Okay. Good morning, everybody. Thanks for joining us at our fireside chat session now with Nexxen. I am Barton Crockett, Internet Media Analyst with Rosenblatt. I am joined by Sagi Niri, Chief Financial Officer of Nexxen, and Karim Rayes, who is the Chief Product Officer. Nexxen is an over $600 million market cap company in the ad tech space that we cover with a buy rating, arguing that the proxy for revenue growth, which is contribution ex-TAC, is growing in the low double-digit range, with margin expansion driving adjusted EBITDA growth faster than that, and that that is not fairly reflected in an EV to adjusted EBITDA multiple near 5x on a US GAAP adjusted basis that we look at. We think it should be re-rated higher. Our $16 price target is up over 50%, assumes kind of a 9x multiple. But the ad tech space is a complicated space with a lot of different kind of puts and takes, so we are really happy to have Sagi and Karim here to kind of talk us through the space and talk us through Nexxen's place in it. So, maybe we will start with that. Sagi, Karim, one of you or both of you, give me a sense of what differentiates Nexxen. Why do clients use it? What is your place in the market? Sure. I can take that. First and foremost, we are a fully integrated platform, a unified end-to-end programmatic platform. So, we operate a demand-side platform, a supply-side platform, a data platform, all powered by AI. So, these platforms sit on differentiated proprietary data, exclusive media assets and technologies, particularly around CTV that we have built out. So, we have full funnel capabilities spanning from audience discovery to planning, to activation, optimization, monetization. So, we are really a one-stop shop for customers. So, this ability to do all of these things and integrate these solutions together help us drive value for customers, minimize their fees and cost, and ultimately, drive return on investment. Increasingly, over the last couple of years, we have invested heavily in AI. So, we have streamlined these workflows and connected these platforms leveraging our AI solutions called nexAI as well that we believe will be a strong growth driver for us in the future. Okay. To drill into that a little bit, as I understand, Nexxen combined the functionality of a DSP with an SSP, which is relatively unique, I think, in this space. Most people will call themselves one or the other, but not both. But there is this kind of crossover happening, which is the DSPs are kind of moving deeper into the SSP space saying, "Hey, you don't need to pay an SSP. You can work directly through us." And the SSPs are doing kind of a similar thing from their side of the universe. You guys offer both. But unlike the other kind of single sources, I think you'll charge a take rate for both sides, versus the other suggesting that one of their efficiencies can be to eliminate the take rates of the other side of it. How do you stay competitive? Do take rates really matter, or is it ultimately the spend driven by return on ad spend? Ultimately, it's really about return on ad spend and the outcomes you're driving for the advertisers, but take rates are part of that calculation as well, right? It comes down to the optimization you'll do, the performance you're driving and, ultimately, what is the cost of media. But when you look at the capabilities as a whole, it's more than just supply and demand, you know, br inging in planning, measurement, data activation, data insights, et cetera, into one platform. We're consolidating many fees that are happening across the ad tech ecosystem into one solution. On take rates in general as well, we have fees on both sides, but it really depends on how customers are transacting. The reason we do that is because we also trade with third-party DSPs, and our DSP also trades with third-party SSPs, so we need to be able to charge fees when it comes to that. When customers are buying end-to-end on our platform, we are unifying those fees and bringing in savings to those customers, so we're highly competitive there when it comes to that as well. Really, it's a combination of the tools and technology we've built driving strong return on investment and very efficient fees that make us extremely competitive in market, and the results speak for themselves. We just, in our last earning, talked about Toyota that saw a 62% reduction in cost per vehicle sold through our advertising solutions, and we have many of these examples. Ultimately, that's why customers are joining our platform. Okay. Can you drill down a little bit deeper and explain why is it that having both a DSP and an SSP in one entity would make a difference? How is it that you're able to deliver some of these better returns like you cited with Toyota? It starts with being able to manage the entire transactions. We have very strong signals across the transaction. We understand the media that's being bought. We have the data around that. We understand the outcomes the buyers are looking to drive, and we're able to unify all of that data into a single solution to drive those outcomes. So, strength of signal, strength of efficiency as well. When you're combining everything in one technology, we're able to combine our OpEx against that as well and bring those savings back to both publishers and advertisers. Ultimately, publishers are seeing unique