You all set? Yes. Awesome. Okay. Hello everyone. Thank you for joining us. My name is Rohit Kulkarni. I'm the internet research analyst here at Roth, and welcome to session number four of our fifth annual AdTech summit at Roth. This is a topic that over the years I have followed directly, indirectly, and kind of sets the tone for how various different components in advertising technology have been moving. I think we have last three sessions, we are seeing that a lot more AI, a lot more agentic AI, and again, as with any tech workflow platforms, I think with data and AI, it becomes even more powerful. That's a very high-level overview. Before I jump in, I'm just going to walk through my disclosures, which I should find. I had it open somewhere. These are important disclosures that I should share per my compliance team. One Mississippi, two Mississippi, three Mississippi. Stop share. With that, as in also another thing which we do with these webinars is an investor poll. I just launched something. People who are online will see a poll of five questions pop up. Would be awesome if you guys can participate at the end. I'll collate all the responses and share the collective wisdom of all the results that we have received in this session so far. That's my long-winded kind of introduction, but most importantly is thank you Steve and Nishant from PubMatic. I've known PubMatic over the years. I would say 15 years back, I first met Rajeev, and I've kind of followed his journey, and I've kind of relied on him on a lot of tech and AI and advertising knowhow. Steve and Nishant, thank you so much for doing this. I guess just to kick things off, maybe at a very high-level for investors who still think of PubMatic as a traditional SSP, maybe just how would you describe the company today? Maybe talk about two or three biggest changes in PubMatic's positioning over the last, call it 18 to 24 months. Great. First of all, thank you, Rohit, for inviting us. We're always very happy to join and talk about all the exciting things at PubMatic. Let me just set the context for investors regarding your question. Clearly, for those who follow us, and I'll underscore for those who haven't, over the last several years we've significantly transformed PubMatic, and we are definitely no longer the SSP that many investors may envision. Today, PubMatic is a unified AI native platform for the open internet. To get there, we've implemented some major changes over the last couple of years. First, we've expanded the platform to serve four key stakeholders, not one. Publishers, of course, buyers, data partners, and Commerce Media companies. What this has done is it's notably broadened our business and diversified our revenue. Second to your earlier question or point that others are making, we believe we're the leader in the space, we put AI native infrastructure at the center of the company. We own and operate our infrastructure, and our longstanding partnership with NVIDIA lets us process massive amounts of data, both structured and unstructured, and that's what's really required in today's real-time decisioning environment. Third, we built an execution layer on top of those assets. We'll get into it in this session, we've launched Activate, our buy-side product, Connect data, AgenticOS, and now most recently we had an announcement on a product we call Decision Fabric. In that broader context, buyers, publishers, Commerce Media companies, and data partners can directly leverage our infrastructure and drive better outcomes, which is ultimately sort of the long-term opportunity vision for the open internet. From our perspective, what makes what we've done so powerful is that each of these pieces reinforce one another. Just to set the stage for investors who don't know us, we have one of the largest scaled transaction platforms on the open internet. We have deep integrations across thousands of publishers. We have 28 of the top 30 global streamers and more than 300 data partners. Why is this relevant? It's more activity generates more data, more intelligence. More intelligence obviously improves performance and better performance attracts more buyers, publishers, and now new partners that we're bringing on every day. That's the flywheel that we built at PubMatic, we are no longer the traditional SSP, but a very different multifaceted company. Importantly, I think one of these aspects that's going to become increasingly relevant, important for investors to understand is we operate on largely a fixed cost base. As that flywheel works, it doesn't just drive revenue, but it also drives margin and cash flow. Okay. Awesome. As we discussed earlier, there are three topics I would love to touch on with each one at each of the session. One is just AI agents and data. I think you touched upon a little bit, but I'll just draw them out a little bit more, Steve. I guess, as of right now, or maybe the last six months, where is AI manifesting in your P&L? As in both, well, first, more importantly, the revenue side, and then maybe we can also talk about the cost side. If investors are to point to one or two things, like this is how AI is manifesting, what would you point them towards? Sure. As we've shared the last couple of earnings calls, AI is showing up across our P&L. The way we describe it is that for PubMatic, AI is a financial lever. It starts with the revenue side. It's really the traditional points of yield optimization, campaign performance, workflow automation. What that has allowed us to do is to leverage the power of AI to help improve all of those areas. Now, where it's manifesting itself in the P&L is, we're seeing