The Maze: AI could drive nearly one in six US online retail sales by 2030. The base-case forecast rises from 3.2% in 2026 to 15.8%, almost five times the share. But this is a wider category than orders placed inside a chatbot. It includes shopping sessions involving a retailer's own assistant. That distinction changes the strategic reading: a bigger AI channel does not automatically mean independent AI platforms take the customer away. Some of that growth can happen inside the stores shoppers already use.
The share grows by 12.6 percentage points. The base case moves from 3.2% to 15.8% between 2026 and 2030. That is a 4.94-fold share, or roughly a 394% relative increase. Those are different ways of describing the same forecast, not separate pools of growth. The denominator is US retail ecommerce, not all retail spending. Both endpoints are projections. Reading the first as a measured baseline and the second as a promised outcome would make the forecast look more certain than it is.
The definition reaches well beyond chatbot checkout. The original evidence includes purchases through general-purpose AI platforms and attributable links, retailer sessions involving an embedded shopping assistant, and specialized third-party assistants. A shopper can therefore remain on a retailer's site while the sale falls inside the AI-driven category. This measures a defined role in the shopping journey. It does not establish that the assistant created demand that would otherwise have disappeared, or completed every step without human involvement.
Retailers are inside the forecast, not merely exposed to it. The report author's explanation expects most AI-driven sales to involve retailer-native assistants, citing Amazon's Alexa for Shopping and Walmart's Sparky. Major external platforms such as ChatGPT and Gemini would originate less than half. No exact split is publicly disclosed here. The implication is conditional but useful: brands need to understand product discovery within retailer assistants as well as recommendations from external platforms. The forecast does not justify treating either route as the whole market.
The missing numbers should stay missing. Intermediate annual percentages and every dollar value are redacted in the public exhibit. The visible endpoints support a large shift in share, but not a precise annual budget schedule or a dollar opportunity calculation. Even in 2030, 84.2% of ecommerce remains outside this defined category under the base case. That is not proof those purchases have no exposure to AI anywhere in the business. It is the remainder after applying a specific commerce definition, and a reminder that a rapidly growing channel can still be a minority of sales.
Why it matters: Treat AI commerce as an attribution and distribution question before treating it as a new revenue pool. Separate sales involving your own assistant from referrals arriving through outside platforms, then test whether either improves conversion or creates incremental demand. The forecast gives a reason to prepare for both. It does not tell a retailer to surrender its customer relationship, or a brand to move nearly one sixth of its budget into chatbots. A share of sales is not a spending instruction.
