Great. Great to be here with you, as always. Great to be here.
Yeah. We decided we would interview each other today, and I'm going to start by asking you the first question. So, Palantir just reported while we were in the back room getting ready, the CFO of Palantir called you. What were you guys — I know without violating any privacy issues — what were you guys talking about?
Look, I mean, to me, they had a robust quarter. I mean, if you look at their growth, it's just phenomenal. I think it's a good example of, you know, what's the validation for this AI revolution, right? All the CapEx spending. Palantir, I think, front and center. And I think...
They grew their US business 100%.
Exactly. Look, and I think what Palantir is doing speaks to — and obviously, it's a big focus of everything that you're doing when it comes to Athena.
Look, I think this is just the beginning of the AI revolution. Like when you think about the software, the use cases, and I think what they're seeing is a lot of the stuff that, you know, you're seeing with Athena, right? I mean, it's, you know, I guess I'll go to you, okay? Like, you know, we talk about Palantir, and I think also a lot of the stuff that I see with Palantir, it reminds me, if I go back a few years ago, right? Where like no one gave them the respect in terms of Palantir, and look what they've done. When it comes to like Athena, it really feels like Athena is changing data.
She is. Yeah, it's funny, because if you look at our quarter, we grew the business by 50% top line year over year at scale. I mean, that's, you know, then you look at Palantir at 100, you know, you feel a little like you're sleeping there, but at 50% growth, I was super, super proud of what we did. I think there was some confusion in the quarter, and you know, listen, we've projected $397 million in EBITDA for this year, and you know, we very much expect to be meaningfully net income positive this year, which I think is causing confusion.
It's good you hit on that, because right there is... Well, a lot of people are talking about that. And the truth of the matter is that you're projecting 397 million in EBITDA. We expect to be, you know, net income positive. And the restructuring from the Marigold acquisition, which by the way, we are just killing with that acquisition. I mean, if you look at the business, we grew 29% organically in the quarter and 50% total in the quarter. That's because Marigold was so far ahead of where we expected it to be. And you know, quite frankly, our organic business is growing because of Athena.
In the first quarter, we saw our agentic volumes go up by 700%. 60% of which was Athena.
But also, I remember like when we sat down at CES in January, and now you think you're at Milken — I can't believe you're doing it. The outfits just get better though. I know. No, you could wear the same outfit. I don't think I could pull that off. But if I go back to CES in January and here we are at Milken in May, I mean, to some extent like the strategy is really playing out exactly, and even I think almost ahead of plan.
Yeah, Athena's way ahead of where we expected it to be. 100% of our enterprise clients have Athena now in their hands and everybody's using it. And quite frankly, if you look at Zeta's return on investment, according to Forrester, for every dollar that's spent on our platform, we return 600% return on investment. So, $1 equals $6 in business. Our clients using Athena are seeing a substantially higher return on investment than that. And that speaks like, look, just like Palantir...
Wait, we're supposed to be interviewing each other. But to some extent, what I was saying is like just like Palantir, what we've seen there, it's really my view that you're actually starting to see not just validation, the actual use cases play out. I mean, that's really what you guys are seeing in terms of when it comes to Athena.
Yeah. The use cases have been really surprising how many there are. And when you look at a 700% step up in agentic utilization in the quarter, and 60% of that being utilized by the platform that you put live in that quarter, right? And you know, we're going to be rolling out V2, V3, V4 of Athena, you know, almost every month or every other month over the next year. Let's talk for a minute about SaaSageddon. Yep. Right? So, you're getting asked this question. You're on CNBC, you know, if not every day, almost every day. What do you think? How does a company like Palantir or us or others, how do we react to something that's so nonsensical?
Look, I think part of the — and I'll almost call it, we call it the AI ghost train in terms of Anthropic, basically that they're going to, you know, structurally break every piece of software and they're going to ruin... I think you have to separate out: are there going to be areas of software disintermediated relative to Anthropic? But of course, for sure. And I think those companies could clearly be exposed, and those are the ones either cause M&A or structure those companies are on the loser category.
