Thank you so much for being here today. Let's dive right in.
Okay.
Nexxen has spent several years integrating assets and building this unified platform with both an SSP and a DSP. As we talked about at your Analyst Day, momentum has really picked up. Nexxen has raised full-year guidance three times. Why is the market more receptive now to your unified kind of end-to-end model versus several years ago? How does consolidating demand, supply data help Nexxen capture more value and lower customer cost? Loaded question, but I wanted to really get started off right.
Sure. No problem. Thanks for having me. I think that, as you said, in the last couple of years, we integrated and consolidated different companies. We acknowledge back in 2019 even, that going end-to-end probably is making the more sense economically and for the clients and our publishers. I know that there are still some companies that are one-sided and saying that they will not go dual-sided because of transparency and conflict of interest and other than that. Having said that, we are going that direction. I think that it took us a while in order to put everything in place, sunset some of the assets, putting everything on the same infrastructure and refine our offering in order for the market to acknowledge and understand exactly what we are offering and what we are doing.
I think that it's making sense because at the end of the day, if you are controlling the full funnel, you are really controlling everything you are doing for both of your partners, i.e. the publishers, the broadcasters, the apps or the clients, which can be agencies, DSPs, direct clients, and other. I think sitting on one infrastructure and having a full functional DSP and a full functional SSP, and in the middle, controlling the DMP, which is being leveraged by first-party data of the demand-side partners and first-party data coming from the supply-side partners and our own exclusive ACR data and other 70 different data publishers that are connected into our own DMP, is making the efficiency and the life of all of our partners much easier.
Instead of getting an activation partner, which is only a DSP, and utilize only that, with us, we are getting a technology partner, which is answering all the full funnel of their ask. We are doing discovery, we are doing planning, we are doing activation, we are doing optimization, we are doing measurement. Without the technology stack and the end-to-end abilities, we couldn't do that. This is the reason I think that now enterprise clients are moving towards us much more because they acknowledge that they need a technology partner and not only an activation partner. They understand the level of efficiency we are getting from working on our end-to-end platform. I give you a good example. For example, Toyota onboarded our platform in early 2025. At the beginning, they were only using our DSP. They did well.
They were happy from the performance, from the metrics they got, and from the measurements that they had. But after that, I think it was Q3 2025, we managed to make them understood that they should utilize more than the DSP itself. They should utilize our data and our SSP and our premium inventory and exclusive inventory. They did that, and what happened? In the 12 months after, they lowered their cost per vehicle by 62%. They are paying only 38% right now in order to get the same result they got before. Their return on advertising spend went up by more than 2.5 x. This is an example of showing you why utilizing everything on the same infrastructure is giving you a lot of advantages when you are utilizing that. I think this is where the market is.
This is what now enterprise clients are acknowledging, and probably there will be a question on AI. I will make sure you understand why AI now is doing all of that much bigger and much streamlined.
Toyota was a good example. First of all, so many things to touch on there. Toyota is a great example of kind of the enterprise focus, which is now your primary go-to-market motion. Your enterprise spend grew 25%, advertisers activated through enterprise customers now over 750, right? As the DSP has historically been Nexxen's entry point into these new kind of enterprise relationships, how has that go-to-market motion evolved to win these larger agencies and brands? How do you move customers up the value chain from DSP onboarding to becoming just a more core technology partner?
Okay, that is a good one as well. I think at the beginning, when we are pitching on the demand side to our enterprise client or independent agencies or whoever, direct clients as well, we understand that the DSP is the entry point into our full ecosystem. This is what we are looking for. At the end, when we are looking for something, they understand that they need someone to activate their campaigns and bring them the desirable action and to get to the audience that they are looking for. Having said that, they are onboarding, and the flow is like, we are pitching, and they are onboarded into the system. It is taking around between three to six to nine months in order to expand, get the right confidence, understand how the ecosystem is working.
After they are acknowledging that, there is no reason why not do everything through the same entry point. It is very easy without doing anything now to buy Inventory over Nexxen SSP. It is very easy, without doing anything, to utilize exclusive data and a lot of data discovery and data audience through our DMP. It only makes sense that when they are onboarded, they are utilizing a very small portion of our abilities. Over time, when we are getting the confident and we are seeing the metrics and the KPIs that are being met, they are moving up the ladder in order to get the full functionality of the ecosystem. Then they are acknowledging that the efficiency that we are getting, as in the Toyota example, is much higher than they can get only if utilizing part of the ecosystem.
