The Maze: Ecommerce has been dying for 26 years. Conveniently, it forgot to stop growing. James Fredley's timeline puts every supposed killer on the same US revenue curve: dotcom bust, Amazon dominance, Shopify, mobile, DTC fatigue, Covid math, iOS privacy, TikTok Shop, and now AI agents. The pattern is useful because it is not really about death. It is about handoffs. Each wave changes who captures margin, who owns demand, and who pays the tax. The market keeps moving.
The visible punchline is that the obituaries sit on an upward line. The source visual starts near the dotcom bust and ends above $1.2 trillion of US ecommerce revenue, with each panic label attached to a higher point than the last. `Amazon will eat all` appears around the early scale-up phase. `Shopify killed custom` and `mobile killed desktop` sit around the mid-2010s climb. `Covid broke ecom math` lands before the steep pandemic acceleration. `iOS 14.5 killed DTC` appears near the trillion-dollar zone. The story is not that each fear was silly. It is that each fear described a real constraint that operators then priced, routed around, or built into the stack.
The category survived by changing the layer where competition happens. Amazon did not kill ecommerce; it professionalized marketplace expectations. Shopify did not kill custom commerce; it made storefront software cheap enough for more merchants to try. Mobile did not kill desktop; it moved conversion pressure into smaller screens and faster checkout. TikTok Shop did not kill Amazon; it turned discovery into a transaction surface. AI agents will likely do the same. They may compress search, comparison, and product discovery. But the order still needs a merchant record, payment authorization, inventory promise, delivery path, and return policy.
The harder question is who gets squeezed during each survival cycle. The post's own examples point to a recurring migration of power. When traffic gets more expensive, DTC brands lose operating leverage. When privacy rules reduce signal quality, acquisition teams pay for weaker targeting. When marketplaces add ad products, sellers pay rent to be seen. When AI agents mediate shopping, brands may face another toll booth between intent and checkout. The channel lives. The margin pool moves.
The caveat is that growth is not the same as platform stability. One visible comment pointed to Ramp's vendor data as a sign that custom and developer-led options are still gaining attention in web hosting and site-building stacks. That does not refute Fredley's point. It sharpens it. Ecommerce can compound while Shopify, custom builds, Webflow, Vercel, marketplaces, and AI interfaces fight over the next control point. The right question is not whether ecommerce dies. It is which layer becomes the new tax collector.
Why it matters: Death narratives are usually lazy strategy. They make a channel shift sound like a category collapse. The better operating question is where the bottleneck moved. In 2026, AI shopping agents probably will not erase ecommerce. They may change who owns the customer relationship before checkout. That is a smaller claim. It is also the one worth budgeting for.
