Key Takeaways
- Secondhand booksellers in the UK, Ireland, the US, Australia and continental Europe are receiving unusually large, non‑thematic bulk orders from opaque buyers.
- The orders span a wide range of titles—from Estonian translations of John le Carré to 18th‑century African agricultural manuals—suggesting a machine‑driven selection process.
- Booksellers suspect AI companies are purchasing these volumes to scan their text for training data, a practice already documented with Anthropic’s Claude model.
- Despite paying “top whack” (full price) for bulk lots, buyers are not seeking discounts, indicating a willingness to absorb cost for data acquisition.
- The trend is disrupting the traditional secondhand book market, which relies on thematic lots and modest pricing, and raises questions about the future availability of pre‑2022 print material for AI training.
Surge of Unexplained Bulk Orders
Over the past three months, secondhand booksellers across the United Kingdom and Ireland have reported a sudden influx of large, varied book orders that defy the usual patterns of trade. “It’s definitely a strange combination of books,” said Stuart Manley, co‑owner of Barter Books in Alnwick, Northumberland. “Normally a book order is on a theme but this is all over the place.” Similar spikes have been noted by sellers in the United States, Australia and continental Europe, pointing to a global phenomenon rather than an isolated local glitch.
Stuart Manley’s Observations
Manley revealed that his shop has sold “hundreds” of random books to three unidentified buyers for a total of about £4,000 since the orders began arriving three months ago. One recent order exemplified the eclectic nature of the requests: an Estonian translation of John le Carré’s The Mission Song, a specific‑imprint copy of Anne Brontë’s Agnes Grey, and the October 1983 edition of the magazine Warship. The lack of thematic cohesion led Manley to speculate that “masses and masses of AI money [are] being squandered” on acquiring these volumes.
Irish Bookseller’s Perspective
Further west, Jim Shaughnessy, owner of MW Books in Claregalway, Ireland, told the Guardian that his business has been receiving “large, varied orders since May.” He described the orders as “sporadic, presumably machine‑driven and have little in terms of thematic currency; books on agricultural implements in 18th century Africa through to biographies of 1950s racing car drivers.” While he refrained from judging customers’ intentions, Shaughnessy admitted the selections were “more than curious” and difficult to reconcile with typical buying habits.
Bath Bookseller’s Experience
David Gower‑Spence of BookLovers of Bath reported a rush of orders between May and July for roughly 200 books from buyers who have also appeared in other sellers’ logs. One of those repeat buyers, the Canada‑based Zoom Bills, describes itself as “North America’s leading book recycling company.” Gower‑Spence noted that the volume was “overwhelming for a while” despite his organisational systems, underscoring the scale of the demand.
Opaque Buyers and Heathrow Link
An anonymous UK‑based bookseller, speaking on condition of anonymity, observed that some random orders could originate from different buyers yet be shipped to the same address. “The aliases behind these orders are very opaque even to experienced booksellers,” the source said. A postcode seen by the Guardian and used by multiple purchasers points to a cluster of freight warehouses near Heathrow airport, suggesting a centralised logistical hub for the purchases. The same bookseller reported receiving orders for 6,000 books from similar buyers since January, worth thousands of pounds, and emphasized that the buyers were paying “top whack” and not seeking discounts despite the bulk nature of the requests.
International Echoes
The pattern is not confined to the British Isles. The Guardian noted that Australian booksellers have also experienced a burst of orders, with one describing the sudden scale of demand as “driving me nuts.” Sellers in the United States and across continental Europe have echoed similar reports, indicating that the phenomenon is widespread and likely driven by a common external factor rather than regional market fluctuations.
AI Training Hypothesis
The leading explanation among booksellers is that artificial‑intelligence firms are acquiring these texts to scan their contents for training data before pulping the books. This theory gained credence when The Washington Post revealed that Anthropic, the developer of the Claude chatbot, had spent tens of millions of dollars acquiring books, slicing off their spines, and scanning the pages before sending the remnants for recycling. Anthropic defended the practice, stating that “Claude is trained on a mix of publicly available web data, commercially acquired datasets, and data we generate ourselves. Sourcing books is a widely used approach for training large language models across the AI industry. None of our data acquisition programs buy and destroy rare or antiquarian books.” Nonetheless, the scale of purchases described by booksellers suggests that at least some companies are pursuing aggressive data‑gathering strategies.
Industry Commentary and Public Data Concerns
Additional fuel for speculation came from a report by the tech news site 404 Media, which cited a books database claiming that “the world’s best AI training data is sitting on a shelf.” The database, ISBNdb, noted that books published before 2022 are especially attractive because their text remains unpolluted by AI‑generated content. An online forum for booksellers featured a post lamenting, “Shame these are all destined for AI training, and will therefore get the covers stripped and ultimately end up being pulped.” Other contributors puzzled over why buyers would request titles freely available in the public domain—such as Gulliver’s Travels (1726)—if the goal were solely to obtain unique, hard‑to‑find material.
Market Disruption and Pricing Trends
The unusual buying behaviour is already reshaping the secondhand book market. Traditional dealers rely on thematic lots (e.g., sport, motoring) and modest pricing to move inventory. In contrast, the mystery buyers are paying full wholesale prices for large, heterogeneous bundles, effectively removing those volumes from the resale pool and driving up acquisition costs for other sellers. One bookseller warned that the trend “is disrupting the secondhand bookselling business because this isn’t typical for the trade.” The opacity of the buyers’ identities and the apparent lack of discount‑seeking suggest a purpose-driven acquisition rather than speculative resale.
Looking Ahead: Questions and Recommendations
As the flow of bulk orders continues, several questions remain pressing: Who exactly are the end‑users of these books? Are they AI firms, data‑brokerage houses, or other entities seeking textual corpora? How will the depletion of pre‑2022 print material affect future AI training sets, and what ethical considerations arise from destroying physical books for data extraction? For the secondhand trade, greater transparency from buyers—whether through verified invoices or clear statements of intent—could help sellers adapt their inventory strategies. Until then, booksellers like Manley, Shaughnessy and Gower‑Spence will continue to watch the strange parcels arrive, wondering whether they are witnessing the early chapters of a new data‑driven economy or merely a fleeting anomaly in the world of old paper.
https://www.theguardian.com/technology/2026/aug/15/uk-ireland-booksellers-suspect-ai-companies-bulk-orders-data-acquisition