If your company holds years of operating records and you've wondered who would pay for them, the answer in 2026 is four kinds of buyer. The frontier AI labs buy directly when the catalog is famous enough. The AI data companies that supply those labs buy from everyone else. Data marketplaces aggregate thousands of companies into one counterparty. And closure marketplaces buy code and archives from companies winding down. Read on to find out who each buyer is from the public record, what they want from a company like yours, and how you get in front of them.
The four kinds of buyer
| Buyer | Who they are | What they want | How you reach them |
|---|---|---|---|
| Frontier AI labs | The companies training the largest models | Famous catalogs and living sources, licensed directly | Only if your catalog is already known to them |
| AI data companies | Firms that source, package and sell data to the labs | Operating records they can turn into training sets, often bought outright | Through their own intake programs and bids |
| Data marketplaces | Platforms that license many owners' data to many buyers | Ordinary companies' operating records, licensed selectively | One agreement, then the marketplace does the selling |
| Closure marketplaces | Platforms that buy code and archives from companies shutting down | Full repositories, workspace archives, documentation | When you're winding the company down |
The first two write the headlines. The last two are where most companies land, and we cover why below. Our guide to data monetization companies and platforms covers the marketplaces, exchanges and brokers in more detail.
The frontier labs, from the public record
The labs buy directly, and the deals that have been reported show who they are. OpenAI licensed News Corp's archive for a reported 250 million dollars over five years, licenses Reddit's conversation data, and pays Dotdash Meredith around 16 million dollars a year. Google pays Reddit about 60 million dollars a year and bid 10 million dollars for Spirit Airlines' operating records in bankruptcy court. Amazon pays The New York Times a reported 20 to 25 million dollars a year. Shutterstock earned 104 million dollars in 2023 from licensing deals with OpenAI, Meta, Google, Amazon and Apple, one of the few audited figures in the market.
Every one of those sellers is a brand-name catalog with a legal team and a data team. That's the price of a direct deal. The labs don't issue a specification for the data they want and most don't know until they see it, so a company nobody has heard of rarely gets a direct conversation. Our guide to how AI data licensing deals work covers what those deals contain.
The AI data companies
Between the labs and everyone else sits a layer of companies whose business is finding data, preparing it and selling it on. Two of them bid against Google for Spirit Airlines' records, Micro1 at 12.5 million dollars and Mercor at 7.5 million. In late 2025 Turing launched a program to acquire full private repositories, design docs, testing logs and internal wikis from defunct startups. Firms like these buy operating records because the labs they supply want the record of how real work gets done, and that record isn't on the public internet.
Reaching them usually means an outright sale through their intake program, which pays once and ends your part in the deal. Our guide to data monetization models covers why a sale pays once and a license pays again.
The data marketplaces
A marketplace signs one license with you and licenses your data onward to many buyers, so you never negotiate with a lab yourself. It handles the rights review, the packaging (with names and personal details replaced by pseudonyms) and every buyer conversation, and you're paid on each sale. Thousands of creators and companies license through marketplaces like Troveo, and typical full-company licensing deals for operating businesses start at six figures.
This is the route for a company with 50 or more people on ordinary tools and no data team, because there's nothing to build and no buyer to find. Our owner's guide to licensing company data for AI covers the process from inventory to payment.
The closure marketplaces
If the company is shutting down, a different set of buyers appears. Closure platforms broker code and workspace archives from companies that are closing, and the public figures put those deals at roughly 10,000 to 100,000 dollars each, the small end of the market, with one codebase-buying platform listing a 5,000 dollar baseline per repository. A closing company can transfer everything cleanly and has every reason to, which is why most of the reported code deals so far involve startups that shut down. Our guide to selling your code and data to AI labs when shutting down covers that route.
What all four buyers want
The buyers differ, and what they pay for is the same across all four. In order, they pay for uniqueness (records that aren't on the public internet), years of history, how connected your systems are, whether the data shows outcomes, clean rights, and what buyers need right now. Volume on its own does little. A 40 person company with ten years of specialized records can be worth more than a large company with generic data. Our guide to how much your company's data is worth to AI labs covers the six factors and the public price points.
How to get in front of a buyer
Buyers don't come looking for a company they've never heard of, so the owner has to start the conversation. Three things make it a short one.
- Know what you hold. List the systems you run, how many years each covers, and whether the data is finished or still growing.
- Know what you can license. Check your customer contracts and employee agreements, and pull out anything that came from third parties or sits under a confidentiality clause.
- Have a sample a buyer can test. Buyers react to a curated slice of real records, and a description on its own rarely gets a reply. If you license through a marketplace, it prepares the sample, with personal details replaced by pseudonyms, and puts it in front of its buyers for you.
Where to start
The quickest way to tell which buyer fits your company is our free data value assessment (ten questions about your systems, history and industry, about five minutes). You get an estimate of what your data could license for, which tells you whether a marketplace license, a direct deal, or no deal at all is the right answer.
Where Troveo fits
Troveo is a data licensing marketplace. We work with more than 40 active buyers, we've paid rights holders more than 20 million dollars across video, audio, gaming, robotics and business data, and we take no fees and no deductions from your payouts. You sign one license, we handle the rights review, the packaging and every buyer conversation, your company is never named publicly, and you're paid on every sale. Start with the data value assessment, read our business data page for how the process works, or talk to us about what your company holds.