But I think for software companies, and it's not just you, it's not just Palantir, it's a Salesforce, it's a ServiceNow, it's everyone. It's trying to figure out, okay, when does the monetization phase start? See, my whole point is if you think about data centers taking two or three years to build and GPUs and the infrastructure, we're only now just starting to get into the software phase.
Yeah, because all of these tech cycles start with infrastructure. Exactly. Right? And then the application layer that sits on top of it comes next, aka software. And that's why I think like what we see with Palantir, what you guys are seeing with Athena, this is actually just the start — Snowflake seeing and others. So my view is that they painted software all with the same brush, and you see it structurally if any sort of thing happens, it takes the stocks down significantly. The reality is that the software use case phases are just starting. I'm sure you see that when you talk to customers. Like when you talk to customers today, how much different is it today than six months ago?
Oh, it's — I mean, and I remember we started working in AI in 2017. So for years and years nobody even cared. Like they would be like, 'What are you talking about?' Now we're getting pulled into use cases. We're, you know, I think Athena is going to solve far more problems for our clients than people expect that she will.
Totally. And then look, I think so to your question, I would just say I think investors — to me, covering tech stocks in the late '90s, it's the most incorrect narrative that I've actually ever seen in my career relative to software and what's happening. It will take time to prove itself out, but the reality is AI-native stocks are never going to unseat core install bases. If you look at the data, if you look at ultimate lab use cases where they're going to play on, LLMs will partner with software companies, and over time LLMs will get commoditized, and you're going to have hundreds or thousands around the world relative to models as they ultimately deploy. Software will be the hearts and lungs of AI.
Well, and data, right? I mean, if you look at it, proprietary data, I think it's going to be the biggest differentiator as you think about where these models are going, right? Because models are only as smart as the data you feed into them. Exactly. And if you look at — and I'll talk about data again. We have, if you think of the moat around our business, we have 555 million people who have opted into our data cloud that we have never fed that data into a large language model ever, nor would we. It simply trains our models, and then we take our clients' first-party data, which includes 51% of the Fortune 100, we're ingesting all of their data into a consumer data platform, we're matching the data together, and it's training the models just for that client.
One of the really interesting things that I think people missed about our company was last year, our net retention rate as a company was 120%. And most software companies are happy at 100, we were at 120. That's because once we ingest our clients' data and we map it with our clients' data, it's almost impossible to fire us because you keep getting better and better return on investment for them, therefore they keep growing with you.
But that really goes to the broader theme here. And I think — well, you saw it obviously in this quarter in terms of those numbers, but I think more and more we are now in the software phase. We are, and I think that's what companies are starting to see like, how do I deploy it? I've spent tens of millions or more on infrastructure. Where are the use cases? Use cases aren't coming from Nvidia, it's not coming from a lot of the infrastructure, it's coming from software. When do you think the street gets that? Or how do we convince the street of that? Like I just — that to me is confusing.
But I'd say, David, I think like the issue that like from a communication standpoint, the street and the SaaS apocalypse, or the AI ghost train — it's not... These are conversations that I see with so many of the best software companies in the world. I think it's going to start — look, it starts with the results. Just like you guys are showing in terms of the actual monetization. And then I think it really starts to — as you start to see these stacks and the infrastructure get built out, I think investors are going to realize it's the most disconnected narrative in the last 25 years relative to the end profits and others, the view that they're going to unseat software companies. And I think it's really — our view is the second half of '26 is where it really starts to play. Because that's where the use cases start to play. I mean, we've seen it, you know, you showed it, you saw it on Palantir, but it's really as we get into the second half of the year, that's where that starts to run. And that's our view like software, and we think software stocks are, you know, potentially they're up double what tech is in the second half of the year.
