Snowflake Marketplace and AWS Data Exchange are both places where a data product is listed, bought and delivered inside a cloud platform, and they do the delivering in different ways. This comparison sets out, for a buyer and then for a provider, what each data marketplace lists, how the data reaches you, how you are billed and what it takes to sell there. Every statement about either service comes from Snowflake's or AWS's own pages, read on 4 October 2026.
It ends with a third route: data licensed direct from its vendor. A listing makes delivery part of the platform. Direct delivery gives you the files themselves, a dated record that you hold. Fokals is delivered direct, by REST API and as bulk files, which you load with the platform's own loader, and the closing sections set that route beside the two marketplaces.
The two side by side
| Snowflake Marketplace | AWS Data Exchange | |
|---|---|---|
| What a listing holds | A share or a Snowflake Native App, with metadata such as sample SQL queries | A data product in the AWS Marketplace catalogue, holding one or more data sets |
| Forms the data takes | The tables and other objects of a share, or an app | Files, an API, an Amazon Redshift datashare, access to the provider's Amazon S3 data, AWS Lake Formation permissions (Preview) |
| How it reaches you | A read-only database in your Snowflake account, with no data copied | By form: exported files, API calls, or read-only access in place |
| Updates | Changes to shared objects are visible to consumers at once | New revisions, each announced by an event you can automate on |
| Ways to offer it | Free, a limited trial, or paid | Public offers, private offers and Bring Your Own Subscription offers |
| Paid pricing | Usage-based, billed in arrears; a subscription billed upfront; or an installment plan | A price for each subscription duration, or pay-as-you-go billed monthly |
| The bill | A Marketplace invoice, separate from other Snowflake invoices, in US dollars | Part of your AWS Marketplace charges, in US dollars, apart from the AWS services you use |
| Who can sell | A full Snowflake account with an approved provider profile | A registered AWS Marketplace seller whose data meets the publishing guidelines |
How the data reaches a buyer
A listing, in Snowflake's documentation, is an enhanced method of Secure Data Sharing that packages a share or a Snowflake Native App with metadata. When you get a listing that holds a share, a read-only database is created in your account from it. Nothing is copied: Secure Data Sharing is documented as copying no data, the shared objects take up no storage on your side, and you pay only for the compute your queries use. The Marketplace is available to Snowflake accounts hosted on Amazon Web Services, Google Cloud and Microsoft Azure, with Microsoft Azure Government the stated exception. When the provider's data sits in another region, a feature called Cross-Cloud Auto-Fulfillment lets the provider have the product replicated to yours.
AWS Data Exchange delivers according to the kind of data set, and its user guide names five.
| Data set type | What the subscriber does with it |
|---|---|
| Files | Exports the files to its own Amazon S3 bucket or downloads them through a signed URL |
| API | Calls an endpoint managed by AWS Data Exchange, which passes the request to the provider's API |
| Amazon Redshift | Queries the provider's datashare, read-only, with no extract or load |
| Amazon S3 data access | Reads the provider's own S3 objects in place, with no copy made |
| AWS Lake Formation (Preview) | Queries the provider's data lake through services integrated with Lake Formation |
The difference shows most clearly when the data changes. A share shows the provider's tables as they are now. A Files product arrives as revisions: the provider finalises one, it cannot then be changed, only revoked, and AWS sends an event you can use to start your own load. A subscriber can have new revisions exported to its S3 buckets automatically.
That has a consequence for research. An exported revision is a dated object that you hold. A share is a live view, so a query you ran last month returns today's answer today. If you need the data as it stood on a past date, ask any provider that delivers by share whether each row carries the time it was observed and whether past rows are ever revised.
It also decides what happens at the end. AWS's page on subscriptions says that when a subscription to a Files data set ends, you keep the files you already exported, and that you should read the data subscription agreement to see whether it requires you to delete them. Snowflake's page on Secure Data Sharing says a provider can revoke access to a share at any time.
How a buyer is billed
On Snowflake a free listing costs only the compute of your own queries. A paid listing follows one of the pricing models the provider chose.
- Usage-based. A charge for each query, a fee for each calendar month in which you run at least one query, or both, and for an application a charge for specific types of usage. Plans with a per-query charge carry a maximum monthly charge, after which further queries in that month are free. You are billed in arrears, for months with usage.
- Subscription-based. One price paid upfront for a term that the provider sets, recurring or for a single term.
- Installment. The provider divides the total price into portions, and you pay one portion at a time.
Snowflake's page on paying for listings adds the mechanics. Purchases are billed in US dollars on Marketplace invoices that are separate from your other Snowflake invoices. Payment is by credit card, by bank transfer or, for eligible organisations and listings, from funds reserved out of committed Snowflake capacity. Paid listings are open only to organisations whose billing address is in a supported country. Under a usage-based plan a query that touches paid data can be billed even when it returns no rows.
AWS says that every Data Exchange product is subscription-based. A public offer lists a price for each duration the provider offers, and a provider can make a private offer with its own terms. The pricing page describes two models: subscription products billed upfront or on a schedule, and pay-as-you-go products billed for the usage of each calendar month. The subscription appears among your AWS Marketplace charges. The AWS services you use to store and query the data are billed separately, and standard transfer rates apply when you export files across Regions or through signed URLs.
Prices, trial lengths, subscription terms and fees change, so read the current ones on Snowflake's pricing models page and AWS's pricing page, both linked above, on the day you buy.
Two features of AWS Data Exchange concern a buyer's contract more than its bill. A provider can turn on subscription verification, so that you request a subscription and state your intended use before you are admitted. A Bring Your Own Subscription offer moves a subscription you already hold with a provider onto AWS Data Exchange, with the billing relationship left between the two of you.
