The verdict: not dead, just misread
The "online arbitrage is dead" narrative assumes every worthwhile product is mobbed by dozens of sellers and dominated by Amazon. The data from 342,931 real products says the opposite. Nearly 44% have three or fewer sellers, and only about a quarter are genuinely crowded with ten or more. Amazon itself only shows up as a seller on 16% of these listings, leaving 84% without that competition.
Demand is not the problem either. 94.8% of the products tracked sold at least once in the trailing 30 days. The opportunity that "died" is still selling, still uncrowded, and still mostly Amazon-free for the sellers who know where to look.
What actually changed is not saturation. It is that finding these products by hand got harder, so casual sellers quit and assumed the model was dead. The data shows the products are still there.
Methodology
Sample: 342,931 products observed across active Amazon seller storefronts monitored by Arbitrage Stalker between March 12 and June 6, 2026, using Keepa data. This is the live inventory of working arbitrage sellers, not a random sample of the entire Amazon catalog, so it reflects the real opportunity set sellers are operating in. "Seller count" is the number of new offers on a listing. "Sold in the last 30 days" means the product's sales rank dropped at least once in the period (Keepa's standard proxy for a unit sold). Amazon competition is flagged when Amazon holds an active offer. Raw aggregate data available to journalists on request.