Analysts argue Zeta Global is fully valued at 30x earnings, citing lack of proven AI monetization and high growth requirements.
Summary
Speakers evaluate Zeta Global as a marketing technology company with a unified data cloud and AI agent, Athena. They acknowledge Zeta's strong data moat and partnerships with Palantir and Snowflake but remain skeptical of its current valuation. The hosts argue that Zeta trades at a premium 30x earnings multiple that requires 25% organic growth to justify. They contrast Zeta's unproven AI revenue streams with competitors like AppLovin and Meta, suggesting the stock prices in too much future narrative.
Episode sections
- 0:00 – 2:03
Intro and Company Overview
Shay introduces Zeta Global as a fast-growing marketing technology company helping brands understand customers and buy ads, while noting its potential as an agentic AI winner due to its identity data. Both hosts mention having 'Trade Desk scars' and aim to evaluate Zeta objectively as its own company.
- 2:03 – 7:27
Zeta Data Cloud and Platform Pitch
Brad describes Zeta as an 'AI marketing cloud' offering end-to-end marketing software with a rich dataset and network effect. He highlights their unified data cloud and identity graph of 245 million American adults, which allows for better customer understanding and outcome prediction across multiple channels.
- 7:27 – 9:53
AI Risks and Vendor Consolidation
Brad argues that customers cannot 'vibe code' a solution to replace Zeta because Zeta possesses superior third-party signals and data. He notes that while OpenAI is a partner, it is not a long-term moat, citing Palantir's skepticism of OpenAI. Shay agrees, emphasizing that Zeta's moat lies in its closed-loop system of prediction, execution, and measurement.
- 9:53 – 11:05
Acquisitions and 'One Zeta' Strategy
Shay discusses Zeta's acquisitions, including LiveIntent to strengthen identity and post-publisher capabilities, and Marigold to expand into loyalty and retention. He introduces the concept of 'One Zeta,' where customers adopt multiple products (email, identity, paid media, Athena), increasing stickiness and revenue expansion.
- 11:05 – 13:19
Athena AI Agent and Partnerships
Brad introduces Athena, Zeta's platform-wide AI agent, which allows customers to build and optimize campaigns conversationally. He highlights partnerships with Palantir and Snowflake as foundational data partners, viewing them as more significant moats than the OpenAI partnership due to their channel partner roles.
- 13:19 – 16:16
Athena Adoption and Economics
Shay notes that over half of super-scaled customers are monthly active Athena users, and adopters account for over 70% of company revenue. He expresses caution about selection bias, wanting to see if Athena drives retention and spend among the broader, average customer base, similar to Palantir's scaling trends.
- 16:16 – 19:43
AI Costs and Specialized Models
Brad explains that AI input costs are falling as companies shift from expensive frontier models to cheaper open-source or small models. Shay adds that Zeta is building specialized models using Fireworks and NVIDIA's Nemotron, claiming 95% accuracy at 40% lower cost per request than frontier models, optimizing for specific tasks like churn prediction.
- 19:43 – 30:59
Hyperscalers vs. Independent Layer
The hosts discuss the risk of AI shifting ad spend away from traditional search. Brad argues Zeta is agnostic to channel shifts because it plays everywhere. Shay warns that while hyperscalers like Google and Meta will remain dominant, Zeta's value lies in being an independent layer for customer intent across platforms, though he questions if it deserves a premium over Meta.
- 30:59 – 32:55
Valuation and Competitive Positioning
Speaker_0 argues Zeta competes with Adobe, Oracle, Klaviyo, The Trade Desk, Google, and Meta, noting that while Zeta offers a unified architecture, competitors have massive distribution. Speaker_0 contends Zeta is fully valued at 30x earnings and should not trade above that multiple, rejecting the idea that it can command a premium based on AI optionality alone.
- 32:55 – 34:08
Athena OS and Partnership Optionality
Speaker_0 discusses Zeta's Athena OS and partnerships with Snowflake and Palantir, stating that while the optionality is real, it does not justify a higher multiple until it generates measurable revenue. Speaker_0 contrasts Zeta's need for AI monetization proof with Meta's existing AI-driven advertising improvements.
- 34:08 – 36:00
Palantir Partnership and TAM Expansion
Speaker_0 highlights the Palantir partnership as Zeta's most interesting long-term optionality, combining Zeta's consumer behavior data with Palantir's enterprise operational data. This integration could expand Zeta's total addressable market beyond ad tech into supply chain and inventory decisions, though Speaker_0 remains cautious due to limited production revenue.
- 36:00 – 37:29
Model Economics and Cost Substitution
Speaker_1 appreciates that Zeta frames its Palantir and Snowflake partnerships as technology collaborations rather than competitive differentiators. Speaker_1 suggests Zeta's goal is to eventually replace expensive models like OpenAI with cheaper alternatives to improve margins, viewing this as a more exciting long-term driver than current model usage.
- 37:29 – 39:53
Bull/Bear Stance and Growth Requirements
Speaker_0 identifies as bearish, requiring 25% organic growth for the next few years to justify the valuation. Speaker_0 lists conditions for a bullish view, including better retention for Athena Dina users, measurable recurring revenue from partnerships, and improving free cash flow per share.
- 39:53 – 45:55
Peer Comparison and Portfolio Fit
Speaker_1 views Zeta as a portfolio name for diversified funds but not a top conviction pick, preferring to see more proof of AI and non-ad tech expansion before buying. Speaker_1 compares Zeta's 30x multiple unfavorably to AppLovin's 20x multiple, noting that Zeta has not yet faced a significant 'oh shit' moment like AppLovin did with a delayed model launch.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| OpenAI | private company | neutral | So I I think I would like Zeta is building a better AI model uh than OpenAI. I wouldn't say that. |
| AlphabetGOOGL | equity | neutral | I think it's really important if you're a Zeta Bull to not think that like most dollars are going to leave Google and suddenly land at Zeta. |
| AmazonAMZN | equity | neutral | obviously, especially with Amazon. |
| Meta PlatformsMETA | equity | mixed | Um what are your is this where your like trade desk goggles come in with Zeta? Because I always think that hyperscalers like um why bother? Like Meta, for example, like we'll talk to this more detail, but like Meta, we just went over it. |
| CrowdStrike Holdings, Inc.CRWD | equity | neutral | And they very intentionally put valuation comps to CrowdStrike and all these other enterprise software |
| Snowflake Inc.SNOW | equity | neutral | And again, that's why to me, um, this Palantir partnership and the and the and Snowflake P partnership that helps them take advantage of this data more intelligently, that gives them access to new relationships to package all this value together and find more demand is more exciting than they're using OpenAI models in their in their platform. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Zeta is a fast-growing marketing technology company helping large brands understand customers, buy ads, and manage relationships.
Unidentified speaker (UNKNOWN) ·0:09
How to verify: Company website and investor presentations; SEC filings for business description. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta reported revenue up 46% to $493 million on November 9.
Unidentified speaker (speaker_0) ·0:29
How to verify: Zeta Global Q3 2026 Earnings Release and 8-K filing on SEC EDGAR. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta's stock price moved from about $15 or $16 a share to about $35.
Unidentified speaker (speaker_0) ·0:49
How to verify: Historical stock price data for ZETA (or ZG if split-adjusted) from financial data providers like Bloomberg or Yahoo Finance. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Zeta calls itself the 'AI marketing cloud' and offers end-to-end marketing software.
Unidentified speaker (speaker_1) ·2:03
How to verify: Zeta Global marketing materials and website. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Zeta's identity graph maps 245 million American adults.
Unidentified speaker (speaker_2) ·3:03
How to verify: Zeta Global investor presentations or earnings call transcripts. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Sell-side analysts expect growth to slow from 27% in 2026 to 20% in 2027 and teens in 2028.
Unidentified speaker (speaker_1) ·4:53
How to verify: Consensus analyst estimates from financial data providers like Bloomberg or FactSet. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Zeta has an OpenAI partnership.
Unidentified speaker (speaker_1) ·7:49
How to verify: Zeta Global press releases or SEC filings regarding partnerships. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta has partnerships with Palantir and Snowflake.
Unidentified speaker (speaker_1) ·11:59
How to verify: Zeta Global press releases or SEC filings regarding partnerships. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceMore than half of super-scaled customers are monthly active Athena users.
Unidentified speaker (speaker_0) ·14:39
How to verify: Zeta Global Q3 2026 Earnings Release. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Athena adopters account for more than 70% of company revenue.