Make sense. On that kind of end-to-end expansion, your long-term target there in terms of makes is 65%, as everyone knows. As you continue to add more capabilities and improve the economics for these customers that have 35% better performance with end-to-end relationships, how do you think about customers consolidating increasingly more activity on your platform and gaining share relative to peers?
I think that within the question you have the answer. First of all, you just mentioned that most of our clients are transparent with us. There is almost no one client that is utilizing only one DSP. Most of them are using at least two in order to benchmark and to understand what DSP is performing better, and that is fine. When we are finding that, of course, we are allocating more spend towards the winning DSP or the winning platform. I think that, as you mentioned, an average of our clients are doing better 35% on performance when they are utilizing our ecosystem. That means either they can take 35% less or buy less media, or they can get much more with the same budget that we are spending somewhere else. I think the answer is there.
When people are acknowledging that and they understand that the efficiency and the KPIs and the metrics that we are looking for is much more positive on our platform, we are allocating more and more towards our platform. I think that we are seeing a shift and a move from managed campaign that were the majority, I don't know, 10 years ago, of course, to now enterprise or what we are calling self-serve clients. We have in the middle even another layer which we are calling hybrid clients, because the movement from managed into self-serve is not easy for anyone. So we are offering a level of service that is higher than the self-serve, but as transparent as the self-serve. For that, they are paying somewhere between managed to self-serve activation. We are seeing a lot of movement from managed clients into this hybrid model.
From the hybrid model, it's much easier for them to go to the self-serve or to the enterprise solution.
Makes sense. Let's switch gears a little bit and talk about CTV.
Okay.
One of Nexxen's largest growth engines. I think you grew 33% this past quarter despite minimal contribution so far from smart home TV screen. What's driving your CTV momentum today? Are you seeing a shift from, call it the reach-based buying towards more programmatic measurable performance?
Yeah. We had an investor meeting before that fireside chat, and we had almost the same question. I think that at the end of the day, we were like, I do not want to say the first one, but one of the first ones to make a foray into the CTV industry or phenomena. I think that after the pandemic or within the pandemic, it got a huge kick and boost into that. People were sitting home consuming content through CTVs all the time. I think that what is driving now the meaningful growth, as you said, 33% Q2 2026 versus Q2 2025, and already 40% of our net revenues are coming from CTV campaigns is, first of all, the enterprise growth. A lot of clients that are coming into the enterprise solution are utilizing our CTV capabilities. We have endless touchpoints within that.
We are not missing anyone. We just brought into the system, I think Universal, Disney, and very soon Netflix. We are all done. We are not missing any major or even non-major streaming service. On top of that, I think that we are differentiated very much because we have an exclusive ACR data with VIDAA, and this is another very important point. We are already invested in VIDAA around $60 million, which reflects almost 6% of their outstanding shares. It is giving us on top, of course, of the partnership and being a shareholder, it is allowing us to enjoy ACR worldwide. VIDAA, which is an affiliate company of Hisense, is spread at around 50 million screens worldwide. We are touching 50 million different users, and this is tons of data in order to facilitate.
ACR is very important in order to target within the CTV industry because it is like a mirror or a glass that you are looking into the TV and know exactly what the people that are watching the TV are consuming, even if it is content or commercials. For example, if someone is utilizing now Netflix is aware what he is watching when he is in Netflix. But the minute he is shifting to Amazon Prime or to Disney or to Paramount, they do not have a clue where exactly he is. Because we are owning this ACR data within the TV, we are totally transparent and totally clear around what this people is consuming. Even if he is going on his Xbox or PlayStation 5, we know exactly what game he is playing. It is endless really. You are looking into the TV and seeing everything. This is very important.
The second part is that we have a North American inventory exclusivity on VIDAA inventory. We have it only in North America, and it is giving us another layer of exclusivity and differentiator through other peers. I think that the most important part is what we are talking now, or we just launched at the beginning of 2026, and now we are evolving that. It is what we are calling on-screen. When you are opening a CTV, you are seeing an on-screen. With the on-screen, you are seeing all the different apps, Netflix, Prime Video, Paramount, whatever you want to see. Surveys are saying that people are around 10 and a half minutes a day in the home screen. I can tell you from my own experience, I am probably over there much more because you do not know exactly what you want to see. You are moving between apps.