very significant growth in what we call emerging revenue streams. In the first quarter, that category grew over 80% for us- Represented 14% of total revenue. This is coming from AI-powered capabilities for publishers and buyers. This is across our Activate product, AgenticOS, Commerce Media. Just the Activate product alone in the first quarter, that was up 3x in terms of spend activity on a year-over-year basis. For us, in terms of AI and revenue, the consistent theme is really around performance. AI makes our platform more effective, and that's obviously what drives more usage. As a reminder, as I just called out, when we do that, we deliver not only more revenue, but also more profit. The other facet that we've clearly been focused on is on internal productivity. We've shared over the last couple of quarters how we've done that internally. Clearly, it's a big driver of our progress in innovation engineering. Over 80% of our code now is leveraging AI. I shared in a recent quarter that we're using AI in finance, legal, customer success, where we're automating work, and we're improving productivity. One objective measure is that in the first quarter, while our overall headcount was down in the single digits, we actually increased investment in our go-to-market organization. Net-net, we're seeing a tremendous financial lever in terms of revenue and productivity in the bottom line. Fantastic. I guess just to double-click on this 80% year-on-year growth in emerging streams, what are the advertiser objectives that you think you're being very successful? Or what pain points is AI helping you solve from a customer perspective that is leading to that kind of accelerating growth? Nishant, I'll let you take that because you're seeing that every day when you talk to customers. I think we're seeing that we're able to solve for a lot of customer use cases, actually. We are able to drive a lot of measurable outcomes for our publishers and buyers even today. I'd say, from a customer activity perspective, I would categorize their work in our platform in two categories. One is efficiency, and the other one is effectiveness. We see efficiency as where we are able to reduce the time complexity, the cost required to execute advertising. Right? Effectiveness is about improving performance and outcomes for the ad dollars that advertisers are spending on our platform, right? We've done is that we've launched 20 agents, these are driving significant workflow automation across soup to nuts, from campaign planning, which is the first step that an advertiser goes through when thinking about achieving a certain objective, such as for a back-to-school advertising campaign. They go ahead and set up the ad campaign, activate it, once activation happens, impressions start delivering to consumers on the internet, the campaigns are optimized. It follows through with the step of reporting. We've launched 20 agents, which are able to work across this entire cycle, soup to nuts. In many of these cases, our customers are reporting that the campaign setup times, for instance, have reduced by 90%. The troubleshooting time is reduced by 50%. That allows their teams actually in being able to spend their time on strategic tasks, not on the cumbersome details of getting programmatic to work, that is something that our customers are really enjoying. From an efficiency perspective, we've made a lot of progress. Similarly, on the effectiveness side, we're seeing that customers are able to improve the amount of working media, the campaign performance is improving. The inventory utilization rate has increased. One of the examples there that we had shared previously as well was our work with Butler/Till, which is an independent agency. Using AgenticOS and Activate, Butler/Till reduced the intermediary fee by almost 40%. While maintaining the benchmarks that they look at internally, as well as the industry benchmarks for both fraud as well as intermediary quality. Looking at those efficiencies, they increased the amount of working media that was available to the campaign, and that resulted in almost a 40% increase in impressions that they were delivering to the intended consumers. That resulted in a 30% reduction in CPMs. These are successful examples of our customers actually completing the entire campaign life cycle. This is not a situation of these being isolated use cases. We are seeing these same themes emerge across publishers as well, where publishers are benefiting from better yield optimization, greater monetization efficiency, and also generating new demand opportunities. On the other hand, buyers are benefiting from some of these things that I described. simplified workflows, faster execution, better targeting, and eventually all of that leads to improved performance. That's why we are so excited about this opportunity. We're having a lot of fun building what we're building because it's actually solving a lot of meaningful customer problems. We're not just automating existing processes. We're really seeing that our mission is to fundamentally improve how digital advertising is bought and sold and optimizing, and that's creating value for every participant in the ecosystem. Very cool. I guess, where do you think we are in this? You must have seen a lot of tech diffusion happen and adoption by your customers. You launch a product and it kind of steadily gets adopted across various cohorts of customers. Is there any evidence that you see that the adoption is much faster with the new products that you are doing today, given all the benefits that you talked about, efficiency, effectiveness and whatnot? Maybe 12 months from