Unidentified speaker (speaker_0) ·14:39
How to verify: Zeta Global Q3 2026 Earnings Release. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Zeta competes with Adobe, Oracle, Klaviyo, The Trade Desk, Google, and Meta in the enterprise distribution and real-time engagement space.
Unidentified speaker (speaker_0) ·30:59
How to verify: Zeta Global's investor presentations or S-1/F-1 filings listing competitors; check company websites for product descriptions. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta trades at a 30 times earnings multiple.
Unidentified speaker (speaker_0) ·31:19
How to verify: Calculate Zeta Global's (ZETA) current share price divided by trailing twelve-month earnings per share (TTM EPS) using SEC filings or financial data providers. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta's Athena OS has eight affirmative customer responses, two operating customers, and one expected to become a 'super speed' customer.
Unidentified speaker (UNKNOWN) ·35:30
How to verify: Zeta Global's Q3 2026 earnings call transcript or press release (filed with SEC EDGAR) detailing Athena OS adoption metrics. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourcePalantir could eventually contribute more than $100 million annually to Zeta.
Unidentified speaker (speaker_0) ·33:15
How to verify: This is a speaker estimate/claim; verify by checking Zeta's investor presentations for partnership revenue guidance or Palantir's earnings calls for Zeta-specific revenue attribution. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
AppLovin is growing top line 50% plus with accelerating sequential growth.
Unidentified speaker (speaker_0) ·41:16
How to verify: AppLovin Corporation (APP) Q3 2026 earnings report filed with SEC EDGAR; verify revenue growth rate and sequential trends. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceAppLovin trades at a 20 times earnings multiple.
Unidentified speaker (speaker_0) ·41:16
How to verify: Calculate AppLovin's (APP) current share price divided by TTM EPS using SEC filings or financial data providers. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceZeta's stock rose 77% year-to-date and 113% over the past six months as of November 2026.
Unidentified speaker (speaker_1) ·44:37
How to verify: Calculate Zeta Global's (ZETA) stock price change from January 1, 2026, and May 2026 to November 2026 using historical price data from Yahoo Finance or Bloomberg. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Meta trades at a 19 times earnings multiple.
Unidentified speaker (speaker_0) ·44:02
How to verify: Calculate Meta Platforms (META) current share price divided by TTM EPS using SEC filings or financial data providers. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceVertiv trades at 25 times EBITDA.
Unidentified speaker (speaker_0) ·38:09
How to verify: Calculate Vertiv Holdings (VRT) Enterprise Value divided by TTM EBITDA using financial data providers. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Nebius trades at 60 times earnings.
Unidentified speaker (speaker_0) ·38:09
How to verify: Calculate Nebius Group (NBIS) current share price divided by TTM EPS using SEC filings or financial data providers. For recent events, verify 2026-08-16 through 2026-11-16. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Verify Zeta Global's Q3 revenue growth rates and Athena adoption metrics via the latest 10-Q or earnings press release.
- Check the current P/E ratio of Zeta Global (ZETA) and compare it to peers like AppLovin (APP), Adobe (ADBE), and The Trade Desk (TTD) using a financial data provider.
- Review Zeta's disclosed customer count for Athena OS and the specific revenue contribution from Palantir and Snowflake partnerships in recent filings.
- Analyze Zeta's gross margin trends to assess the impact of shifting from frontier models to cheaper open-source models.
- Read Palantir's (PLTR) and Snowflake's (SNOW) investor presentations to evaluate the depth and commercial stage of their partnerships with Zeta.
- Compare Zeta's organic growth guidance against the 25% threshold cited by the speakers to determine if the valuation is supported by forward estimates.
Full transcriptexpand
- 0:00
- Unidentified speaker (speaker_0): Welcome back to the deep end. Today we are diving deep into Zeta. And I think this is probably one of the more interesting names we've covered because
- 0:09
- Unidentified speaker (UNKNOWN): I
- 0:09
- Unidentified speaker (speaker_0): don't think anyone really knows what bucket to put this company in. Uh you can look at Zeta and say this is basically a really fast-growing marketing technology company helping larger brands understand customers, buy ads, and essentially manage relationships. And if that's all the business ever becomes, I think the valuation debate's gonna be really interesting. as we're gonna talk about for uh later in the episode.
- 0:29
- Unidentified speaker (speaker_0): But there's also another version of the story where man, if all the identity data, customer signals, and infrastructures data spend years building becomes a lot more valuable to deploy AI agents, this is an energetic AI winner. But before we get into it, I mean uh how how are you doing about this uh topic? So I know uh you and I are both trade
- 0:49
- Unidentified speaker (speaker_0): desk owners. Uh we've I I've specifically been burned in the space where I won't touch any company now that's any at all associated with ad tech. It's just not my lane. I have that TTD scar, never again. So, like, what are your thoughts before we get into the company?
- 1:03
- Unidentified speaker (speaker_1): Yeah, so I definitely have trade desk goggles on uh when looking at ad tech, and it's important for me to kind of not punt punish it or penalize these other brands and business models for another company. that had a really charismatic and convincing founder that just kind of talked out of his behind for a few years. And not not extrapolate that to mean the other companies are too. But
- 1:24
- Unidentified speaker (speaker_1): man, made a ton of money on Trade Des for a couple years and man did most of that go away very quickly. So yeah, the the uh the scar for me is also there. Um so yeah I I think it's just have to be objective here have to have to look at Zeta as its own company and not uh both of us carrying our baggage and our our prior
- 1:43
- Unidentified speaker (speaker_1): um like bad pick, which would trade desk, and we're both gonna, I think, do a good job doing that. But I know I'm biased, but you guys will be the judge of whether we do a good job doing that as we dive into this episode.
- 1:54
- Unidentified speaker (speaker_0): Yeah, I mean, let's get right into it. Brad, uh, before we get into Athena and the valuation of Zadia, Zaya, want to talk about what Zeta is actually doing for the customers. today?
- 2:03
- Unidentified speaker (speaker_1): Yes. So um they call themselves the AI marketing cloud. Uh so basically end-to-end marketing software, uh rich data set to help optimally acquire customers. And that's really, I would say that data set is is their network effect. Um so one of the unique things about the Zeta Marketing platform is they don't fish in um third-party third party channels or own channels on their
- 2:23
- Unidentified speaker (speaker_1): customers' behalf uh exclusively. They do everything. Um so they have a demand side platform like the trade desk. They have uh an an email marketing management platform, they have a customer data platform. And and we'll get into why vendor consolidation is a big part of their value proposition as well. But what that means from a data point of view is that their customer data profile that is unified
- 2:43
- Unidentified speaker (speaker_1): across all of these channels um is richer and has uh outcome history from more places than than just uh a a streaming placement um or a mobile web placement or something like that. So they they their their pitch, one of their big pitches is that uh this Zeta data cloud, and it's I should say data cloud and not use my Midwestern accent because the Zeta data kind of kinda sounds funny,
- 3:02
- Unidentified speaker (UNKNOWN): but uh but
- 3:03
- Unidentified speaker (speaker_2): Zeta
- 3:04
- Unidentified speaker (speaker_1): data cloud. Uh so basically this is their identity graph, and that is uh their They're opt all always opt-in, but their map of 245 American adults uh that allows them. to understand the customer better and provide all of this data data to their customers so that they can pitch things like better return on ad spend, um, better
- 3:24
- Unidentified speaker (speaker_1): reporting uh or better top of funnel, the bottom of funnel connection uh with building brand awareness and connecting that to sales and offering um reliable outcome data that will tell you when that's going to happen more times than not. And and I think
- 3:37
- Unidentified speaker (speaker_1): we're gonna going to get into AI and for the entire sector, not just Zeta, it's a polarizing topic for is this a tailwind or is this a headwind. But really when we're talking about can customers build their own AI agents and and can they uh vibe code all that stuff, which is still a lingering risk, although it's died down a lot, the the most important thing for Zeta is
- 3:57
- Unidentified speaker (speaker_1): that when you enter the platform, you aren't just getting customer or you aren't just organizing
- 4:03
- Unidentified speaker (speaker_1): and and gearing your customer data, your customer data around optimal placements and best return on ad spend, but you're gaining access to their gigantic platform and all the third party data from similar customers that are trying to do the exact same thing. So you're not operating on an island, you're not all by yourself. Uh you have the power of the platform and you have all of these point solutions tied
- 4:23
- Unidentified speaker (speaker_1): into a unified platform that allows you to cut costs. have again better better access to information on your customers, augment campaigns so that you're making more money on them, accelerate growth. That that's really their pitch. That's what they do very, very well right now. And now I I think that the two things that we'll get into next are what they're doing with AI and what they're
- 4:43
- Unidentified speaker (speaker_1): doing outside of marketing, because I think those are really the two uh big unlocks uh for uh for for Zeta earning. Um
- 4:53
- Unidentified speaker (speaker_1): we'll say that sell side analysts, well, I'll just be blunt right now. Sell sell side analysts do not think this company is gonna sustain the growth that they think they're gonna sustain over the coming years. So they have them going from 27% growth in 2026 to 20% in 27 and then teens in in 2028. Um, and that that is a byproduct of skepticism surrounding whether or not they can expand beyond um
- 5:13
- Unidentified speaker (speaker_1): this marketing core that they've done such a great job with. And I think the signs so far in both of these categories, we'll get into it, don't want to steal the entire episode for one segment, but the signs in both categories are quite good. And I would say the signs need to be quite good for this to turn into a story where uh growth is more durable than we expect, um, and markets are less cyclical than we expect, and
- 5:33
- Unidentified speaker (speaker_1): Zeta earns the kind of durable profit compounding and multiple expansion that gets c investors paid from human. here after the stock's gone from I don't know like 15 or 16 bucks a share to this episode we're doing, it's it's at 34 right
- 5:45
- Unidentified speaker (UNKNOWN): now.
