A famous forecast says ad-driven shopping beats AI-shopping-agent revenue tenfold — but most of that gap comes from guessed profit margins, not market size.

This explores how much the 'retail media wins' conclusion depends on the margin percentages behind it. The corpus has one note that does the calculation directly, and its arithmetic shows something easy to miss. Lipsman applies a 60% operating margin to retail media and a 10% margin to agent-driven ecommerce, and gets $29 billion versus $2.8 billion by 2029, a tenfold gap Does retail media or agentic commerce offer bigger margins by 2029?. Work backward and the revenue behind those figures is about $48 billion for retail media and about $28 billion for agent-platform sales. That is less than a twofold difference. Most of the tenfold gap comes from the margin assumptions, not from the size of the markets. If you assume agentic commerce earns 20% instead of 10%, the gap halves. So the real question is what kind of business agentic commerce turns out to be.

The 10% figure treats agentic commerce as ordinary retail: you sell goods and keep a thin slice. Other notes suggest the profits may sit somewhere else. BCG's survey projects $1.3 trillion in spending shifting to agent-assisted shopping, with consumers and the platform networks capturing most of the gains while merchants face a structural disadvantage Who wins and loses as AI agents reshape shopping?. Waller argues the market will split into open and closed systems, much like payment networks Will agentic commerce platforms adopt open or closed models?. In a closed system, whoever runs the agent could take a toll on every transaction, and that kind of business looks much more like the high-margin side of the comparison. A single blended 10% hides the fact that agentic commerce may be low-margin for retailers and high-margin for platforms.

The comparison also assumes retail media and agentic commerce are two separate businesses. The corpus suggests they could merge or eat into each other. Hedges warns that shopping agents funded by advertising have built-in reasons to favor paid placements Do advertising-funded shopping agents favor paid product placements?. If that happens, ad-style margins move into the agent layer. The weakness also runs the other way. Retail media's 60% margin depends on ads that sway shoppers. Agents act on very different cues. They show strong position biases that persist even in text-only interfaces Do AI shopping agents show persistent position bias in text interfaces?. They react to price and rating changes far more sharply than humans do Do shopping agents respond more strongly to price and rating cues?. And a platform's ranking rules can combine with sellers' tuned product descriptions to more than double consumer harm Can platform ranking and seller descriptions separately harm AI shoppers?. 'Media' sold to agents may be a different product from media sold to people, and it may not hold a 60% margin.

There is also a theoretical reason to expect margins to shrink. A formal model finds that an optimal recommendation algorithm's welfare edge over seller competition gets smaller as more sellers compete, approaching zero in crowded markets Does algorithmic recommendation lose its edge as sellers compete more?. Data portability works in the same direction. If regulators treat it as basic infrastructure, as the CMA argues, consumers can switch agents easily, which limits how much any platform can charge Should data portability be infrastructure for agent-driven markets?.

The corpus has only one note that runs the margin numbers. Everything else bears on which assumptions are believable, not on alternative forecasts. The useful takeaway is that 'retail media is ten times bigger' mostly reflects the assumed margins, not the size of the markets. Those margins depend on who controls the agent and whether agents can be advertised to the way people can.