Amid Advertising Week New York, LUMA’s Kawaja Says AI is Turning Adtech Into Cage Match
Artificial intelligence may make advertising more efficient, but Terence Kawaja doesn’t think efficiency is where the real money will be made.
Kawaja, founder and chief executive of investment bank LUMA Partners, said AI will increasingly automate media buying and other advertising workflows. The bigger prize will go to companies that combine AI with proprietary data to make better decisions about where marketers should spend their money.
“The industry’s getting this one a little bit wrong,” Kawaja said in an interview with Pooja Midha, reporting on behalf of Beet.TV and the New York Stock Exchange. “What differentiates is decisioning.”
That distinction could determine who emerges from what Kawaja predicts will be a messy restructuring of advertising technology. AI may lower the cost of operating the machinery, but it also could expose which companies have something valuable under the hood.
Efficiency is nice, but effectiveness pays
Kawaja expects much of media buying to migrate from programmatic systems toward what he called “agentic rails,” with AI handling more of the routine workflow. But if everyone gets access to roughly the same productivity improvements, efficiency itself won’t provide much competitive advantage.
Instead, Kawaja said companies need proprietary signals that can be combined with algorithms to produce better outcomes. He contrasted incremental efficiency improvements with the potentially much larger gains that come from effectiveness.
“That’s good, but it doesn’t differentiate,” Kawaja said of workflow improvements. “You combine proprietary data or signals and the right algorithms and you can get better decisioning, which means better outcomes.”
Investors may already be placing their bets. Kawaja noted that the largest technology platforms have been early beneficiaries of AI and enjoy valuations that advertising agencies and publishers can only admire from several tax brackets away.
Speaking from the New York Stock Exchange, Kawaja also nominated a rather obvious candidate for the world’s best prediction market.
“Everyone says Polymarket and Kalshi are the best prediction markets. No. The stock market is the best prediction market,” he said. Technology companies command some of the market’s biggest valuations and highest multiples while agencies and publishers have lost ground, he noted.
Adtech still has too many cooks
AI also could help force a long-awaited consolidation of ad tech, according to Kawaja.
The sector has gone through the usual explosion of company formation but never completed the less enjoyable part of capitalism in which many of those companies disappear, merge or discover that a logo on the LUMAscape isn’t necessarily a business model.
Kawaja described adtech as suffering from “persistent fragmentation,” with thousands of companies contributing to inefficient economics and an opaque supply chain. Those conditions create what he calls “dark corners” where arbitrage, fraud and made-for-advertising inventory can flourish.
Ultimately, he expects the industry to end up with fewer companies handling larger volumes at lower take rates and delivering better quality.
“It’s inevitable,” Kawaja said.
Welcome to the cage match
Getting from thousands of companies to fewer companies, naturally, may not resemble an orderly queue at the DMV.
“AI will create a UFC-like cage match,” Kawaja said.
Agencies, demand-side platforms, supply-side platforms and retail media networks increasingly overlap as they chase the higher-value intelligence layer of advertising. Simply executing a media buy is becoming a lower-margin activity. Figuring out what should be bought, for whom and why is where companies hope to defend their margins.
Kawaja pointed to tensions between Publicis and The Trade Desk as an example of companies that historically occupied different parts of the advertising ecosystem increasingly competing with one another.
He even claims to have warned everyone, albeit musically.
Three years ago, Kawaja made a parody video called “Don’t Stop Competing” about the trend. “I swear to God, you could take my parody videos about the industry, take the harmony away, take the music away and just read the lyrics, and I’d stand by that as a thesis,” he said.
Investment banking research departments everywhere can presumably stand down.
M&A separates the story from the storytellers
The changing competitive landscape also has implications for mergers and acquisitions, a subject Kawaja knows rather well from LUMA Partners’ advisory work.
“The harsh reality is that 90-plus percent of startups fail,” Kawaja said, defining failure as not generating meaningful positive returns for investors.
He sees three broad categories emerging: companies that produce little or no return, those that wind up in middling consolidation or “cleanup” transactions and companies that achieve genuine strategic exits. Some subscale public companies also are candidates for rationalization because, in his view, they simply shouldn’t be public.
The companies that attract strategic buyers tend to identify important trends early and execute against them. Kawaja added another requirement on the spot: they need to explain themselves.
“The narrative on that story has to be given and understood by a broad audience and be compelling,” he said.
Apparently even in the age of artificial intelligence, telling a good story remains annoyingly resistant to automation.
Independents may get another shot
The giant technology platforms aren’t about to disappear. Kawaja said the walled gardens have dominated digital advertising because they offer scale, ease of use and performance.
Still, with more than $1 trillion flowing through the broader advertising ecosystem, the portion outside those platforms remains substantial. AI could give independent ad-tech companies better tools to compete.
Agentic workflows may simplify an industry notorious for its thickets of vendors. Meanwhile, independents with valuable proprietary data could combine those signals with increasingly powerful algorithms to improve performance.
“If that happens, we’ll have a more democratic ecosystem, which I think will accrue benefits to many,” Kawaja said.
Whether democracy is the natural result of a UFC cage match remains to be seen.
Next stop: DMS
Kawaja will have another opportunity to test his thesis when LUMA Partners hosts its Digital Media Summit on Nov. 9.
The event is expected to draw about 400 people, including more than 300 CEOs, to discuss how AI is accelerating changes across media, commerce and marketing and how those shifts are translating into corporate strategy and dealmaking.
His closing forecast required considerably fewer charts.