- 5:46
- Unidentified speaker (speaker_0): Yeah, I mean that I think that's the tricky thing because I mean at the center of Zeta is super graph, obviously, because that's basically the identity layer trying to connect really uh useful things for the customer. I mean, email addresses, phone numbers, device, a customer record into one persistent view of the person. So a good example I always talk about. when I mention Zeta is like imagine if you're visiting a
- 6:06
- Unidentified speaker (speaker_0): retailer's website. You browse for a pair of shoes, then you get an email later, watch you watch a connected television ad, then eventually buys you buy it through the mobile app. Uh, because without identity resolution, those can look like several completely separate interactions or even separate several several different people, which makes it really much harder to know what's actually
- 6:26
- Unidentified speaker (speaker_0): influencing the purchase. Zeta is trying to connect those signals into one customer profile. So it's super valuable in the future because a business now gets to understand that this is the same person and they can do a much better decision about what actually makes sense next. And the moat that is potentially going to develop for Zeta, that if you're buying the
- 6:46
- Unidentified speaker (speaker_0): stock at 30 plus, what Brad's talking about, it is, I agree, it's a full valuation right now based on. what the estimates are showing. But the surprise factor could be that the moat that Zeta's develop developing right now is they take the identity, predict who's likely to buy, execute the campaign, and measure what happens next, and then feed that result back into this next prediction. That's where
- 7:06
- Unidentified speaker (speaker_0): it becomes super valuable because every additional interaction makes the next decision better. Then the data becomes even more valuable than just owning a big database. So uh a question for you, Brad. Like the broader pitch that customers can replace a bunch of different marketing products with Zeta, like, do you think that's true? Because that is like what they consist consistently say
- 7:26
- Unidentified speaker (speaker_0): right
- 7:26
- Unidentified speaker (UNKNOWN): now.
- 7:27
- Unidentified speaker (speaker_1): Uh no. Um, but and and I
- 7:29
- Unidentified speaker (UNKNOWN): will say,
- 7:29
- Unidentified speaker (speaker_1): like, yeah, no, I don't think that's true. Just because models are Those racing towards all being a lot more similar to, I won't say commoditization, but being more and more more and more similar to what they've been in the past. Um, so we're all using the same models in our work, and data and experience and governance is what separates leaders from from the pack. I think that was true during the first Spocalypse vibe
- 7:49
- Unidentified speaker (speaker_1): coding um help or uh episode, and I think that's true now. But I will say it them having an OpenAI partnership, which they do, which will for their AI engine that we'll we'll soon get into, and I I keep teasing, but I probably promise we'll get into it soon. Um, but they just just I mean, just just look at what just look at Figma and
- 8:09
- Unidentified speaker (speaker_1): ask them if partnerships with AI companies meant that these companies are loyal to them and won't turn around and do something to be direct competitors in the future. Because they are, I mean, the these companies need to fund these massive training, uh, the massive model training and infrastructure. inference. And how they're doing that is expanding into enterprise
- 8:29
- Unidentified speaker (speaker_1): software use cases and starting to make those more efficient and and make those prettier and more modern and tie everything together a little bit more conversationally so that we can act more efficiently in those areas, just like we've done with our consumer chat bots and chat GPT for for a few years. So I I think OpenAI being their partner is great news, but their other partner,
- 8:49
- Unidentified speaker (speaker_1): Palantir, would tell you. you shouldn't trust OpenAI and they're just stealing your data and and eventually going to turn around and do the exact same thing. So um I will say no, I do not think um c companies can vipode a solution to Zeta that's nearly as effective uh because Zeta knows their customers as as well as they do, and they know their customers better
- 9:09
- Unidentified speaker (speaker_1): than they do because they have all these third-party signals that customers can't enjoy or that their clients can't enjoy. Um but I I I would say know that they they can't be vibe coded away, but know these this these partnerships aren't uh long-term proof that there is never going to be an an AI-based, an AI native-based disruptor trying to take um their
- 9:29
- Unidentified speaker (speaker_1): market share or Applevin's market share or Trade Desk's market share, any of these companies.
- 9:33
- Unidentified speaker (UNKNOWN): I
- 9:33
- Unidentified speaker (speaker_0): think I think that's also the risk to the story because Zeta is fully valued right now, but fully valued. alongside uh having a potential crack due to AI is always a little worried. So before we get into the next session about Dina and AI, we should also talk about the acquisitions that uh Zeta has made because they have made a few. Uh I think
- 9:53
- Unidentified speaker (speaker_0): it's really important to kind of initially look at these MA strategies of what Zeta's trying to become. You have live intent, that's obviously been their strategy to strengthen identity and post. publisher side, while Marigold expanded Zeta into like loyalties, retention, and um subscription-based marketing. And also brought more than like, I think it was like a hundred enterprise
- 10:13
- Unidentified speaker (speaker_0): brands into the ecosystem, which was a huge ad. And I also think the bigger opportunity is what Zeta calls one Zeta, because if somebody can come into the relationship for email, maybe they eventually adopt the identity graph. paid media, analytics, and Athena that we're about to talk about. So you can now see the increased revenue inside an existing
- 10:33
- Unidentified speaker (speaker_0): relationship rather than having to kind of pay to acquire a brand new customer every single time. And that's one of the things I think you should be watching very closely going forward because the best enterprise platforms usually become more valuable as customers use more products, which makes the relationship harder to replace and gives the vendor much more room to expand over time. But now we are at the AI part of the
- 10:53
- Unidentified speaker (speaker_0): episode, Athena. And let's get into Dina because this seems to be where
- 10:57
- Unidentified speaker (UNKNOWN): I
- 10:57
- Unidentified speaker (speaker_0): don't know the market starts getting a lot more excited about Zeta, uh, if you follow a lot of the uh influencers who cover the stock. So I'll pass it to you, Brad.
