My attention was drawn to a troubling trend: at one of Amazon's logistics facilities in Las Vegas, a mass digitization of books is underway, which, in essence, means their physical destruction. This is not about simple scanning, but a barbaric procedure: bindings are cut off the editions, they are separated into individual sheets, and passed through industrial scanners. Such books cannot be restored after processing.
How the scheme was uncovered
The starting point was anomalous orders at used bookstores. Sellers recorded wholesale purchases of hundreds of books on a wide variety of topics and years of publication, including genuine rarities. In July, one store received an order for roughly a thousand editions through the Biblio platform from an anonymous buyer. The seller agreed to track the fate of the batch by hiding an Apple AirTag tracker inside one of the books.
The signal led directly to an Amazon facility in North Las Vegas, where the VGT3 division is located. Analysis of employee posts and their personal testimonies confirmed: this is not a warehouse, but a workshop for processing printed materials. Workers describe a clear conveyor line: registration of copies, scanning of barcodes and ISBNs, cutting of bindings, and subsequent digitization of pages.
What the corporation is hiding
Notably, Amazon had not previously publicly advertised this activity. In response to a request, the corporation only confirmed that it purchases printed editions through commercial channels and uses them for "developing and improving products and services." However, key details—purchase volumes, program timelines, and the specific AI models for which the data is being prepared—remained undisclosed.
In early 2026, supply disruptions were even observed on the platform: books occasionally ran out, which indirectly indicates the scale of the operation. It is obvious that Amazon has joined the global trend of AI companies buying up physical books as a source of high-quality textual data for training language models.
My analysis: This situation is a vivid symptom of a systemic problem in the industry. We are witnessing how corporations, in the race for data, destroy cultural heritage without offering authors and publishers adequate compensation. As long as the AI industry follows the path of least resistance, buying up and grinding down rare books, the question of the legality and ethics of such an approach will remain a ticking time bomb for the entire market.