Interesting. You know, it's funny, at Zeta, this was our fifth consecutive quarter of accelerating organic growth. And it's all coming from Athena now. And you know, it's one of the things I find interesting because we are a native AI company. We re-architected our entire platform to put AI as native and data as native, but foundational to the application layer. It's been so hard to explain the AI. When we go into a client now and talk about Athena, it's almost like we've been driving this Ferrari. The AI's in the engine, but you can't see it, and now Athena is the big hood ornament sitting right on top. We're going in with a full Athena pitch to all of our clients. That's really what we're doing.
It almost — Athena's really changed the company in terms of like for sure the go-to-market strategy. And at a time where customers across verticals are looking at how they're going to monetize AI, what else are you seeing in the markets outside of this stuff right now? I mean, I think like right now, and I just got back from Asia for a few weeks. So like I could tell you...
I don't know how you do it. Like I called you last week just to say hi and you're like, 'I'm going to go to Korea.' I think part of it is like the only way you could understand what the demand looks like, you have to see it globally. So I think it's very easy while the bears, they're negative on AI in the 35th floor in a New York City office building. So you don't see AI when you're in Metro North. And I think the reality is that the demand that we see, it's 12 to 1 demand and supply. Wow. And obviously you're going to get to be at Milken, he'll be talking about the massive AI...
in about 20 minutes. But the reality is like the chip demand is unlike anything we've seen. It's a super cycle in terms of memory and overall chips. But guess what? In terms of that CapEx build out, it's all getting built out for what the use cases are going to be. And I think more and more you're seeing — I'm seeing from CIOs in so many companies, they're now going into that next phase. And I think there's an arms race, right? Like which are the stack players that are going to benefit? Like who are some of the niche software companies where some of them might get acquired? Some of them might fall apart. And I think it's those that stay on the treadmill will get passed by.
Well, that's — I mean if you look at guys like Palantir, right? Didn't they go through something like this a few years ago? So look, Palantir is a good example, like and obviously you know being such a supporter of them over the years, you know so many times...
Wait, watching you talk to like the guys who run that company while we're in the other room. These are your boys, right? Like you get along really well with them. Well, part of it is like I'm a huge believer in what Karp and Palantir have built. But it's very easy to look at them today, just like it's very easy to look at Nvidia today. You don't see a few years ago. You don't see where like they were building it all on government. They were building the infrastructure. No one believed in them. No one gave them a multiple, and look where it is today. No, so I'm saying like that's why like I said, like it's a lot of the stuff that you're going through that I think other software companies that are thriving are going to — it's very similar to what Palantir a few years ago where it took time for them to execute, show the vision, and then investors ultimately rewarded them. And that's why I just think right now we are going through this massive prove-me phase for the rest of software. Because I think right now it's like, okay, there's Palantir, and they're maybe put up here. And everyone's trying to figure out who else is going to join them. Is it Snowflake? Is it MongoDB? You know, is it Zeta? So everyone's trying to figure out who those software — where does Oracle fit? Where does ServiceNow? Where does Salesforce? And I think that's why right now we're in that phase.
Interesting. So, we're sort of in the — we've got infrastructure and software in the same sort of concentric circle growing together right now.
Well, because like right now like every investor, they're focused on like hardware, chips. I don't care if it's easy or year up here. You own memory stocks. You own chips. Software is still a do-not-enter zone. I think that's the wrong trade, but it's really because like you're not getting paid to own these stocks right now because the more... And I think that creates the massive opportunities.
Right. If you look at the reactions like ServiceNow and others, but if you look at how it plays out, okay, as companies go from like the prove-it phase or maybe the skeptic phase from an investor to here, that's the huge multiple re-rating. And that's what's going to separate the companies where we'll be here a year from now. And there will be stocks that you look back on and be like, what? How did no one think? Because it was the prove-it phase. And I think that's — we're going through that now, not in chips. In software.