- 11:05
- Unidentified speaker (UNKNOWN): Uh
- 11:05
- Unidentified speaker (speaker_1): yeah, and I think excitement is is well placed here. So um this is this is their AI agent. This is their platform wide AI agent. So if you're familiar with CrowdStrike, and Charlotte AI or some of these other agents being built around the enterprise software platform, the or enterprise software universe, the concept is similar. So basically take all of the use
- 11:25
- Unidentified speaker (speaker_1): cases that they've built that are made more effective with rich data and all those other things we talk about, and make them conversationally accessible so that customers can build campaigns. They can optimize these campaigns with their words instead of needing
- 11:39
- Unidentified speaker (speaker_1): complex data science experience or or coding experience or anything like that. So OpenAI is the foundational model partner again. Palantir and Snowflake are the foundational data partners. They and and these partnerships aren't just, I wouldn't I would say these are not just a press release. Like they're, they're more, they are much more involved in in the the technical
- 11:59
- Unidentified speaker (speaker_1): collaboration, uh, the go-to-market collaboration Palantir and Snowflake. Um, so I think that that, I mean, I'm more way, way, way more encouraged by those partnerships than the OpenAI partnership, I would say, because they can work with OpenAI or Anthropic or Muse or whoever the heck they want to work with as these models get more similar. Um, those companies can turn around
- 12:19
- Unidentified speaker (speaker_1): and leave whenever they want to. But for Snowflake and Palantir, these are seen, or this is seen as basically a a channel partner that can support um growth for. their existing products and support the AI platform for Palantir and everything that Snowflake's working on in data engineering to actually support their business in a way that they're not looking to turn
- 12:39
- Unidentified speaker (speaker_1): around as a plant. So that to me, reading about those partnerships made me a lot more interested in the company, uh made me more encouraged by the realness of their we target. better and we report better, um, uh aside from the 6X return on ad spent and the time to market benefits that they uh pitch over the competition. Um,
- 12:59
- Unidentified speaker (speaker_1): so that really was exciting to me. And I think uh building an AI platform for for marketing software, um, you question can it be cutting edge? You you question can it be best in class? You don't question it when it's built in tandem with Snowflake and Palantir. So um that is legitimacy, right? reputation building, whatever you want to say. That that was um a head turner for me for sure.
- 13:19
- Unidentified speaker (speaker_0): Yeah, and I think that Dina is like definitely the most exciting part of the story because they're trying to become the natural language layer on sitting on top of everything that Zeta has already built for years. So the best way to think about it is like instead of a marketer manually pulling data, building an audience, sending that audience to another platform, running a campaign, and then waiting for that somebody to essentially
- 13:39
- Unidentified speaker (speaker_0): analyze the result, the customer can now increasingly just ask the system what they're trying to accomplish. So you can imagine again asking which customers are most likely to churn next month, for example. What offers is most likely to keep them and where the next marketing dollar should go. And rather than Athena simply returning an answer like a some kind of chatbot via Chat GPT, like
- 13:59
- Unidentified speaker (speaker_0): Zeta wants to build an audience and actually help execute the action. Intent behavior, outcome based behavior. Really great to see. That's kind of the language you need to have in the new AI economy because that's where I think the architecture gets a lot more compelling because AI isn't sitting next to the workflow as another chatbot anymore. That's fully 2023, 2024, and
- 14:19
- Unidentified speaker (speaker_0): parts of 2025. Now you need AI to kind of become the interface into the workflow itself. You're seeing the early numbers are. generally strong. More than half of the super scaled customers are now monthly active Athena users. Great to see. Athena adopters account for more than like, I think it was 70% of the company revenue. Great to see
- 14:39
- Unidentified speaker (speaker_0): because the whole story about Zeta is Athena. And the top 25 Athena users are growing revenue with Zeta Data by more than like 70% in Q3. So uh really great numbers. Like Athena's working. And I think the net revenue. retention is also more than I think it goes 500 basis points above the company average. So uh the issue though, like I'm saying all these great numbers, like there's
- 14:59
- Unidentified speaker (speaker_0): an asterisk of it because I'm not giving Athena full credit for those numbers yet, because um the most engaged customers are probably naturally the first ones to adopt it. So there is some selection bias inside the data, in my opinion.
- 15:11
- Unidentified speaker (UNKNOWN): And
- 15:11
- Unidentified speaker (speaker_0): I really want to see whether the same thing happens when Athena moves to. deeper into the broader customer base. Like if an average customer adopts Athena and their net their retention improves and spend goes higher and that customer starts using more channels, then it becomes harder to argue that Athena is simply correlating with the best customers rather than actually causing the expansion.
- 15:31
- Unidentified speaker (speaker_0): Like Palantir is a great example. You mentioned Palantir, like the more they're scaling their commercial business, their revenue growth is sequential sequentially increasing quarter after quarter, but so is their net revenue retention. So like as it keeps scaling, so are every KPI metric that matters.
- 15:46
- Unidentified speaker (UNKNOWN): I think it got
- 15:47
- Unidentified speaker (speaker_1): 120, 125% for 2025, like up five points, I think. Like really impressive. Sorry, go ahead.
- 15:53
- Unidentified speaker (speaker_0): Yeah. No, no. So I think I think that um you're seeing great momentum, Athena. And the obvious though issue, like this is a good time to talk about the Gentic AI, because the obvious issue with Asia. Asians is usually just can explode while inference cons costs also eat the economics of it. So how how do you think Zeta thinks about that? Because that's like kind of the biggest question of
- 16:12
- Unidentified speaker (speaker_0): AI good, AI bad.
- 16:16
- Unidentified speaker (UNKNOWN): Yeah.
- 16:16
- Unidentified speaker (speaker_1): So from a cost point of view, I think they uh it it that they are part of the cliche. Um so everyone raced to install workflows or around the absolute best models they could possibly find. Um, cost was not a consideration. And now that everyone's figured out that 90% plus of this work can be done by very
- 16:36
- Unidentified speaker (speaker_1): cheap, increasingly cheap open source or or or small, smaller models from Meta or Google or other companies like that. Um, there is much, much more mixing and matching happening. Um, and you're seeing it's not not just from Zeta, but companies like Shopify talk about uh just vastly ramping their sidekick usage, which is their AI agent, and seeing overall costs fall.
- 16:56
- Unidentified speaker (speaker_1): So um I don't I don't think we should be expecting this to be a massive profit driver in the near term, but I think there are clear signs that input costs, which they're they're they're a software company, so AI is an input cost for them, um, are continuing to come down very quickly. And if any technological cycle in our in human history is any indication at all. um that process will just keep
- 17:16
- Unidentified speaker (speaker_1): playing out and they will just keep finding more leverage here as they find more scale. And I mean we all know how good that is for for margins. Um and and that that is the trend um that I see playing out for um the Athena AI agent. Uh and and I mean you I think you said seven X usage when they immediately after moving it from limited to broad availability, uh commentary in the call
- 17:36
- Unidentified speaker (speaker_1): was really positive, which which happened um
- 17:39
- Unidentified speaker (speaker_1): what day is it? It's been a long few weeks. Uh but but but the earnings call happened kind of yesterday. I'm sorry. Um so uh commentary was uniformly positive around that and and the adoption curve they're seeing there. So yeah, two two very good trends um usage up cost down I think that just continues for a long time and supports the eventual profit
- 17:59
- Unidentified speaker (speaker_1): inflection for this specific profit product.
- 18:02
- Unidentified speaker (UNKNOWN): I
- 18:03
- Unidentified speaker (speaker_0): also think it's really important to talk about like what like Zeta's building Athena inference models with fireworks using NVIDIA's nemotron technology. And I think that's really important as the open model uh momentum continues because management also claimed that specialized model achieves something like 95% accuracy on selected evaluation while
- 18:22
- Unidentified speaker (speaker_0): operating around like 40% cheaper per quote qu request, and like like you mentioned, 17%. seven times faster than uh frontier model comparison. So I I think I would like Zeta is building a better AI model uh than OpenAI. I wouldn't say that. Uh but for
- 18:37
- Unidentified speaker (UNKNOWN): their
- 18:37
- Unidentified speaker (speaker_1): maybe for some of their very hyper-specific needs, I would say that they're building a product
- 18:42
- Unidentified speaker (UNKNOWN): better
- 18:42
- Unidentified speaker (speaker_1): suited for them. Yeah.
- 18:44
- Unidentified speaker (speaker_0): It completely misses the point if you try to compare Zeta's AI model with OpenAI's model, because if the task is predicting churn, building an audience. or deciding which offer a customer should receive, you probably don't need to pay for the largest and greatest frontier model every single time. And I think that's kind of becomes one of the most important themes across enterprise AI more broadly, we've seen the past
- 19:04
- Unidentified speaker (speaker_0): couple months, where the future probably isn't one giant model doing every single task. It c it's companies routing workloads towards whatever intelligence can perform the job at the lowest point. possible acceptable cost. If Zeta can use smaller specialized models for a lot of Athena's interactions while reserving frontier models for the more difficult work, then
- 19:24
- Unidentified speaker (speaker_0): AI usage it can increase dramatically without inference expenses consuming all of the incremental revenue. And I think that's when Athena starts becoming a lot more interesting on an ROI basis rather than having like a really fancy user. interface. But I think uh this is also a good time to talk about
- 19:43
- Unidentified speaker (UNKNOWN): uh kind
- 19:45
- Unidentified speaker (speaker_0): of like brands versus the hyperscaler. Uh what happens if search changes, for example. Like management also thinks AI can ship advertising dollars away from traditional search. I mean, how much do you buy that, Brad?
