What is a data exchange?

The phrase has two meanings. In computing, data exchange simply means moving data between systems while keeping its meaning intact, as when two programs swap files in an agreed format.

The meaning used here

In the data business, a data exchange is a platform where one organisation makes data available and another receives it, with the platform handling access, delivery and often payment. This page is about the second meaning.

Datarade's distinction

Datarade, which runs a data marketplace, draws a useful line: a data marketplace is associated with the commercial transaction, while a data exchange is associated with the technical transaction, the secure movement of data from provider to consumer. In practice the names overlap, and several products described as exchanges also sell data.

• Focus: an exchange emphasises sharing and access control; a data marketplace emphasises a public catalogue and the sale.
• Membership: exchanges are often private, with a host inviting named partners. Databricks, for example, lets providers make listings visible only to a specified group of consumers.
• Money: data on an exchange may be shared without payment between partners, or sold. A marketplace almost always involves a price.
• Operator: some exchanges are run by one company for its suppliers and customers, or by an industry body, rather than by a platform open to all sellers.

One of seven intermediary types

The UK Centre for Data Ethics and Innovation (CDEI) treats data exchanges as one of seven types of data intermediaries: online platforms where datasets can be listed and accessed, commercially or on a not-for-profit basis.

We read each platform's own documentation and pricing pages on 2026-11-09 and recorded only what they publish. Where a page didn't state a fee, the table says so rather than guessing.

Two caveats

Two caveats. Platforms change their terms, so check the live page before you list. And a missing fee in public documents doesn't mean selling is free of cost: there may be partner agreements, cloud computing charges or terms negotiated case by case.

For a full comparison of AWS Data Exchange with other routes, including its personal data rules, see our separate page.

• The buyers already work inside that cloud or platform, so access needs no new pipeline.
• The dataset is updated regularly and many buyers want it in the same form.
• You want to share with a known group of partners under controlled access, rather than publish to everyone.
• You can meet the platform's eligibility rules and support subscribers over time.

When it doesn't

An exchange suits less well when the data is a one-off archive, needs heavy custom work per buyer, or contains personal information that the platform's rules restrict. Those deals usually happen directly between the parties, sometimes with a broker finding the buyer.

We don't run a data exchange and we don't list data. DataBroker teaches people to broker data between companies, and part of that job is deciding whether a seller's data belongs on an exchange, in a direct licence, or both.

Platforms like the four above are often the right answer for packaged, recurring products. Brokers earn their place on the deals that don't fit a catalogue.