Right. Because chips have already seen that. We've already seen that. And right now, but go back — I could go back in terms of like these periods of AI revolution. Go back a year ago. Or go back like two years ago. There was a video — it can only be a video. AMD can't be successful. Memory stocks...
Intel can't win. Intel can't win. Micron can't benefit. SK Hynix, no, it's the memory stocks. They're... What about SanDisk? They're not — no one's going to pay a multiple like that. So, the point is like now it's consensus and that's where the crowd is. Software, it's still... No, it's the — I believe as an investor that so many of them are structurally troubled. But I disagree with that because ultimately the hearts and lungs of these use cases will be software. We're going through from May to let's say like December is the prove-it phase for software companies.
Well, it's interesting because if you look at most of the infrastructure and I mean, listen, they've built some insane companies. Like these are great, great, great companies. But they're effectively selling the infrastructure once. Yes. When we sell a client, it's a subscription for a very long time. And if you look at those use cases, you know, you would think software companies that win it... I want to go back to your original premise, right? Because I do agree they're going to be what I call workflow management companies that have built platforms that if they don't pivot, they're going to have some real problems. Sure. Because I think agents are going to disintermediate them. As well as seat-based models.
Yeah, of course. Well, the seat-based model people were beating the living crap out of us for charging a subscription fee and a utilization fee five years ago when we went public. And somebody was asking me the other day, well, it seems like you've done really well on the transition to subscription. I'm like, we didn't transition. Like we're doing — we've been — we're 60%...
That was the premise. It was a hurdle before everything. That was the whole thing. And they beat the living out of us at the IPO for that. Like nobody wants a utilization model. You got to have a seat-based model. You got to have a CPU model. And we just didn't think that model scaled.
But it looks like now everybody's trying to move to sort of your subscription fee plus your utilization fee. Because of the worries, too. And again, this also goes back to a lot of AI alarmist, job growth — jobs going to — everyone's going to lose their job. And I think a lot of that is more of a dystopian type narrative. I think a lot of that is going to be — I'm not saying there's not going to be areas around that, but the view — and I think you're starting to see changes now even in the job market and some of the ones, you know, that we've seen. I'm just saying I think that's also playing into this narrative about seats going down, job growth, pricing power's going...
Well, if you don't need the employees, why do you need seats? No, but that's... Right. But then to you, like the question I would say for a lot of software companies like, are you losing pricing power? Are you? And like what would you say to that? Like you know, there was like because you know, there...
Well, we grew 50% last quarter. And if you, you know, looked at the business, we're adding operating margin this year and getting more profitable this year. The interesting thing when you think about sort of how this all plays out, right? This morning I did a — I was invited to a panel at the Milken Global Conference with some of the largest CEOs in the world. And I then ended up having Coke or diet Coke with the... Wait, because you only drink diet... I only drink diet Coke. I only drink — I love — I have — I love diet Coke. My wife doesn't love me drinking it. I probably shouldn't be announcing this on a live stream, but I had a diet Coke with a senator, a US senator today. And what I can tell you is whether it's a global CEO of which there were I'd say 15 to 20 Fortune 500 CEOs in the room, or the senator, the job repercussions of AI are the third rail discussion. Nobody wants to have it. But unless we start having it, we're not going to figure it out, right? And it's something like when you look at sort of AI, I look at Athena. Last quarter, not only did we increase our agentic utilization by 700%, but clients who adopted Athena used 78% less labor to manage the platform. Now, what I say lovingly is and they can use those employees to be repurposed into other functions in their business. Productivity, yeah. Right? And I think that's going to be a big part of it, but I think there's another part of it that we're going to have to think about where do some of these jobs go? But then I would say and I would tell you like my whole career coming from Asia to here and so many times I'd be in Taiwan and you see, you know, you see a fab 18 hours a day. You know, the amount of production and I land in New York Airport there's a fist fight at Dunkin' Donuts. And then you're like there's a reason...
