See if your data qualifies

Guides · Published

Best AI training data licensing platforms (2026)

Compare six places to license AI training data in 2026, from rights marketplaces and stock libraries to a free matching service for internal business records.

The best AI training data licensing option in 2026 depends on what you have to license. Marketplaces such as Troveo suit rights holders with large media catalogues, stock libraries such as Depositphotos and Shutterstock suit visual content, and Smelted suits US companies that want to license anonymized internal business records for free.

We make Smelted, so it's on this list; we judged it on the same criteria as the others.

How we picked

We judged each entry on six criteria: what it licenses, who buys, what it costs the company, how much control the owner keeps, how exclusivity works, and which company types it fits. Where a provider's pages did not state a fee or term, we say so rather than guess.

We relied on each company's own pages and on the market guides listed in Sources. We did not test any service or talk to its customers.

The list

1. Smelted

Smelted is a free matching service that licenses anonymized internal business records, such as support tickets, CRM history, SOPs, project files and code, to AI data buyers.

Best for: US companies with 50 to 500 employees and years of records in tools like Zendesk, Jira, Salesforce, Bullhorn, AutoCAD, Revit or Benchling.

Strengths: The company pays nothing; the buyer pays Smelted's fee. The company keeps ownership, which is a license to a copy, chooses what is included and which buyers it meets, and names, contact details and IDs are removed before anything is shared. Smelted introduces the company to 2 to 3 matching buyers.

What to watch for: Value depends on headcount, years of records and how connected the data is, so no figure is given before a buyer reviews a sample. Payment comes a few weeks to two months after signing, so plan on 1 to 3 months from first call to payment. Buyers who take smaller companies pay less, so companies under 30 employees should expect a lower offer. Start at the 2-minute form.

2. Troveo

Troveo runs a licensing marketplace where rights holders license content for AI training. According to Troveo, more than 7,000 licensors have signed agreements that explicitly cover AI training.

Best for: rights holders with large catalogues across video, audio, gaming, robotics and business data.

Strengths: Troveo says 95 percent of its licensors have signed exclusively, and it reports more than 20 million dollars paid out to rights holders so far. A marketplace spares each owner from negotiating separately with every lab.

What to watch for: Exclusivity is a real trade-off, because an exclusive licence can limit where else the same material is sold. Troveo's own page notes that reported deals vary widely. Troveo puts them at a few million dollars a year for mid-size publishers, and Quartz reports that OpenAI's five-year News Corp deal is worth more than $250 million. Large publisher deals are not a guide to what a mid-size company will receive.

3. Depositphotos

Depositphotos sells rights-cleared visual datasets. It describes AI data licensing as the legal authorization of datasets used to train machine learning and generative AI models (Depositphotos).

Best for: teams that need images or video with clear rights for model training.

Strengths: The rights position is the core of the offer, and the company describes custom plans for buyers who need specific data.

What to watch for: Its catalogue is stock media, not internal business records. A company with support tickets or engineering files will not find a fit here.

4. Shutterstock

Shutterstock runs a data licensing programme alongside a contributor fund. Credtent uses it as its example of a collective AI training licence in its market map, and the Shutterstock page itself covers the data licensing and contributor fund.

Best for: photographers and contributors to a large image library who want to share in licensing income.

Strengths: A collective licence pools many small contributors into one deal, which is simpler than negotiating individually.

What to watch for: The model is built around contributors to a stock library. It does not fit a company that wants to license its own internal records.

5. Opendatabay

Opendatabay publishes a guide to licensed AI training data types that compares commercial and general licences for buyers and sellers (Opendatabay).

Best for: buyers and sellers who want to compare licence types before they sign.

Strengths: The guide sets out the difference between commercial and general licences, which helps a first-time seller read a contract.

What to watch for: The guide explains licence types. Check each listing's terms directly, because the guide does not replace a specific contract.

6. micro1

micro1 is one of the expert data and human feedback providers that Troveo groups with Mercor, Surge AI and Handshake AI, as opposed to annotation vendors such as Scale AI, Appen and Labelbox (Troveo).

Best for: teams that need expert human data and feedback work rather than licensing of existing records.

Strengths: It sits in the expert-data group that Troveo describes for frontier-scale and human feedback programmes.

What to watch for: This is a service for producing new expert data, not a way to license records a company already holds. The pages we reviewed do not state its fees or licence terms.

Comparison

ProviderWhat it licensesBest forCost to the companyControl over what's shared
SmeltedAnonymized internal business recordsUS firms with 50 to 500 employees and years of recordsFree; the buyer pays the feeCompany chooses content and buyers; names and IDs removed
TroveoVideo, audio, gaming, robotics, business dataRights holders with large cataloguesNot stated on the pages reviewedLicensors sign; 95% exclusive per Troveo
DepositphotosRights-cleared visual datasetsTeams needing clean image or video dataCustom plans; no fee statedSet by the custom plan
ShutterstockStock content via data licensing and a contributor fundContributors to a stock libraryNot statedNot stated
OpendatabayGuide to licensed data typesBuyers comparing licence termsNot statedNot stated
micro1Expert human data and feedbackTeams buying expert workNot statedNot stated

Which should you choose?

Choose a marketplace if you hold media or rights-heavy content at scale, and a stock library if your content is visual and contributed. Choose a matching service if your value sits in internal records such as support tickets, CRM history or engineering files. Read any exclusivity clause before you sign, since it can limit where else the same material is sold.

FAQ

How do you sell your data to AI companies?

You either license it through a marketplace or a stock library, or you use a matching service that introduces you to buyers. With Smelted, a company licenses an anonymized copy of its internal records, keeps ownership, and chooses what is included and which buyers it meets. Payment comes a few weeks to two months after signing.

How much money do companies make selling their data?

It varies widely. Troveo reports deals from a few million dollars a year for mid-size publishers up to $250 million over five years for News Corp and OpenAI, and Quartz reports the same News Corp deal at more than $250 million. Those are large publisher deals. For a company licensing internal records, value depends on headcount, years of records and how connected the data is.

How much is 100 GB of data worth?

There is no fixed price for a number of gigabytes. Value depends on what the records contain and who wants them. Smelted gives a ballpark range only after it has reviewed the company's details, and a real figure comes after a buyer reviews a sample.

Can a company license its data without giving it away?

Yes, if the licence is limited. A licence to a copy leaves ownership with the company, and the company can choose what is included. Check whether a deal is exclusive, because exclusivity limits where else the same material can be sold.

Sources