---
title: "Global Buyout of Used Books: How a 2025 Court Ruling Sparked a 'Gold Rush' for AI Training"
description: "📚 **Global Book Buyout for AI**: Bookstores worldwide are receiving anomalous orders for used books. The cause is a 2025 US court ruling that legalized the use of books for training neural networks. 🤖 📉 **Market in Shock**: Orders cover everything—from Latin to westerns. But there is a dark side: some books are destroyed after digitization. What do collectors think, and how will this affect culture?"
date: 2026-08-17T23:34:40.000Z
lang: en
url: https://xab.info/en/posts/global-book-buyout-for-ai-training-2026-en
tags: [artificial-intelligence, publishing, anthropic, copyright-law, book-industry, data-scraping]
publisher: "XAB.info"
---

# Global Buyout of Used Books: How a 2025 Court Ruling Sparked a 'Gold Rush' for AI Training

![Shelves filled with old paper books, symbolizing raw material for AI training after the 2025 court ruling](https://xab.info/media/2026/08/18/global-book-buyout-for-ai-training-2026/global-book-buyout-for-ai-training-2026-1.webp)

## 🎯 Key Points

- Bookstores are recording an unprecedented surge in orders for used books.
- A 2025 US court ruling in the Anthropic case legalized the purchase of books for AI training.
- AI companies are seeking rare and diverse texts to improve language models.
- Ethical disputes are arising due to the potential destruction of books after their digitization.

In August 2026, the book industry faced an unprecedented phenomenon: bookstores around the world began receiving anomalously large orders for used books, which were then shipped to remote warehouses. Stuart Manley, owner of the historic Barter Books in Northumberland (England), who has been in the business for three decades, told the BBC in an interview that he had never seen anything like it. According to him, one recent order from a Canadian company included between 2,000 and 3,000 copies—a volume comparable to the entire store's weekly sales.

### The 2025 Court Precedent as a Market Catalyst

Analysts link this surge in demand to a key court ruling issued in the US in 2025. Judge William Alsup ruled that the use of books acquired by the company Anthropic to train its Claude chatbot is "substantially transformative" and, therefore, does not violate US copyright law. This ruling effectively legalized the mass purchase of physical media for subsequent digitization and training of large language models (LLMs). As a result, AI development companies received the green light to purchase massive datasets, including rare and obscure texts.

### Diversity as the New Data Currency

Orders are not limited to bestsellers or classics. Buyers are showing interest in the widest range of topics: from obscure Latin works to westerns and specific technical literature. Experts note that this diversity is a key factor for modern AI. Rare and unusual texts provide unique training data that cannot be obtained from publicly available internet sources, allowing models to better understand context, linguistic nuances, and specific fields of knowledge.

### Contradictory Data

Despite the positive economic impact for sellers, the situation raises serious ethical concerns. Recently released court documents caused a stir as it was revealed that some books were destroyed during the training process of Anthropic's chatbot. The company confirmed that purchasing textbooks is a common practice and stated that their project is not aimed at destroying rare or valuable books. However, collectors and bibliophiles remain concerned: the risk of damage or loss of unique copies after digitization remains a real threat. While store owners see this as a chance to sell off stagnant stock, the cultural community fears the loss of physical heritage.

### Economic Prospects and Risks for the Industry

For bookstore owners, such as David Tobin of Walden Books in North London, this trend has a dual nature. On the one hand, the ability to sell books that have been sitting in warehouses for years is a great opportunity for business. On the other hand, the thought that this knowledge could be destroyed after digitization causes deep regret. The industry is at a crossroads: on one side—technological progress and new sales markets, on the other—the threat of losing material culture for the sake of virtual data.

## ❓ FAQ

### Q: Why are companies buying old books?
**A:** Companies use physical books for digitization and training AI language models because they contain unique data not available in the open internet.

### Q: Is this legal?
**A:** Yes, according to a 2025 US court ruling, the use of acquired books for AI training is considered "transformative" and does not violate copyright.

### Q: Are books destroyed after purchase?
**A:** There are concerns and reports that some books are destroyed after digitization, although companies state they try to avoid destroying rare copies.