In short: Guest (Nick, 42% of his portfolio, ~$18 avg cost, +~50%): an "agnostic" marketing cloud built on a proprietary first-party identity graph (552M global / 240M US IDs refreshed with 6,000+ signals a day via owned data rails — Disqus, LiveIntent, its pixel) that pairs with a client's data in a clean room to predict return on ad spend before a dollar is spent. NRR 114→120 (128 with Marigold, top spenders 132–134); wallet share of Fortune-100 clients 1.7% and rising ~0.5–0.68pt a year; a Palantir go-to-market channel (2 of a 20-deal pipeline signed); SBC falling while FCF rises. Accumulated heavily sub-$20, would sell ~$46. Carlson: "a really good pitch" — 36% growth at a 26 PE and <5× sales "is not a super overvalued company"; frames it as "the Palantir of the demand side," with the caveat that "there hasn't been any next Amazon."
Zeta sells big companies a single software hub where they keep their customer data, design ad campaigns and push them out to wherever people actually are — Facebook, Google, email, streaming TV, even ChatGPT. That kind of hub is called a "marketing cloud," and older versions from Salesforce or Adobe were stitched together from acquisitions, so data has to hop between systems. Zeta's pitch is that its one-piece design reacts in milliseconds.
The part the guest thinks is special is that Zeta owns its own giant database of about 240 million Americans' anonymized profiles, fed by services it bought (the Disqus comment system, the LiveIntent newsletter network). When a brand joins, Zeta matches the brand's customer list against that database to learn what a typical buyer looks like, then finds more people like them — and can estimate the return on an ad before the brand spends anything. The brand's own data is never shared with other clients, and the longer it stays, the more useful the system gets, which makes leaving expensive.
The investment case is growth without new customers: Zeta already works with about half the Fortune 100 but gets under 2% of their marketing budgets, and existing clients keep spending more (every $100 of last year's revenue from a group of customers became ~$120). Add a new sales partnership with Palantir and a young business-intelligence product, and the guest thinks Wall Street's forecast of sharply slowing growth is too low. He owns a very concentrated position (42% of his portfolio). The risks: a privacy crackdown that limits data collection, the Marigold acquisition flattering growth, and the Palantir channel not delivering. Carlson liked the story and the price (about 26 times earnings for a company growing 36%) but noted how hard the business is to understand.
In short: "Zeta closed up 12% at a 52-week high as analysts raised price targets… KeyBank increased the target to 27 from 22. The company had early traction for Athena, especially among customers with over 1 million in ARPU, which is a growth engine for them."
Zeta sells marketing software that helps large companies target customers. It closed up 12% at a 52-week high as analysts raised targets — KeyBank went from $27 up from $22 — on early traction for Athena, its newer platform, particularly among its largest customers (those spending over $1 million a year), which is where the growth compounds fastest.
Full passage: premium transcript (PDF).
In short: Pro pick #2 (the transcript garbles the ticker as "ZA"). New York ad-tech with net cash, still founder-led, proprietary databases built "back when AI was called machine learning"; sells to enterprises and the agencies (Publicis/WPP/Omnicom) and to companies that don't want to hand proprietary data to Meta/Google. Just did a JV with Palantir — "a very nice endorsement." "We think it can double or triple in size."
Zeta helps companies spend their advertising money better. It owns large databases about consumers that it built up over years, and it combines those with a client's own customer data to work out which ads to send to whom. It sells both to big companies directly and to the advertising agencies (Publicis, WPP, Omnicom) that buy media on their behalf.
Morrison's screen was deliberately not "who sells AI" — he thinks that trade (Nvidia and the other "picks and shovels") is over. He asked his team for companies that use AI on data nobody else has. Zeta fits: it has more cash than debt ("net cash," so it can't be forced into trouble by lenders), it's still run by its founder, and it benefits from a trust problem — plenty of advertisers don't want to hand their customer data to Google or Meta, so they'd rather work with a neutral party. A newly announced joint venture with Palantir is the outside validation, though he's careful to say Zeta won't ever be valued like Palantir. He thinks the business "can double or triple in size."
54:58So Zeta is, headquartered in New York, has net cash on their balance sheet. They've built proprietary databases, back when AI wasn't called AI and it was called machine learning. Yes. — Right. and they sell to enterprises and they sell to agencies like the Publixes of the world and WPPs and omniccoms who are desperately trying to figure out ways to add value or add ROI on advertising spend.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.