- 19:55
- Unidentified speaker (UNKNOWN): Well,
- 19:56
- Unidentified speaker (speaker_1): I think that's why it's so important that they play in every channel because I I do think like human human eyeballs and human human traffic isn't gonna shrink. I don't think we're just gonna spend more time on things that we want to do instead of what we need to do. So maybe we'll watch more TV or YouTube videos or or whatever that is, whatever we find enjoyable.
- 20:15
- Unidentified speaker (speaker_1): Um, so I don't think AI automating tedious work means screen time and impressions from consumers go down. I do think the market share of where those impressions go does change a lot. Um over the coming years, but that's why it's so convenient that they play everywhere because they're kind of agnostic to where those impressions go as long as as long as people are still
- 20:35
- Unidentified speaker (speaker_1): online and looking at things, which again, I don't while while we use AI to automate all the stuff we don't want to do, we're still gonna we're still gonna be twiddling our thumbs and doing something. There's They're still 24 hours in a day. Um we still humans still don't do. super well when we're extremely bored for long stretches of time. Um so I
- 20:55
- Unidentified speaker (speaker_1): think it it it's not an overall impression threat to them, but it is a channel threat to a lot of people and they're they're the company that kind of guards against that because they'll go wherever they the the consumers
- 21:06
- Unidentified speaker (UNKNOWN): go. Yeah
- 21:07
- Unidentified speaker (speaker_0): I think it's really important if you're a Zeta Bull to not think that like most dollars are going to leave Google and suddenly land at Zeta. Like that's that cannot be part of your bull thesis because Google, Meta, OpenAI, and everybody else building their own advertising and commerce systems around this transition, like they're going to be massive players in this space. But I do think there
- 21:27
- Unidentified speaker (speaker_0): is an argument that the value of being an independent layer that helps companies understand consumers' intent across multiple platforms rather than just tying the whole strategy to one closed ecosystem. makes sense. That's where my scar tissue comes out because of TTD. That was a trade desk, is the whole thing. Uh that blew up,
- 21:46
- Unidentified speaker (speaker_0): obviously, especially with Amazon. Uh, but I also think that's why the senso acquisition gets really interesting because Zeta is trying to move into the generative engine optimization and helping companies kind of understand how they appear inside Chat GPT, Cloud, Claude, and Gemini. So if an agent kind of increasingly decides which products can make the con consideration
- 22:06
- Unidentified speaker (speaker_0): set, then marketing changes pretty dramatically, in my opinion. Because if a company is no longer only trying to influence the consumer, it may need the influence of an agent acting on the behalf of the consumer. Uh, could that actually become a threat if OpenAI or Google owns the interface? I think so. And that's the risk of owning Zeta at full valuation because that's
- 22:27
- Unidentified speaker (speaker_0): why this moat simply cannot be Athena, Athena as a conversational interface, because every frontier model can eventually build one. The moat has to be that it sits in the proprietary customer data, the identity identity graph, and the execution layer that already lets enterprises do something with the answer. OpenAI OpenAI already owns a chat box while
- 22:46
- Unidentified speaker (speaker_0): Zeta owns a customer intelligence and execution underneath it. But that's plenty of room for. Zeta to kind of create value in. Um what are your is this where your like trade desk goggles come in with Zeta? Because I always think that hyperscalers like um why bother? Like Meta, for example, like we'll talk to this more detail, but like Meta, we just went over it. Like Meta has a
- 23:06
- Unidentified speaker (speaker_0): similar organic growth rate as uh Zeta, both around 30%. Well, meta's trending training at 19 times earnings, Zeta's trending training at 30 times earnings. So why bother? with the independent layer when the king profit engine is trading at a 50% discount? I'll pass it to you now.
- 23:21
- Unidentified speaker (speaker_1): Sure. Yeah. Well, I I we should say that Meta's doesn't have a free cash flow multiple right now because they're spending so much on CapEx.
- 23:27
- Unidentified speaker (UNKNOWN): So for your people for free cash
- 23:30
- Unidentified speaker (speaker_1): flow traditionalists and nothing else matters, then there you go, I guess. But um I I am of the opinion that it's it's gonna be hard for companies over the next five and ten years to consistently convince their customers that you need us to find your customers um effectively and and and and and do
- 23:50
- Unidentified speaker (speaker_1): so in in a way that's better than what you can do on your own. And I'm not saying they're gonna just snap their fingers in vibecode, but I am saying they can in install models and install workflows and um AI FDEs and figure out how to actually target customers in a more intelligent way that makes them less reliant on platforms and maybe cuts into pricing power a little bit. But I think the antithesis
- 24:10
- Unidentified speaker (speaker_1): to that is an absolute need to play, and we'll keep keep going back to the exact same thing, is an absolute need to have, like you're saying, better customer data profiles, more complete customer data profiles. And
- 24:22
- Unidentified speaker (UNKNOWN): I
- 24:22
- Unidentified speaker (speaker_1): mean, hyperscalers have have better than everyone else. But when you look across pure play, ad tech companies and where they are willing to play, where they are willing to fish on a per customer basis, again, more channels, more data, more interactions per customer, knowing them better. And that is really where the sustainable differentiation can come from. That is where they can continue to sprint
- 24:42
- Unidentified speaker (speaker_1): to stay ahead as everyone else sprints behind them. If they take this data network effect, they continue rapidly growing and they continue effectively using it to make their models better. That is the flywheel. that can keep them ahead. And that is what can make them a reliable and valuable partner that can fetch their take rate for a very long time to come. Um and the signs of that I think are very good for Zeta.
- 25:02
- Unidentified speaker (speaker_1): Um, I don't think they're great for a lot of the the smaller point solution names in in the space. Um, but I do think there is room for pure play plate platforms. There is room for, um, and I don't want to sound like Jeff Green because I never know we all have too much trade desk PTSD, but there is room for a company that does not own impressions and does not have that conflict of interest inherently embedded
- 25:22
- Unidentified speaker (speaker_1): in their in their um in their model. There is room for a company that is solely determined well, they do out the I mean m or they do have an SSP or a supply side platform, but they are predominantly placing through Magnite and other publishers. Um there is room for companies that are predominantly focused on buyers um and and that don't have um Instagram impressions or YouTube impressions or their
- 25:42
- Unidentified speaker (speaker_1): Amazon marketplace impressions to worry about um getting top dollar for. And that's where Zeta can sustainably stand out, I think.
- 25:49
- Unidentified speaker (speaker_0): Well, I also think like I think it's a good time to talk about the Zeta versus like valuation a bit because I think one of the things I commonly see with a retail-friendly retail friendly name, like Zeta's a very retail-friendly retail friendly name.
- 26:02
- Unidentified speaker (UNKNOWN): I
- 26:02
- Unidentified speaker (speaker_0): think there's definitely exciting parts. Obviously, if you heard this episode so far, like Brad and I are talking about a lot of the exciting AI parts of the Zeta story. But I think what loses me is when people act like being smaller independence and having an AI story means that you deserve a premium over the king. That doesn't make sense to me. And that's why I
- 26:22
- Unidentified speaker (speaker_0): that's where buying Zeta at $35 loses me on like why buying. bother at 30 times earnings. When you have Meta growing the same organic rates as Zeta for a company of its size is bonkers. Yes, free cash flow negative, but if you're bullish on AI for Zeta, you're bullish on what Meta's doing then on a CapEx front.
- 26:42
- Unidentified speaker (speaker_0): Like you just are. You can't have one of those two things being truth and the other being wrong. Like Meta has billions of users. the most profitable ad engine business in the world, and they're already seeing proof of acceleration due to AI and engagement and monetization inside this existing business. So I'm paying a lower
- 27:02
- Unidentified speaker (speaker_0): multiple for the king at a similar growth rate. Why pay 30 times earnings for Zeta when really all you're paying up for is the hope that if it's smaller with an AI story that they're gonna change their growth curve in a couple years. When I haven't seen proof of that organically yet,
- 27:22
- Unidentified speaker (speaker_0): really. I've seen like there's momentum. I just haven't seen proof. Like if Zeta can organically grow 25%, for example, I would be like that's where I'll be okay. If they can grow 25% organically for the next three years, that's something that Southside is wrong on. But where where's the proof? I I mean, you're gonna say something, sorry.
