Hightouch raises $150M Series D at $2.75B to become the only on-brand AI platform for marketers

Speaker 2: you can't keep asking John to do to do that. It's

Speaker 1: I

Speaker 2: What's wrong, Jordan? I had to How do you drink your diet Coke? I had to step away for a second because I was gonna

Speaker 1: Well, our next guest is in the waiting room. Let's bring him in to the TVP and UltraDome. How are you doing? Let's welcome to the show.

Speaker 9: Hey, Jordan. Hey, John. Thanks for having me. Thanks for joining.

Speaker 1: Introduce the company, introduce yourself, and then tell us the news. And I have a million questions here, so very excited to talk to you. But give us a little introduction first.

Speaker 9: Sure. I'm Kashish Gupta, co founder and co CEO of High Touch. Mhmm. We built AI platform for marketers. Mhmm. And today, we're announcing our series d, which is a 150,000,000 at two point seven five post led by Goldman Sachs.

Speaker 2: Two point seven five.

Speaker 1: Absolutely massive. Yes, sir. I I feel like I don't know. I mean, category, fantastic industry, fantastic megatrend with AI. This feels a little under the radar to me. Like, how what what is the prehistory of the company? Like, how have you built to this point so successfully?

Speaker 9: Yeah. So we started in 2020. Okay. For the past four years, we've helped marketers use data to personalize their market. Sure. Now, helping them use AI Yeah. To actually build on brand content such as ads and emails. Okay. So, we believe and we know that we're the only on brand AI for marketers. If you use Gemini,

Speaker 1: it will hallucinate things about your brand whether whereas if you use High Touch, Sure. We will get everything by your brand correct. Okay. I I I what level do you want to operate at? What level have you operated before? Because a marketing team can be thinking about everything from like what podcast the CEO should go on to a Yeah. Billboard campaign to a Super Bowl ad to like a micro optimization in the Facebook algorithm or their Instagram ad delivery or creative. Like, where or and then you could dial really deep into just one niche or vertical. Like, how have you thought about rightsizing and creating like a correct solution for the marketer?

Speaker 9: Yeah. So, I mean, marketing is very heterogeneous. If you think about it, most marketers are spending their time on strategy and branding. Yeah. With AI, we now actually have kind of unlimited labor at our disposal. Mhmm. So the bet is that a single marketer can now do the work of so many. Mhmm. In the past, we were kind of stuck doing things like segmentation and like microscopic ads in emails. Sure. Now, with our AI platform, we're letting marketers think of new strategy, generate those images, and try stuff in market very, very quickly. Yeah. So with High Touch, the marketer can launch something like a 100 campaigns in a week, get feedback on all those, and then launch new campaigns based on what's working and not working. Yeah. You said images. Are you doing video?

Speaker 1: What's the status on AI and video? Video as well, especially UGC. And UGC is working extremely well. Yeah. And do you want to be model agnostic and then set up like a harness that makes sure that everything's on brand? Or do you want to train foundation models, diffusion models, image generation models, video generation models at some point? Model agnostic. Okay. So we find right now that for example, images two point o from Tatchi can see is the best at skin. Yeah. We use other models like Gemini as well. Yeah. And so as models get better, we want to make sure that our harness chooses the best model for each task. And then there's also a world where you generate an image or a series of images with one model and then you animate them with a different model and there's a whole bunch of different or do text overlay or motion graphics with a different model or maybe, you know, puppeteer some actual code or or or like an Adobe's tool. Interesting. Jordy?

Speaker 2: Yeah. How how is the role of a how is the role generally of a of a media buyer changing? Obviously, hopefully, they're using your tool and that's changing their workflow. But it's interesting that I it feels like media buying has always been, like, very high leverage. Like, somebody that's good at at Facebook ads could manage 20,000,000 of budget on one account 40,000,000 on Like another they're very very very it's just a very high leverage role. Is it is it just helping them be be better? Do you think marketers in the future will be able to manage, you know, more spend or more accounts? Like, what what are what are the different ways that it goes?

Speaker 9: I think there's two main shifts. So as you mentioned, Jordy, there's about imagine AI is a twenty four seven analyst that's helping you figure out your opportunities for improving your media spend, which channels do you spend more, is TikTok working better than Snapchat? Those kinds of decisions now you can fully delegate to AI, and our platform enables that. But the second and much more important trend, in my opinion, is that if you're a media buyer, you can only buy media against content that you have. Most brands don't have automated content. They only have maybe like 50 to a 100 pieces of really good content. Yep. With AI, you can now generate unlimited amounts of content. So for example, localizing to different languages, doing different geographies. A great example of this is one of the top marketplaces for home services uses HITUCH, and now they can do different marketing for owners versus homeowners versus renters. Yeah. And because they've kind of now they have unlimited content building capability at their disposal, that same media buyer can test all three against the market, see which works which one works against which demographic, and then optimize it.

Speaker 2: Do you guys care about finding and and eliminating like wasted spend? That feels like a solution that there's kind of an interesting thing where, you know, let's say like let's take meta platforms for example. They want you to get good results in general so that you spend more money so that your business grows, so you can spend, you know, more money etcetera etcetera and have that flywheel. But at the same time, if you're running an ad and it's not performing well, like, on a on a sort of like local level or micro level, they don't really care that much if you're spending in a in a way that's like not super efficient. And so maybe Mhmm. There's an opportunity to have like a platform that kind of sits above that and says like, hey, this is like to basically like turn off and like try to identify wasted spend. Is that part of what you guys are doing? Is it something you would do in in the future?