Why are you in Dunkin' Donuts? At New York Airport. Okay. And then you're like there's a reason we're 17th in math. At that point for so much of my career, now for the first time US is ahead of China when it comes to tech. So, I can make the other argument, the amount of data centers going to get built, the ripple effect, engineer, private companies, innovation. So that's why like I do think the view of AI, you have to address it, yeah, but I do think going back to like even weaving into like software and private credit and seats and pricing model, I do think we're going to look back and be like the bark was a lot worse than the bite.
100%, 100%, 100%. And by the way, we're at, you know, 4 and a half percent unemployment for the third or fourth year in a row. Three years ago, and this is I think I'll show how incredibly old I am. I think this is my 25th or 26th Milken Global Conference. Three years ago, the entire conference all...
When you started to go to Milken when you were 14 years old? I loved you. That's why I love you. So, 14 years ago was your first Milken? That's a me... Well, in all seriousness, three years ago, all anybody could talk about was aging global population, and we're never going to have enough people for all the jobs we're creating. Now, all we're talking about is AI is going to disintermediate. As you said, pendulums swing hard, but they're almost always wrong. They overcompensate, and I think we are going to find great, great jobs for people. But when you look at the difference between the winners and the losers, I think productivity accelerants are going to be major. Tomorrow, I'm moderating the lunch in the main ballroom, which is... Yeah, it's huge. It was nice of Mike to ask me to do that. They obviously they probably asked 10 people and you know, nine said no, and then they ended up on me. But it's all about is this going to be the step function in the productivity curve, right? Because we've been stalled for a few years. We had this massive productivity that started with the computer. Then we had this massive productivity that started with the internet. And for the last 5 to 10 years, we've been very stalled in productivity. The question is, is AI the next step function? Or is it just going to keep us sort of growing at 2 or 3%? And I don't think anybody's going to know the answer to that for a few years.
But the use cases will be the key. And now we enter the software phase, where obviously that's a huge part of you know, where Athena and Zeta Plus... Well, the one good thing about our stock right now is it's buy one get one free. And you know, we bought back I think 25 or 30 million dollars worth of our stock last quarter. We're going to continue to aggressively buy back our stock. And you know, as long as it continues to be, you know, buy one get one free, we'll keep doing that.
Also part of it, too, is that you can just execute on the plan, show the numbers, build it out, and everything else, you know, takes care of itself. Damn, what else do you want to talk about tonight? I mean, I would just say like who... when we talk about like SaaS apocalypse and just everything that we're seeing from... Can you just like maybe like without obvious... I asked you the question. Now you're asking me a question. But without naming customers, can you just explain how the conversations specifically have changed? Like if you just go back to...
So, back to what we were talking about. Athena has changed the narrative with our customers in ways that I didn't even think possible. So, I think a lot of our customers just looked at us as a marketing platform. And we did a good job, and we had a good return on investment. And when Athena was first sort of introduced at Zeta Live and then really we got behind it at CES and then she was born, you know, in the first quarter we did that big birth announcement... birth announcement... You know, when she was born, clients started looking at us more like an innovation engine and less like, oh, they can just help us with return on investment for marketing. Yeah. And it's opened up entirely new doors and new use cases. One of the things that I thought was the coolest statistic we reported in the first quarter was that clients who use multiple use cases on our platform, which is the single biggest driver of revenue unlocked, went up 50% year over year. Clients who use us for, you know, customer acquisition, retention, monetization went from doing one to two or more went up 50% and that is a flywheel that's going to continue for many, many years.
I mean, I would just say to that point from my perspective, I think the main thing that software companies right now are struggling with is like a lot of acronyms, a lot of great marketing stuff. But we call it roadmap AI. It's like we're seeing a lot of roadmap AI nowadays. But customers don't want roadmaps. No, so that's the — that would be the one thing that I would say is that I'm seeing a lot of companies that I meet with whether it's private or public, it's like they have a lot of — everybody, but they have a lot of roadmaps. But right now customers want action.