demand from us, and lower fees when that demand's coming from our platform. And same for the advertisers on the other side, seeing direct access to publishers at a lower cost. Again, all that tied to the optimization we do on top. What we're seeing in the market today, let's say for us, we see it as a validation of our strategy that we've put together over five years ago now, in terms of bringing these technologies together. I think the market is trying to catch up to that and build the technologies on the other side today. We already have it, so that gives us a leg up in the market and the ability to go quickly. Again, we're seeing this in our results. Advertisers want access to unique data. They want premium direct media. We're able to offer that to them today with a full suite of technology that an SSP offers, whereas DSPs are building that technology today, they're playing catch up there. You could say the same on the supply side as supply partners, supply platforms are looking to build buy-side technology. I would say that technology today is rudimentary compared to a full stack DSP and what that's able to do. So, really having the strength, technology on both sides is a huge differentiator for us. Okay. We are certainly dealing with a market that is digesting what it means to live in an age of artificial intelligence, right, where t here's so much investment, so much change. I think ad tech has been, you know, the sector generally has been subject to a lot of questions around this. One of the questions is: What is really, at the end of the day, kind of drive durable value for ad tech entities? I think one of the theses out there is value will accrue to those with exclusive, meaningful data. I wanted you to talk a little bit about what do you think about that kind of thesis, and what does Nexxen do in that regard? Before you go into the answer, I did want to also say one thing to the audience, which is, if people want to put some questions in, there is a Q&A kind of function in your webcast kind of interface. If you put something in, I'll see it pop up on my screen, and I can try and work it into the conversation. Yeah, so, what do you think about the idea of exclusive data, and that being important in the age of AI and Nexxen's positioning in that regard? Sure. I mean, AI is transformational for our industry, will drastically change how people are buying and selling media, but AI alone won't be the differentiator. Value really comes from the data you're training the AI on, as well as the tools that are underlying that AI. You could look at it as AI drives a lot of the orchestration, so it helps us surface helpful information, helps us drive some automation around services we've built and all that, b ut the technology that sits under that is extremely important. This is a combination, again, of unique data assets. We have a platform called Nexxen Discovery, that gives us great audience insight and activation tools. We sit on proprietary TV data as well. Again, because we see the entire transaction, we have all of the data across the flow of that transaction feeding into our system as well, and AI is leveraging all of that to drive this automation and these outcomes. The true differentiators, if you look years out, really, it's not going to be so much the AI tools sitting on top. I think that will become a commodity over time. It's going to be on what technology is this AI working, and what is the core technology behind it. Again, f or us, having that strong foundation, a platform that optimizes to outcome extremely well, that's driving results, direct access to that premium media, direct access to differentiated data, all powering the AI, is really how we see ourselves differentiating. Okay. I think one of the other questions in this sector is around mix. I think on the positive side of mix, there's some meaningful enthusiasm, I think growth embedded in the CTV sector, connected television, which I think industry-wide is easily a teens grower. I think for some players, including you guys of late, even faster, your CTV line in the second quarter was up 33% year-over-year, I think to 40% of your total programmatic, which is the bulk of your revenues. I guess two questions embedded in this. One, what's driving this outsized growth in CTV for Nexxen? Do you see a future where Nexxen is majority CTV? Karim, you want me to take it? Please, yeah. Yeah. Okay. I think that we saw the growth in CTV in Q2 all around. It was driven by increased spend across our CTV technology as Karim mentioned, our proprietary TV data and our premium data. We are like closing small gaps that we still have with missing premium media, like NBCUniversal and Disney, which like almost are the final missing parts within our total aggregate inventory that we are catering to different partners, of course. Enterprise customers scaled their spend, and of course, as I mentioned, new publishers onboarded, of course, contributed to the increase in CTV in Q2. I think that we have long-term opportunities around on-screen, native ecosystem, and performance-based CTV, which we are already existing, like the capabilities and the abilities and the technology, but we are still commercially, we are still at early days, but we will see it scaling up, in Q3, Q4, and mainly, in 2027 and inwards. I think that Home Screen, as I said, wasn't a