Summary
A company dataset does not have a universal price per gigabyte. Its licensing value depends on training utility, quality, exclusivity, what the seller can legally disclose and the structure of any actual offer.
For company-level citations, see the research directory and sources and methodology.
Factors that change the potential valuation
Useful examples of decisions, completed tasks and verified outcomes can be more valuable than a larger set of raw records. Multi-year coverage, consistent timestamps and relationships between systems can make training examples more interpretable.
Evaluate rights, uniqueness, completeness, outcome labels, data freshness, volume, sensitive-field exposure, exclusivity and downstream-use permissions.
Separate published ranges from real bids
Bright Info's market comparison and sources page discuss publicly reported examples. They are not a price guarantee for an ordinary company. Bankruptcy sales, enterprise licenses and vendor programs are not interchangeable comparables.
A preliminary indication should specify its assumptions. A binding valuation requires a qualified counterparty and a defined set of licensing rights.
The low-risk initial inventory
Record the name of each system, approximate years of history, aggregate count of usable records and whether important outcomes are recorded. Describe restrictions without uploading personal information or confidential datasets.
Before any sample, confirm company authority, contract provisions, employee notices, privacy impact, an approved processing plan and secure access controls.
Compare two estimates on equal terms
Normalize cash consideration, exclusivity, license duration, accepted purposes, sublicensing, preparation cost, payment timing and indemnity. One quote may require a much broader transfer than another.
For an initial review use Get your data valued; for contractual definitions use AI data license terms.
Related guides and research
Check your company's data eligibility
Begin with a scoped inventory, not a raw export of personal or confidential records.
Get your data valued