AI Agents vs Amazon's Ad Toll
Stuart Kearney, co-founder of Vetted, on AI agents in shopping
The core point is that Amazon is no longer just a store, it is a toll road where brands pay multiple layers of fees to reach the shopper. Amazon books revenue not only from marketplace commissions and fulfillment, but also from sponsored listings that sellers buy to win placement inside search results. That turns product discovery into a paid auction, and pushes a much larger share of each order back to Amazon than plain retail margin alone would suggest.
-
Amazon reported $156.1B of third party seller services revenue and $56.2B of advertising services revenue in 2024. Those are separate lines, but both are tied to marketplace activity, which shows how much seller spend now sits on top of the product transaction itself.
-
For a seller, the money stack is concrete. A brand can pay a referral fee, FBA fulfillment and storage fees, then sponsored product spend just to stay visible for a search like blender. In practice, seller guides often frame total Amazon fees around 30% to 40% before the brand’s own product costs.
-
This is the opening Vetted is aiming at. If an AI shopping agent decides what to buy before the shopper reaches Amazon search results, the brand may need less marketplace advertising. The battle shifts from paying for placement inside Amazon to winning recommendation upstream, where trust and relevance matter more than bid price.
The next phase of commerce will be a fight over who controls the starting point for purchase decisions. If more shopping begins in AI assistants, marketplaces keep fulfillment and checkout, but they risk losing the highest margin layer, which is paid discovery inside the marketplace itself.