Open in interactive viewer → charts, metric popovers & call review

Zeta's AthenaOS Land Grab: The Marketing Cloud Grew a Second Act

A 21st straight beat-and-raise was the headline; the real story is a headless-architecture reversal, eight Palantir yeses, and Google search budgets finally moving.
ZETA · Earnings Call · 2026-11-09

The Beat Was Loud; the Narrative Changed

Zeta's Q3 was a blowout by any measure: revenue $493 million, up 46% (30% excluding Marigold and political candidate revenue), adjusted EBITDA $121 million at a 24.6% margin, free cash flow $82 million, and GAAP EPS of $0.14. Management called it the 21st consecutive beat-and-raise and raised full-year revenue to $1.855 billion. But the numbers were not the surprise. The surprise was how far CEO David Steinberg pushed the story beyond marketing software. return on marketing spend remains the engine, but the new pitch is enterprise intelligence — using Zeta's Data Cloud and AthenaOS to solve problems in inventory, supply chain, fraud detection, even cybersecurity. The Q3 call introduced an AI scorecard that is now central to how management wants investors to value the business. AI scorecard metrics showed over 50% of super-scaled customers were monthly active Athena users, up from just over 40% in Q2. Athena adopters accounted for more than 70% of revenue. The top 25 Athena users grew revenue more than 70% in Q3 versus 46% company-wide. And net revenue retention for Athena adopters ran over 500 basis points above overall Zeta. This is not boilerplate AI enthusiasm; it is a disclosed usage funnel that management says it also uses internally as a forecasting tool.

Athena Is No Longer a Demo

A year ago Athena was a promise. In Q2 2026, management was still selling the early innings. On that call, Chris Greiner noted that “heavy users had a 400 basis point higher net retention rate than our other clients for Athena today.” — David A. Steinberg, Chief Executive Officer (CEO) · 2026-08-04 Three months later the premium is 500 basis points. That progression matters. It suggests Athena is not just a UI layer but a consumption expander. Steinberg described the dynamic bluntly: clients who adopt Athena are “moving budget faster than at any point in our corporate history” — David Steinberg, Chief Executive Officer · 2026-11-09. Client campaign development more than doubled quarter over quarter, which management treats as a top-of-funnel indicator for activation revenue. The financials confirm the operating leverage. In the last filed quarter (Q2), revenue was $443 million, up 44% year over year, while operating margin reached 3.8% and net income turned positive at $8 million. The Q3 call pushes those trends further: adjusted EBITDA grew 55%, free cash flow grew 73%, and GAAP EPS was $0.14. The old knock on Zeta was that it grew fast but not profitably. That knock is getting stale.

The Headless U-Turn

The most revealing exchange came near the end, when Scott Berg asked about Zeta's new headless MCP server. In Q2, Steinberg had dismissed headless architectures: “I don't think it becomes a competitive force for us, and I don't think it is something that we would move into quickly from our front.” — David A. Steinberg, Chief Executive Officer (CEO) · 2026-08-04 On this call, he reversed himself.

yes, this is a pivot for us. I have always been we need to control everything, and now I'm saying the UI can actually exist in a third-party platform.

David Steinberg, Chief Executive Officer · 2026-11-09
That is a genuine strategic change, not a messaging tweak. Athena will plug into ChatGPT, Claude, Gemini and other frontier models, while Zeta keeps its data cloud and activation layer proprietary. The UI becomes someone else's; the intelligence and execution stay Zeta's. Why does it matter? Because it turns the large language models from a threat into a distribution channel. Zeta does not need every enterprise to live inside its interface. It can meet customers inside the tools they already use. That is a bigger addressable market and a lower-friction land motion, but it also puts Zeta in a more crowded interface layer. The moat now has to be the data and the activation, not the dashboard. This is Zeta's enterprise version of the market's Personal AI agents wave.

Palantir, Braze, and Google: Three External Shocks Zeta Is Monetizing

Three outside events are converging on Zeta's P&L. First, Palantir. In Q2, Steinberg said the first two joint deals closed at a “100% hit rate. Net with 2, close 2.” — David A. Steinberg, Chief Executive Officer (CEO) · 2026-08-04 On this call he said Zeta is up to 8 yeses and that the first Palantir pilot will become a super-scaled client this month. His line: “we have not left without a yes yet.” — David Steinberg, Chief Executive Officer · 2026-11-09 Management still expects the partnership to add more than $100 million in incremental annual revenue in the years to come, and it is not in the pipeline count. Second, the Salesforce-Braze deal. Steinberg argues that neither platform is AI-native and that integrating Braze could take up to two years, right in the middle of a replacement cycle for legacy marketing clouds. Zeta already wins more than 50% of RFPs it enters; RFPs were up more than 100% year over year. The Braze acquisition gives Zeta a window to argue that customers should not wait for a two-year integration. Third, Google. This is the most underappreciated thread. Steinberg said 68% of Google queries are now answered on-platform by Gemini and clicks have fallen to roughly 32% from 90%-plus. That is a structural problem for search-dependent brands — and a budget shift toward Zeta. “for the first time that we've been doing this... we're starting to see meaningful budget move to Zeta from what was in search originally.” — David Steinberg, Chief Executive Officer · 2026-11-09 The Google dynamic is not new to Zeta's calls, but the tone has changed from tailwind to budget is actually moving.

The Price of Admission

The stock has noticed. Zeta is up 60.4% over the last 90 trading days and sits within 4.5% of its all-time high, a sharp reversal from the 57% drawdown earlier in its history. Yet the valuation is not stretched by its own history: price to revenue is 3.1x, versus a 7.6x peak in 2024. The market is paying up for growth, but not euphorically. The tension is in the guide. For 2027, Chris Greiner signaled the same 20% to 21% organic starting point, with the usual 2% to 5% cushion. That is below the 30% organic growth just delivered. Management's message is that the core business is durable and that AthenaOS, Palantir, OpenAI and Braze displacement are upside not yet in the number. If that is right, the current multiple is reasonable. If the partnerships stay in pilot purgatory, the stock is already discounting a lot of enterprise-intelligence hope. For now, the evidence — eight Palantir yeses, a doubled campaign-development metric, 500 basis points of NRR uplift — leans toward the former.