This guide explains how AWS Data Exchange works for a buyer: the two ways data reaches your account, the five types of data set, what a subscription entitles you to, and how revisions arrive in Amazon S3. It then covers company data that comes direct from the provider. Fokals is one such source: it is delivered direct, by REST API and as bulk files, which you load into your own AWS account with the loaders of the services you already use. Where this guide describes AWS Data Exchange it states what the AWS documentation says on 4 October 2026, and it links the page.
Two ways data reaches you
AWS describes the service as a way to share and manage data entitlements from other organisations. There are two units of exchange. A data grant is created by a sender, who needs only an AWS account, and sent to a receiver's AWS account; the receiver accepts it to gain access. A product is published by a provider in the AWS Marketplace catalogue after AWS reviews it against its guidelines, so a buyer meets it as a data marketplace listing, and the buyer subscribes to it. A product holds one or more data sets, and its offers define the terms a subscriber agrees to.
Both units hold the same five types of data set, and the type decides what you do with the data. All the data sets of one product or grant must be in the same AWS Region.
| Data set type | What you receive | What you do with it |
|---|---|---|
| Files (Amazon S3 objects) | Revisions made of files | Export to an S3 bucket, or download through a signed URL |
| API | Access to the provider's Amazon API Gateway API | Call it through endpoints that AWS Data Exchange manages and proxies to the provider |
| Amazon Redshift | A datashare of tables, views, schemas and functions | Query it read-only without loading it; access ends with the subscription |
| Amazon S3 data access | Read-only access to the provider's S3 objects | Use them in place through an S3 access point that AWS Data Exchange provisions |
| AWS Lake Formation | Data permissions in the provider's data lake | Query through services such as Athena; the documentation marks it Preview |
A Redshift datashare has one further condition: the documentation tells you to check the date, time and duration of the cluster's maintenance window, because during it you have no access to the datashare.
Subscriptions, offers and entitlements
All products are subscription-based. An offer states the price, the duration, the payment schedule, the data subscription agreement and the refund policy, and by subscribing you agree to those terms. A public offer can carry several price and duration pairs. A provider can also send a private offer to your account, or a Bring Your Own Subscription offer that moves an existing subscription onto AWS Data Exchange. Some products require subscription verification: you submit information about your identity and intended use, and the provider reviews it before approving.
When a subscription is active, the product's data sets appear as entitled data sets, a read-only view of the provider's own, listed for each Region on the Entitled data sets page. You can choose auto-renewal where the provider allows it, and if the offer terms have changed by the renewal, the new terms apply. If the provider unpublishes the product, your access continues while the subscription is active, but the subscription cannot renew. When a subscription ends you can no longer view or export the data sets. For a Files data set you keep what you already exported, and the data subscription agreement says whether you must delete it.
A paid subscription appears on your AWS bill as part of your AWS Marketplace charges, on the payment schedule in the offer, and AWS bills the services you use to work with the data, such as Amazon S3 or Athena, separately. The documentation also states that AWS may share transaction information, including payment terms and usage metrics, with the seller, and that it might share your contact information with the provider. That matters if the fact that you subscribe is itself confidential.
Revisions: what arrives and when
A data set is a series of revisions, and a revision is a container of assets. The documentation says each revision may hold all the data for the data set or only the data added since the previous revision, and that this is up to the provider. An offer can give you access to some, none or all historical revisions, and to future revisions made during the subscription. A finalised revision cannot be changed except by revoking it, and a revoked revision carries a comment that tells subscribers why.
For company data that last point matters. A buyer who tests a strategy or reports on the data needs to know, before subscribing, whether a revision is a snapshot or a delta, whether each row carries the time it was observed, and how a correction arrives: as a revoked revision, or as changed rows in a later one. The glossary entry on point-in-time data explains why.
For Files data sets you can export each revision by hand, or turn on auto-export, which writes every new revision to up to five S3 buckets of yours. The bucket needs a policy that lets AWS Data Exchange write to it. Object names follow a key pattern that you set from references such as ${Revision.Id} and ${Asset.Name}, and it must include the asset name or the asset identifier. Each resulting name must be unique, or an export overwrites an existing object. A pattern such as vendor/${Revision.Id}/${Asset.Name} therefore keeps each revision in its own folder, while a pattern with nothing specific to the revision lets a file in the next revision overwrite a file of the same name in the last.
