Research Finder
Find by Keyword
Databricks OpenSharing Extends Data Sharing Into Governed AI Asset Exchange
OpenSharing builds on the foundation established by Delta Sharing and reflects a growing requirement to connect AI systems to information that will remain in existing enterprise environments
6/15/2026
Key Highlights
- Databricks OpenSharing extends the company's sharing strategy beyond traditional data collaboration and into broader AI consumption scenarios
- The participating ecosystem supports enterprise information where it already resides and where organizations are seeking greater AI value
- Linux Foundation governance gives OpenSharing a stronger claim as an open framework, but adoption beyond Databricks-led integrations will determine its industry role.
The News
Databricks announced OpenSharing, an open framework designed to provide governed access to enterprise information across a growing ecosystem of storage and infrastructure platforms. The initiative builds on Delta Sharing and is intended to make data and AI assets available to analytics, applications, and agentic AI workloads. The company simultaneously announced a Databricks Storage Ecosystem that includes support from MinIO, Everpure, and Qumulo at the time of the announcement, with additional participation planned from VAST Data, NetApp, HPE, Nutanix, Commvault, Cohesity, and Rubrik. For more details, read the official Databricks news release.
Analyst Take
Delta Sharing established a framework for governed collaboration across organizational and platform boundaries. OpenSharing extends that effort into a new phase of enterprise AI adoption. Organizations increasingly need AI systems to consume information that remains distributed across the enterprise, yet many continue to struggle with the complexity required to connect information to production AI initiatives.
The technical significance is that OpenSharing moves the sharing discussion beyond tables and dashboards and into the assets that increasingly surround AI systems: models, agent skills, unstructured information, and the authorization patterns required to expose those assets safely across platforms. Production AI does not only need access to governed data; it needs reliable access to the context, tools, metadata, policies, and reusable skills that allow agents and applications to act on that data without creating a new layer of brittle, one-off integrations.
HyperFRAME Research's State of Enterprise AI Stack 1H 2026 study finds that 78% of organizations consider AI strategically important, while only 37% have established a structured process for evaluation and deployment. Just 23% of AI and machine learning projects launched during the previous year reached production with measurable ROI. In addition, HyperFRAME Research's State of Infrastructure & Operations 1H 2026 study shows that only 14% of organizations report having an AI-ready data architecture, while 62% identify complexity as a significant barrier to infrastructure deployment and expansion.
These survey results echo conversations across our vendor briefings and customer discussions. Enterprise information remains distributed because it supports the systems, processes, and business activities that organizations depend upon every day. AI initiatives inherit that environment. The challenge increasingly centers on connecting AI systems to existing information, context, and reusable AI assets without introducing another cycle of migration projects, data movement, and custom integration work.
OpenSharing is designed on the premise that information will remain distributed. Through this framework, information can participate in analytics and AI workflows while remaining under existing governance and management models. That extends beyond storage interoperability by addressing the effort required to make enterprise information accessible to AI systems and analytics services.
Snowflake is pursuing a similar enterprise requirement through secure sharing, marketplace services, native applications, Horizon governance, and its broader effort to make governed data and context available to AI applications. The distinction is architectural. Snowflake’s approach is more platform-centered, while Databricks is positioning OpenSharing as an open protocol and ecosystem framework anchored in Delta Sharing, Unity Catalog, and Iceberg interoperability. Both strategies are credible, but neither eliminates the hard work of mapping permissions, lineage, policy enforcement, performance expectations, and cost controls across distributed environments.
The Enterprise Information Estate Comes Into Focus
The ecosystem reflects where enterprise information already resides, but the practical value of OpenSharing will depend on how quickly partner implementations move from structured and tabular access into the broader unstructured estates that matter for enterprise AI. NetApp, HPE, Nutanix, Everpure, and Qumulo bring established enterprise environments into the framework. VAST Data contributes an AI-focused perspective shaped by large-scale information environments that support retrieval and inference workloads. Commvault, Cohesity, and Rubrik highlight the growing importance of protected information estates as potential sources of AI value. Their inclusion reflects a broader shift in how organizations view information that has traditionally been associated with recovery and retention.
MinIO warrants attention because it is currently the only generally available OpenSharing implementation. The company demonstrates that OpenSharing is already progressing beyond roadmap discussions and into customer deployments. Its participation also reinforces a recurring theme throughout the announcement: organizations want to maintain direct control over information while making it accessible to AI and analytics services.
That distinction is important because much of the enterprise AI opportunity sits outside clean, structured datasets. Documents, images, support archives, engineering files, media assets, and backup repositories are often the raw material for retrieval, fine-tuning, and agentic workflows. Databricks is directionally aligned with that requirement, but enterprises should separate the OpenSharing framework from the maturity of each partner implementation. A standard reduces integration friction only when the connectors, governance mappings, performance characteristics, and operational controls are production-ready.