Look, I think it goes back to like part of why I think Palantir has been so successful is like if you work with them as a customer and you do their own co-... they could show results. And I think but like to some extent like you guys are almost like there is a lot of similar DNA in terms of like this is not top like in the you're seeing...
We've because of you we're spending more time with them and getting to know them better and it's been really interesting. But what I would say is the thing that the Palantir guys do just so incredibly well is they solve problems and they execute. And that's sort of key to success in almost any business. Solve your clients' problems and execute. And Palantir does that. I mean, they do that as well as anybody. And it's — we feel like we do it in all the world as well. But it is crazy to see sort of — we'll go — it's funny. If we're on a pitch, we might go into a room where seven other companies have said they do the exact same thing we do. And then you actually...
Like they're just making it up. And then we come in and they're like, 'Oh, we...' I said I can say this to a client. Great. Test us. If I don't cut your cost to create customers by at least 50%, I will give you your money back. Even on a test.
Because there's so many things like that's where like where you're going to thrive and obviously have a massive install, but a lot of it comes down to that, right? It's like proof of concept. It's proof of concept. And if you think about it, that's one of the reasons our net retention rate's so high. Because if you land with a client and then you massively expand with the client, your net retention rate's going to continue to go up. We've given long-term guidance of 110 to 115% net retention rate. We were at 120 last year and we were well ahead of that, the top of the range in the first quarter of 2026, which is traditionally our lowest net retention rate quarter of any year for a whole host of reasons, and we were still well above the 115. That's because we come in and say, 'We'll prove ourselves.' And when you do that, clients move more budget, which then naturally drives your net retention rate. But I do think that it's crazy how much roadmap AI is out there versus the companies that are really delivering and really executing.
Well, that's the — I think that's the separation. And you're seeing like from cybersecurity to software to that is like a huge...
Are good cybersecurity companies at this point?
To me it's like when you look like CrowdStrike and Palo Alto, they're almost playing a different game than everyone else in cybersecurity because of — I could argue like Palo Alto in terms of like CyberArk acquisition, what George has done at CrowdStrike in terms of building out the AI suite and what they've been able to be successful because right now when you think about cybersecurity, that's going to be a huge piece of every...
Oh, for sure. You know? For sure. So I just think like what we're seeing in cyber is very similar to what we're seeing across other areas of software. And looking like physical AI is just starting, right? Like when you think about like autonomous and robotics and what we see with physical like that phase right now is just starting to play...
I think world AI is going to be very big. And I think you're seeing a lot of that come down the pike.
And then when you think world AI, it's like US, you still have less than 5% of companies gone down the AI path. Asia ex-China less than 2%. Middle East has a huge build out in regard obviously, you know, conflict now, but if you look at the data centers that are getting built in Middle East...
Yeah. Europe continues to build...
And once the conflict is done, that'll come right back.
Data centers outside the US are happening in the Middle East, which just shows... Well, they're the closest to the energy, right?
Well, and also and who's really front and center of all of those deals? US tech. Right. So I just think and then also what we're seeing on the government side is just it's — I think this next phase right now where you're going to see just massive government spending on AI. It's a huge part of the administration that we've seen in the beltway. So, you would sum it up by saying you think the back half of the year, how does that look? Oh, I think that — I mean we've said it, we think tech stocks are up, you know, another 12 to 15% second half of the year, but I think stars of the show will be the software companies that could produce, show the use cases. We're in a skeptical phase now, but I think that this is now — we're now going into the software phase of AI.
Well, and I would tell you for Zeta in particular, with in addition to everything else we announced coming out of the first quarter was the highest pipeline we've ever had. Our pipeline was up 40%. So I see it in the adoption and I see it in the use case.
And the timing. So, look guys, first of all, thanks everyone for coming and as always...