“Buckle up,” Kawaja said. “It’s gonna be fun.”
Artificial intelligence may make advertising more efficient, but Terence Kawaja doesn’t think efficiency is where the real money will be made.
Kawaja, founder and chief executive of investment bank LUMA Partners, said AI will increasingly automate media buying and other advertising workflows. The bigger prize will go to companies that combine AI with proprietary data to make better decisions about where marketers should spend their money.
“The industry’s getting this one a little bit wrong,” Kawaja said in an interview with Pooja Midha, reporting on behalf of Beet.TV and the New York Stock Exchange. “What differentiates is decisioning.”
That distinction could determine who emerges from what Kawaja predicts will be a messy restructuring of advertising technology. AI may lower the cost of operating the machinery, but it also could expose which companies have something valuable under the hood.
Efficiency is nice, but effectiveness pays
Kawaja expects much of media buying to migrate from programmatic systems toward what he called “agentic rails,” with AI handling more of the routine workflow. But if everyone gets access to roughly the same productivity improvements, efficiency itself won’t provide much competitive advantage.
Instead, Kawaja said companies need proprietary signals that can be combined with algorithms to produce better outcomes. He contrasted incremental efficiency improvements with the potentially much larger gains that come from effectiveness.
“That’s good, but it doesn’t differentiate,” Kawaja said of workflow improvements. “You combine proprietary data or signals and the right algorithms and you can get better decisioning, which means better outcomes.”
Investors may already be placing their bets. Kawaja noted that the largest technology platforms have been early beneficiaries of AI and enjoy valuations that advertising agencies and publishers can only admire from several tax brackets away.
Speaking from the New York Stock Exchange, Kawaja also nominated a rather obvious candidate for the world’s best prediction market.
“Everyone says Polymarket and Kalshi are the best prediction markets. No. The stock market is the best prediction market,” he said. Technology companies command some of the market’s biggest valuations and highest multiples while agencies and publishers have lost ground, he noted.
Adtech still has too many cooks
AI also could help force a long-awaited consolidation of ad tech, according to Kawaja.
The sector has gone through the usual explosion of company formation but never completed the less enjoyable part of capitalism in which many of those companies disappear, merge or discover that a logo on the LUMAscape isn’t necessarily a business model.
Kawaja described adtech as suffering from “persistent fragmentation,” with thousands of companies contributing to inefficient economics and an opaque supply chain. Those conditions create what he calls “dark corners” where arbitrage, fraud and made-for-advertising inventory can flourish.
Ultimately, he expects the industry to end up with fewer companies handling larger volumes at lower take rates and delivering better quality.
“It’s inevitable,” Kawaja said.
Welcome to the cage match
Getting from thousands of companies to fewer companies, naturally, may not resemble an orderly queue at the DMV.
“AI will create a UFC-like cage match,” Kawaja said.
Agencies, demand-side platforms, supply-side platforms and retail media networks increasingly overlap as they chase the higher-value intelligence layer of advertising. Simply executing a media buy is becoming a lower-margin activity. Figuring out what should be bought, for whom and why is where companies hope to defend their margins.
Kawaja pointed to tensions between Publicis and The Trade Desk as an example of companies that historically occupied different parts of the advertising ecosystem increasingly competing with one another.
He even claims to have warned everyone, albeit musically.
Three years ago, Kawaja made a parody video called “Don’t Stop Competing” about the trend. “I swear to God, you could take my parody videos about the industry, take the harmony away, take the music away and just read the lyrics, and I’d stand by that as a thesis,” he said.
Investment banking research departments everywhere can presumably stand down.
M&A separates the story from the storytellers
The changing competitive landscape also has implications for mergers and acquisitions, a subject Kawaja knows rather well from LUMA Partners’ advisory work.
“The harsh reality is that 90-plus percent of startups fail,” Kawaja said, defining failure as not generating meaningful positive returns for investors.
He sees three broad categories emerging: companies that produce little or no return, those that wind up in middling consolidation or “cleanup” transactions and companies that achieve genuine strategic exits. Some subscale public companies also are candidates for rationalization because, in his view, they simply shouldn’t be public.
The companies that attract strategic buyers tend to identify important trends early and execute against them. Kawaja added another requirement on the spot: they need to explain themselves.
“The narrative on that story has to be given and understood by a broad audience and be compelling,” he said.
Apparently even in the age of artificial intelligence, telling a good story remains annoyingly resistant to automation.
Independents may get another shot
The giant technology platforms aren’t about to disappear. Kawaja said the walled gardens have dominated digital advertising because they offer scale, ease of use and performance.
Still, with more than $1 trillion flowing through the broader advertising ecosystem, the portion outside those platforms remains substantial. AI could give independent ad-tech companies better tools to compete.
Agentic workflows may simplify an industry notorious for its thickets of vendors. Meanwhile, independents with valuable proprietary data could combine those signals with increasingly powerful algorithms to improve performance.
“If that happens, we’ll have a more democratic ecosystem, which I think will accrue benefits to many,” Kawaja said.
Whether democracy is the natural result of a UFC cage match remains to be seen.
Next stop: DMS
Kawaja will have another opportunity to test his thesis when LUMA Partners hosts its Digital Media Summit on Nov. 9.
The event is expected to draw about 400 people, including more than 300 CEOs, to discuss how AI is accelerating changes across media, commerce and marketing and how those shifts are translating into corporate strategy and dealmaking.
His closing forecast required considerably fewer charts.
“Buckle up,” Kawaja said. “It’s gonna be fun.”
