404 Media tracked a rare book to an Amazon warehouse that cuts bindings off to scan pages

404 Media hid a tracking device in a rare book and followed where it went. The signal ended at an Amazon warehouse in Las Vegas.

What the reporters found

Employees at the site told the outlet that the unit does one thing. It receives massive shipments of printed books, cuts the bindings off so the pages can be scanned faster, and destroys the physical copies in the process.

The team has a name, VGT3, and a logo: a dinosaur holding a book. The investigation was published on August 17, 2026.

Amazon told 404 Media it “purchases books through commercial channels” to improve its products and services.

Why printed books became valuable

The appeal is in what these texts are not. Books that predate the AI boom are clean human writing that never passed through the internet, which makes them premium training material with no risk of feeding a model its own output back.

As the open web fills with generated text, that property gets harder to source and more valuable to own. A printed book from 2004 is a guarantee of provenance in a way that no web crawl can be.

What separates this from scraping

Web scraping copies. This consumes. Some of the volumes moving through the warehouse are rare or out of print, which means a destroyed copy may have no replacement anywhere.

The company that started as a bookstore is now processing books as raw material, and the arithmetic works only in one direction. A scan can be made from a book. A book cannot be made from a scan.