The Contest Moved Upstream
For twenty years, retail's digital battle was a fight for the click. Whoever captured the search query, the product page view and the add to cart button captured the sale. That era is quietly ending. As AI shopping assistants and their more autonomous cousins, agentic commerce systems, insert themselves between the shopper and the transaction, the decisive moment moves upstream, away from the click and toward the context in which the agent forms its recommendation. Amazon and Walmart both understand this, and their July maneuvering shows two of the largest companies in the world repositioning for a contest whose rules they are still writing in real time.
The shift matters because context is a more durable asset than a click. A click is a single, contested event that any competitor can bid for. Context, the accumulated understanding of what a household needs, when it reorders, what constraints govern its choices, is a compounding advantage that gets harder to dislodge the longer it is held. Whoever owns the context in which an AI agent operates effectively owns the purchase decision before it is consciously made. Amazon and Walmart are racing to become that context layer, and they are pursuing it from opposite ends of the strategic spectrum.
Amazon's Vertically Integrated Bet
Amazon's approach is to own everything. Its model stitches together the assistant, the marketplace, Prime, the payments credential, fulfillment, reviews, advertising and post purchase service into a single vertically integrated stack. The company is using Alexa for Shopping to help consumers find deals, compare products, track prices, set alerts and, most tellingly, automatically buy items when they hit a target price. That last capability is the quiet revolution. It captures consumer intent before a transaction exists, converting a future purchase into a standing instruction that Amazon alone is positioned to fulfill, because Amazon controls every layer the instruction touches.
The logic of vertical integration is control. By owning the full stack, Amazon can guarantee that its agent operates in an environment it designed end to end, with no dependency on a third party interpreting shopper intent. The company held 9.3 percent of US consumer retail spending in the first quarter of 2026, up from 8.6 percent a year earlier, and the ambient, always on nature of its assistant driven model is built to extend that lead. When the agent lives inside the same system as the catalog, the payment and the delivery, friction approaches zero, and zero friction is the whole point of agentic commerce.
Walmart's Distributed Counter
Walmart is playing a different game, and deliberately so. Rather than building a closed stack, it partnered with Google to connect Gemini's conversational interface to Walmart and Sam's Club products, store inventory, membership benefits, account history and fulfillment options. The bet is distributed rather than integrated. Walmart is wagering that it can win situational commerce, the dinner that needs planning, the household that needs restocking, the event that needs supplies, the urgent need that must be met today, by meeting shoppers wherever they already are rather than compelling them into a proprietary assistant. Its model spans stores, clubs, grocery, local inventory, Walmart plus and third party AI discovery.
The distributed approach trades control for reach. By plugging into Gemini, Walmart accepts that Google's agent, not its own, mediates part of the shopper relationship, and that is a real concession. But it gains presence in an AI surface that hundreds of millions of people already use, without having to convince them to adopt yet another assistant. Walmart held 7.8 percent of US retail spending in the first quarter, and its physical footprint gives it a situational advantage Amazon cannot easily replicate, the ability to fulfill from a store a few miles away. The question is whether reach through a partner beats control through ownership. Neither company knows yet.
The Behavior Is Already Shifting
This is not a speculative future being fought over prematurely. The behavioral data shows the ground already moving under retailers' feet. Forty seven percent of ecommerce shoppers used AI during their most recent purchase, a figure high enough that AI assisted shopping can no longer be treated as an early adopter curiosity. More striking still, ChatGPT's share as a product research tool climbed from 2 percent to 30 percent in just two years. That is a fifteen fold increase in how often a general purpose AI, owned by neither retailer, sits between a shopper and a buying decision. The context layer is being contested precisely because shoppers are already using it.
For brands and smaller retailers watching the giants maneuver, the 30 percent figure is the one that should command attention. When nearly a third of product research runs through an AI assistant, visibility inside that assistant becomes as strategically important as shelf placement or search ranking ever were. The shopper who once typed a query and scanned results now asks an agent and receives an answer, often a single recommendation. Being the product that the agent names, rather than one of ten links it might have shown, is a fundamentally different and more winner take all competition. The intermediary has changed, and so have the stakes.
The Choice Every Retailer Now Faces
The Amazon versus Walmart divergence crystallizes the strategic choice confronting every retailer and consumer brand. One path is to own the agent layer directly, building or controlling the AI interface through which customers shop, as Amazon is attempting. The other is to distribute through third party AI platforms, accepting mediation in exchange for reach, as Walmart is doing with Google. Each carries a distinct risk. Ownership demands scale and capital most companies lack and risks building an assistant no one adopts. Distribution risks disintermediation, ceding the customer relationship and the margin to whoever owns the agent, and waking up as a commodity supplier to someone else's interface.
There is no universally correct answer, which is exactly why this is interesting to watch. Amazon's scale makes vertical integration viable in a way it simply is not for a mid sized retailer. Walmart's physical network makes distribution a strength rather than a surrender. The right move depends on assets a company actually holds, not on which giant's strategy looks more impressive in a headline. What is not optional is having a strategy at all. The context layer is being built now, by these two companies and by the AI platforms sitting between them and their customers. Retailers who do not decide where they stand will find the decision made for them.
