Dwight Evers

A Fortune 500 semiconductor manufacturer generates hundreds of terabytes of production data across facilities in two international regions. That data must stay on-premises for regulatory compliance, latency sensitivity, and economics of scale — but the data science team depends on Databricks for the AI modeling that drives yield improvement and quality assurance. Rather than replicating data to the cloud or building custom ETL pipelines, they deployed MinIO AIStor with native Delta Sharing embedded at the storage layer. Databricks queries live Delta and Iceberg tables on-premises through Unity Catalog with zero data movement — no copies, no pipelines, no additional infrastructure. In this session, hear how the solution was architected, what decisions were made and why, and what the team learned operating it in production. Includes perspectives from MinIO and Databricks, a technical walkthrough, and a live demo of AIStor Table Sharing connected to a Databricks workspace.

Organizations need to give Databricks direct access to on-premises data without copying it or building pipelines. Traditional approaches introduce duplication, complexity and risk, especially when sovereignty or latency requirements prevent cloud movement. AIStor Table Sharing is MinIO's native Delta Sharing protocol implementation embedded in the AIStor binary, delivering streamlined administration, consistency and full governance while enabling secure, read-only access for consumers such as Databricks. Key highlights: - Native Delta Lake and Apache Iceberg table support - Authentication via bearer tokens and OAuth2/OIDC - Zero-downtime share, schema, and table administration - Bare-metal, containerized, and hybrid deployment support - Practical guidance deploying with Databricks Attendees will learn how AIStor Table Sharing eliminates data movement, reduces egress costs, enforces granular security and enables hybrid analytics while keeping data under organizational control.