Adam Foroughi's bet: LLM ad products will cannibalize Google Search, not discovery platforms

Adam Foroughi, co-founder of AppLovin, places LLM-based advertising directly in competition with Google’s search business. Not adjacent to it, not upstream of it. The claim is that the loop a user closes when querying a large language model with purchase intent is the same loop Google Search has monetized for two decades, and that the two will therefore compete for the same advertiser dollars.

The reasoning behind the call is structural. Search advertising derives its value from intent: a user arrives with a specific question or need, and an ad placed at that moment reaches someone already in motion toward a decision. Foroughi’s argument, stated plainly, is that large language models now replicate that intent-capture mechanism. When a user asks an LLM what to buy or where to go, the query carries the same commercial signal that has made search inventory the most valuable real estate in digital advertising. If LLMs can close that loop, they sit in the same category as Google Search, not in the category of social feeds or display networks, where the audience is browsing rather than querying.

That framing has a specific competitive implication. Discovery-oriented advertising, the kind that reaches users before intent has formed, operates on a different dynamic. Foroughi’s call does not suggest LLMs threaten that market. The threat, as he states it, is narrower and more pointed: it goes to the segment of advertising that depends on capturing expressed intent at the moment of query.

External reporting from the last year gives some texture to where things currently stand. Analysis published by M+C Saatchi Performance in April 2026 notes that ChatGPT now draws roughly 800 million weekly active users, describing it as fertile ground for intent-driven search. The same analysis observes that OpenAI has not yet mounted a serious challenge to search ad dominance. A piece from Search Engine Land from March 2026 describes Google as likely to monetize its own AI Mode lightly and gradually, grinding down LLM competitors on price while keeping high-value transactional queries inside its own environment. Early LLM ad products, as PYMNTS reported in late 2025, launched without functioning targeting, attribution, or conversion metrics, which kept the highest-value commercial queries inside Google’s ecosystem for longer than the competitive narrative suggested they would.

None of that history contradicts Foroughi’s call. It sets the baseline from which the competition he describes would have to develop. The structural argument does not require LLMs to be competitive with Google Search today. It requires that the mechanism is the same, and that advertisers will eventually follow intent wherever it concentrates.

The falsifiable element of the call is the “almost exclusively” framing. Foroughi is not saying LLMs will pick off a slice of search budgets. He is saying the gravitational pull runs almost entirely toward that category and away from discovery advertising. A reader who follows ad budget allocation over the next two to three years has a clear test: if significant LLM ad revenue comes primarily at the expense of search spend rather than social or display spend, the call lands. If LLM advertising grows as an upper-funnel, discovery-adjacent product without meaningfully compressing search budgets, it does not.

What makes the bet worth watching is that most of the industry conversation has framed LLM advertising as additive, a new layer sitting above search rather than a direct substitute. Foroughi’s position, which comes from running a large advertising business, goes in the opposite direction. He is not describing a complementary market. He is describing a collision, and naming Google Search as the thing that gets hit.