Why did Amazon block Meta's shopping agent while opening its own?

Ayeshha argues that the opposite moves by Amazon and Shopify toward AI shopping agents are not, as "most people" frame it, a fight between an anti-AI company and a pro-AI one, but "a dispute about who gets to stand between you and the buy button, and charge rent for the spot." On September 20, 2026, Amazon began blocking Meta's agent Muse from shopping on amazon.com, citing, per reporting, that Meta never disclosed Muse would shop the store, that the agent does not identify itself while browsing, and that it "appears to capture and store customer credentials." Eight days later, Shopify extended its WebMCP standard to checkout itself, giving agents three tools — get_checkout, update_checkout, complete_checkout — to inspect, modify, and place an order once the buyer authorizes it, and its CEO announced Shopify was "partnering deeply" with Muse to enable agentic checkout via Shop Pay.

The essay locates the real stakes in Amazon's ad revenue: Amazon earned "more than $68 billion from advertising in 2025," a business that depends on shoppers scrolling past sponsored placements, and "an agent that skips straight to checkout never sees the sponsored shelf" — a risk Fortune had already flagged. Ayeshha treats Amazon's stated security objections as real but incomplete, since Amazon's own cross-site buying agent, Alexa for Shopping's "Buy for Me" feature (launched May 13, 2026, replacing the Rufus chatbot), completes purchases on other retailers' sites using the customer's stored address and payment method — the same behavior Amazon objects to when Muse does it. The author also notes Amazon has "done this dance before": ChatGPT, Gemini, and Perplexity's Comet are reportedly blocked too, and Amazon won and then lost an injunction against Comet after an appeals court ruled on August 4 that it was "Comet's users, not Perplexity, doing the accessing" — leaving, in the author's words, "the legal question of whether you can delegate your own shopping to a robot ... genuinely unsettled."

This confirms and extends Can an AI agent serve both merchant and user interests fairly?, which already names the Buy-for-Me asymmetry — Amazon is happy to be the agent, so long as it's the one collecting rather than losing the fee. Ayeshha adds two facts Azhar's account does not carry: Shopify's countervailing move to open checkout to agents industry-wide, not just Muse, and the unresolved legal status of agent-delegated shopping raised by the Comet litigation — evidence that the gatekeeping fight is being contested in courts and commercial deals, not settled by any single platform's policy. It also sits in contrast to Can a protocol split agent access from agent incentives?: where that protocol proposes a negotiated, cross-industry standard for agent-business access, Amazon's blocking of Muse shows a dominant platform enforcing access unilaterally and selectively — shutting out a competitor's agent while running its own.

The piece is a single Medium essay built from secondhand reporting — Bloomberg, CNBC, TechCrunch, Fortune, and The Information are each cited for a specific fact, but none is independently verified here, and the author is explicit about an incentive to find Amazon's wall reassuring: "I make money when a human clicks an Amazon link and buys something. A robot can't do that." The excerpt does not establish that Amazon's stated security concerns are pretextual, only that its enforcement is inconsistent with its own agent's behavior elsewhere; the better-supported claim is that platform policy toward shopping agents tracks who captures the transaction's revenue at least as much as it tracks any stated safety rationale.