Million-dollar Amazon sellers surge past 100,000 as the marketplace consolidates

New data shows the number of Amazon sellers generating over a million dollars a year has climbed past 100,000 today, up from roughly 60,000 four years earlier, even as the total count of active sellers fell by 10 percent. The marketplace is consolidating: fewer sellers, higher revenues for those who remain.

The concentration is striking at the top. More than 230 sellers now generate over 100 million dollars annually, up from 50 four years ago. Traffic per seller is up 31 percent since 2021 thanks to reduced competition, and in the US just 2 percent of sellers account for over half of marketplace revenue. Among million-dollar US sellers, 57 percent are China-based and 39 percent American, while the 100-million-dollar tier remains mostly US-based.

Walmart, meanwhile, announced an AI shopping agent called Sparky designed to replace the traditional search bar entirely, handling tasks from building weekly grocery lists to furnishing an apartment on a budget. The deployment spans 230 million weekly customers, and unlike Amazon's walled-garden approach, Walmart is designing its agent to interoperate with external AI assistants, a bid to become the hub of AI-mediated shopping.

PayPal's chief executive added his voice to the agentic commerce chorus, predicting AI agents will drive 25 percent of e-commerce spending by 2030. The company is releasing agent toolkits and protocol servers, with partnerships across major AI platforms, positioning itself as the payment infrastructure for machine buyers.

On the AI model front, ChatGPT now fields 2.5 billion prompts daily, a volume that has more than doubled since December 2024, and its next major model release is expected imminently, promising automatic tool selection, a much larger context window and stronger coding and reasoning. For sellers, that means AI shifting from assistant to operator.

A content-strategy framework also made the rounds: instead of spreading across every social platform, map the 31 platform-and-format combinations, cut everything without algorithmic discovery, find each platform's hero content type, match it to where your customer actually spends time, then run one engine with daily consistency for six months. One jewelry business applying it grew from 800 to 45,000 followers and lifted monthly revenue fourfold, with email doing the converting. The trivia answer closing the issue: a mere 1 percent of users follow a link embedded within an AI search summary.