Data licensing is a business model in which a company sells or rents access to data it collects or compiles, usually by subscription, to customers who use it in their own work. The data is gathered once and sold many times, so the margin depends on how hard it would be for a customer to collect it alone.
RewiredAI rewires it: the reader of the data is increasingly software, so the product shifts from screens built for people to feeds built for AI agents.
How it makes money
A data business collects something once and sells it many times. Prices, filings, estimates, ownership records: gathered, cleaned, checked and kept current, then rented to everyone who needs them. The customer pays for not having to build the same thing alone.
For most of its history, the product was a screen. An analyst sat at a terminal and read the data. AI changes who is reading. More and more of the work is done by software that pulls the data, combines it and acts on it, and software has different needs: clean structure, consistent identifiers, and a way in that does not involve a person clicking.
A product built for machines
FactSet's latest 10-K describes the shift in its own terms. "We also deliver our content in forms designed for machine consumption, including the metadata and symbology that allow AI agents and other automated systems to interpret and act on it." Alongside its desktop product, it offers "application programming interfaces ("APIs"), and model context protocol ("MCP") servers," the standard way for an AI model to connect to an outside source.
The company's stated aim is to expand delivery "through other methods, including MCP, to advance our position as a leading enterprise data provider for our clients in the AI age." It is also building a platform for clients to manage AI agents, whether built by FactSet, by the clients or by others.
Where the growth came from
At August 31, 2026, FactSet's organic annual subscription value was $2,568.2 million, up 7.0%. "Organic ASV growth was mainly driven by data solutions and, to a lesser extent, workstations and middle office solutions." Users grew faster than clients: users rose 9.8% to 260,671, clients 2.9% to 9,258.
How it fails
A data licensing business fails when its data becomes easy to get elsewhere, or when the customer can no longer tell the difference. If an AI model can assemble a good enough answer from public sources, the licensed feed loses its premium. The defense is data that is hard to collect, checked to a standard customers trust, and delivered in the form their systems need.
Our read
AI does not make licensed data less valuable. It makes it more valuable and more exposed at once. More valuable, because an AI agent is only as good as what it reads, and a firm that acts on wrong numbers pays for them. More exposed, because the screen that used to bind the customer to the provider matters less when software is doing the reading.
That moves the competition from the interface to the data itself: how complete it is, how quickly it is corrected, and how easily an agent can use it. For a data seller, the question is whether agents are wired to its feed by default. For a buyer, it is which data your AI is reading, and whether you would stake a decision on it.
Questions
- What is data licensing?
- A company sells or rents access to data it collects or compiles, usually by subscription, to customers who use it in their own work. FactSet, for example, licenses financial data to investment firms.
- How is AI changing data licensing?
- More of the data is read by software rather than people. FactSet now delivers content in forms designed for machine consumption, including metadata for AI agents, through APIs and model context protocol servers.
- What is an MCP server?
- Model context protocol is a standard way for an AI model to connect to an outside source of data or tools. A data provider that runs an MCP server lets its customers' AI agents query its data directly.
Sources
- FactSet Research Systems Inc. (SEC EDGAR), Form 10-K for the fiscal year ended August 31, 2026 (2026-10-20)
Every figure and quotation on this page traces to these filings. “Our read” is Stratrix’s judgment, drawn from them. Analysis and data, not advice.
More from Stratrix
“Should you license the thing that makes you special, and to whom?”
See the pack — $499 →The Business Model Wizard
Make the seven choices that define your model, test your numbers against real companies, and meet the rival AI would build against you.
Free · about twelve minutesBuild your Model Thesis →Test a decision, free
Set a live decision against the companies that made it before you.
AcquisitionsThe Distance TestDivestituresThe Divestiture TestDiversificationThe Diversification TestMake or buyThe Make-or-Buy TestStrategic Decisions
The big calls companies keep facing, told through the 819 companies on record that made them.
24 decisionsFind your decision →New analyses, by email
The defining moves in business, checked against the record, as they publish.