- 27:41
- Unidentified speaker (speaker_1): No, I I think multiple expansion for this name or continuing to fetch a 30x multiple for the next next few years as as growth curve slows is is slow. I I don't know how how meaningful that slowing is going to be, but assuming that it doesn't just keep accelerating in perpetuity as as we lap the MA and all that good stuff. But um, I would say them
- 28:02
- Unidentified speaker (speaker_1): in in in and this gets back to Athena AI, and they're branching out into the Athena operating system or Athena OS, and their push to pitch themselves as not a marketing software platform, but as a true enterprise enterprise software platform. And they very intentionally put valuation comps to CrowdStrike and all these other enterprise software
- 28:21
- Unidentified speaker (speaker_1): leaders in their slide deck because they are really trying to re-reposition the message. messaging for this company behind we help customers target and report target customers and report outcomes more reliably. To we have this thing called Zeta Business Intelligence built on their data cloud, built on all this consumer pro these
- 28:41
- Unidentified speaker (speaker_1): customer profiles that they have, but branching out into things like customer service or market analytics or real estate optimization. So they they tell customers this is where the next location should be. to minimize um cannibalization and maximize average unit volume, things like that. So business intelligence um generally speaking, and that is that that is something
- 29:01
- Unidentified speaker (speaker_1): that that could fetch a multiple like 35, 40, 45 X and and and really give that that upward surprise in in growth over the next few years versus what people are currently expecting. Um, Again, super, super early here. The this the early signs are promising, but but too early to really get all that excited about. But but everything that
- 29:21
- Unidentified speaker (speaker_1): needed to look good so far has looked good. And and to me, this is how they move from um then
- 29:28
- Unidentified speaker (speaker_1): I almost said the next trade desk. I'm not I'm not calling it the next trade desk, but like this is how they move from basically being the next pure play because trade desk was an ad tech darling for two decades. So that's that's really the part that I'm comparing them to. That's how they move from being the darling for the next two decades in just ad tech to being uh I'm struggling for the right comp because Salesforce is
- 29:48
- Unidentified speaker (speaker_1): growing at 9% top line right now, but to being that generalist enterprise software partner that helps them in so many other use cases besides just specifically marketing, that is where five to ten points of a multiple expansion can happen.
- 30:00
- Unidentified speaker (UNKNOWN): I
- 30:00
- Unidentified speaker (speaker_0): think you also brought up a great point though on Salesforce. These bottom They just bought Braze. Like that's competition. Like I think my issue that I think that's a pure ad tech name. I I know a lot of I see a lot of comments here that this could be a data play infrastructure name that it deserves a valuation with that. No, I I totally disagree with that. I think they have an
- 30:20
- Unidentified speaker (speaker_0): absolute premium for being an ad tech play because of their data advantage, but they should nowhere in fact be trading at a multiple like 50 times earnings. I think a lot of people are naturally comparing it to Palantir for whatever reason. Maybe they're same similar retail crowd. I don't know what. But this is not going to be the next Palantir. Pound tier is a totally different
- 30:39
- Unidentified speaker (speaker_0): business that's t outcome based, that's kind of removed from the recessionary risk of having an ad tech company and hyperscalers can't do what Palantir does. Um spoiler alert Palantir's not an LM rapper. That's why I said that. But uh I I think like Salesforce combining with Braze that puts a lot of pressure in my opinion
- 30:59
- Unidentified speaker (speaker_0): on inner enterprise distribution, for example, next to a strong real time engagement platform. Like Zeta's competing with Adobe, Oracle, Clavio, the Trade Desk, still Google, Meta, and I don't know, potentially the Frontier AI platforms themselves. Like I think I know Zeta argues that these big integration integrations can create uncertainty and given an opportunity to go
- 31:19
- Unidentified speaker (speaker_0): into accounts with a much more unified architecture. And I think there's probably some truth in that. But I would never build the thesis around Salesforce or Adobe being unable to comp compete because those companies have a massive, enormous distribution and they're investing heavily too. So I I think again at 30 thirty times times earnings, earnings, like like I don't
- 31:39
- Unidentified speaker (speaker_0): think it can have a multiple expansion story. I think they're at peak earnings right now. They're never gonna they shouldn't trade above 30 times earnings. But the question now is going forward.
- 31:49
- Unidentified speaker (UNKNOWN): You mean you mean you mean
- 31:50
- Unidentified speaker (speaker_1): peak earnings multiple, not peak earnings, right? Just uh
- 31:52
- Unidentified speaker (UNKNOWN): peak earnings
- 31:53
- Unidentified speaker (speaker_0): multiple. Yeah, that's
- 31:53
- Unidentified speaker (UNKNOWN): yeah. Okay, cool.
- 31:55
- Unidentified speaker (speaker_0): They're never gonna trade over thirty to thirty times earnings multiple.
- 31:57
- Unidentified speaker (speaker_1): We talk on
- 31:58
- Unidentified speaker (UNKNOWN): T
- 31:58
- Unidentified speaker (speaker_1): we talk on behind microphones hours every day. We're gonna have slip slip-ups, ups, it's inevitable, but but I yeah. but no, you're
- 32:03
- Unidentified speaker (UNKNOWN): good. No,
- 32:05
- Unidentified speaker (speaker_0): but I mean
- 32:07
- Unidentified speaker (speaker_0): I still don't uh I still have a very difficult time with the premium. I just don't understand it. Like it's so full. Like I see a lot of people who just reported their earnings report, for example. Like, why is it red? They beat and they re they raised a little. Well they did, but like what changed change really? Nothing, it's fully valued, and I think the Zeta management.
- 32:27
- Unidentified speaker (speaker_0): team has done a great job making sure they are part of the syngetic AI world. Uh, they have an absolute premium, but in ad tech, again, trade desk goggles on, you can see how aggressive a compression can happen once there's a crack in the story. It's excessive, it's quick.
- 32:46
- Unidentified speaker (speaker_0): It's just the same thing as consumers. Like, if you're a Celsius owner, for example, shareholder, you see saw how quickly the sentiment could change. And I think that
- 32:55
- Unidentified speaker (UNKNOWN): like
- 32:55
- Unidentified speaker (speaker_0): Athena OS is interesting. We didn't talk, I guess, too much about that. I think the relationship they have with a Snowflake and Palantir is great. They have a lot of optionality, but I don't think AI optionality gives you a free multiple. I think like their optionality is very real. I just don't think the optionality should automatically be capitalized like revenue because. that's where you really
- 33:15
- Unidentified speaker (speaker_0): get into trouble. Like Palantir could eventually contribute more than $100 million annually, but today that partnership is very early. Athena OS can take data into supply chain, fraud, and much broader uh enterprise workflows, but the product is still moving towards general availability. OpenAI can have a massive, huge distribute distribution channel, but we don't really know
- 33:35
- Unidentified speaker (speaker_0): exactly how much like of the margins that Z Zeta's is going
- 33:38
- Unidentified speaker (UNKNOWN): to.
- 33:38
- Unidentified speaker (speaker_0): gonna capture from that relationship. I think it was a pure PR relationship. I actually don't think it's that much of a benefit. I could be wrong. Uh but again, I just think that
- 33:47
- Unidentified speaker (UNKNOWN): my
- 33:48
- Unidentified speaker (speaker_0): meta goggles might be on. I just meta doesn't require me to imagine what AI monetization could look like because AI is already improving their existing advertising machine today at an insane scale. So yeah, Zeta Zeta needs the optionality. to become real faster than the market expects. But man, why bother?
- 34:08
- Unidentified speaker (speaker_0): You have the discount on the king. All right. Well, we're gonna circle back towards uh one of the more exciting partnerships actually with Zeta is Palantir. And that obviously caught my eye because I'm a big Palantir bull, uh, as many are aware. Uh, and I think it's the most interesting long-term optionality for Zeta because Zeta understands what's happening outside the enterprise through. customer beh consumer behavior.