big contributor in our scaling up in Q2, so we still have some more room to grow. I think that we said in our Analyst Day like two months ago that we are aiming on the long term to get to a 40% TV spend generation out of our total revenue generation. I think that we are already there. I think when you are asking if it can be our major channel, I think it is our major channel. We are doing on mobile, you know, the same, around 40% as well. So, I think on the future, probably CTV will become the main distribution channel, not only for Nexxen, maybe we will be some of the leaders, but it will become the major channel of distribution for most of the ad tech players because this is where the trends are and this is where the end users and the audience are consuming their content. Okay. Yep. All right. Now, it's great we have a couple of questions in from those watching. Let me see. I'll try and work them in here a little bit. I think one of the questions, it's kind of a long question, but I'll try and synthesize it here. I think that there is a question about how do you operate this integrated end-to-end stack, both the DSP and the SSP, and prove neutrality to clients who are worried about the fox guarding the henhouse? Also, how do you structure the data flows and auction mechanics to ensure that the auction's running properly, that everyone can be satisfied that there's, given that you're on both sides of it, that there's not anything happening that's untoward in the auction? Sure, I can take that. If you think about it, as an SSP, ultimately your customer is the publisher, and they're looking for maximum outcome there. So, our SSP teams are working on publisher growth and revenue tied to those publishers. Then, we have separate DSP teams that are working on behalf of the advertisers. Only a few of us sit on top of all of that and integrate these things together. So, we have targets split between supply and demand side when it comes to revenue, EBITDA, et cetera. In terms of auction, we run a neutral auction regardless, and our advertisers can audit that very easily in our platform. They can see what they're paying across the SSPs for the same media, et cetera. Ultimately, if it's cheaper to buy from elsewhere, that's where the money's going to go. It's all automated algorithms that will drive that. There's not a differentiation in terms of weights or priority or anything like that when it comes to the media. It's all competing equally in that auction. If our system thinks it's going to drive a better return on investment, it'll spend there. What you see organically happen, though, is as you're consolidating those fees and reducing that overall take rate, the algorithm will optimize to that and drive because it is seeing a better outcome there. So, ultimately, that's what we focus on, what will drive the best outcomes for our clients, whether those are publishers and advertisers. If we weren't doing that, we wouldn't be successful with them, we wouldn't be growing there. So, that's really how we manage the business. Okay. All right. So, I mean, the business is grounded in trust, and proof's in the pudding that you're growing and clients trust you, would be how you position it. Okay. Now, Sagi, you were talking a little bit about this, and I want to drill into this a little bit more, which is the Nexxen TV Home Screen. Karim or Sagi, either of you can address this, but just describe what you are doing there, what the opportunity is, and when you think we will begin to see it become impactful. I can take the first part. Yeah. I will let Karim take it since it is more product-wise these days and less commercial. Yeah. So, what we have built essentially is an industry-first programmatic solution for native home screen TVs. What does that mean? Essentially, we are connecting with TV OEMs and bringing their media into the programmatic ecosystem. This media goes into our own DSPs for activation there, but we also have the ability to bring that media to other DSPs as well. We have announced a deal with The Trade Desk, as an example, a few months ago around this. If you look at the native home screen, the challenge there is like these are bespoke units. Each TV and each OEM operates slightly different specs. In programmatic in general, we are talking about standardized specs for video, for display, et cetera. So, the challenge was around there, is how do we standardize those formats, so we are able to run one-to-many campaigns, meaning being able to run a campaign across multiple OEM? That is what we have built. We have built a combination of those integration with multiple OEMs. We have built standardization around streamlining the workflow to activate those campaigns against all of those OEMs. The benefit this brings is twofold. One, advertisers are able to run these campaigns where they currently operate today, whether that is our platform or another platform, another programmatic platform. They do not have to go and essentially work directly with the OEM and their ad server to launch those campaigns. That enables them to unify their strategy, but also to layer in the optimization that they are looking for and the outcome logic that they are looking for on the buy side. When we are talking about performance CTV, this is actually one of the primary units we expect to be able to drive that performance. It