Enterprise Data Licensing Shifts Toward AI Agent Access

Summary
S&P Global and Research Solutions say customers' AI agents now pull paid data directly, and contracts adding that access are growing faster, funded from AI budgets.
AI agent data licensing is starting to show up in contract values at data and research-content vendors. Between October 27 and November 17, 2026, S&P Global (SPGI) and Research Solutions (RSSS) said on their earnings calls that customers are letting in-house AI programs pull paid data and full-text research articles directly, that contracts which add this access are larger than before, and that the money is beginning to come from AI budgets.
The customer's AI program, not the employee, is now doing the lookups
Both companies sell content their customers cannot assemble on their own. S&P Global sells company financials, credit data and commodity data to financial institutions and corporations. Research Solutions buys scientific articles on behalf of corporate R&D departments and universities and offers an AI literature tool called Scite. The old pricing was simple: data terminals were priced per user seat, one account per employee, and articles were bought one at a time out of library or content budgets.
What changed is that customers are building their own AI agents, meaning AI programs that can search, compare and summarize material over many steps on their own. These programs read licensed data and article full text repeatedly at machine speed, and a per-person account cannot measure how much they use. So customers instead buy company-wide licences, API access or MCP connections (MCP is an interface standard that lets an AI assistant connect directly to an outside data source), or they pay per call. S&P Global described one customer that first bought S&P's retrieval agent for datasets it already licensed, then, seeing that the agent worked better across more datasets, moved to license new ones [1]. Any content vendor that sells by the seat or by the article runs into the same question once customers deploy these programs internally.
Customers using AI are growing their contracts faster
On its third-quarter call, S&P Global gave two sets of figures. API calls rose about 4,000% year over year. In Market Intelligence, annual contract value at customers using its AI products grew 1.8 times as fast as at other customers, and in its Energy business 3 times as fast [1]. Management also said customers have accepted the standardized AI-rights terms the company now uses across its businesses, and that for agent products able to query across several datasets it sees potential for consumption-based pricing [1].
Research Solutions is far smaller but reports the same direction. At the end of September, 49 business customers had AI access written into their contracts, up from 42 the prior quarter, mostly existing customers adding it at renewal or expansion. AI-related annual recurring revenue was about $1.1 million, up 38% from the end of the prior fiscal year [2]. In November it signed a Scite enterprise agreement valued at $1 million with a Fortune 500 biopharma company [3]. Management said this long-time customer's article-purchasing revenue had been roughly flat for more than five years, so the deal is a material step up [2]. Management also explained why large deals are closing: corporate and academic AI budgets are growing while library and content budgets are flat or falling, and the new buyers joining the conversation are used to numbers of this size [2].
Pricing is moving from head count to machine access rights
Taken together, the two disclosures suggest that the basis for pricing in the data business is shifting from "how many people use it" to "what the customer's AI can read, and how much." AI rights are written into standard contracts, and access is priced per enterprise or per call. Once connected, customers license additional datasets so their agents can query across them, which makes the same existing customer's contract larger. The budget owner is changing as well, from library and desk budgets to AI budgets that are currently growing.
Monetization is still early, and higher usage does not automatically mean higher revenue. Clarivate's AI API calls this quarter were 2.5 times the level of a year earlier, yet its organic annual contract value grew only about 1.8%, mostly from roughly 3% price increases [4]. Research Solutions' per-article transaction revenue fell 8%, and AI replacing some single-article purchases cannot be ruled out; its $1 million contract also took a year to negotiate [2]. Three numbers are worth tracking from here: the count of contracts with AI access written in, AI-related recurring revenue, and the gap in contract value growth between AI-using customers and the rest.
Companies this change could affect
- Moody's (MCO): Its analytics business delivers research and data into customers' AI platforms through MCP, and management lists consumption-based charging as one way to monetize it [5], so its pricing faces the same move from seats toward machine access.
- Nasdaq (NDAQ): Its data business sells enterprise licences for market and investment data to institutions and charges more for data delivered through MCP connectors [6], so changes in customers' demand for AI access would show up in its licence pricing.
- Wiley (WLY): An academic publisher. Aggregators such as Research Solutions pay publishers royalties when they deliver articles to companies, so if corporate AI reading turns into ongoing purchases it would land in Wiley's recurring AI revenue, although its AI revenue today comes mostly from model-training licences [7].
Sources
[1] Drillr · S&P Global · 2026-10-27 · Earnings call
[2] Drillr · Research Solutions · 2026-11-17 · Earnings call
"AI budgets, both across corporate and academic, I would say are growing. Whereas, you know, historically library budgets and content budgets are kind of flat or decreasing."
[3] PR Newswire · Research Solutions announces agreement with Fortune 500 biopharma company valued at $1 million · 2026-11-11 · Company announcement · https://finviz.com/news/420910/research-solutions-announces-agreement-with-fortune-500-biopharma-company-valued-at-1-million
[4] Drillr · Clarivate · 2026-11-03 · Earnings call
[5] Drillr · Moody's · 2026-10-21 · Earnings call
[6] Drillr · Nasdaq · 2026-10-22 · Earnings call
[7] Drillr · Wiley · 2026-09-03 · Earnings call
This article only helps you spot industry changes and companies that may be overlooked. It is not a stock recommendation.