Updated 16 Oct 2026
Who buys AI training data, and what each one publishes.
Much of what AI labs spend on outside data reaches the companies that hold it through vendors that buy for their own account and resell. Around them sit marketplaces that pay a share of each sale, intermediaries, wind-down services and programmes that pay for work to order. These 37, across 5 kinds, publish a way in. Open any one for its terms, in its own words.
37 buyers, largest published figure first
- Not offers.
- A published rule is not an offer: each buyer evaluates the data before it names a price. None of these buyers is our client, and we are not the agent, partner or representative of any of them.
- Introduction fees.
- Several buyers pay whoever introduces a holder. In a sale we run we decline that money, tell you beside each offer what that buyer pays introducers where it publishes it or tells us, and ask it to put that amount into your price instead. If you choose instead to apply to one buyer through its own referral programme, that buyer pays us its published payout as well as our fee, and you see the amount in writing before you choose. An opened row says what each pays.
- Not listed.
- The AI labs themselves. They license data directly, and most publish no price or policy on brokers, so they are approached through a named role, first with a description that does not name the holder, then with the datasheet the holder approved. Vendors that say they buy data but publish no programme are on the vendor register.
The same data, shown to one buyer, gets one price.
Many of these buyers resell what they buy. One was reported to earn gross margins as high as 80 to 90 percent on off-the-shelf datasets it resells, and another says it will beat a qualifying written offer for the same data and rights, which shows how far apart their prices can be.
A holder that answers the first email sees one number. A holder that puts the same datasheet in front of several buyers at once, with its rights file ready, can compare their offers. Going direct, or through us: what each costs you.
In this topic
Buyers
- Start hereWho buys AI training data, and on what terms?The labs, data vendors and brokers that buy or resell data, what each publishes about price and contract terms, and how they source.
- GuideHow Do AI Companies Get Their Training Data?AI companies get training data six ways: crawling the web, licensing content, paying experts, buying company records, building environments and using user data.
- GuideAI Training Data Companies: Who Does What in 2026AI training data companies in 2026 by category: expert data platforms, content marketplaces, records buyers, environment builders, annotation tools and brokers.
- Guidemicro1, Handshake, Turing, Mercor: Data Programs Comparedmicro1, Handshake AI, Turing, Mercor and others publish programs to buy company records. Prices look alike; the licences, payment terms and leverage do not.
- GuideWhich AI Data Buyers Take Holders Outside the US?Few AI data buyers shut out companies abroad, but few say they want them, and their own estimators price the same company far lower outside the US.
- RegisterVendorsWho sells data, RL environments and evaluations, in their own words.
- RegisterData on offerEvery broker's lots on offer, side by side, read daily.
- RegisterCompaniesEvery organisation in the market with what is recorded about it.
- PageAcquire dataFor buyers: data sourced to your specification.
Cite this
AIData.Broker, “Who buys AI training data, and what each one publishes”, reviewed 16 Oct 2026. https://aidata.markets/buyers
Free to reuse under CC BY 4.0 if you credit AIData.Broker and link to this page. Each row names the publication that stated it, and words quoted from a source keep that source's own terms.