AI Licensing
Our clients have access to our network of internal and external SMEs to answer their ad-hoc questions that drive revenue and mitigate risk in this next era of AI transformation. Lately a spate of questions about AI licensing… the answers are priceless.
In the Outsell Accelerate A! Benchmark™ there are five dimensions we measure. One is all about new revenue from licensing. Still an evolving category, the leaders are winning the race driving revenue from big tech LLMs and for customer purposes as well. S&P’s last earnings cited this phenom. Wiley has been public about its $100m plus revenue stream since 2024.
Recently one client asked about licensing terms for a deal they were close to signing and we saved them from big exposure with an exclusivity term that was fraught.
Another client asked about editorial vs data-oriented content licensing. From Outsell’s expert Hugh Logue in answer:
…I agree the visible deal flow skews to editorial and publishing assets. The disclosed market is dominated by news, magazines, journals and books because those rightsholders had the most acute infringement exposure and the strongest motivation to monetise quickly.
That said, the Outsell licensing tracker does capture some genuine data-company transactions, Reddit and Stack Overflow licensing structured user-generated corpora at guaranteed annual fees (Reddit pulling circa $60m a year from Google, and a similar amount from OpenAI), Yelp and Photobucket monetising listings and user-uploaded assets, and the LexisNexis and Wolters Kluwer alliances with Harvey treat primary legal sources as structured data infrastructure rather than editorial commentary.
Overall, licensing deals are becoming overshadowed by MCP. The commercial logic differs sharply from the news and publishing deals: the data sits with the licensor, and an AI agent, such as Claude, pulls it on demand, and usage is metered.
S&P Global integrated Capital IQ financials and earnings call transcripts into Claude in July 2025 through Kensho, its AI subsidiary. LSEG followed with a similar arrangement covering its Workspace and Financial Analytics products. Moody’s embedded its full platform as a native app inside Claude, covering credit ratings and risk data on more than 600 million companies.
Anthropic also brought FactSet, Morningstar, PitchBook, Daloopa, MT Newswires, Aiera and Chronograph into Claude for Financial Services, alongside new connectors for Verisk, Dun & Bradstreet, Experian, IBISWorld, Third Bridge, Fiscal AI, GLG and Guidepoint. Meanwhile, Perplexity has built a parallel finance network through Benzinga, FactSet, Morningstar and Quartr.
These data-company deals follow a different logic. Models absorb editorial content at training time because that is how they learn to construct language and argument, the writing itself has to be internalised.
Structured data does not work that way: a model does not need to memorise every credit rating, only to query and interpret them in context. MCP integrations meet that need at inference time, with the data staying on the licensor’s side and getting pulled on demand. The commercials for pure data lean more towards metered channel partnerships rather than corpus sales.
Just like the MasterCard commercial it’s priceless. One call helps make a great deal and minimizes a bad one. One subscription paid for in an answer. This is why we tell our clients for their businesses and ours — don’t measure usage. Measure outcomes. The alchemy is in the ROI and the ROE — money saved, money made, and risk mitigated.
Inquiries are confidential and off the record and those that ask are never identified in or outside their company. Questions? Contact us.
Need a full licensing strategy framework? We can do that too. Whether custom projects or quick answers via inquiry privilege. If you haven’t taken the Outsell Accelerate A! Benchmark™ yet get in touch with Michael. We have your back.