is a very high-attention unit. You can think about it as a viewer is leaned in when they are seeing that unit. It is when they open their TV and when they are making a decision on what to watch. So, they are highly engaged versus a commercial break. When we are looking to drive outcome, whether that is recommending a different show or promoting a product, et cetera, we see this as a very strong opportunity. We have launched this in market now a little over a quarter ago. We are seeing activation on our platforms. Buy-side platform, essentially, need to integrate and finish their scale-up of the solution, and there are many actively integrating with us today. When we are saying we expect to see revenue in Q4 and then more so in 2027, it is really about that ramp-up and getting these integration done, et cetera. That is what we are working through now. We see this as a strong driver for us because it ties in really well with performance moving to CTV and ultimately, advertisers looking for new ad impact inventory, in an age where AI took away a lot of that web inventory that was traditionally available on browsers. We see this as a strong replacement for those budgets. Okay. I want to drill into this a little bit more and also get to a question that we have around this. First, kind of on a related, just to level set. You have a minority investment in an entity now known as V, formerly a thing called VIDAA, that is essentially a TV interface, originally kind of created within, or kind of backed by Hisense. Could you talk about why you have this investment? Is it financial? Is it strategic? Is it both? How is that playing into your kind of push into the Nexxen TV Home Screen? Yeah, sure. I think, to your question, probably it is both. It is an investment, and it is financial investment, and it is a strategic investment. I think that we made this investment in order to have some commercial relationship, and as Karim mentioned, some of our differentiated, and as you asked, to have some kind of exclusive data and exclusive inventory is something that making us unique and making people coming into our ecosystem and utilizing our data and inventory, or one or both. We have exclusivity with VIDAA on ACR globally, and we have exclusivity with VIDAA on their North American inventory, so it is very meaningful, and it is giving us grasp with things that others can't facilitate. I think by the end of Q3, we are going to get to an invest of aggregation of $60 million, which represent around 6% of V equity. Of course, it deepens the relationship we have with V, and of course, strengthens the access to our differentiated CTV data and media. It's very unique in our industry. I don't think that anyone else like this kind of exclusive commercial and strategic partnership with any OEM. All the other OEM are considered to be like closed garden, as Samsung and LG, they are facilitating their own ecosystem. They are not allowing anyone to enjoy their ACR data. So, if you want to utilize their inventory, you are working through them, but you are not enjoying their ACR capabilities. Of course, it supports our broader CTV strategy and our anticipation that CTV will become the main channel for distribution in ad tech. Of course, it's giving us, as Karim mentioned, the first foray into this home screen and TV monetization opportunities because it's not programmatic these days, and we are the only ones that offer this kind of activation. Of course, the unique data and media gets us into the room with prospects and larger customer we wouldn't have otherwise. I think that all of that, of course, it's making the investment with V financial. I think that over time, they have big plans for themself, and probably they go to an IPO sometime in the future, and we will enjoy our equity stake in them. On the other side, of course, we are enjoying the special partnership and the commercial agreement we have with them in order to scale in our own business. Okay. Just to add to that, from a product standpoint, that data is extremely important as part of many of our tools. So, we leverage that TV data in our TV and media planning. We leverage it for audience planning and audience activation. We do TV measurements, so that's TV tune-ins, et cetera, measurement and outcomes tied to our TV campaigns we sell. So, it's an integral part of our portfolio and a strong revenue driver for us today already. Okay. I think one of the questions here is to try and understand the footprint a little bit, in a little bit more detail around the Nexxen TV Home Screen and to the extent that V feeds into that. Give us a sense of how many screens are you serving with the Nexxen TV Home Screen, and also, how large is the footprint that you're accessing data from and able to use the data for via your V relationship? You want to take that, Karim? The footprint you might need to take. Yeah. You have it? Yep. I think that VIDAA is the operating system of Hisense, and as that, we are enjoying to install their operating system into some of Hisense TV screens. Hisense is using most of the players out there, their own VIDAA operating system, but they can install Amazon Fire or Roku or whatever they want. Of course, they have the best interest in order to grow their own operating system to do that. They have around 50 million screens worldwide. They are very, very strong