Snowflake Opens Iceberg Data Sharing to External Tools
Snowflake AI & Data Sharing now lets providers share supported Apache Iceberg tables with consumers using Iceberg REST Catalog compatible clients, including Apache Spark. The company says users can work with shared data in their existing tools and compute environments without scheduled exports or additional copies.
- Snowflake enables Open Data Sharing through the Iceberg REST Catalog.
- Providers can share supported Iceberg tables with external IRC-compatible clients.
- Apache Spark is among the tools that can access shared data.
- Snowflake provides identities and controls to manage or revoke access.
- The feature complements Marketplace, reader accounts and third-party accounts.
Sharing data through an open protocol
The new Open Data Sharing capability is powered by Apache Polaris and uses the Iceberg REST Catalog, or IRC, as its interface for external consumers. Clients that implement the API can connect to shared Iceberg data, allowing consumers to analyze it outside Snowflake.
Snowflake describes IRC as an interoperability layer rather than a Snowflake-specific protocol. According to the company, more than 40 platforms and tools currently read from or integrate with an IRC ecosystem, based on the official community matrix from the Apache Iceberg vendors page and independent Iceberg compatibility matrices.
For providers, the service adds management features around that open interface. These include Snowflake shares and listings, dedicated external consumer identities, and controls to grant, manage and revoke access. Snowflake says the sharing experience uses IRC exclusively.
Part of Snowflake's broader move toward open exchange
Snowflake positions Open Data Sharing as the latest step in a series of additions focused on interoperability. In October 2025, the company enabled sharing of open table formats such as Apache Iceberg and Delta Lake across clouds and regions. In August 2026, it added databases linked to catalogs including Apache Polaris, Unity Catalog and AWS Glue. Open Data Sharing followed in October 2026, extending provider-to-consumer sharing to external IRC-compatible clients.
The company says open formats make data portable, while open protocols provide a common way for compatible systems to connect. Open Data Sharing does not replace Snowflake's other sharing options. Snowflake Marketplace remains a route to consumers within the Snowflake environment, while reader and third-party accounts remain options for external consumers who want a managed experience without IRC.
Use cases for providers and external consumers
Providers can share supported Iceberg tables and dynamic Iceberg tables. Snowflake says other supported table types can be prepared for sharing by materializing them into a dynamic Iceberg table. Those tables can also support filtered data, including data based on customer-specific requirements.
For example, a provider could make a customer-specific Iceberg table available in a partner's Apache Spark environment. The partner would connect through an IRC-compatible client and read the data without adopting Snowflake, while the provider could manage and revoke access through Snowflake.
This gives organizations distributing data products to customers, partners or other external consumers an alternative to scheduled exports, FTP workflows and legacy Delta Sharing, which Snowflake says was recently rebranded as OpenSharing. For users, the stated benefit is access to current shared data in existing environments, while providers retain centralized controls over external relationships.
Snowflake: Open Data Sharing: Share Live Data Beyond Snowflake
