got escalated to a federal case and prosecuted. So these are the type of impact. So we're looking at weeks 2 minutes.

Got it. Andrew, I think we're over time. So I want to be respectful of everyone's time. So I don't know if you want to end it there or if you want to keep going or whatever.

I think, Brian, we want to thank you for hosting this today. We hope this was helpful for everyone, and we'll look forward to just continuing to stay in touch because I think we're very excited about the products and the technology we're building, and we'll look forward to just [indiscernible] other windows to share more about the amount of innovation that we're bringing to market.

All right. Sounds good. So Chris, Shiv, Evyatar, you too, Dave, and Andrew, thank you so much for joining us, and thank everyone else on the line as well.

Thank you, guys.

Thanks a lot.

Thanks, take care, everyone. Take care. Bye.