- Published AI data prices range from 2 cents per image in very large archives to $4 per minute of unpublished video.
- Many data types, such as code, chat and support data, are priced per deal rather than per unit, so the tool shows comparable deals instead.
- Real prices depend on scarcity, quality, exclusivity and rights clarity.
How the estimate works
The tool multiplies your volume by the lowest and highest published unit prices for that type of data. Every price comes from a named source on our AI data prices page. Where no unit price has been published, it shows comparable reported deals.
Why your real price may differ
- Scarcity. Data no one else has, in a specialist domain or language, sells above published ranges.
- Quality and structure. Labels, outcomes and metadata add value.
- Exclusivity. Exclusive licenses usually command a premium.
- Rights clarity. Buyers discount data with uncertain rights or heavy personal information.
- Volume. Very large archives often sell at lower unit prices.
Read how to value a dataset.
Read next
Questions
Is this a quote?
No. It applies published reference prices to your volume. Only a buyer can make an offer, and the price depends on factors the tool cannot see.
Why does code show comparable deals instead of a number?
No per-line or per-repository price has been published. Reported deals, such as about $25,000 for one startup's codebase, are the best reference.
Sources
- Reuters via The Star: Inside Big Tech's underground race to buy AI training data (April 2024)
- Bloomberg via The Star: YouTubers are selling their unused video footage to AI companies (January 2025)
- MIT Technology Review: AI labs are buying data from dead startups to train models (October 2026)
- TravelPulse: Google buys Spirit Airlines' data for $10 million (August 2026)