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

Nexxen's First EBITDA Grow in Four Quarters Comes With an Agentic Asterisk

Record CTV and enterprise numbers hide a 9% blended growth rate, a wide Q4 guide, and a home-screen promise that only just started paying
NEXN · Earnings Call · 2026-11-12

The quarter that was mechanically meant to happen

Nexxen's Q3 was, on paper, the best quarter it has ever printed on the top line: contribution ex-TAC of $100.3 million (+8%), programmatic revenue of $97.8 million (+9%), and a third consecutive quarterly record in CTV revenue at $34.1 million, up 39%. Adjusted EBITDA rose 22% to $34.5 million — a 34% margin against contribution ex-TAC versus 30% a year ago — and operating cash flow jumped to $46.1 million from $35.8 million. Cash stood at $152.9 million with no long-term debt. The most important sentence on the call, though, was the least promotional. After “several consecutive quarters of year-over-year decline, adjusted EBITDA returned to growth and margin expanded in Q3” — Sagi Niri, CFO or similar financial executive · 2026-11-12. Sagi Niri then explained why that was always going to be true this quarter, regardless of execution:

It's the first quarter that we are really reporting on an apples-to-apples basis. So a lot of partnerships and infrastructure investments around hosting and data that we did in the second half of 2025 are now fully in place.

Sagi Niri, CFO or similar financial executive · 2026-11-12
That is a real inflection — the comp base stops being hostile from here, and management guides to a >40% adjusted EBITDA margin in Q4 and roughly 34% for the whole of 2027. But investors should be clear-eyed: part of this "turn" is arithmetic catching up with a spending cycle, not a step-change in demand.

Two speeds inside one nine

The composition underneath that 9% is the actual story. Enterprise contribution ex-TAC — customers running the Nexxen DSP self-serve or platform-managed — grew over 40% year-on-year. Data products rose 73%, PMPs 18%, display 11%, mobile 9%. Against that, desktop fell 29%, and management was blunt that this is structural, not cyclical. Niri expects desktop to keep shrinking: “people are finding them less and less into websites and more and more into different LLMs. So that's something that we are preparing to and understanding and knowing that it will go lower in 2027” — Sagi Niri, CFO or similar financial executive · 2026-11-12. That is the same open-web traffic migration the market has been pricing into the whole ad-tech complex — the market's own keyword board carried Legacy Search into its top-75 this past spring, and Ziff Davis-type publishers have been flagging AI overview impact all year. Nexxen is not discovering a theme; it is disclosing, in dollars, how much the theme is costing it. The whole strategy — CTV, home screen, mobile web retreat into mobile in-app — is an explicit flight away from formats that LLMs erode. On the same call, Ofer Druker noted in-app now represents the majority of mobile activity, up from a minority a year ago. The scale of the escape hatch matters too. Enterprise and strategic are roughly the same size today in contribution terms; Niri's guess is that “probably in the next 3 years, enterprise will be 70%, 80% of our net revenue” — Sagi Niri, CFO or similar financial executive · 2026-11-12. If that happens, blended growth converges toward the enterprise rate. If it doesn't, 9% is the run rate.

The longest-running promise finally pays — a little

Nexxen has been selling the smart-TV home screen for four quarters. On the May call, Druker sized the prize at ten-and-a-half minutes of daily user exposure to the operating-system interface and called the TAM “huge” — Ofer Druker, Chief Executive Officer · 2026-05-13. The keyword history shows how central the pitch became: Nexxen TV Home Screen entered the company's own keyword lexicon as a top-five theme, and "Home Screen" was its single highest-momentum term in the June quarter. For three quarters, revenue was essentially zero. On this call came the underwhelming-but-honest admission: after minimal onscreen revenue through Q3 2026, there has been "a notable uptick in adoption and spending in Q4," with Q4 called the first quarter of notable revenue and 2027 the real catalyst year. That is the fourth consecutive call in which the home screen is the next thing rather than the current thing — and it is also the first in which it is no longer zero. Worth noting what faded. "The Trade Desk" was once a top-three keyword for this company and sits in its momentum decliners list; this quarter the former marquee home-screen launch partner is one name inside a list that also includes Google's DV360, StackAdapt and Epsilon. Likewise the VIDAA/"V" relationship — the subject of a $60 million investment now equating to roughly 6% of that business — has quietly left the keyword board even as Nexxen committed a further $15 million in the quarter. Partnerships are being normalized into plumbing, which is what maturity looks like.

Agents, acquisitions, and a buyback on hold

The genuinely new material is agentic. Druker made it the frame of the entire prepared remarks:

Advertising will be transacted more and more by AI agents in the future, and Nexxen is well positioned to help enable and define their future.

Ofer Druker, CEO · 2026-11-12
Eight agency and holding-company partners are now testing or using Nexxen inside their own AI environments via MCP, and management expects most of its platform workflow to be executable agentically by the end of Q1. This is an escalation, not a debut: in May, Druker was publicly skeptical that agents would disintermediate people — “I am not sure it will replace people that are salespeople because at the end of the day, need people to create a relationship in this stage” — Ofer Druker, Chief Executive Officer · 2026-05-13. Six months later the agentic solutions pitch is the whole equity story, with a deliberately vague "there will be a few business models" on monetization. And it is not company-unique. Across the last five days of reporters, agentic language appeared at 6098.T, PUBM, TIT.MI and RPAY alike, and it has been a fixture on the global keyword board for multiple quarters — the agentic AI economy framing, personal agents, agent stacks. Nexxen is riding a wave, not creating one; its differentiated claim is the end-to-end transaction data that feeds the agents, which is credible but unproven in revenue. Capital allocation carries its own small tension. The $40 million buyback authorization was extended, yet zero shares were repurchased in Q3, with management preferring to stay liquid while hunting bolt-on acquisitions — the first M&A push in roughly four years, and pointedly aimed at mobile in-app capabilities rather than the "we have everything we need" technology stance of a year ago. Meanwhile the long-flagged non-programmatic tail is being dismantled: the RhythmInfluence influencer business is wound down, and the remaining non-programmatic line is under strategic review.

What to watch

Two analysts pressed on the Q4 guide, which implies a wide band of growth despite an October that management called its best month ever — “October was our best month ever” — Sagi Niri, CFO or similar financial executive · 2026-11-12, Niri said, adding that the range was kept wide because "you can never know what will happen." Nexxen has raised guidance three times in 2026 and is now reaffirming a ~12% contribution ex-TAC year at the midpoint. A record quarter that produces a reaffirmation rather than a fourth raise, paired with a cautious geopolitical tension caveat, tells you management does not yet trust its own momentum. The bull case now rests on three things that only show up in 2027: home screen revenue becoming material, enterprise compounding toward 70-80% of net revenue, and agents actually being paid for rather than demoed.