- 34:28
- Unidentified speaker (speaker_0): Obviously, if you heard this video, you know that's what they do. But Palantir understands what's happening inside the enterprise through operational data. And if you connect those two things, you can start making decisions that go well beyond advertising. That's the margin expansion opportunity that a more multiple expansion opportunity that Zeta might be able to have. Because imagine a consumer products company where Palantir
- 34:48
- Unidentified speaker (speaker_0): understands the inventory manufacturing capacity. warehouse warehouse availability, and the cons transportation constraints, while Zeta can tell you where consumer demand is accelerating and which groups are more likely to buy. So now the system can go potentially help decide how much the product to produce, where the inventory should go, and how
- 35:08
- Unidentified speaker (speaker_0): much promotional s promotional spending sh it should support. And that's a very different TAM than at tag. tech. So I think that if you are a bull on Zeta at thirty five dollars, thirty times earnings multiple, you need to believe this is going to become a reality. Again, though, I'm
- 35:24
- Unidentified speaker (UNKNOWN): I'm
- 35:25
- Unidentified speaker (speaker_0): still cautious because we don't have enough production revenue yet. Management has talked about like
- 35:30
- Unidentified speaker (UNKNOWN): I
- 35:30
- Unidentified speaker (speaker_0): think it was eight affirmative customer responses, two operating customers, one expected to become super speed.
- 35:36
- Unidentified speaker (UNKNOWN): No one
- 35:36
- Unidentified speaker (speaker_1): said no yet, so that's cool.
- 35:37
- Unidentified speaker (speaker_0): That's true,
- 35:38
- Unidentified speaker (UNKNOWN): but it.
- 35:38
- Unidentified speaker (speaker_1): yeah. High batting
- 35:39
- Unidentified speaker (UNKNOWN): out.
- 35:40
- Unidentified speaker (speaker_0): It's still it's still so little though. Like it's so tiny that you can't justify the 30 times multiple on that. So I think it's genuinely exciting. I'm not putting the entire um business segment into what I think is gonna happen, but otherwise, there's again, there's a lot of future business being priced in today's valuation. Now, is there anything else you want to add to the Zeta and
- 35:59
- Unidentified speaker (speaker_0): Palantir? Pounds here?
- 36:00
- Unidentified speaker (speaker_1): No, I would just say them. I mean I I appreciate that they don't not just not to throw one of your one of your companies under the bus, but I I appreciate that they don't pitch this as the source of their competitive differentiation um and and and simply as like a more of a a technology and go to market partnership. And I also think that them partnering with Palantir
- 36:20
- Unidentified speaker (speaker_1): to going going back to AI costs and and that kind of idea, it just tells you exactly where they're going in in model. land, especially with the work that they're already doing. It just how how much how much more value there is to place around the data and the harness than the model. And again, that's why to me, um, this Palantir partnership and the and the and Snowflake P partnership that helps them
- 36:40
- Unidentified speaker (speaker_1): take advantage of this data more intelligently, that gives them access to new relationships to package all this value together and find more demand is more exciting than they're using OpenAI models in their in their platform. And I mean I hope one day they don't have to use OpenAI models in their platform because they figured out a way to replace all of all of those workflows with something way, way cheaper so their margins go
- 37:00
- Unidentified speaker (speaker_1): higher and and and the stock price can go higher because there's more profit. Like that that's that's the good outcome to me. Um them continuing to work very closely with with Snowflake and Palantir and them um being the ones to say goodbye OpenAI when they figure out how how to sub substitute them with models and and um in the in the places they need to, which I don't think we're that far
- 37:20
- Unidentified speaker (speaker_1): away from some companies figuring out.
- 37:23
- Unidentified speaker (speaker_0): Yeah, really well said. I think uh it's good to pivot towards our last section, bullish or bearish.
- 37:29
- Unidentified speaker (UNKNOWN): Um
- 37:29
- Unidentified speaker (speaker_0): again. It's such a cop out for me to say I'm not bullish, but I'm also not bearish because that's uh you have to pick one. I'd say bearish. I think at 30 times earnings, a lot of future d demand or narratives being priced into today's reality. But for me to go bullish, I guess, is talk about like what would cause me to change.
- 37:49
- Unidentified speaker (speaker_0): Because I'm very open. Like maybe I just don't understand the space too great. I have my trade desk goggles on, maybe I have my meta goggles on too much, or maybe the recent AI drawdown we've seen in the market is providing me a lot more better opportunity, like acts Axon on a 20 at times. EBITDA, you have Metastone at 90 times earnings, 19 times earnings. Um, you got Vertive closer
- 38:09
- Unidentified speaker (speaker_0): to 25 times EBITDA as well. Like Nebius is now 60. Like this list goes on. There's a lot of opportunity out there. Zeta, in my opinion, because they've had a multiple expansion story past couple months, while a lot of these AI names have multiple compression stories. story, It makes them look a bit more expensive uh at the moment. But what would change my mind? Well, organic growth. You guys heard
- 38:29
- Unidentified speaker (speaker_0): me saying that has to be a twenty five percent for the next couple of years to justify the current valuation. And that will tell me that Dina and one Zeta is actually keeping the growth algorithm elevated rather than I don't know the company just settling around twenty percent and hoping the uh future narrative story justifies a premium. I also want Athena Dina to kind of Athena Dina users to
- 38:49
- Unidentified speaker (speaker_0): consistently show better retention. and higher spending as adoption broadens. Like it's still very early, tons of momentum, but it's going to be a lot harder once the early enthusiasts, early enthusiasts aren't the everything in the momentum. I also uh the Palantir and OpenAI relationship is interesting. Like it's a really interesting partnership. I need
- 39:09
- Unidentified speaker (speaker_0): to see more measurable recurring revenue because that proves that Za Z Zeta's intelligence actually has value outside of the traditional marketing uh space.
- 39:18
- Unidentified speaker (UNKNOWN): Uh,
- 39:18
- Unidentified speaker (speaker_0): I also think
- 39:20
- Unidentified speaker (UNKNOWN): I
- 39:20
- Unidentified speaker (speaker_0): don't know
- 39:21
- Unidentified speaker (UNKNOWN): free
- 39:21
- Unidentified speaker (speaker_0): we didn't talk about this much, but free cash flow per share like isn't growing that quickly. I I I think there's uh that could have definitely
- 39:31
- Unidentified speaker (UNKNOWN): like
- 39:31
- Unidentified speaker (speaker_0): it would make me a little more enthusiastic about the operating leverage levers they that you can pull while they scale. Uh dilution is still the thing that we didn't talk about too much on there. Um that needs to keep coming down. Uh, either way, like
- 39:43
- Unidentified speaker (speaker_0): all of these things have to come true, in my opinion, for Zeta to justify having a premium to meta. Um, and I just that's it's a tall, tall task. Uh, what about
- 39:51
- Unidentified speaker (UNKNOWN): you?
- 39:53
- Unidentified speaker (speaker_1): Yeah, um, so I mean
- 39:56
- Unidentified speaker (speaker_1): I I think similar answer to the AMD episode. I think if I was running a portfolio with 50 or 60 names, um and they everything got one to two percent, then this would be a name in there and I and I wouldn't um and and I'd feel pretty confident about that. But as I pick my favorite five, 10, 15 best ideas, I I I
- 40:15
- Unidentified speaker (UNKNOWN): uh
- 40:15
- Unidentified speaker (speaker_1): I I mean if you if you gave and obviously if you cut it in half and told me I could buy it there, like yes, I'll buy it there. But so that was about to be a really stupid comment, comments. so But never mind. But um I want to see basically the same things. I want to see the fact that that Athena has made their platform more accessible, and that naturally bolsters engagement and demand and volume because now there's less
- 40:35
- Unidentified speaker (speaker_1): friction associated with uh creating successful outcomes and successful campaigns on that platform. And I want to see it I want to see Athena OS um and and Zeta or Zeta Business Solutions really branch out beyond advertising. And that to me is what can make them an enterprise. software company, a general enterprise software company instead of an act ad tech company, which they
- 40:55
- Unidentified speaker (speaker_1): are right now, which I don't think um more I mean, just just looking around at the peer group, you you've got, I mean you've got app 11 growing top line 50% plus and accelerating sequential growth and just gaudy, gaudy margins. And you're paying, I don't think they're anywhere near a 30 thirty times or earnings rings multiple shy, if you wouldn't mind pulling that
- 41:12
- Unidentified speaker (UNKNOWN): up quickly. Yeah, let
- 41:13
- Unidentified speaker (speaker_0): me let me look
- 41:14
- Unidentified speaker (UNKNOWN): at it. while I
- 41:14
- Unidentified speaker (speaker_1): keep But I mean
- 41:16
- Unidentified speaker (speaker_0): twenty uh they they have twenty times earnings multiple app lovin'. Again.
