Databricks has announced the launch of OpenSharing, a new open-source protocol designed to change how organizations share artificial intelligence assets and data. Developed as the next version of the Delta Sharing protocol, OpenSharing is now a project under the Linux Foundation.
Delta Sharing, launched by Databricks in 2021, focused on open data sharing across platforms. OpenSharing extends that foundation to cover AI-era assets like agent skills and AI models, adds support for Apache Iceberg IRC clients, and enables on-premises and private-cloud data sources to connect directly to cloud platforms without data movement.
The protocol arrives during a shift toward agentic AI, where autonomous systems require easy access to tools and information. By making the protocol open source, Databricks aims to create a universal standard for the industry.
“Delta Sharing proved the industry would choose open over locked-in,” said Matei Zaharia, Co-founder and CTO of Databricks. “OpenSharing extends that principle to the full AI stack, while expanding the cross-platform ecosystem to Iceberg recipients and on-premises providers. The agentic era deserves an open foundation, and OpenSharing delivers it.”
Sharing AI Assets Across Organizations
A major feature of OpenSharing is its ability to handle AI models and agent skills. Previously, companies had to build custom integrations or use specific marketplaces to share these tools. This often led to high costs and technical barriers.
With OpenSharing, enterprises can publish agent skills via a single protocol. Partners can then discover and access these assets using standard APIs. This “zero-copy” approach means data and models do not need to be manually moved or copied, which improves security and efficiency.
OpenSharing also addresses the issue of vendor lock-in for structured data. While the original Delta Sharing protocol allowed collaboration across several platforms like Spark, Oracle, and Snowflake, OpenSharing adds support for Apache Iceberg APIs.
This expansion allows data providers to reach a wider audience. Companies using Iceberg-native tools can now receive shared data through the same protocol used by the rest of the ecosystem. This makes it easier for providers to publish their data once and reach users on different platforms.
“We believe in open AI ecosystems and are excited to collaborate with Databricks on providing a standard, secure way to discover and authorize access to AI assets,” said Alexander Embiricos, Head of Enterprise Product at OpenAI.
Connecting Cloud AI to On-Premises Data
For many organizations, moving sensitive data to the cloud is not an option due to regulations or security concerns. OpenSharing introduces native integrations with on-premises storage partners like MinIO and Qumulo.
These integrations allow cloud-based AI and analytics platforms to connect directly to data stored on-premises. This gives companies the ability to use modern AI tools without moving their information. Several other partners, including NetApp and Dell, are expected to add support in the future.
The project is currently available to the public. Interested developers and organizations can access the OpenSharing protocol on GitHub or visit the project’s official website to learn more about implementation.
