Beyond the Walled Gardens: OpenSharing Signals a New Era for AI

SAN FRANCISCO, CA – June 10, 2026 – For the past few years, the narrative around artificial intelligence has been one of explosive growth, but also one of digital feudalism. The great AI platforms have been building taller walls around their gardens, hoarding the data and models that represent the new gold. Today, the Linux Foundation, in partnership with Databricks, just fired a cannonball at those walls. Their announcement of the OpenSharing Project isn't just another piece of open-source software; it's a declaration of independence for the coming agentic era.

This isn't merely a technical update for data engineers to debate over coffee. This is a fundamental shift in commercial strategy. We're moving from a world where competitive advantage came from owning proprietary assets to one where it will come from the ability to securely and seamlessly collaborate. OpenSharing provides the grammar for that new conversation, and for anyone trying to understand the 2026 consumer experience, this is the essential, unseen plumbing that will define it.

From Data Hoarding to Digital Handshakes

The central problem of the current AI landscape is friction. As enterprises rush to deploy autonomous AI agents, they face a crippling lack of standardization. Sharing a new AI model, a valuable dataset, or a reusable "agent skill" with a partner requires bespoke, costly, and insecure point-to-point integrations. The alternative has been to capitulate to a single vendor's marketplace, trading flexibility for convenience and inviting the very vendor lock-in that the cloud era was supposed to eliminate.

OpenSharing proposes a radical, yet simple, solution: a universal, open, and vendor-neutral protocol for exchange. Think of it as a digital handshake. Instead of physically moving and copying massive, sensitive datasets, the protocol allows organizations to grant secure, temporary, and direct access to assets right where they live. It’s an evolution of the widely adopted Delta Sharing protocol, but supercharged for the new demands of AI.

"OpenSharing addresses a critical need for a common, vendor-neutral framework that enables organizations to exchange AI assets securely and interoperably across platforms and ecosystems," said Jim Zemlin, CEO of the Linux Foundation. By housing the project, the foundation aims to "foster open collaboration, broad industry participation, and the shared governance needed to accelerate AI innovation at scale."

This is about creating a common language not just for data, but for the very building blocks of intelligent systems: AI models, unstructured data volumes, and—critically—reusable Agent Skills. It’s the kind of foundational standard that, like TCP/IP for the internet, could unlock an entirely new wave of innovation by allowing disparate systems to communicate fluently.

The 'Why' Behind the Open-Source Bet

The strategic genius of OpenSharing lies in its architecture and its governance. By contributing the protocol to the Linux Foundation, Databricks is making a powerful statement. "Delta Sharing proved the industry would choose open over locked-in," explained Matei Zaharia, Co-founder and CTO of Databricks. "OpenSharing extends that principle to the full AI stack... The agentic era deserves an open foundation, and OpenSharing delivers it."

This isn't pure altruism; it's smart business. By championing an open standard, Databricks positions its platform as a central, but not restrictive, hub in a thriving ecosystem. The protocol’s "zero-copy" approach is a cornerstone of this strategy. Assets remain in the provider's storage, and the sharing server simply provides temporary, scoped credentials for direct access. This elegantly solves major security, privacy, and cost headaches associated with duplicating petabytes of data across organizations.

Furthermore, the project shows a keen awareness of the existing landscape. Rather than forcing a "my way or the highway" approach, OpenSharing is designed for interoperability. It builds on Delta Sharing's existing connectors while adding support for Iceberg IRC clients, a nod to another popular open table format. This move to unite, rather than divide, the open-source data community is crucial for achieving the broad adoption needed for a standard to become a true standard.

Where the Digital Rubber Meets the Road

While the protocol's design is elegant, its true impact is measured by its adoption. The list of initial supporters reads like a who's-who of data-intensive industries, demonstrating immediate, tangible demand.

For the financial world, this is a game-changer. Ron Lefferts, a Divisional CEO at LSEG, noted that OpenSharing aligns with their strategy to "deliver trusted, AI-ready financial data and AI assets wherever our customers work, regardless of which cloud, tools, or AI model they want to use." Similarly, Emily Sands, Head of Data and AI at Stripe, emphasized that the protocol allows users to "securely and effortlessly unlock advanced analytics and AI capabilities on their customer, billing, and transaction data."

This isn't just a cloud-native phenomenon. One of the most significant aspects of OpenSharing is its ability to bridge the on-premises world with modern cloud AI platforms. For sectors with immense data gravity or strict regulatory constraints, like healthcare and finance, moving data is a non-starter. AB Periasamy, Co-Founder of MinIO, explained that native OpenSharing support "opens up access to data that cannot move," unlocking value from sensitive data that has been trapped on-premises.

This is echoed in healthcare, where Jeff McDonald, CEO of Kythera Labs, sees the protocol as a way to combat the chronic fragmentation of data. By delivering analysis-ready data directly into existing environments, it reduces friction and accelerates the journey from raw data to life-saving insights. In real estate, Cotality's Mark Weaver sees it as the framework for enabling collaboration around property intelligence and risk modeling.

The Unseen Architecture of Our AI Future

For the average consumer in 2026, the term "OpenSharing protocol" will likely remain unheard. But its impact will be felt everywhere. The seamless, context-aware AI assistants we've been promised are not going to be the product of a single, monolithic company. They will be orchestrators, assemblages of specialized skills and data streams from countless providers.

When you ask your AI agent to plan a vacation, it will need to securely query an airline's model for flight availability, a hotel chain's system for room pricing, and a review platform's database for recommendations. OpenSharing is the standard that makes this complex dance possible without every interaction being a clunky, pre-negotiated partnership or a privacy-violating data transfer.

This initiative isn't just about making data scientists' lives easier. It's about laying the foundational grammar for a future where AIs from different creators can communicate, collaborate, and create value together. It represents a maturation of the AI industry, a recognition that the next trillion dollars of value will be generated not by hoarding data, but by enabling its secure and intelligent flow across the entire economy.