- 41:19
- Unidentified speaker (UNKNOWN): Okay. So
- 41:20
- Unidentified speaker (speaker_1): so that's
- 41:20
- Unidentified speaker (speaker_0): that that feels like like the tier one multiple for the like ad space is twenty times earnings.
- 41:25
- Unidentified speaker (UNKNOWN): And
- 41:25
- Unidentified speaker (speaker_1): and two quarters ago, they gave you the scare with a delayed model launch that sounded exactly like what Trade Desk had with Coke Eye. So that freaked me out personally. And then last quarter, they delivered
- 41:37
- Unidentified speaker (UNKNOWN): other
- 41:37
- Unidentified speaker (speaker_1): masterpiece. Absolutely killed it where no, it was a delay. Um, we got everything back on track, and that was only a partial launch, by the way, and we're just going to keep getting better and better from here. So to me, you've gotten through a quarter of oh the potential oh shit for App Loven. And you haven't had that yet for Zeta with all this positive momentum happening. So I almost want,
- 41:57
- Unidentified speaker (speaker_1): I mean, I I I almost think an imperfect quarter, which companies have the though, every single company has. those.
- 42:05
- Unidentified speaker (speaker_1): Drumming up some of this multiple contraction as you get more proof that AI and non-ad tech um business is expanding and and compounding nicely at a at a profitable clip. That is when I think this goes from my fortieth or fiftieth best idea to my 10th or 15th best idea and could
- 42:21
- Unidentified speaker (UNKNOWN): make
- 42:21
- Unidentified speaker (speaker_1): its way into the portfolio. Um but I'd rather do that at $50 fifty dollars a share with more certainty and all in of those areas than I would at $35 a share today. Um and yeah, I I I will say there there are two uh there are two impressive companies in pure play ad tech. Um I know Trade Desk beat by a zillion percent yesterday, which we're recording on November tenth, so
- 42:41
- Unidentified speaker (speaker_1): that was November ninth, but that's fifteen million in in revenue. They Knew literally said the 10 Q's Q is not going to come back. And then they are still shrinking year over year and they just absolutely disgustingly sandbag guidance and beat a lowered bar. So to me, there are two healthy names in this space. There are two names that provide value, uh, like a six X return on ad spend, three
- 43:01
- Unidentified speaker (speaker_1): million in annual savings, and fifty percent faster time to value, things like that that will justify take rates and premiums and growth for a very long time. Um, that's apple-in AppLeven first. And and good job, app Applevin loving on the last quarter, because I would not have put you first to two weeks ago. Um, and then Zeta second. So those are really the only two names that if you want exposure to this space,
- 43:21
- Unidentified speaker (speaker_1): fundamentally speaking, that those are the two deserving exposure um from the companies that don't actually own their impressions. Now, this is not including Amazon or Google or Meta in that conversation because they do own their impressions, and I think that provides a lot more um visibility and certainty as AI disruption takes place. Uh, but those are the two standing out from everybody else, and and I don't really think
- 43:41
- Unidentified speaker (speaker_1): it's close.
- 43:42
- Unidentified speaker (speaker_0): Yeah, and I think again, I think valuation is a real issue with what we're talking about here. Like they're doing great on the AI story, they are doing an exceptional job right now. But man, like App Loven and Meta, they're both at twenty times earnings. That should tell you what the premium tier should be in that space. Uh and meta's like got a ton of other things as well.
- 44:02
- Unidentified speaker (speaker_0): It's like, so we won't get into that too much. But Za I, there's a real scenario where Zeta doesn't move for the next year and a half. I actually think that's a very real reality because they really need that pound tier partnership to really start scaling up. And Athena OS as well. And I just think they're they're too experimental right now at this current moment, moments and they won't be big
- 44:22
- Unidentified speaker (speaker_0): enough to actually move the needle on the total grand scheme of. the Zeta umbrella. So I think that be cautious. Like just because like I what's their performance for the uh this year? I think it's year to date's
- 44:37
- Unidentified speaker (speaker_1): pretty good I think.
- 44:39
- Unidentified speaker (speaker_0): 77%
- 44:40
- Unidentified speaker (speaker_0): in the past year.
- 44:41
- Unidentified speaker (UNKNOWN): Go
- 44:42
- Unidentified speaker (speaker_1): Bulls. Good job, guys.
- 44:43
- Unidentified speaker (UNKNOWN): Like seriously
- 44:44
- Unidentified speaker (speaker_1): that's that's five
- 44:45
- Unidentified speaker (UNKNOWN): years of great
- 44:46
- Unidentified speaker (speaker_1): returns and in in in 10 ten months. Like great job.
- 44:49
- Unidentified speaker (UNKNOWN): That
- 44:49
- Unidentified speaker (speaker_0): past six months is a hundred 113%, actually. So
- 44:53
- Unidentified speaker (UNKNOWN): af
- 44:54
- Unidentified speaker (speaker_0): when you have a move like that that's so dislocated from what Growth Land has been, you have to ask yourself what's being priced in. I think a lot in this name. So for that reason alone, I don't see
- 45:06
- Unidentified speaker (UNKNOWN): the
- 45:06
- Unidentified speaker (speaker_0): purpose of adding. So I am bearish.
- 45:09
- Unidentified speaker (speaker_1): I thought
- 45:09
- Unidentified speaker (UNKNOWN): you were
- 45:09
- Unidentified speaker (speaker_1): gonna get hit us with a shark tank, and for that reason, I am out. I
- 45:13
- Unidentified speaker (UNKNOWN): am out. Next time next
- 45:14
- Unidentified speaker (speaker_1): time.
- 45:15
- Unidentified speaker (speaker_0): Wait, wait. Let me let me rewind. No.
- 45:18
- Unidentified speaker (speaker_0): Well, um, all right. Well, uh again, everyone. I hope you guys enjoy this episode. Um, I think Zeta is a new fun AI story in a uh increasingly moving landscape of ad tech.
- 45:31
- Unidentified speaker (UNKNOWN): It's
- 45:31
- Unidentified speaker (speaker_0): not the next trade desk, but it might not even be the next great story that a lot of people think it is. So
- 45:37
- Unidentified speaker (UNKNOWN): uh
- 45:38
- Unidentified speaker (speaker_0): let us know what you. guys think. Um I can obviously be wrong. Brad can obviously be wrong, and maybe our trauma is guiding our uh uh our decision making on Zeta. But either way, let us know. Uh, and thanks for tuning
- 45:51
- Unidentified speaker (UNKNOWN): in.
More from The Deep End: Stock and Earnings Analysis from Futurum Equities
AMD Had the Best Quarter in Its History. Then It Fell 17% in a Day | Ep 14
Episode discussion: AMD reports record earnings but faces valuation skepticism amid supply and regulatory concerns
Speakers’ claims: Speakers analyze AMD's record Q4 revenue of $13.4 billion and aggressive guidance, noting a structural shift where data center AI revenue surpassed server revenue. They highlight AMD's strategic moves, including the Helios rack-scale systems and acquisitions of Talas and World Labs, aimed at capturing agentic AI and physical AI workloads. Despite strong fundamentals, the stock fell 17% due to market fears regarding regulatory risks, peak capex narratives, and widening credit spreads. The speakers contrast AMD's high valuation multiples with NVIDIA's lower multiple, citing concerns over free cash flow and customer concentration. Ultimately, the hosts view the sell-off as a political overreaction to a strong earnings report, though they maintain a cautious stance due to execution risks.
Palantir Might Be the Most Expensive Bargain in the Stock Market
Episode discussion: Futurum Equities analysts argue Palantir is undervalued based on PEG ratios, margin expansion, and AIP adoption.
Speakers’ claims: Hosts Shay and Tom Nash of Futurum Equities argue that Palantir Technologies is currently undervalued, citing a PEG ratio of 1.3x compared to peers trading at 4.5x-4.6x. Nash highlights Palantir’s operational leverage, noting operating profit margins expanded from -107% five years ago to 43% in 2026, while stock-based compensation dropped from ~50% to 14%. The speakers emphasize the 'boot camp' deployment model and the AIP platform’s ability to integrate proprietary data, which drives net revenue retention above 150%. They contend that cheaper AI models act as a tailwind via the Jevons Paradox, increasing demand for Palantir’s governance layer. Finally, Nash presents a DCF model yielding a $917 price target and expresses confidence in Palantir reaching a trillion-dollar market cap.