outside of North America. They are very strong in Japan because they hold Toshiba as well. They are strong, of course, in APAC, and they are very strong in Europe, in the U.K., in Germany, in Spain. In North America, we have couple of millions of screen. But as Karim mentioned, in our Home Screen ecosystem, we are onboarding more and more OEMs, which are looking for a platform that can monetize their unmonetized home screen inventory. If I am taking like the all announced one and the one we didn't announce, we'll probably get into the home screen ecosystem to somewhere around 10 million- 20 million screens that we can activate. That's. On ACR, Karim, you want to take the ACR part of the data that we are getting from the screens? Right. That's an integral part of this. We collect the data specifically on VIDAA TV, but also on Hisense TVs beyond VIDAA as well, where the media comes from VIDAA TV when it comes to exclusive media. But as Sagi mentioned, we're also integrating non-exclusively with other OEMs as well. But given we're the first to market solution here, it's de facto kind of exclusive right now. Ultimately, we expect that mix to continue to grow on the Home Screen, and that's the beauty of programmatic. You're able to integrate multiple supply sources together against a single campaign activation. That's the push here. The data is tightly tied to these campaigns. I'll give you an example. A lot of what you'll see on Home Screen is different content recommendation, different show recommendation, et cetera. If you think about a campaign against that, we use the TV data to then measure how many people actually tuned into that show from the advertising. As we're selling that media, we're leveraging the data in combination to drive those campaigns and, one, optimize to the outcome, but then measure the outcome the advertiser's looking for. That enables us to grow and get more budgets around this. So, data plus media is really tightly connected together around these activations, and anything around performance data is very important, so I'd say ACR as a starting point, but also, our other data sets we have within Nexxen now are part of those activations. Okay. When you say ACR, automated content recognition, just describe, you know, so we understand, what are you recognizing? I know there's some other solutions out there, I mean, Viant has something, IRIS.TV. I think Nielsen has a solution. To what degree are they similar kind of functions or different? Right. We do that as well. The way I could describe it is automatic content recognition essentially detecting what the customer is watching on TV. We're building sessions around that. So, think of the shows that are being watched, the commercials people have been exposed to. All that data gets collected and then leveraged again for planning, activation, insights, measurement, et cetera. When we're talking about solutions you've listed here, and we have the same, that's content targeting. We're getting that from the bidstream from the publisher. This is more in real time, "Hey, I want to target that show." You're going to get that signal from the publisher to say, "Well, this is the show. You should buy this inventory." Content recognition is across anything watched, regardless if a campaign is running or not. So, we have the full view of what's playing on the glass. That is technically what is used for data collection versus straight show activation. The first one will be tied to content targeting. ACR can be used for content retargeting, meaning, someone that watched the Super Bowl, you can target on a different show using ACR, where on content targeting, while the Super Bowl's playing, you could be targeting them there. You can think about a lot of strategies tied to that as well in terms of being able to reach an audience in a less premium environment where the media is more affordable. But you now, but have a message, let's say, you were an advertiser on the Super Bowl, and you want to retarget those customers and continue pushing your message, you're able to leverage ACR to do that. Okay. Now, I want to move ahead a little bit to mobile, which in your second quarter report, was also a strong grower, up 23% year-over-year. Excuse me. I was wondering if you could talk about what you see as the opportunity there, how big mobile, I think you've already talked about this a little bit but make sure we understand how big that is in your mix, what you think of the opportunity. I think you've also spoken about some M&A acquisitions around this, including potentially buying your own SDK. Talk about why you would want to do that as well. Okay. I think revenue, as you said, grew like 23% year-over-year in Q2. It's driven mainly by our execution against our mobile in-app strategy. I think that when AI came into the market, which seems like a long time ago, but it wasn't a long time ago, it, of course, hurt some media channels, as, you know, web domains and desktop traffic. I think that we realized at that point of time that other than CTV, which we're very focused on and have a lot of capabilities and assets around that, in-app mobile probably is more resilient to AI because people are still using their mobile in order to consume, a nd mainly, we are doing it through different apps, which we are not going into the LLMs in order to ask them questions. I