now, where do you see this kind of diffusion of new products with AI underpinning to be at? I can take that. Yeah. I think, Steve, you touched upon this a little bit, right? I think Steve called out that we are already generating revenue from our AI solutions. When we go into customers, even today, and I reference this from a timeline perspective, we launched some of our initial agents for our customers more than 18 months ago, in late Q4 2024, early Q1 2025. A couple of quarters ago, we announced AgenticOS in Q4 2025, I believe late Q4. Even though we've had such a steady cadence of releases and innovation from an AI perspective that we've brought to market, we still get into customer conversations and customers tell us, "Hey, this is really cool. We're not seeing this kind of agentic lean in and this kind of innovative AI solutions coming for publishers as well as ad buyers. We've been truly a market leader in agentic, and that's really exciting for us. What's equally exciting also is that we are further along from an adoption perspective also. We're no longer talking about AI as a future concept, something that will happen in future. We've facilitated more than 1,000 AI-powered publisher deals. For buyers, we've enabled more than 30 live Agentic campaigns with many customers who are renewing, expanding usage, engaging with us for figuring out how they can scale their campaigns further. All of this is in the backdrop of how much efficiency increases these customers are seeing. They're seeing campaign setup times reduced by as much as 90%. That's a very meaningful workflow automation, and that's leading to better performance outcomes for each side of the transaction for publishers as well as ad buyers. The interest level is very high and the adoption is ramping up. In that backdrop, we continue to bring more innovation to drive up that momentum, build on that momentum, and continue to build scale. Recently we announced our buy-side agent, which is called a Detailed Reasoning Agent. Not a very creative name, but symbolizes what it does. What the Detailed Reasoning Agent does is that it drives greater transparency. Our intention here is that with that transparency in how the technology works, we're going to be able to accelerate adoption of Agentic buying. One of the biggest questions that buyers have around AI and this level of automation is that why are these recommendations being made? Why is the platform thinking about how decisions are being executed? What the Detailed Reasoning Agent does is that it addresses that ask from customers directly, and it provides visibility into the recommendations and AI-driven decisions that are being made in campaign setup or during the optimization phase later on. That transparency is really important. Okay. Good point. Buyers can see not only what the AI is recommending, but they also see the rationale behind those recommendations, right? That level of insight actually builds trust, right? When they have trust in the decisioning, they feel like they're remaining in control. Right? The feedback that we received when we launched this is that this is actually incredibly encouraging for them. One of our major holding companies actually told us that this PubMatic approach to transparency via an innovation like this is actually essential for agentic advertising to work at scale. Right? My point is that the innovation momentum around agentic continues to come out, and as a market leader and now experts in this area, we are building agents that establish not just the core functional workflow, but also the guardrails and governance around these agentic campaigns. Right? That will ensure that these buyers are able to operate within their campaign objectives and the requirements. Interesting. Right. Just to wrap it up, what we are hearing from our customers is that PubMatic reality is increasingly meeting the customers' expectations. This is much bigger than just a technology upgrade for us. We think that we're transforming workflows, bringing in efficiency, and driving performance outcomes. When buyers have this level of transparency and control, the adoption that we're seeing continues to be very encouraging for us. Very cool. I do want to also address kind of what, Steve, you mentioned about internal efficiency. It feels like the flavor of the month keeps swinging around token maxing and engineers spending way too many tokens and CFOs running through their budgets in the first two months of the year and whatnot. I guess, maybe just talk about how are you as a CFO measuring the ROI of AI on efficiency gains internally, and kind of is there a reason to believe that structurally companies like PubMatic are going to be much more profitable if they kind of use AI internally in a much kind of efficient manner? I mean, from our perspective, it starts out with our mindset. Once those tools started to become available, we embrace them. We're now on cycle two, cycle three in terms of year cycles, where we've been working with various tools, and it really comes down to like any other investment, what are the metrics that you're going to measure in terms of success, and then tracking those. We go through a constant cycle of putting more dollars behind positive ROI projects. There's really no magic to it. It's really just the same metrics you would with any investment. The discipline comes with what are you trying to do? How are you going to measure it? And then, you measure it, and then the learnings from that feed back into the cycle for that next round of investment. From our perspective, it's absolutely central to everything that we're doing today, and