Industry story
Kawaja: 90%+ of Adtech Startups Fail; Narrative Matters for M&A Exits
ai-in-adtech antitrust m-and-a programmatic
Kawaja offered a blunt M&A outlook for ad-tech: more than 90% of startups will fail to generate meaningful returns for investors. He segmented the field into three buckets — companies returning little or nothing, those absorbed in low-value 'cleanup' consolidations, and the minority that achieve genuine strategic exits. He also flagged that some subscale public ad-tech companies lack the scale to justify being public and are candidates for take-private or roll-up deals.
For companies hoping to land in the strategic-exit bucket, Kawaja said early trend identification and strong execution are necessary but not sufficient — they must also be able to tell a compelling story to a broad audience. The implication for JWX and peers is that proprietary data assets and a clear narrative around AI-enabled decisioning are table stakes for attracting premium acquirers.
Analysis
Showing the shorter version.
Terry Kawaja of LUMA Partners said at Advertising Week New York what he says every year: more than 90% of ad-tech startups won't return real money. The new addition is the prescription. To land a premium exit, you need early trend-spotting, strong execution, and a story a broad audience buys.
On the narrative point, Kawaja is at least partly selling his own services. Strategic acquirers run diligence teams. They don't miss a good business because the pitch deck was thin. They pass because ARR growth stalled or gross margin is weak. A compelling narrative on a mediocre asset gets you a slightly better cleanup price, not a strategic exit. Where narrative does move things is on the borderline assets, which happen to be most of the market. If your numbers are strong, the story is a tiebreaker. If your numbers are marginal, the story is the difference between "interesting acquisition" and "we'll take it off your hands."
The structural picture is what matters more. Subscale public ad-tech names, Criteo, Innovid, Viant, and Zeta Global, trade at depressed multiples because public investors won't pay to hold middle-of-the-stack infrastructure that looks like it's waiting to be acquired. The Google antitrust remedy adds uncertainty for every intermediary sitting between buyer and seller, which pushes those multiples down further. A scaled buyer or PE shop can come in at one to two times revenue, strip the public-company overhead, and fold the asset into something bigger. That's the trade. The longer a subscale company holds out for a strategic multiple it may not deserve, the more leverage shifts to the buyer.
Two questions decide where an operator lands. Do you own data a buyer can't get elsewhere? Is your recurring revenue still growing? If both are yes, you have time and options. If either is no, narrative buys you a marginally better cleanup price. Before spending a dollar on exit storytelling, pressure-test the data moat the way an acquirer's diligence team will. If it folds under an hour of scrutiny, skip the narrative work and go sell the assets while they still have a bid.
Our call: At least two of Criteo, Innovid, Viant, and Zeta Global announce a take-private or strategic acquisition by mid-May 2027, because their multiples already price in the outcome and buyers with clean balance sheets can time offers into the weakness. Medium confidence; deal timing depends on buyers moving, not on whether the pressure is real.
Terry Kawaja of LUMA Partners stood up at Advertising Week New York and said what he says every year: more than 90% of ad-tech startups won't return real money. The new wrinkle is the advice attached. To land a premium exit, you need early trend-spotting, strong execution, and a story a broad audience actually buys. The question for operators: does narrative move the exit price, or is that just LUMA selling its own bankers?
What's being decided: Nothing, by you, today. This is a read on where the ad-tech M&A market is heading and what subscale companies should do about it. Easy to undo, because no one's committing to anything on Kawaja's say-so. What sets the clock is the next funding round or the next strategic buyer walking in the door, whichever comes first.
The Market Analyst
Kawaja is describing a buyer's market, and the buyers already know it. Subscale public ad-tech names (Criteo, Innovid, Viant, Zeta) trade at depressed multiples because public investors won't pay to hold infrastructure that's waiting to be acquired. That's the take-private and roll-up trade: a scaled buyer or a PE shop buys the whole company at one to two times revenue, strips the public-company cost, and folds it into something bigger. The Google antitrust remedy adds uncertainty for every intermediary in the middle of the stack, which pushes multiples down further. In plain terms: if your company sits in the middle and can't say why it survives Google's reshaping, the market has already marked you down.
The Skeptic
Kawaja has been calling 90% failure for a decade. It's a structural truism, not a forecast, and dressing it in "narrative matters" is convenient for a firm that sells narrative. The advice gets the causality backwards: strategic acquirers run diligence teams. They don't miss a good business because the deck was weak. They pass because ARR growth stalled or gross margin is thin. The story follows the numbers; it doesn't rescue them. A great narrative on a mediocre asset gets you a slightly better cleanup price, not a strategic exit. For a non-specialist: good businesses sell themselves, and a nice slideshow doesn't turn a weak one into a strong one.
