📅 November 5, 2026 • The Register
SAP is opening its data layer in both directions. Per The Register, the company announced that SAP Business Data Cloud, the data foundation of the SAP Business AI Platform, now supports native Apache Iceberg tables alongside a universal catalog built on Apache Polaris. The move is framed as a bidirectional data flow: data can enter the SAP environment in open formats, and data held by SAP can be consumed outside it without proprietary conversion. For a platform whose gravity has historically come from the ERP suite, publishing the data layer around open formats is a notable repositioning.
The practical consequence is that the same tables become reachable by two very different audiences. Human analytics teams query Iceberg data with the tools they already use, and the Joule AI agent platform reads from the same open layer. In an agentic architecture that alignment matters: an agent answering a business question and an analyst building a dashboard operate on one shared table rather than on divergent copies that must be reconciled. A shared substrate also cuts the reconciliation work that typically appears when analytics stacks drift away from operational systems.
SAP also promises zero-copy data exchange with Databricks, Snowflake, Google, AWS and Microsoft — data is shared by reference instead of being duplicated across platforms. The open-table push is reinforced by the acquisition of lakehouse vendor Dremio in May 2026: its SQL query engine is being embedded into the stack and is designed to run queries against sources such as SQL Server, MongoDB, Oracle and Postgres, extending open access beyond the Iceberg layer itself.
The strategic angle is decoupling. AI and analytics workloads no longer have to follow the ERP upgrade path: customers still running legacy ECC systems can attach a lakehouse without migrating to S/4HANA Cloud. Open table formats are becoming infrastructure for AI agents, and lock-in pressure around where data lives is decreasing. It also gives partners and customers a common table format to build on, from dashboards to agent workflows.
IMPACT: for data teams, native Iceberg means lakehouse analytics and Joule agents run against open tables with no proprietary export step. For architects, zero-copy exchange with Databricks, Snowflake and the major clouds reduces duplication cost and eases multi-platform designs. For SAP estates still on ECC, this is a way to modernize analytics ahead of any ERP migration decision.
Sources: theregister.com
Continue the conversation: https://t.me/AIAutomationArchitecture · Bluesky
Top comments (0)