What it takes to sell on each
Snowflake's page for providers asks for a full account, a role with provider privileges, acceptance of the Snowflake Provider and Consumer Terms, and a provider profile that Snowflake has approved. To charge, a provider needs a billing address in one of the countries the page lists and must set up payouts through Stripe. Before creating a paid listing for the Marketplace it must contact its business development partner at Snowflake, or submit a case with Marketplace Operations if it has none, which the page calls required for listing approval. The pricing models page adds that trials are required for listings offered publicly on the Marketplace. Making the product available in other regions through auto-fulfillment adds cost for the provider.
AWS's guide to providing data products starts with registering as a provider and with data that meets the legal eligibility requirements and the publishing guidelines. The user guide adds that providers must be registered as AWS Marketplace sellers. The provider creates a data set, a revision and its assets, then a product and an offer. The offer carries the data subscription agreement, for which AWS supplies a template that the provider can replace, together with the prices by duration, the refund policy and the choice of auto-renewal. AWS reviews the product before it enters the catalogue, and AWS Marketplace distributes the funds. The pricing page lists what a provider pays: tiered fulfilment fees on the revenue AWS collects for new subscriptions, and storage fees for products that contain files. The same page says Bring Your Own Subscription offers are fulfilled at no additional cost.
Both services also have a route that skips the public catalogue. On Snowflake it is a private listing. On AWS it is a data grant, which any AWS account can send to another, and for which the sender pays for each hour the grant is active.
The same table by three routes
Suppose a daily dataset with one row per company and closed day, the shape of Hiring Activity in the Fokals hiring dataset. The work a buyer does each day depends on the route by which such a dataset arrives.
| Step | A Snowflake listing | An AWS Data Exchange Files product | Direct from Fokals |
|---|---|---|---|
| Get access | Get the listing, and a database appears | Subscribe, and an entitled data set appears | Sign the licence and receive an API key |
| Receive a new day | Nothing to do | A revision arrives, and you export it to S3 | Call the export endpoint for the closed day |
| Make it queryable | Query the shared database | Query the files in S3 or load them into a warehouse | Load the file with the loader of the platform you use |
| Hold history as delivered | What the provider keeps in the shared tables | Each revision you export, unchanged once finalised | Each file and its manifest, written once and never revised, which you keep |
The third column is how Fokals data arrives, on any platform. Each daily and weekly dataset is written once, after the period closes, and is never revised, and every record carries the time it was observed. An export package comes with a manifest naming its period, label versions and licence, so each delivery you hold is a dated record that reads the same on any later day. The delivery page describes the API and the exports. On Snowflake the load is a stage and COPY INTO, covered in loading company data into Snowflake. On AWS the files can stay in your bucket and be queried there, as querying company data on Amazon S3 with Athena shows.
Which fits, by use
- Your analytics run in Snowflake and you want no load. A listing puts the data beside your tables, and you pay for queries. The buyer's side is in Snowflake Marketplace: a guide for buyers.
- You need files, or an API, or your estate is on AWS. AWS Data Exchange offers all five forms its user guide lists. See AWS Data Exchange: a guide for buyers.
- You need each delivery kept as it was. A Files product on AWS Data Exchange gives you revisions you can keep.
- You sell a product that consumers query in Snowflake. Listing a data product on Snowflake Marketplace covers that side.
- The vendor delivers direct. License the data by agreement and load it with the platform's own loader. Data marketplace vs direct licensing weighs the two ways of buying.
If the product you would list is built on data licensed from someone else, including Fokals, check that the licence permits it. Fokals licenses internal use, embedding in a product or redistribution by written agreement.
Frequently asked questions
What is the difference between Snowflake Marketplace and AWS Data Exchange?
Snowflake Marketplace delivers a share or a Snowflake Native App into a Snowflake account: a read-only database appears, no data is copied and you pay for the compute of your queries. AWS Data Exchange's user guide lists five types of data set: files you export, an API you call, a Redshift datashare, access to the provider's S3 data, and Lake Formation permissions in preview.
What types of data products does AWS Data Exchange support?
Its user guide lists five types of data set: Files, API, Amazon Redshift, Amazon S3 data access and AWS Lake Formation, the last in preview. A product in the AWS Marketplace catalogue holds one or more data sets. Files are exported to your own bucket, and the other types give access to the provider's API, datashare, bucket or data lake.
How is a paid listing on Snowflake Marketplace billed?
The provider chooses a usage-based plan, a subscription or an installment plan. Usage-based plans charge for each query, for each calendar month with at least one query, or both, and are billed in arrears in months with usage. Subscriptions are paid upfront for a term the provider sets, and an installment plan divides the price into portions paid one at a time. Charges are in US dollars on a Marketplace invoice separate from other Snowflake invoices.
How am I billed for a product on AWS Data Exchange?
AWS's user guide says every product is subscription-based. You choose a duration from those the provider offers, and the subscription appears among your AWS Marketplace charges according to its payment schedule. The pricing page also describes pay-as-you-go products, billed for each calendar month's usage. The AWS services you use to store and query the data are billed separately.
Can I load files from AWS Data Exchange into Snowflake?
A Files product on AWS Data Exchange can be exported to your own Amazon S3 bucket, and Snowflake documents loading from an external stage over S3 with COPY INTO. The mechanics therefore allow it. Whether you may is a question for the data subscription agreement of the product, which sets the provider's terms of use.
How do I use Fokals data in Snowflake or on AWS?
Fokals is delivered direct, by REST API and as bulk files in CSV, JSON or JSON Lines, under a written agreement. In Snowflake you place the files in a stage and load them with COPY INTO. On AWS you place them in your own Amazon S3 bucket and query them there or load them into a warehouse. Each daily and weekly dataset is written once and never revised, and an export package carries a manifest naming its period, label versions and licence, so the tables you hold are your own dated copy.
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.