Amazon’s World Product Graph plan aims to catalog nearly every product on Earth
Amazon logo (public domain) via Wikimedia Commons
An internal document reviewed by Reuters describes a 2027 target to keep shoppers researching on Amazon instead of Google or AI agents
Amazon wants to become the place you look up a product, even when Amazon doesn’t sell it. An internal planning document reviewed by Reuters describes an effort called the World Product Graph. The goal is to gather detailed information on almost every item for sale anywhere.
The document sets an aspirational public launch date of October 1, 2027. Amazon, for its part, has described the document as an outdated brainstorming memo. The company says it does not reflect its current strategy.
What the plan reportedly involves
Amazon would assemble a database covering nearly every product sold on the planet. It would cover items whether or not Amazon carries them.
To build it, the plan relies on two main sources. The first is data feeds supplied directly by brands and retailers. The second is web crawling, the automated process of scanning websites to collect information at scale.
Some big consumer brands are reportedly already on board. According to the document, Coca-Cola and Bayer are contributing data. Procter & Gamble took part in a trial summit tied to the effort.
Why Amazon cares so much about product data
The stated aim is to stop shoppers from drifting elsewhere to do their homework. Right now, people often research products on Google, online forums and brand websites before buying.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
There is also an AI problem. The internal document reportedly says flawed AI answers and high costs make richer catalog data critical.
That points to Rufus, Amazon’s AI shopping assistant. Rufus has reportedly fallen behind competitors in accuracy and functionality.
People familiar with the matter told Reuters a broader product database could also blunt AI shopping agents and protect advertising revenue. Those agents are autonomous tools that can browse, compare and potentially buy on a user’s behalf.
The search battlefield
Amazon already starts from a strong position. Figures included in the research on the plan put Amazon’s share of direct searches at 41%. Google sits at 23%, with Walmart at 18%.
What this means for retailers, brands and investors
For brands, the calculation could get complicated. Feeding data to Amazon may improve how products appear in its search results and AI answers. It also hands a powerful retailer an even deeper view into the market, including goods sold through rival channels.
The participation of Coca-Cola and Bayer suggests at least some large brands see more upside than risk.
The caveats are just as clear. Amazon has called the document an outdated brainstorming memo. The October 1, 2027 launch date was explicitly labeled aspirational.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

