What is your company's data worth?
The factors that set the price, with the figures that have been reported publicly.
Updated September 30, 2026
There's no public price list for company data, and anyone who quotes you a number before seeing what you have is guessing. What we can tell you is what has been reported publicly and which factors move the price up or down.
What has been reported
| Deal | Reported value | Source |
|---|---|---|
| Closing startups selling Slack, email and Jira archives, through one shutdown service | $10,000 to $100,000 per company, about 100 deals in a year | (Gizmodo, reporting on Forbes, April 17, 2026) |
| A closed software company selling every Slack message, internal email and Jira ticket | "Hundreds of thousands of dollars" | (Slashdot, summarizing Forbes, April 18, 2026) |
| Google's license to train on Reddit posts | $60 million | (CBS News, February 23, 2024) |
| OpenAI's content deal with News Corp | More than $250 million over five years | (Variety, May 2024) |
The first two rows are the closest match to a typical company: archives of internal work, sold once. The last two are platform-scale content deals, included to show the ceiling rather than to set expectations. An operating company with years of well-kept records in a specialized field sits somewhere in between, and usually above what a small startup gets for its leftovers.
What moves the price
| Pushes it up | Pushes it down |
|---|---|
| Complete tasks with recorded outcomes | Fragments with no beginning or end |
| Several years of history in one system | A few months, or scattered across tools |
| Specialized work few companies can supply | Generic work many companies could supply |
| Clean rights and a clear redaction plan | Customer contracts that restrict use |
| Ongoing refreshes as new records come in | A one-time export |
| Non-exclusive terms you can repeat with other buyers | Exclusivity given away without a premium |
How deals are usually paid
- A fixed price for a defined package, paid on signing, on delivery, or split between the two.
- Milestone payments, released as the buyer accepts each delivery. Common for large or phased datasets.
- Refresh fees for sending new records on a schedule after the first delivery.
How not to undersell
- Get more than one offer. Buyers value the same records very differently depending on what they're building.
- Price exclusivity separately, and put an end date on it.
- Keep the first delivery focused. A well-cleaned, well-described set sells better than everything at once.
- Know your costs: legal review, internal staff time and the cleanup all come out of the price.
Once we know what you have, we can give you a realistic range. The assessment is the place to start, and the licensing agreement checklist covers the terms that protect the price once you have it.
Common questions
Is there a price per record?
Occasionally, but most company datasets are priced as a package. Buyers look at the whole collection (how complete, how specialized, how many years) rather than counting rows.
Should we take an upfront payment or a larger amount spread over milestones?
Upfront money is certain; milestone money depends on the buyer accepting deliveries. If most of the price sits behind milestones, make sure the acceptance criteria are specific and that you keep the right to walk away if they aren't met.
Sources
- Gizmodo, reporting on Forbes, April 17, 2026: Failed companies are selling old Slack chats and email archives to train AI
- Slashdot, summarizing Forbes, April 18, 2026: Shuttered startups are selling old Slack chats, emails to AI companies
- CBS News, February 23, 2024: Google strikes $60 million deal with Reddit for AI training
- Variety, May 2024: News Corp inks OpenAI licensing deal potentially worth more than $250 million
Find out what your records are worth to a buyer
Answer a few questions about your company and your records. It takes about five minutes, and you don't send us any files.