it's our vision for our company over the next five years, to be that AI native performance digital advertising platform. Again, Nishant said it very well, we're now looking at AI as a technical innovation. It's just fundamentally transforming how we think about the opportunities, then changing how we do things when we have those learnings. Okay. I guess just to ask you a more direct question, are you guys worried about this kind of token maxing internally, or is this something that is more kind of controlled kind of spend of tokens by engineers at PubMatic? We believe we manage it very efficiently. It's always sort of a function of like what are we trying to accomplish and then measuring that, and then where we have situations where we let's say hit some threshold, we take a look at the driver behind it, and we assess is that something that can be planned or worked around to get more efficiency out of it. It's really, I think, it's not about trying to constrain the resource, it's how do you direct the resource towards the greatest outcome? Yep. Okay. That's a fair answer. I guess going back to Agentic and AgenticOS, maybe level set. I believe you guys had a pretty good announcement around AgenticOS. I know, Nishant, you referenced 20 agents to do a bunch of things. Maybe talk about kind of what does that launch include and how far are we where kind of you have agents stitching together and data and governance and all compliance problems would be solved, where agents would actually kind of go from beginning to the end of an advertising workflow within PubMatic. Like when does that happen, if that happens? Yeah, I think, Rohit, if I'll start addressing your question and if I haven't addressed it fully, please call me out on that portion. Yeah, sure. From our perspective, like that transformation is already happening for us. We are delivering campaigns. I just shared how we announced that we have 1,000 deals, 30 campaigns which are already live producing revenue and solving customer problems at a very high level of efficiency, right? For us, it's the reality, right? The most important transformation that agentic AI is doing is that I think it's fundamentally changing the traditional open internet advertising environment. One aspect of that large change is that the role of the user interface in media buying is completely being redefined, right? Historically, what we had seen is that a significant portion of the value in advertising platforms came from the UI itself. The buyer spent years learning how to operate very complex systems, and that created a form of a UI lock-in. In an agentic world, what we are seeing is that that lock-in is actually going away. The interface is increasingly becoming the LLM, right? What we are seeing is that our buyers are using their tools that they have adopted in their organizations, like Claude or ChatGPT or maybe even a proprietary agent, and they can simply describe the outcome they are trying to achieve. The AI handles campaign planning, activation, performance, going onwards to steps that come later in a campaign life cycle, such as troubleshooting and eventually reporting, right? There's no special training which is required. As a result of that, what we are seeing is that the adoption curve is actually very flattened out. There's really no adoption heavy lift which is required, right? It's because there's no dependency on learning a specific workflow interface, right? You like a certain tool, like Claude, and you can log in and start creating campaigns with the PubMatic platform, right? These AI agents, they're performing advertising workflows which would have traditionally required people operating multiple software interfaces, right? With agent to agent, those interfaces aren't necessary. As a result of that, the need for infrastructure and scale is perhaps rising even more. Yeah. Agents need inventory access. They need data access. They need decisioning logic. Ultimately, all of that, when it is made available, leads to performance outcomes. That is where we are seeing AgenticOS comes in. AgenticOS provides that operating layer across all of these functional areas so that AI agents are able to interconnect and work with the PubMatic's inventory scale, data scale, and leverage all of our activation infrastructure, including the NVIDIA infrastructure that is available to our customers. That's why we are so excited about agentic advertising. We believe it shifts value away from that UI interface layers and opens up the opportunity for advertisers to leverage the innovation that comes out of platforms like PubMatic. Really extract the value of that infrastructure, inventory, data connectivity, and all of the execution capabilities that power outcomes for our customers. Those are assets we've spent years building at PubMatic, and it's great to see that with Agentic, we are able to use Agentic AI as an unlock for our customers to derive more value here. Okay. I was just going to suggest, Rohit, we recently had an announcement called Decision Fabric, a product that we shared publicly within the last two weeks. It might be worth just digging into that a bit because it really underscores how we have built an AI native infrastructure, because it, I think, will give the investors sort of a clear picture on the trajectory that we're on. Nishant, do you want to just briefly summarize what the product is intended to do? Yeah. I can do that. Steve, I think we're very excited about Decision Fabric. We announced it just last week. We've had this ability to bring together inventory. The scale inventory, including the top streamers globally that we have on