You are not sure. Everyone is saying, "Go to Netflix." At that 10 and a half minutes a day, there's a premium inventory that has not been utilized until today. Why it wasn't utilized? Because every OEM and every TV set has different standards. Because it has different standards, it was only consumed or transacted manually. A lot of salespeople went and tried to sell it to Coca-Cola or McDonald's or whoever. It's not something that can grow and can scale. Because we are an end-to-end solution and have the end-to-end stack and have the DSP and the SSP, we made our own standardization within that, and we are now allowing every partner that can be a different third-party DSP, a direct client and agency to transact it programmatically on this on-screen.
This is like an opportunity which can be medium, can be large, can be extra, extra large. Until now and within our Q2 earnings, it didn't materialize to a lot of money. We are seeing it coming strongly in Q3 and after that in Q4 and onwards into 2027. This is a huge opportunity that from one side, we are seeing the traction from the advertisers. To this on-screen, let's call arena, we have already onboarded from the demand side, The Trade Desk and StackAdapt, and soon Amazon and others, and Basis. From the supply side, we have already, of course, Hisense and Toshiba and VIDAA. On top of that, we have TCL and TiVo, and hopefully soon LG. For everyone, it's incremental, either revenue for the demand partners or revenue for the supply partners.
No reason for them not to come into our-- We are the first in market to allow that or to service that, and we hope that it will do a lot of money going forward.
Integrating all these partnerships, layering on the ACR VIDAA data, that's the key unlock. On the specific ramp, I think you talked about it contributing more meaningfully in Q4 and then obviously 2027. How do you think about the milestones for ramping, not just financial contribution, but call it advertiser count and budgets from there?
I have to tell you the truth, because it will transact programmatically, the minute we have StackAdapt, The Trade Desk, Amazon, and let's say another, I don't know, DV360, which hopefully will come soon to the system as well. We don't need a lot of other partners in order to transact and to utilize it. The vast majority of this inventory will be utilized because these are the big players which are buying in tens or hundreds or billions of dollars. From the other side, I think that having the OEMs connected, VIDAA, Toshiba, TCL, TiVo, hopefully soon LG, we are not missing a lot of the inventory out there. So in my eyes, commercially wise, we are already in a good place because it's really active only for a couple of months. It's taking time, and by the way, now you can show only a native banner.
Very soon, we are deploying an opportunity to show a video over there within the banner, which will make it much easier and probably much more demandable. So I think we are already there, and hopefully very soon it will do a lot of money.
Makes sense. Parallel to this home screen opportunity is mobile. Mobile advertisers increasingly using CTV to drive downloads and other, call it more measurable outcomes. How does your presence across mobile in-app and CTV position you to capture more of that cross-channel spend?
That's a good one. So I think AI is changing the industry, okay? People understand now, I don't know, around 60% of people are going into sites not by clicking a Google search. So it's changing a lot. I think that the two most resilient devices or media channels that are less affected by AI are CTV and mobile in-app. This is where people are now consuming their content other than asking an LLM something. So we acknowledged that again, I don't know, two years ago. In 2025, we went dramatically and signed a lot of partnership with in-app mobile SDK networks in order to have the right inventory. The biggest one is Unity, but we have a lot of other partnership with a lot of other SDK networks.
I will say another thing here, probably the question will come later, but we understand that you need to control the SDK itself. The partnership itself is not enough. It is good, but not good enough. So on the M&A side, we understand that we need and want to buy our own SDK network and to own like a Nexxen SDK. So that is on a different side. Because now our revenue is coming, 40% is generated through CTV and 40% is generated through mobile, like 80% is already there. I think that the cross-sell between the mobile and the CTV is very important. Because the CTV is connected to the home network and to the home Wi-Fi, we understand exactly when we are seeing the Wi-Fi network, what other devices are connected in the same household.
We know this is the CTV of this person, and probably this is his mobile. Then we can, what we are calling cross retargeting from CTV to mobile and from mobile to CTV. So let us take the, I do not know, a McDonald's example. So you are seeing, I do not know, their new Big Mac or something on the TV. 10 seconds after, I am showing you the same ad with a QR code on your mobile app and click or scan or whatever you do in order to order a Big Mac. So I think cross-targeting is very, very strong. It is making the engagement much scaled and much higher, and of course, it is affecting the performance and the efficiency and the return on investment for the advertisers and for the publishers.
