The Brief: Amazon's Logistics, Private Labels Get a Makeover, and AI Giants Want to Own Your Shopping Cart
In today's Azimuth brief: Amazon officially turned its logistics empire into a product any business can buy. Walmart and ALDI are rethinking what their store shelves look like from the inside out. And Anthropic, OpenAI, and Google are all trying to solve the agentic commerce puzzle with completely different pieces.
Today's stories:
1. Amazon opens its entire logistics network to any U.S. business
2. Why major retailers are redesigning their private-label packaging
3. Where Anthropic, OpenAI, and Google each stand in the agentic commerce race
Amazon Opens Its Logistics to Everyone
Amazon launched Supply Chain Services this week, opening its logistics infrastructure to any U.S. business, not just sellers on its marketplace. The offering bundles freight forwarding, customs brokerage, warehousing, transportation, and last-mile delivery into a single package. P&G, 3M, American Eagle, and Lands' End are among the early adopters.
The network Amazon is putting to work: 200-plus fulfillment centers, 80,000 trailers, 24,000 intermodal containers, and 100 cargo aircraft. All of it was originally built to run Amazon's own retail operation. Now they're monetizing it.
FedEx dropped 9%, and UPS fell 10% on the news, and the concern is straightforward. Amazon can price aggressively because the fixed costs of this infrastructure are already covered by the core business. Competing with that as a traditional shipper is a tough spot.
For retailers and brands, the real question is a trust one. Handing your logistics over to a company that competes with you on the shelf is a trade-off every supply chain team is now going to have to think through. The cost savings might make sense. The strategic exposure is the part that takes a little longer to work out.
Private Label Gets a Packaging Refresh
Walmart, ALDI, and Target have all done significant redesigns of their private-label packaging recently, and the rationale is the same across all of them: the products are good, but the packaging does not reflect that.
Walmart's overhaul of Great Value is the most ambitious of the group. The brand spans nearly 10,000 grocery and consumer products, and this is its first full design refresh in over 30 years. Walmart's own customer research found that shoppers liked the quality and the price, but "didn't particularly feel proud to display it in their home or with their families." That is a useful thing to know about your flagship store brand.
ALDI went through a similar refresh to unify the look of all its private-label products under a consistent visual identity. Target updated its Up&up line in 2024 with cleaner, bolder packaging. The pattern is not a coincidence. These retailers are treating their store brands less like budget fallbacks and more like actual brands they want shoppers to pick on purpose.
Private label has been gaining share for years, mostly driven by price. The interesting shift now is that the competition is moving up the value chain. When a store brand looks good sitting out on the counter, the old calculus of "name brand vs. store brand" starts to blur. National brands have been dealing with the price challenge for a while. The design challenge is a newer front.
Anthropic, OpenAI, and Google Are Running Very Different Plays in Agentic Commerce
Digital Commerce 360 put out a solid breakdown of where the three major AI labs stand in agentic commerce, and the most notable thing about it is how different their approaches are from each other.
Anthropic ran an internal experiment called "Project Deal," where 69 employees used AI agents to negotiate with each other in a test marketplace. The bots closed 186 deals totaling more than $4,000. Small dollar amounts, but it is the clearest proof of concept yet for agent-to-agent commerce working in practice. Anthropic also donated MCP, its Model Context Protocol, to the Linux Foundation, which is a significant move on the infrastructure layer. (If you want more background on where this is all heading, I've written about the agentic AI opportunity in retail here.
OpenAI moved away from its original in-ChatGPT checkout experience and pivoted to a ChatGPT app model, where retailers integrate directly into the platform. The Agentic Commerce Protocol it built with Stripe, is still active, and Target, Sephora, Nordstrom, Lowe's, Best Buy, and Home Depot are already plugged in. It is a different architecture than where they started, but the retailer list is serious.
Google is going deeper into enterprise retailer partnerships. Ulta Beauty and Macy's both have live AI shopping agents built on Google's technology. It has multiple active protocols in the market: UCP for product discovery (built with Shopify), AP2 for payments, and A2A for agent-to-agent coordination. Google bets that its existing retailer relationships and ad infrastructure give it an edge that the others cannot easily replicate.
None of these approaches has won yet. The fact that each company is building its own protocols means the underlying infrastructure of agentic commerce is going to be fragmented for a while. For anyone in ecommerce, understanding which protocols your tech stack supports is going to matter sooner than most people expect.
That's it for today.