our platform, combine that with the 300 or so datasets that we have in our platform, connect the two together. To Activate. Now we are bringing in the ability for our customers to bring algorithms and decisioning and apply that together with those other assets. That was previously not possible in the ecosystem, right? That activation layer is called Decision Fabric. What DSPs, curators, Commerce Media networks can do is that they can run their algos directly inside PubMatic's infrastructure. That allows advertising decisions to be made closer to inventory and data, and it happens in real-time, right? That's been our vision that we are creating a unified ecosystem where things don't run in silos. The scale of our ecosystem is fully delivered effectively when we are able to combine these things in which now we've added this decisioning layer also. Okay Unification eliminates unnecessary hops, signal loss, latencies lowered. All of those things result in improved performance as well as efficiency, right? This really, Steve referred to this, right? This is really creating a powerful flywheel for us and for our customers, right? When you deliver better performance, it drives more usage. More usage will create more data. That data will be used to improve the models, and those improved models will result in better performance, which drives more advertising activity. Hmm. I guess, just to ask a more direct and simple question, does that kind of collapse the entire pipe into a single platform, if you will? Why would intermediaries be there, rather brands or directly commerce platforms should be using the Decision Fabric along with its AgenticOS integration? Is that the final vision that you anticipate how this product evolves? Yeah, I'll take that. From our perspective, what it really reflects is our vision of where the ecosystem is going. We've called it out as you have. There's a lot of change and evolution going on, and ultimately, where we see the ecosystem evolving is in the open internet is pure delivery of performance, transparency, and product innovation like the Decision Fabric allows that to happen. The key point is, it's not like a bolt-on. It's how we've architected our platform and how we're going to continue to evolve our business. Ultimately, what it represents is that buyers want these things. They want to be able to deliver high ROI, they want to have control, they want to understand what's going on, and they want flexibility. What we've done as a company is to put all of those capabilities onto a platform, the point that I made right at the outset. We have now multiple stakeholders that can leverage this unified platform because it's basically delivering on what the buyers want. For publishers, this is going to be a significant upside opportunity because it's going to create more demand and incremental U.S. dollars for their inventory. Okay. I know we are past 30 minutes. Quick answers, please. On connected TV, live sports, can you talk about where are we with that and where does PubMatic fit into that as a growth driver going forward? Absolutely. From our perspective, we bring a number of key differentiators into the CTV opportunity. Clearly, a big part of that is our ability to take advantage of the live TV sports opportunity, which is rapidly evolving and emerging. Right now, we're embarking on sort of the World Cup, and it's just sort of a burgeoning opportunity. CTV, it's had great growth, but it's really quite early innings for it. We see live sports as one of those really big unlocks for incremental growth. Okay. I guess if you think through the next couple of years of incremental growth, exactly what you mentioned, Steve, what other growth layers should investors think of when it comes to PubMatic, where there is the traditional part of your business model, and that is getting agentified, and then probably there are other incremental growth drivers. Do you think about your business that way? If so, what would be those layers? How would they look like? Yeah. Absolutely, Rohit. That's exactly the way we think about our business opportunity. Let me summarize briefly. We've been investing in developing multiple growth engines. We touched upon a couple: CTV, Activate, Connect, our data product, Commerce Media, AgenticOS. For investors, I think the key point to take away is that these are not standalone businesses. From our perspective, we've intentionally built our platform where innovation in one area helps accelerate growth in another area. I think that sort of goes back to the flywheel effect and sort of this great innovation cycle that we're on. The point that we're underscoring here is all of these growth engines are built on top of the AI native infrastructure. That is not possible without the 15, 20 years development of our publisher relationships, the considerable data ecosystem that we built. From my perspective as CFO, what makes this so interesting is that our growth is compounding. Meaning, I said it right at the outset, more data improves the AI. Better AI drives better performance, more activity, et cetera. The growth engines that I described, while each in their own right are meaningful opportunities, what is very powerful is the fact that they're compounding and reinforcing. I'll just reference very briefly, as I called out in the first quarter earnings call, as I look ahead to the second half of the year, these growth engines are what are gonna allow us to return to double-digit growth. Nice. Yeah, as in maybe fine-tune that, like what gives you greater confidence that your second half acceleration in growth will manifest the way people expect it