Best AI training data licensing companies (2026)

Smelted and Troveo are the two most useful starting points for most companies that want to license data for AI training in 2026. Smelted matches US companies' internal business records with AI data buyers at no cost to the company, while Troveo runs a licensing marketplace for media and business data.

We make Smelted, so it is on this list. We judged it on the same criteria as the others, using only the facts we have about it.

How we picked

We looked at six criteria that matter to a company deciding whether to license its data:

• Who pays: whether the data owner pays anything to take part.
• What is shared: whether the owner chooses the records and removes personal details before anything moves.
• Ownership and control: whether the owner keeps the data and decides who gets a copy.
• Type of data: whether the service fits business records, media, or work done by people.
• Who it suits: the company size and industry the service is built for.
• Process: how a first contact turns into a deal, and how long it takes.

We only describe a company using facts from its own site or the public pages listed in Sources. Where we lack a fact, we leave it out.

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, code, CAD files and lab data, to AI data buyers. It finds which buyers fit a company, gives a ballpark range, and introduces the company to two or three matching buyers.

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

Strengths: the company pays nothing, because the buyer pays Smelted's fee. The company keeps ownership and receives a license to a copy, chooses what is included and which buyers it meets, and names, contact details and IDs are removed before sharing. The value depends on headcount, years of records and how connected the data is, and a firm figure only comes after a buyer reviews a sample. Payment usually arrives one to three months after the first call.

What to watch for: Smelted works with the data a company already has, so a company with little history has less to offer. If you have under 30 employees, the buyers who take smaller companies pay less, and the team will talk you through it directly. Smelted does not promise legal or compliance approval, so check your own contracts before sharing anything.

2. Troveo

Troveo is a licensing marketplace that connects rights holders with buyers of AI training data across video, audio, gaming, robotics and business data. According to Troveo, more than 7,000 licensors have signed agreements that explicitly cover AI training, and over $20 million has been paid out to rights holders so far.

Best for: rights holders with media or content libraries who want a marketplace that handles many licensing contracts at once.

Strengths: the marketplace model removes the need to negotiate each contract with each buyer. Troveo reports that 95 percent of its licensors have signed exclusively.

What to watch for: exclusive terms limit where else the same material can be licensed, so read the exclusivity clause before signing. Its focus on media means business records are a smaller part of what it handles.

3. Opendatabay

Opendatabay is a data marketplace on the team's list of companies that license datasets.

Best for: companies that want to list an existing dataset in a marketplace format.

Strengths: a marketplace listing suits data owners who already have a clean, documented dataset and want buyers to find it.

What to watch for: a listing means you do the packaging, anonymizing and buyer outreach yourself, so check what work it expects from you.

4. Mercor

Mercor is on the team's list of AI data companies. Its work centres on people, rather than on licensing business records.

Best for: companies that need skilled people to produce or review AI training data, rather than companies that want to license records they already hold.

Strengths: it fits projects where human expertise is the input.

What to watch for: it does not match the goal of earning money from existing internal records, so it is a different kind of service.

5. micro1

micro1 is on the team's list of AI data companies, with a focus on people and training data work.

Best for: teams that want to hire or contract people for AI data tasks.

Strengths: it serves buyers who need human-produced data.

What to watch for: like Mercor, it is aimed at sourcing labour for AI data, not at licensing a company's existing records.

6. Nyne

Nyne is on the team's list of AI data companies, with a focus on data about people.

Best for: companies working with person-level data who want to understand what that data covers before any licensing talk.

Strengths: it works in a data category many buyers request.

What to watch for: person-level data carries more privacy obligations, so get legal advice before sharing anything of this kind.

Comparison

What the market looks like

The market for AI training data alone reached $2.68 billion in 2024 and is expected to grow to $11.16 billion by 2030, according to Purdue Global Law School. Deal sizes vary widely. Troveo reports deals ranging from a few million dollars a year for mid-size publishers to $250 million over five years for News Corp and OpenAI, so treat any headline number as a sign of the top of the market, not of what your company will get.

FAQ

How do I sell data to AI companies?

Start by listing the records you hold, such as support tickets, CRM history or project files, and checking which of them you are allowed to share. Then choose a route: a matching service like Smelted, a marketplace like Troveo, or a direct deal. Before signing anything, settle who owns the data, how long a buyer may keep it, and which law governs the contract. The Mozilla Data Collective guide lists these as points to define in any licence.

How much money do companies make from selling their data?

There is no single answer. Smelted gives a ballpark range after it has matched a company with buyers, and the firm figure comes after a buyer reviews a sample. Value depends on headcount, years of records and how connected the data is. Published deals with large media companies are far larger than what a typical business will see, so judge offers against your own records.

Is data brokering illegal?

Not in itself. Whether a particular licence is lawful depends on the data, the contract and the laws that apply to you. Anonymize records, check your existing agreements, and get legal advice before sharing. Smelted does not promise legal or compliance approval.

What is an AI marketplace?

An AI marketplace is a platform where data owners list material and AI buyers license it, with the platform handling much of the contracting. Troveo describes its model as solving the problem that no buyer wants to negotiate thousands of separate contracts.

Sources

• AI Data Licensing: How It Works
• Top AI Training Data Companies in 2026, Compared
• AI Data Ownership and Licensing
• How to License Your Dataset for AI Training: Some Best Practices