I think because 80% of our revenue is already in the mix, it is giving us a lot of advantage in order to find the right user, to build the right audience, and to attract the right end user.
Perfect. Let us talk about the future of agentic transactions. One of your more important product catalysts. You have got six agents already live, additional agents in development. You have got the DSP UI, nexAI DSP assistance. What nexAI capabilities are you seeing the strongest customer adoption today, and what gets you most excited about the product roadmap here?
That's a good one. I think at the end of the day, AI for us is a huge enablement on our system. It's helping us to onboard clients much faster. It's lowering around 50% the time that we are onboarding and activating an enterprise client, which is amazing because we can run in a double speed, and this is the reason we went from 400 using our services clients within our enterprise to more than 800 only in H1 and probably to grow to 1,600 by the end of the year. I think this is the main catalyst for that. In addition, I think that at the end of the day, AI is strong as the data that is sitting on. This is the difference within our ecosystem, again, from our peers.
If you are only a DSP or SSP, you are enjoying only one side of the data. When you are sitting on the full funnel and the full value chain, you can enjoy both. Then the AI can make the right decision all along the funnel. So from onboarding to planning, to activation, to optimization, to discovery and to measurement. Our AI is different to our peers because it's giving you the full funnel. As you said, we have six active agents already running all the time on the funnel. Very soon, I think it's now seven, if I'm not mistaken, but very soon it will be 10, and it's the future. Amazingly, AI agent can connect to a different AI agent. So AI agent of the client, which is utilizing Claude or ChatGPT or whatever, can utilize our AI agents through an MCP.
You just need to put our MCP code into your AI agent. You are logging in, getting the credentials from Nexxen AI infrastructure, and you are transacting by not using any person. First of all, for our customer, it's amazing. They can utilize the system without increase their staffing. Crazy. The level of efficiency they will get through our agentic solution, much higher than we will get from a one-sided peer.
Makes sense. On that subject, planning and reporting agents are becoming more widely available across ad tech. How does your ownership of the end-to-end DSP, SSP, DMP enable nexAI to deliver outcomes that maybe a standalone agentic layer could not?
Yeah. That's a good one. Let's say that you are a DSP and you have an AI agentic features agent, whatever you want to call it. You can go to the DSP and say, "Hey, what is my campaign doing versus last week and versus the target?" You are getting an answer. It's doing 85%, it's lower on that, and then you are asking, "Okay, on what does it run? On what media channel? On video, and give me the exact allocation," blah, blah, and all around that. But then it stops. It's running somewhere. It's running through the DSP to a third-party SSP, and there it's gone. With us, you can tell now the agent, which is the SSP agent, "Hey, I want this campaign to run on this inventory. I want this inventory.
Tell me exactly what is the win rate. What should I do in order to allocate more? Where should I allocate less? What is the most performing channel? What is the most performing device? Where exactly? What is the most performing publisher?" Then you can do a lot of different optimization within the same agent and with the same system. If you are using a DSP, you don't have this ability. Measurement, okay? Now you have the outcome of the campaign. You want to measure it. It stops within the DSP. You don't have any transparency or any layer into the SSP and what it did over there. I think our AI probably, and I've been, I don't know, with five different clients on a pitch.
Everyone that saw our AI solution, in live, not like a video or something like that, was astonished because they really can get all the services and all the reporting and all the measurement and all the information and all the data that they want in order to understand what they should do differently or what they should do the same. It's really crazy. Hopefully, now I think by the end of October, all the 10 different agents should be running live, and then you can do everything.
How far along the adoption curve do you think we are in terms of that becoming a scalable channel?
I think we are already there.
Yeah.
I think that everyone, existing clients, maybe because we are used to the old UI and to the old platform, they are still doing more of the same, but we are educating them in order to move to that direction. The new ones that are coming are using it instantly, and this is one of the reason within the pitch that they are coming into our system because we want to utilize it. At the end of the day, past days, in order to be an enterprise client, you needed to have the right staff in place. You have to write the right bandwidth in order to facilitate it. Now, with the AI, you don't need a lot of people. You need probably one or two in order to facilitate what, I don't know, 15 did in the past. Everyone that is seeing that wants to embed it immediately.
