Have data from your company? It could be more valuable than you think.

No, you're licensing it. A sale is permanent. A licence lets an AI company train on your data under terms you approve, and you keep ownership.

Licences don't have to be exclusive, so the same data can be licensed to more than one buyer and earn more than once.

No. You decide what goes in and what stays out, deal by deal. If you're not comfortable licensing something, it stays out.

No. There are no public listings and no open marketplace. Every deal is private and direct with the buyer, under strict NDAs, and you can see everything that happens with your data.

It stays out. Personal information is excluded or anonymised before anything is prepared for licensing, and data you hold under confidentiality agreements is never included.

If other parties hold rights in your data, we'll flag it, and your own counsel should look at the agreements before you decide. Nothing on this page is legal advice.

Original material they can't scrape from the open web, with clear rights, good quality and plenty of depth in one field. That could be recordings, images, documents, logs or structured records. The format matters less than how clean and how rare the data is.

You earn a royalty every time your data is licensed. There's no public rate card, because value depends on how much you hold, its quality, how unusual it is and how clean the rights are.

The estimate above shows the market range. A short call gets you a real number.

Anything you don't fully own: client work where the client holds the rights, third-party material licensed to you without AI-training permission, anything under a confidentiality agreement, and personal data.

Buyers pay for clean rights, so we're strict about this.

A lot of it has been, and licensing is the answer to that. Labs increasingly need data with a clear source and real consent behind it, which scraping can't offer.

With licensing, the owner agrees to the deal and gets paid for it.