SAP to acquire Dremio to bridge data fragmentation in agentic AI workloads

SAP SE has reached an agreement to acquire the high-performance data lakehouse platform Dremio. The acquisition is designed to integrate SAP and non-SAP data via an open Apache Iceberg-native architecture, aiming to eliminate the data engineering bottlenecks that currently hinder the scalability of enterprise AI initiatives.

SAP SE has finalized an agreement to purchase Dremio, a leading provider of open data lakehouse technology, with the transaction expected to close in the third quarter of 2026. According to official disclosures, the deal aims to transform the SAP Business Data Cloud into a unified environment capable of processing both proprietary and third-party data without costly format conversions.

SAP is a global leader in enterprise application software, while Dremio specializes in high-performance, serverless data architectures based on open-source standards. This development is significant because the primary obstacle to successful AI implementation remains fragmented data silos. By adopting Dremio’s Apache Iceberg-native foundation, SAP provides a single governed platform that preserves business context across disparate sources.

This move directly addresses the high failure rates of AI pilots by ensuring that data is accessible and explainable, which is essential for meeting increasingly stringent regulatory and compliance requirements. For the information technology and data management sectors, this acquisition reinforces the industry-wide shift toward open table formats and universal catalogs. Impacted stakeholders, including data engineers and CIOs, will likely see a reduction in manual integration workloads and a shift toward elastic, serverless economics that scale with demand.

Strategically, for investors and enterprise clients, the integration of Dremio into the SAP Knowledge Graph provides a competitive edge in developing agentic AI. This approach allows businesses to bypass traditional data movement hurdles, significantly accelerating the time-to-value for real-time analytical workloads and high-density computing projects.