That's very encouraging. I've got a lot more questions, but wanted to open it up to the audience. Any questions out there?
Don't be shy, guys.
Perfect. I will keep going here.
Okay.
Perfect topic for you. Let's talk a touch on the financial model. Raised revenue outlook three times this year, but maintained EBITDA guidance. Can you walk us through, under that context, the areas you're investing, and then looking out to 2027 and beyond, where do you expect to see hopefully greater leverage to get to the path of that 40% EBITDA margin target?
Yeah, that's a good one. I think 2026, as you said, we raised the guidance three times already. I think that this is an investment year in my eyes in Nexxen. It's a transformation year from moving from, I don't want to say old legacy DSP into an AI agentic DSP. We invested a lot in our infrastructure. At the end of the day, AI is enabling you to do much more, but you need the infrastructure in order to do that. If we are doing that through AWS, it's costing a lot of money, and it's killing the extra or the incremental profit that you are generating. So we invested a lot in our infrastructure. We invested a lot in data partnership, in hosting partnerships, in SDK network partnerships.
Having said that, this is probably one of the reason for the growth engine we are seeing in 2026, and it will continue into 2027. The cost transformation will stop at the end of 2026. We are already seeing the efficiencies getting into place. It's not material. It's couple of millions that will go through Q3 and Q4 but will be much more substantial in 2027. We didn't even start at 2027 budget process yet, but looking from an holistic point of view and from high level, I really think this year we are supposed to do like 32% adjusted EBITDA margin that we can move to 34, even 35% already in 2027. Then the way to the 40-ish is very easy because it's only scale.
Perfect. On that subject of finding those greater efficiencies, how do you think about the decision to use frontier models for performance relative to maybe some open weight models? Does that fall into your decision-making process?
I am not sure I understand the question, but if I got it, I think that we are seeing now performance as something that is moving from mobile in-app into CTV. Okay? Because we came at the backbone of our history of Nexxen, we were a performance company, so we know how to do that, and we are now utilizing more and more. I think Moloco is doing a great job of utilizing a performance DSP. People are now looking not only for branding campaigns, but for actionable results that they can measure and do something with. We are moving up this path, in our ability through the SDK networks into our CTV. So I am not sure I-
That is helpful color. I was asking more on using leading frontier models, the likes of an OpenAI, Google, et cetera, relative to using maybe some open weight models that are lower cost. Think of the Meta Llama's of the world, et cetera. But if you do not have a perspective there, that is-
Yeah, no, I have that perspective. So for now, we are using the higher cost LLM.
Yeah.
It's not saying that we will not change it going forward because it's becoming quite easy. I think because we wanted to run very fast and because they are, to some extent, supporting you much more in a lot of different instances, we went to that. Our token cost is not sky-high as other companies or other CFO that I'm talking to. I think that we are very efficient in our way of thinking, in our DNA. This is what we are using for now, and we are relying, of course, it's a fully transparent and agent to agent through MCPs and whatever, so everyone can be connected to our system. We are working according to market standards.
Perfect
and we are there. But maybe
Yeah, and I mean, gave the margin framework, so that's very helpful. Maybe wrapping up in the last minute, capital allocation. We hit on a couple of these subjects. The VIDAA investments, you've repurchased 40% of shares over the last couple of years. You talked about potential tech and M&A, but if you could just help us rank order priorities there.
Yeah. So I think we acquired almost 40% of our company in the last three and a half years. We allocated for that almost $250 million . We still have authorization to do buyback, and we are considering that as well. It's part of our capital allocation. Having said that, after being silent on active M&A from the end of 2022, we acknowledged that in order to grow the system and on top of our organic growth, we need to allocate some important M&A into our end-to-end solution. The first one is SDK network, the most important one. The second one is maybe some AI capability that we are missing in order to enhance everything much faster.
The third one is maybe a performance capability that we are missing within our DSP and can be a big facilitator because we are seeing now a lot of shift of money from branding into performance, and we want to capture that. We are part of this industry, but it is being transacted through third party on our SSP. We want to be to all the full funnel.
Yep
We are moving that direction. This is our main capital allocation areas right now.
Perfect. Thank you so much for joining us today. Really appreciate the discussion. That is a good place to wrap.
Thank you very much.