Key Takeaways
- SAP has integrated Dremio's Apache Iceberg-native lakehouse capabilities into SAP Business Data Cloud.
- The architecture is designed to reduce data copying while extending governed access across SAP and non-SAP systems.
- Open catalogs, semantic context, and elastic execution support moving agentic AI from pilots into production environments.
SAP has completed its acquisition of Dremio, expanding SAP Business Data Cloud with an open data lakehouse platform intended to support analytics and agentic AI across distributed enterprise information.
The transaction closed on July 6, 2026, after the initial agreement was announced in May. Financial terms were not disclosed. According to SAP, the combination lets SAP and non-SAP data coexist on an open foundation without requiring organizations to move or convert the data into another format.
Enterprise data is frequently divided among SAP applications, external databases, cloud object stores, operational systems, and specialized analytics platforms. Copying this information into a central repository introduces storage costs, transformation pipelines, governance hurdles, and operational delays.
Dremio addresses these friction points through Apache Iceberg, an open table format that allows different processing engines to work with the same underlying data. Dremio processes Iceberg data natively and supports federated queries across external sources, allowing organizations to analyze information where it resides rather than relying exclusively on repeated extraction, transformation, and loading.
While SAP HANA Cloud continues to provide in-memory processing for transactional and operational workloads, Dremio adds lakehouse management and federated analytics. This creates a broader architecture for workloads crossing application and platform boundaries.
Broader data access does not automatically produce reliable AI outcomes; an agent must also understand what the information means, how records relate to one another, who holds permission to use them, and whether the data is current. Dremio's Open Catalog directly addresses this challenge. Built around Apache Polaris and the Apache Iceberg REST Catalog API, it provides discovery, access controls, lineage, and semantic information to SAP and non-SAP processing engines. Dremio has contributed to Apache Iceberg, Apache Polaris, and Apache Arrow, and SAP has stated it intends to continue investing in these open-source projects.
The catalog is expected to support the SAP Knowledge Graph to connect business definitions, organizational structures, regulatory classifications, and data relationships across systems. When an AI agent encounters conflicting definitions of revenue, inventory, or customer status, a lack of shared semantics can lead it to choose the wrong input while still generating a plausible response. Integrating a catalog with a knowledge graph establishes a consistent context for retrieval and reasoning, mitigating these risks for autonomous agents recommending operational decisions.
The acquisition also brings serverless, elastic execution into SAP's data portfolio, allowing computing resources to expand when analytical demand rises and contract when activity falls. This model suits agentic workloads, which often arrive unpredictably and generate bursts of queries across multiple systems. Dremio supports access controls extending from users and AI agents to underlying data sources, alongside Model Context Protocol interfaces for connecting agents to enterprise information. This combination of variable compute, federated querying, and policy-based access lets organizations deploy additional agents without provisioning every environment for peak demand.
With this integration, SAP Business Data Cloud sits directly alongside unified data offerings from Databricks and Snowflake. All three vendors are positioning governed enterprise data as the foundation for generative and agentic AI, although their application footprints, storage strategies, and processing models differ. Research from the Futurum Group frames the SAP and Dremio combination around openness, interoperability, and reducing unnecessary data movement.
Enterprise adoption will likely depend on how consistently SAP applies governance across heterogeneous sources, how catalog metadata travels between engines, and whether federated performance holds up under demanding production workloads. By acquiring technology for open tables, distributed queries, semantic discovery, and elastic execution, SAP strengthens SAP Business Data Cloud as a control layer for organizations scaling AI agents over governed business data without rebuilding their entire information estate.
⬇️