Short answer
Publicly reported AI data licensing deals range from about $60 million a year for a large forum's content, to a news publisher agreement reported as potentially worth over $250 million across five years, to a $10 million winning bid for one bankrupt airline's internal archive. Public transaction figures are examples, not predictions of what an operating company can receive.
Most AI data licensing deals are private. A handful have been reported in enough detail to show how this market developed and what different kinds of data have fetched. This page collects the best-documented examples. Each figure below comes from the cited report, and most were reported by journalists rather than confirmed by the companies involved.
Key takeaways
- The first wave of deals, from 2023 to 2025, was for published content: forums, news archives, images.
- The current wave is for internal business records: chat, email, tickets and files.
- Publicly reported figures vary widely by dataset, scope and rights.
- Scale, uniqueness and exclusivity explain most of the spread.
- Reported figures describe other companies' deals. They are not a prediction of what any dataset will fetch.
The reported deals at a glance
| Year | Data holder | Counterparty | Reported value | Type of data |
|---|---|---|---|---|
| 2024 | About $60 million a year | Forum posts and comments | ||
| 2024 | News Corp | OpenAI | Potentially more than $250 million over five years | News archives and current content |
| 2023 | Shutterstock | Several AI developers | $104 million in AI licensing revenue for the year | Images, video and metadata |
| 2026 | Spirit Airlines, in bankruptcy | $10 million winning bid, proposed sale | Internal emails, chat messages and records |
Content deals: the first wave
Reddit and Google
In February 2024, Reddit agreed to make its content available to Google for training AI models. Bloomberg and then Reuters reported the deal to be worth about $60 million a year. Neither company disclosed the terms. The appeal was obvious: years of real people answering one another's questions on every subject.
News Corp and OpenAI
In May 2024, OpenAI and News Corp announced a multi-year agreement giving OpenAI access to current and archived content from News Corp publications. The Wall Street Journal reported the deal could be worth more than $250 million over five years, in cash and credits. The companies did not disclose the terms.
Shutterstock
Shutterstock's chief executive said in 2024 that the company earned $104 million in 2023 from licensing its library to AI developers. That showed a single well-organized archive could be licensed to several buyers at once.
Company archives: the second wave
Spirit Airlines' internal data
Frankfurt Kurnit reported in October 2026 that Google had won an August bankruptcy auction for Spirit Airlines' internal business data with a $10 million bid. Its analysis describes a proposed sale and objections from the carrier's flight attendants' union and a software vendor. A winning bid and proposed sale should not be described as proof that payment and transfer have completed.
The deal was unusual in size, but it showed plainly that internal communications had become an asset with a price.
Defunct startups
Frankfurt Kurnit describes startups that are shutting down licensing chat archives, email and tickets. Its warning is about rights and risk, not a standard payout: a company still needs to review confidentiality, privacy and the transfer method. A wind-down is also different from an operating company with ongoing customer relationships.
Operating companies
Frankfurt Kurnit describes approaches to operating companies as well as startups shutting down. DataRelai's homepage describes advertised program ranges of $100,000 to $2 million or more. Those are marketing ranges, not confirmed transactions or offers for your company.
What the examples show
The asset shifted from content to operations. Early buyers wanted finished, published material. Current buyers want the record of how work gets done. See why AI companies pay for business data.
Scale drives the headline numbers. The largest deals involve archives built by millions of contributors or tens of thousands of employees. A midsize company's archive is smaller by several orders of magnitude, and so is its price.
An archive is not a price list. Companies differ in rights, history, volume and usefulness. A public deal from a different business cannot value your records.
The same archive can be licensed more than once. Where a license is not exclusive, a data holder can license the same material to several buyers. See exclusive vs. non-exclusive data licenses.
Objections follow deals. Employees and unions have pushed back where workplace communications were included. A careful process addresses that before signing. See employee privacy when a company licenses workplace data.
How to read these numbers for your own company
Use the reported deals to understand the market, not to price your data. Three cautions apply.
- Most figures are reported, not confirmed. They come from journalists' sources, and terms such as exclusivity and duration are usually unknown.
- Advertised ranges are marketing. A program that advertises up to $2 million will offer many companies far less.
- Your data's value depends on your data. Years of history, the number of systems, variety and clean rights all matter. See how much your company's data is worth.
The only reliable way to find a price is to get written offers and compare them. DataRelai qualifies your data, finds buyers and brings you their offers. We can't promise any buyer or any price. See if your company qualifies.
Frequently asked questions
What is the biggest AI data licensing deal?
Among deals with reported figures, the News Corp and OpenAI agreement, reported as potentially worth more than $250 million over five years, is one of the largest. Many deals are never disclosed.
Have ordinary companies licensed internal data to AI buyers?
Yes. Reporting in 2026 describes both defunct startups and operating companies licensing chat, email, CRM and ticket archives.
Why are startup payouts so much smaller than publisher deals?
Scale. A publisher or forum holds content from millions of contributors over decades. A small startup holds a few years of records from a handful of people.
Are these deals sales or licenses?
Most are licenses. The data holder grants defined use and keeps the data. See data license vs. data sale.
Sources
- Engadget, "Reddit is licensing its content to Google to help train its AI models" (February 2024)
- Associated Press via WSLS, "OpenAI to start using news content from News Corp. as part of a multiyear deal" (May 23, 2024)
- PetaPixel, "Shutterstock Made $104 Million Licensing Assets to AI Devs Last Year" (June 4, 2024)
This guide is general information, not legal, tax or financial advice. Whether a particular dataset can be licensed depends on what it contains and on obligations your company already has.