The expanding ecosystem connects Databricks to environments where enterprise information already resides. In our view, the success of OpenSharing will depend on how effectively it reduces the effort required to connect those environments to AI systems while preserving the governance structures organizations already rely upon.
What Was Announced
OpenSharing extends Databricks' sharing strategy through an open, vendor-neutral framework designed to provide governed access to enterprise information and AI assets across platforms and organizations. Building on Delta Sharing, the framework supports secure sharing of AI models, agent skills, unstructured information, and other assets required by modern AI applications. Together with Unity Catalog, Delta Sharing, and Apache Iceberg, OpenSharing advances Databricks' strategy of making enterprise information and AI assets more portable, accessible, and governable. The announcement also reinforces Databricks' continued support for Apache Iceberg as a common table format for enterprise analytics and AI workloads.
Databricks announced the framework alongside a Storage Ecosystem intended to connect information residing outside the Databricks platform while preserving existing governance and management controls. The company specifically highlighted environments where information remains outside traditional cloud analytics architectures because of regulatory requirements, sovereignty concerns, latency considerations, economic factors, or existing infrastructure investments. The program currently includes immediate support from MinIO, Everpure, and Qumulo, with additional integrations planned for VAST Data, NetApp, HPE, Nutanix, Commvault, Cohesity, and Rubrik.
Databricks also emphasized Linux Foundation governance for OpenSharing. This positions the framework as an open industry initiative and reflects the company's objective of establishing a common approach for exposing enterprise information and AI assets to analytics and AI services.
Looking Ahead
OpenSharing's long-term significance will depend on whether the industry views it as an ecosystem initiative or shared infrastructure. The Linux Foundation governance model suggests Databricks is pursuing broader adoption than a traditional partner program. Future ecosystem participation will provide an indication of whether the framework is evolving into a common industry standard or remaining primarily associated with Databricks deployments.
The next phase of this discussion will likely extend beyond information access. As organizations move AI initiatives into production, they are creating assets that carry business value in their own right. HyperFRAME Research expects growing attention around how enterprises govern, distribute, and consume the context, intelligence, and decision frameworks that support AI systems. OpenSharing provides an initial framework for governed exchange of data and AI assets, but the market opportunity extends further as organizations seek consistent approaches for sharing and governing the assets that surround AI deployments.
The ecosystem itself is also likely to evolve. Databricks already maintains deep relationships with hyperscale cloud providers and analytics environments where significant amounts of enterprise information reside today. Additional participation from information management vendors, governance technologies, AI infrastructure providers, and cloud ecosystems would expand the reach of the framework and provide stronger evidence that the industry is converging around a shared model for governed access across distributed enterprise information estates.
We will be watching whether OpenSharing becomes an important interoperability framework for enterprise AI or remains primarily a collection of integrations. The answer will depend on ecosystem growth, customer adoption, and the industry's willingness to embrace governed access as a foundational architectural principle for AI.
The real test will be whether OpenSharing becomes part of the control plane for AI systems, not just another interoperability announcement. For CIOs, the question is not whether assets can technically be shared; it is whether those assets can be discovered, authorized, governed, monitored, revoked, and consumed by AI applications and agents without creating new security gaps or cost surprises. If OpenSharing can help standardize that control plane across distributed data and AI assets, it becomes strategically important. If adoption remains concentrated around Databricks-led integrations, it will still be useful, but less transformative.
Don Gentile | Analyst-in-Residence -- Storage & Data Resiliency
Don Gentile brings three decades of experience turning complex enterprise technologies into clear, differentiated narratives that drive competitive relevance and market leadership. He has helped shape iconic infrastructure platforms including IBM z16 and z17 mainframes, HPE ProLiant servers, and HPE GreenLake — guiding strategies that connect technology innovation with customer needs and fast-moving market dynamics.
His current focus spans flash storage, storage area networking, hyperconverged infrastructure (HCI), software-defined storage (SDS), hybrid cloud storage, Ceph/open source, cyber resiliency, and emerging models for integrating AI workloads across storage and compute. By applying deep knowledge of infrastructure technologies with proven skills in positioning, content strategy, and thought leadership, Don helps vendors sharpen their story, differentiate their offerings, and achieve stronger competitive standing across business, media, and technical audiences.
Stephanie Walter | Practice Leader - AI Stack
Stephanie Walter is a results-driven technology executive and analyst in residence with over 20 years leading innovation in Cloud, SaaS, Middleware, Data, and AI. She has guided product life cycles from concept to go-to-market in both senior roles at IBM and fractional executive capacities, blending engineering expertise with business strategy and market insights. From software engineering and architecture to executive product management, Stephanie has driven large-scale transformations, developed technical talent, and solved complex challenges across startup, growth-stage, and enterprise environments.