Speaker 9: It's part of what we're doing and why we think we can do it uniquely is because all of the consumer brands we work with, companies like DoorDash, Aritzia, and so on, have given us access to their customer data. So we know which customers are actually converting, and we know the ground truth of whether the ad is working or the media campaign is working.

Speaker 1: That makes sense. How do you think about your business model? I mean, there's so many huge companies in the marketing industry because if they're able to insert themselves in the pipeline of a billion dollar marketing budget, it's very easy to justify spending millions of dollars on something that increases performance by one or 2%. At the same time, there's a lot of marketing tools out there that are just seat based and they might give you a call and say, hey, you're pretty big. Can we get some more money out of you, sell you some premium features? But this this seat based pricing versus consumption based, performance based pricing, like how have you geared the company for success? Obviously, growth is fantastic and the company is very, you know, like, the valuation is going up and you're raising money. So I imagine that you've solved it, but how have you thought about solving

Speaker 9: Yep. Any transparently, company that's still on seat based pricing is not looking forward to the rest of 2026. That's correct. We have never been on seat based pricing. We've always charged our customers by the number of key things that we power. Yeah. And our belief is that with AI, our customers will be able to run significantly more campaigns. Mhmm. And we really leave it up to the customer. If you keep the campaign on, it must be working, and therefore you should pay us. If you turn the campaign off, then it's alright. So it's simple. Yeah. That makes a ton of sense. Well, how much did you raise? I wanna talk about the did we already hit the gong? I think we already did. Did we already hit the gong? Yeah. It's been a big day.

Speaker 1: We've we've rung the gong a bunch of The gong is Everyone is gong abuse today. Amazing news. Well, congratulations

Speaker 2: Yeah. Very, very the progress. On the last question because I've asked some other people in in MarTech this. Are we headed to a future where every ad that a customer sees is generated in real time on the fly. And and the the example I always give is like, historically, advertisers would make one campaign, they'd put it in a magazine or they'd put it Mhmm. You know, out of home or the Super Bowl and they just knew they were gonna they had to make it and a ton of people were gonna see it. And then you get, the the era of like really really precise, targeting and and the explosion of of meta platforms and now you can be like, I'm just gonna serve this ad to like 10,000 people. And then, it feels like we're headed to a world where the creative is generated on the fly. It's really just, like, meant for one individual person, and that's only gonna be possible because you can just generate the creative, like, in real time on the fly based around, you know, that person's unique interests or their purchase history or where they are, etcetera?

Speaker 9: I I think it's an exciting vision. Will most content be AI generated in the future? I believe so. Yes. Will it be one to one or on the fly time? I think no. And the reason for that is because most consumer brands of any scale have to go through certain approval processes. And the amount of trust people have in AI still is not on a one to one level. On the other hand, could it be You don't think AI can get superhuman?

Speaker 2: You don't think it can be Well, better than the so so but then at what point is it better than the the best creative director and the best person that the best CMO that actually approves the creative to go out. Right? Like, could have multiple agents reviewing each other's work.

Speaker 9: I think we we share a vision. So I hope that we get there. Yeah. And I think AGI will come before marketers do one to one with AIs. Yeah. But I do think that there is an appetite for this and that the more practical way to make AI actually used in marketing. So we talked to pretty much every CMO there was all of last year. Very few CMOs today are using AI market. You ask them why, and it is because of this kind of random Google process. So we believe the way to get AI to actually work is to do one of many, maybe one piece of content for every 10,000 or a 100,000 consumers.

Speaker 2: Yeah. And then over time, create more pieces of content. Mhmm. Very cool. Alright. Well, great. Great having you on. Congrats on on the progress. Around, and we'll talk to you soon. Thank you. Have a good one. Bye. Thanks.

Speaker 1: I wanna pull up this post from David Bazooki, friend of the show, CEO of Roblox. He says we're on the verge of real time photorealistic generation with what are called world models. These require a fair amount of expensive compute, but costs will come down over time. Today these interactions are essentially real time dreams. They lack persistent shared logic that turns a video into a multiplayer game or concert or classroom or the holodeck. We believe the ultimate architecture for gaming in the holodeck is Roblox reality. It's a hybrid architecture. We've been talking about this for months. I'm so excited it's here. It's a hybrid engine that marries the structured data and logic of the Roblox engine and Roblox cloud with the generative power of video world models. The Roblox engine provides the underlying synchronized ground truth, the score, physics, multiplayer sync, etcetera, while our video model acts as a super upsampler to layer on photorealistic detail. We believe this will ultimately remove barriers to high fidelity creation. And if you look at these videos it makes so much sense. Roblox has always been blocky but style transfer is the most mature generative AI technology. We all saw this during the Studio Ghibli moment. Take a cartoon, make it photo real. Obviously image generation is at an incredible place and so being able to do that at 60 frames a second with Roblox is just incredibly bullish for the amount of creativity that can go into this world. I am extremely excited about this because Roblox has always felt like a kid's game. It's always felt like it didn't deliver at the triple a level for so many of those experiences. But now you're going to get the creativity of the Roblox ecosystem married with the triple a graphics that everyone demands.

Speaker 2: I've always I've always felt that Roblox was super well positioned as it becomes easier and easier and easier to make games. Yeah. And it's because they have the infrastructure, they have the they they honestly, you know, we've in talking with David, he's talked about just how intense it is to keep