Evaluating a product before you subscribe
A provider can attach a data dictionary and samples to a product, and you can view and download both before you subscribe. A dictionary lists the columns of a data set and what they mean, and a sample can be any file type that S3 supports. If you cannot find a product, the console lets you request a recommendation from the AWS Data Exchange Data Discovery Team, and the documentation says you should hear back within two business days.
For a company-data product, put these questions to the provider and read the answers against the offer:
- Which identifiers does each row carry: a ticker, an ISIN, an LEI, a FIGI, a stable company key? The guide to joinable identifiers says why each matters.
- How far back do the offered revisions go, and are they snapshots or deltas?
- Does each row carry the time it was observed, and is a closed period ever rewritten?
- Which Region holds the data sets, and does it match your account?
- What does the data subscription agreement say about exported data when the subscription ends, and about use inside a product?
- What did you declare in subscription verification, and does the agreement allow it?
Fokals states its own answers on its site. Every company has one stable company ID, and a listed company carries its ticker, MIC, ISIN, LEI and share-class FIGI. Every observation is dated: daily and weekly datasets are written once after the period closes and never revised, and each record carries its source and the time it was observed. A breaking change to labels or scores ships as a new version with at least 90 days' notice.
Loading company data delivered direct
A provider that delivers direct sends the data by its own route, and you do by hand what auto-export does for you. Fokals is delivered by REST API and as bulk files in CSV, JSON Lines or JSON, with a manifest. Download the export, or page the API from the last cursor you stored, write the files to your own bucket under one prefix for each table and period, keep the manifest under a separate prefix and load from there. Two guides take it on from the bucket: loading company data into Amazon Redshift and querying company data on S3 with Athena.
A provider's revisions give you a record of what arrived and when. With files you fetch directly, that record is yours to keep: store each manifest, note the period and the label versions it names, and compare the periods you loaded with the periods the manifests name. The delivery page describes how Fokals data reaches you.
AWS Data Exchange fits when procurement needs the charge on your AWS bill, when you want revisions written to your bucket on the provider's schedule or a Redshift datashare you can query without loading, and when the provider you need publishes there. A direct licence fits when you want data delivered straight from the provider under a written agreement, which for Fokals covers internal use, embedding in a product or redistribution. The comparison of marketplaces and direct licensing and the comparison of Snowflake Marketplace and AWS Data Exchange set the routes side by side.
Frequently asked questions
How do I subscribe to a data product on AWS Data Exchange?
Sign in to the AWS Data Exchange console with an AWS account, browse the catalogue and filter by data set type, and open a product's page to read its offers, data dictionary and samples. Choose a price and duration pair and whether to auto-renew, then subscribe. Some products first need a subscription verification request that the provider approves. When the subscription is active the data sets appear as entitled data sets, and Files data sets can be exported to S3.
What is the difference between a data grant and a subscription on AWS Data Exchange?
A data grant is created by a sender, who needs only an AWS account, and sent to a receiver, who accepts it to gain access. A subscription is a buyer's agreement to a provider's offer for a product published in the AWS Marketplace catalogue, with a price, a duration and a data subscription agreement. Both give access to the same five types of data set.
What happens to the data when an AWS Data Exchange subscription ends?
You can no longer view or export the data sets, and access to an Amazon Redshift datashare ends. For a Files data set you keep the files you already exported, and the data subscription agreement states whether you must delete them. If the provider unpublished the product, access continues until the subscription expires, and the subscription cannot be renewed.
How is AWS Data Exchange billed?
A paid subscription appears on your AWS bill as part of your AWS Marketplace charges, on the payment schedule in the offer, and an upfront commitment is invoiced immediately. The services you use with the data, such as Amazon S3 or Athena, are billed separately. Prices are set by each provider in its offer.
Can I use company data delivered direct with AWS services?
Yes. AWS services read data from your own storage whatever its source. Fokals is delivered direct, by REST API and as bulk files: you download the files, write them to your own S3 bucket and load them into Amazon Redshift or query them with Athena, as the guides on this site show.
What this page says about the products it names was checked against their public documentation on 4 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.