to? Okay. A couple things. One. Sorry to put you on the spot. I know we were talking about. No, of course. I'd love to double-click on this. First and foremost, how we're going to get to double-digit growth in the second half is the core of our business is actually already growing double digits. A stat that I shared in the first quarter was if you exclude the large legacy DSP buyer impact from mid-2025, the rest of our business, which is about 83% of all of our revenues, grew 13%. That's obviously a very significant scale part of our business. Where is this growth coming from? It's of course, the growth engines that we've discussed. CTV, excluding the legacy DSP buyer, grew 18% in the first quarter. Mobile app on its own grew over 20%. Which is a very powerful growth engine for us. It's about over 20% of all of our revenues. I'd referenced already sort of this strong growth that we are seeing in emerging revenues, which represented about 14% of revenues in the first quarter, and that grew over 80%. All told, we have a core business excluding the legacy DSP growing double digits. We expect to lap that impact from the legacy DSP buyer being in the third quarter. With the growth engines lapping the DSP buyer, we are on a trajectory to deliver double-digit growth in the second half. Fantastic. I guess your shares have been one of the good performers year-to-date, and almost have doubled from the end of March trough. In your opinion, what is still misunderstood in the shares despite the run that you've had so far this year? I think we touched upon it briefly a moment ago. I think the top reported revenue line obfuscates sort of what's really going on with the business. Okay. I think that is important to keep communicating to the street in terms of what's going on beneath that surface, all the growth engines that we have. The other facet is that, clearly we are a leader in AI. We're generating revenues today. We are learning from the process of pushing the technology and the workflows. We have great learning cycles that we are feeding back into our product. That's going to continue to sort of percolate, but doesn't really show up in the total numbers yet. It's embedded in our emerging revenues. Of course, the subtext is in the street consensus, there's really no factoring in of the Google remedies. We're well down the path of Judge Brinkema deciding what those remedies are going to be. As a reminder to everybody, Google was found to be a monopolistic party in the AdTech ecosystem. What investors may not fully appreciate is that there isn't sort of anything that fundamentally we need to do to take advantage of the opportunity when remedies get put in place. We already share basically all the same customers, publishers with Google. As soon as that, I'll call it, the field is made fair and level, we are going to deploy all the resources that we've been talking about, and we're going to win our fair share of the auctions. A way for investors to dimensionalize what that could be is every incremental one percentage point of market share that we gain as a result of having a level playing field translates to about $50 million-$75 million in incremental revenue. As I called out right at the outset, because of our leverage cost structure, the vast majority of that revenue will be dropping to the bottom line. What investors really don't understand or envision today is the impact of that. Yep That alpha. It's an asymmetric outcome. Yeah. Any sense on timing of where the remedies and when the remedies could be decided and when they could be implemented, like those two steps? We have no specific insights, of course. Judge Brinkema has demonstrated historically in decisions that she's made that she moves pretty efficiently through the system, but she obviously controls the timing, the outcome, et cetera. I think the real takeaway is, there doesn't have to be a miracle for us to benefit- Yep from what happens with those remedies. When it happens, there will be a clear positive impact on our business. Awesome. I know we are over 40 minutes. I wanted to be careful of your time. I'm going to share my screen very quickly on this survey that I do, and I'll share the collective results by Monday night with everyone. Few more sessions remaining. As in, I'll just run through this super fast. One is, how worried are you about a recession? Slightly positive bias, neutral to positive. People are not very concerned about a recession right now. Excited about which is the incremental growth driver, connected TV comes up, and then ChatGPT ads comes up. We did talk about that a lot. Then in terms of AI applications to advertising, ad targeting has come up, and then ad attribution and measurement. These two are kind of the top selected ones. Which area of AdTech do you feel there would be meaningful M&A or consolidation? I think identity, data collaboration, clean rooms and measurement has been one of the maybe AI workflow and creative optimization. Last is which company do you think is best positioned in the AI-driven future of advertising? Again, I think it's kind of similar to what we had on the top, connected TV and data. That's that. Most importantly, I think, Steve and Nishant, you guys did a fantastic job going through the story, going through some of the AI and agentic things as well as your story to investors. Thank you for spending the last 45 minutes with us. We'll have a research note out and, again, we have three more sessions remaining. They're all on our microsite. Thank you- Great Steve and Nishant. Thank you, Rohit, for hosting us. Really appreciate it. Thank you. Bye-bye. Thank you, Rohit. Bye-bye.