think we understand that, we realized that we were under-indexed to mobile in-app advertising, and we invested around it through major partnership with SDK network. We don't have our own SDK, and we'll talk shortly about M&A. We did a couple of big partnership with SDK networks, where we are bringing our own demand into their inventory. They are not exposed to this demand, so for them, it's incremental. We are doing an amazing job on that, and we are scaling up our revenue and their revenue, and everyone is happy. I think that the durability and resilience to AI-driven distribution in mobile open us to that front. Again, we understood that mobile is probably going to stay for a long period of time. I think that direct SDK integration is expanding across the industry. You're seeing AppLovin and Unity and all of these guys and what we are doing. I think that this kind of advertising is scaling up performance advertising as well because within the in-app industry, a lot of performance advertising is happening, and we want to be part of that. As Karim mentioned, we want to take this performance advertising from mobile into CTV, and probably, the best place to do that is within our Home Screen ecosystem. I think that today, mobile is around our CTV footprint, around the 40-ish%. Regarding M&A opportunities, we said to the market long time ago, and we are keeping saying that we are going now much more actively on M&A. We recognize that probably, SDK network into all their own SDK, which we are controlling, probably is making sense to us. This is where we are looking for our next M&A. We believe having our own SDK will enable us to accelerate growth in this AI-resilient channel. Of course, it make our top line trajectory and durability much better. Within this SDK network, we are aiming for somewhere around, I do not know, do not catch me in the word, but around $50 million of deal. It can be financed through cash and through equity as well. Okay. Just to flesh out the M&A commentary. I think on your second quarter call, you were talking about potentially $100 million-ish or so of tuck-in M&A in coming quarters, where SDK network would be one of the opportunities, AI capabilities and other, and performance CTV a third. You exited the second quarter with $132 million of cash and no debt. With your free cash flow generation, that could be $180 million in a couple of quarters, helping kind of fund the tuck-ins. Now, switching gears a little bit to enterprise. So, ad spend on your Nexxen via enterprise clients, I think, has been an area of strength as well, up 25%, with the brands activated nearly doubling to 750. What is driving this? I mean, what has been the change in Nexxen, and what is behind the growth there? Karim, I will let you lead on the product side. Sure. I think, overall, we made the acquisition of Amobee several years ago now. We finished integrating that platform and unified our solution. I think we are starting to see the rewards of that and the benefits of that as customers are testing a platform and seeing the outcomes that we discussed earlier. They are joining and growing on the platform. We are seeing both, really. We are seeing current customers grow and scale, and we are seeing a lot of new customers joining the platform. We have nearly doubled the number of brands that are live on the platform from last year to this year. If you look from a product perspective, we have launched a new UI to the DSP that we have built over the last couple of years that is heavily powered by AI and has our nexAI Assistant deeply integrated into it. We have built a lot of new performance models and optimization. We have unified our workflows between the media side and the activation side. We have brought in a new unified ID solution, which is our identity graph. When we are talking about CTV, when we are talking about mobile, et cetera, that is extremely critical as we are able to measure outcomes from one screen to the next and move budgets around. So, it is a combination of many products that we have launched and integrated over the last two years and strong execution from the team going to market with a unified, simpler story and winning new business. Ultimately, that is what is driving that growth. We are seeing that growth accelerate, and we expect that to be a meaningful part of our business. When we say enterprises for everyone in the room here, we are talking about self-service DSP customers. These are customers that are hands-on keys on our platform, managing their own campaign. What is great about this is that it is a lot more predictable revenue than the transactional revenue you would see selling an IO or selling a Deal ID as an SSP, et cetera. These are customers that, typically, once you get them on your platform, and they deeply integrate their solution with your AI, with your data, et cetera, stick around, and you have very little churn on that business. For us, that revenue mix shifting to enterprise is extremely critical to be able to have a growing, predictable business. Okay. I want to switch gears a little bit and talk more about how you guys are engaging with AI. I think that you are using AI internally, and I think you are also doing engineering AI increasingly into your product set. Tell us a little bit about what you are doing, and are your efforts, things that require big