Real prices paid for company data
Most deals are confidential, but these numbers are on the public record:
| Deal | Price | Source |
|---|---|---|
| Shuttered startups' Slack, email and ticket archives (100 deals in a year) | $10K–$100K each | Fast Company |
| Cielo24's Slack, email and Jira data | "Hundreds of thousands" | Fast Company |
| HVAC companies' data | $150,000 | The Information |
| Enterprise program: qualified partnership | $100K+ | Published program terms |
| Enterprise program: large-scale or multi-team | $500K+ | Published program terms |
| Enterprise program: highly unique, proprietary | $1M+ | Published program terms |
| Spirit Airlines records (bankruptcy sale to Google) | $10M | The Next Web |
| Reddit content license to Google | ~$60M / year | Engadget |
What drives the price up
Buyers list five factors: size and volume, workflow complexity, data quality, domain expertise and uniqueness. In practice, that means:
1. Depth beats volume
A thousand connected support tickets, each with the SOP that governed it, the QA score it got and the outcome, are worth more than a million disconnected emails. Buyers want full workflows, not fragments.
2. Years of history
A CRM with seven years of deal history shows how decisions change over time. A two-year-old wiki shows a snapshot.
3. Specialized expertise
Work that takes trained judgment, like underwriting, claims review, legal review or complex dispatch, is harder for AI to learn from public data. That makes it worth more.
4. Multiple teams
Data spanning support, sales, finance and operations lets buyers model how a whole business works. The $500K+ published tier is built around this.
5. Clean rights
If you clearly own the data and can license it without client consent problems, it's easier to buy, and buyers pay for easy.
What brings the price down
- Data mostly in people's heads instead of written down
- Non-English documentation
- Heavy personal or regulated data that needs extensive removal
- Records you hold for clients rather than own
- Generic content that looks like what's already public
Get a quick estimate
Our 60-second estimator maps your company to the published tier that fits it best. It's indicative only. The real number comes from the buyer's evaluation after you apply.
Frequently asked questions
How much is my company's data worth to AI companies?
Published enterprise payouts start at $100K and exceed $1M for large or unique datasets. Smaller archive sales have closed at $10K–$100K. Value depends on depth, history, expertise, multi-team coverage and clean ownership.
Is there a calculator for company data value?
DataPaid's free estimator maps your company to the closest published payout tier in about 60 seconds. It's indicative; final pricing is set by the buyer after evaluation.
Is selling data a one-time payment?
It depends on the agreement. Buyers mention both one-time datasets and ongoing participation across teams, which can pay more over time.
- Published enterprise data program terms (2026)
- Fast Company — Shuttered startups are selling old Slack chats and emails to AI companies (Apr 17, 2026)
- The Information — Data-labeling startups paying HVAC companies $150,000 for data
- The Next Web — AI data company commits $1bn to buy company data (Oct 9, 2026)
- Engadget — Reddit is licensing its content to Google (Feb 22, 2024)