The Operator
Whether your data is actually yours is the question that decides your exit bucket. Kawaja's "cleanup consolidation" category is where you land when you wait too long and the acquirer sets the price. So the pressure lands now, on BD and product teams, to crystallize AI-enabled differentiation before the next round closes. The casualty is the pure-play SSP or data-management platform with no proprietary data moat, stress-tested on why it exists at all once Google's remedy reshapes the pipes. If your answer to "why do you exist" is "we're cheaper plumbing," you're a wind-down candidate. For a non-specialist: companies that just move other people's data between buyers and sellers, without owning anything unique, are the ones in trouble.
The CFO
Opportunity cost is what the math comes down to. Every quarter you spend polishing a strategic-exit story is a quarter you're not spending on the two things that actually set the price: growing recurring revenue and holding margin. Kawaja's own buckets tell you the base rate. Most companies return little, a slug get absorbed cheap, a minority exit well. If you're not already in that minority on the numbers, the rational move is to stop burning cash chasing a premium and take the cleanup deal while a buyer still wants the assets. Waiting costs you leverage. The seller who waits lets the acquirer name the price.
Where they split
The Skeptic and the Strategist genuinely disagree on one thing: does narrative change the exit price, or just describe it? The Skeptic says acquirers buy numbers and the story is window dressing. The Strategist says AI-enabled decisioning is now the minimum pitch, and a company that can't tie proprietary data to optimization doesn't get a seat at the premium table at all.
Both are right about different companies. If your numbers are strong, the story is a tiebreaker. If your numbers are borderline, the story is the difference between "interesting acquisition" and "we'll take it off your hands." Narrative doesn't move a great asset much. It moves the borderline ones, which is most of the market.
The second split is on timing. The Market Analyst and the Operator both see the consolidation clock running now, set by cost of capital and Google uncertainty. The CFO says that clock is exactly why you don't wait: the longer you hold out for a strategic multiple you may not deserve, the more leverage shifts to the buyer.
What it hinges on
Two facts decide an operator's move. One: do you own data a buyer can't get elsewhere? Two: is your recurring revenue still growing? If both are yes, you're in the minority and you have time. If either is no, narrative buys you a marginally better cleanup price and nothing more. The council leans toward the Skeptic and the CFO: the numbers set the bucket, and the story only moves you within it.
Before you spend a dollar on exit storytelling, pressure-test the data moat the way an acquirer's diligence team will. If it folds under one hour of scrutiny, skip the narrative work and go sell the assets while they still have a bid.
Prediction: At least two of Criteo, Innovid, Viant, and Zeta Global will announce a take-private or strategic-acquisition deal by the end of Q1 2027 earnings season (mid-May 2027).
Confidence: Medium — the structural pressure is real, but deal timing depends on buyers moving.
Why: These four trade at depressed multiples that make them cheaper to buy whole than to build around, and public investors have stopped paying to hold middle-of-the-stack infrastructure that looks like it's waiting to be acquired. Kawaja is naming out loud what the multiples already say: subscale public ad-tech lacks the scale to justify being public, which is the exact profile PE roll-ups and scaled strategics target at one to two times revenue. The Google antitrust remedy adds enough uncertainty to intermediaries that a buyer with a clean balance sheet can time an offer into the weakness. The opposite outcome (all four stay independent through spring 2027) requires buyers to sit on their hands while assets sit at distressed prices, which is not how cost-of-capital-driven consolidation usually plays out.
Revisit by 2027-05-15: We're right if at least two of Criteo, Innovid, Viant, or Zeta Global announce a definitive take-private or strategic-acquisition agreement by the close of Q1 2027 earnings season. We're wrong if one or none of them does.
Also covered this issue
-
Apple iOS 27 WebKit Block List Threatens Programmatic Ad Ecosystem
adexchanger
Apple's undisclosed WebKit block list permanently impairs Safari web identity without user consent, forcing operators to verify vendor exposure before iOS 27 ships broadly.
-
Omnicom Media Partners Rembrand for AI-Powered In-Content Streaming Ads
digiday
Omnicom's Rembrand deal tests whether agencies can control premium CTV placements directly, bypassing publisher auctions and squeezing out independent identity vendors.
Comments