investments in GPUs, or otherwise, how are you integrating the AI capabilities? Sure. First and foremost, we're building essentially AI agents that perform multiple tasks within our platform. We have an agent for audience research, we have an agent for reporting, we have an agent for troubleshooting for QA. We're soon launching agent for planning. All of these sit on top of our core technology, and we're building agents on the data side, on the demand side, as well as the supply side. All these agents together essentially can be leveraged in two ways. The first is within our own walls, within our own platform, they power the assistance we have across the platform. For a customer looking for an in-app experience, essentially, they're able to integrate and engage with our agents there. But we're also making these agents interoperable, meaning, they can be accessed within our customer's AI ecosystem. Whatever solution they're using, they're essentially able to leverage Nexxen capabilities within their walls for their own planning, et cetera. Of course, for us, what that does is essentially, we help them plan campaigns, we help them find efficiencies, et cetera, and those dollars then flow to us for activation and creates a more stickiness with those customers. That's on the product side. We keep essentially releasing these core agents that are then deployed within our own platform and within our interoperable layer so they can be leveraged by our customers. Internally, beyond that, we're leveraging AI very heavily, both operationally, so our operations team are leveraging AI to pull reports for clients to automate their workflows for trading, et cetera. Over time, essentially, we're driving a lot of efficiency there, and it's not just in cost-save, also in terms of how much a person within the company is able to do. Of course, in engineering as well, we're leveraging AI heavily for software development. I think the number at this point is over 90% of our code is written leveraging AI tools. So, we brought a lot of efficiency there as well. In terms of tokens and so on, what's great about building these agentic workflows rather than, I'd say, these raw AI calls, we are able to manage very effectively the usage of tokens against our workflows. If you look at nexAI and our agents there, the numbers are really minimal in terms of our AI spend and in terms of cost around that. We also manage tokens for development, et cetera, with engineering. Of course, this is where essentially, it's not so much in GPUs, it's more in tokens we use with various LLMs we leverage. So that's Anthropic, OpenAI, Google, et cetera. We manage that very carefully and manage our costs around that very carefully and make sure we're efficient there and using the right tool for the right solution. But what we see here in terms of our go-to-market and the tools we bring to market, we think our costs around that will be very well- managed and fairly minimum, as we are able to essentially build that orchestration and manage that usage. Okay. On the topic of leveraging AI, I think that does play into what you guys see as a margin expansion opportunity. When we think about your margin, we think about that as a percentage of contribution ex-TAC. I think your guidance for this year, at the high end, would have an EBITDA margin in the 33% range. I think you've spoken about a percentage point or so kind of improvement over following year 2027, potentially into 2028. I was wondering if you could talk a little bit about how you have confidence in the ability to put up some margin expansion, and how long you think the runway is on margin opportunity at Nexxen. In 2026, we already upgraded or raised our top line three times. I think that in H1, EBITDA revenue growth outpaced the original plan we had. Intentionally, in 2026, we invested around acceleration and maintaining the growth because we are seeing big opportunity. I think that the current margin reflects what I'm calling a transformational year, when we are investing a lot across AI, data infrastructure, go-to-market, and platform capabilities, as Karim just mentioned. All of these capabilities, of course, needs resources and head count in order to invest in. I think that expect operating leverage as revenue scales across our existing platform and enterprise customers is increases our end-to-end utilization. I think that we will see the fruit of all the investment that we are doing in 2026 already in 2027. According to our initial plan, we will get somewhere around the 34% EBITDA margin in 2027, and I think it will grow at least 1% or 2% in the coming years, and we will get to the 40-ish% EBITDA margin that we disclosed at our last Analyst Day. I think that the combination of our discipline, the cost management, and improving programmatic mix, we remain very confident that we can reach that in the long term. Okay. Just to level set here, I think in the first half, EBITDA was down 30% in the first quarter, down 8% in the second quarter, but your guidance would have very robust growth in the back half of this year, to give us I think up to 15% growth in your guidance to get us to that 33% kind of margin. Okay. Well, that's great. I think we're at the end of the time here. Sagi, Karim, thank you very much for joining us, and thanks everyone for listening in. Thank you, everyone. Thank you.