Bloomsbury holds forecast as AI licensing lifts results
TL;DR:
- Bloomsbury Publishing maintained its annual forecast, citing continued AI licensing revenue and strength in academic and professional publishing.
- The company expects adjusted profit in line with analyst expectations of £49.9 million ($66.9 million).
- It struck an AI partnership with Google last September and has benefited from licensing revenue in the first four months of its new fiscal year.
Bloomsbury Publishing has kept its annual guidance steady, and the reason it gave says something about where trade publishing now finds its margin: AI licensing revenue, alongside strength in its academic and professional business. The publisher of JK Rowling’s Harry Potter books and Sarah Maas’s A Court of Thorns and Roses series said it expects adjusted profit in line with analysts’ average expectation of £49.9 million ($66.9 million).
Licensing as a line of defence
The AI revenue is not incidental. Bloomsbury struck a partnership with Google last September and said it has continued to benefit from licensing income in the first four months of its new fiscal year. Its academic and professional division has grown across all territories so far this year — the quieter half of the business doing steady work while the consumer market stays volatile.
The wider context arrived the same week, and it is unusually clean. On Tuesday, three publishers — Hachette, Cengage and Elsevier — sued Google in New York alleging that books supplied for limited services were repurposed to train Gemini without permission or payment. Bloomsbury’s arrangement is with the same company the others are suing. Two strategies, one counterparty: licence the corpus and book the revenue, or litigate and seek damages.
Neither path is obviously right yet. Litigation has produced both a defeat for authors against Meta and a $1.5 billion settlement from Anthropic, which suggests the value of a training corpus is real but unsettled. Licensing converts that uncertainty into recognised revenue now, at whatever price was agreed before anyone knew what the corpus was worth — the trade-off being that a publisher which licenses early has priced an asset in an immature market.
For UK readers the relevance is that Bloomsbury is a British company demonstrating a commercial model rather than a legal position. It is one of the few concrete data points on what AI licensing does to a UK P&L, in a debate mostly conducted through lawsuits and open letters. An MP’s bill would separately require AI scraping bots to identify themselves, which would make the difference between licensed and unlicensed use easier to see.
Looking forward
Bloomsbury will report against this guidance later in the year, and the AI licensing line is now something investors will look for specifically. If it keeps growing, expect more UK publishers to treat their back catalogues as licensable assets rather than litigation exhibits.