Data marketplaces compared: Snowflake, Databricks, AWS

Data marketplaces compared: buyers, listing rules and fees

Fees and requirements are as each company publishes them, checked November 5, 2026. "Not published" means the figure is not on any official page we could find.

How to sell data on Snowflake Marketplace

Snowflake requires a full account; trial accounts can share data with named consumers but cannot list on the Marketplace. An organisation administrator must accept the Snowflake Provider and Consumer Terms, and your provider profile must be approved before you offer paid or Marketplace listings. Paid listings need a billing address in one of 20 countries, including the United States, United Kingdom, Canada and Australia, and a Stripe Express connected account for payouts.

You attach one pricing plan to a listing: usage-based, charging per query, a monthly fee or both, or subscription-based for a set term. The provider is the seller of record. Snowflake invoices buyers in US dollars on your behalf and pays out 30 days after the buyer pays in full, minus a transaction fee whose schedule is shown only inside your account. You stay responsible for collecting if a buyer does not pay.

Partners earned more than $100 million in gross bookings on the Marketplace in the first half of 2026, up 277% year on year, across more than 1,700 transactions, Snowflake says. Its named data providers include S&P Global, FactSet and Nasdaq. Its policies allow personal data in products only if you are legally authorised to share it, ban sensitive personal data in public products and expect replies to buyer inquiries within three business days.

Source: Snowflake: use listings as a provider

How to list data on Databricks Marketplace

Databricks Marketplace lists datasets, AI models, notebooks, apps and MCP servers. Datasets are usually catalogs of tabular data, with non-tabular data supported as volumes. To list, you need a Databricks account on the Premium plan or above and a workspace enabled for Unity Catalog. Public listings require an application through the Databricks Data Partner Program; private exchanges, visible only to member buyers, are self-service.

Each listing is either instantly accessible, which requires a share, or available on request. If you need a business agreement first, as with paid data, you choose the request option and manage the agreement with the buyer outside Databricks Marketplace. Databricks' documentation does not publish a marketplace fee. Buyers can use up to 10% of their Databricks Universal Commits on eligible third-party products in the Marketplace.

The provider policies require all necessary rights to the data, a reply within three business days to access requests, dataset updates at the most frequent logical refresh rate, and at least 30 days' notice before you stop supplying a product. Listing pages may not include personal data; products may include it only if you are legally authorised to share it, and never sensitive personal data.

Source: Databricks: become a Databricks Marketplace provider

How to sell data on AWS Data Exchange

To become a provider, you register as an AWS Marketplace seller and are then qualified by the AWS Data Exchange team through a support case. You need a defined customer support process, a way to keep data regularly updated, and residence or incorporation in an eligible jurisdiction such as the United States, United Kingdom or an EU member state. Paid products also need tax and bank details.

Products can contain Amazon S3 files, APIs and Amazon Redshift data shares, with subscriptions of 1 to 36 months; API products must run through Amazon API Gateway. The catalog holds more than 3,500 data products. Publishing guidelines bar data that can identify a person unless it is publicly available, outside the Extended Provider Program, and require sensitive categories to be aggregated or anonymised.

The listing fee is 3% on public offers. Private offers cost 3% under $1 million, 2% from $1 million to under $10 million and 1.5% at $10 million or more; renewals cost 1.5%. AWS pays monthly, only after it collects from the subscriber.

Source: AWS Data Exchange: getting started as a provider

How to sell data on Google Cloud Marketplace

On Google Cloud you sell data through BigQuery sharing (formerly Analytics Hub), linked to Google Cloud Marketplace, with paid, free or trial subscription plans. You create a BigQuery sharing listing, choose List on Marketplace, then onboard it as a data product in the Producer Portal for review. Publishers and buyers must be in a supported Cloud Marketplace jurisdiction.

New vendors sign up, complete an onboarding validation form for every listing and submit each pricing configuration for review; an approved listing can take up to two days to appear. Pricing is free or a flat subscription fee, with custom quotes for specific customers. Data products cannot contain "personally identifiable sensitive information" as defined in the Protecting Americans' Data from Foreign Adversaries Act of 2024.

Since April 21, 2025, Marketplace transactions follow Google's Vendor Net Revenue Schedule where it applies: eligible vendors keep 97% on standard offers and on private offers under $1 million, 98% from $1 million to under $10 million, and 98.5% at $10 million or more.

Source: Google Cloud Marketplace: offer data products

Datarade and Neudata: selling beyond one cloud

Datarade is a B2B data marketplace for organisations searching for commercial data, including AI training datasets. It accepts registered businesses with the legal right to commercialise their data and usually reviews applications within 1 to 2 business days. Its Commission-only plan is free with a 30% commission; Bronze costs $6,000 a year plus 20%, Silver $12,000 a year plus 15%, and Gold is custom. It bans unanonymised sensitive personal data and illegally scraped datasets.

Neudata is a data scout rather than a marketplace. Listing is free, it takes no cut of sales, and it lets you run the buyer relationship yourself. Listings have historically averaged 2.6 free buyer introductions a year. It says 16 of the top 20 discretionary hedge funds use it, and it now also helps AI companies source training data; it does not sell datasets or earn commission on them.

For investment-fund buyers specifically, including data scouts and matchmaking events, see where to sell data to hedge funds.

Source: Datarade: apply to become a data provider

Which data fits a data marketplace?

Marketplaces work best for one kind of data. Compare yours against both columns.

When does a direct license to an AI developer fit better?

When the records are internal and the buyer wants them for AI training. AI developers license tickets, email, documents and code through their own programs and negotiated deals, and most of those sales are private, according to Origin's deals tracker. Marketplace rules on personal data and regular updates rarely fit a one-time archive of internal records.

A direct license lets you set the scope, permitted uses, de-identification and exclusivity for one buyer. In Origin's buyer programs study, the typical published range was about $50K to $843K per company, with published payment timing from 2 to 4 weeks to about 60 days. These are ceilings, not offers.

A marketplace fits better when you have a clean, structured dataset you can refresh, buyers already on that cloud, and a team to support subscribers. You can do both: list a structured product and license internal records separately. See how to sell data for all four routes.

Source: Origin Data Partners: what AI buyers publish for company data, 8 programs compared (Oct 2026)

Next step

Decide which column of the fit table your data sits in before you build a listing. If it is internal records, start with a company-level description: the system, years covered and volume. See what company data is worth.

Origin Data Partners works with companies that own data and helps get it bought. We are independent of the services on this page, and a buyer may pay us when a deal closes (how Origin is paid). See what your data is worth.