SAP Acquires Dremio to Deepen Apache Iceberg Strategy

SAP has acquired Dremio, a data integration and analytics platform built on Apache Iceberg, in a move that significantly reshapes the enterprise data landscape. The deal positions the ERP giant to bring external data sources directly into its analytics and AI agent-building ecosystem — and notably reduces its prior dependence on Databricks for open data lakehouse capabilities.

While financial terms of the acquisition have not been publicly disclosed, the strategic implications are substantial. SAP is effectively betting that owning the Iceberg integration layer will give it a competitive edge as enterprises race to unify structured ERP data with the sprawling, unstructured data lakes that fuel modern AI workloads.

Key Takeaways at a Glance

• SAP acquires Dremio, a leading data lakehouse platform built natively on Apache Iceberg
• The deal reduces SAP's reliance on Databricks, which previously served as its primary integration bridge to external data
• Dremio's technology enables high-performance SQL queries directly on data lake storage without copying or moving data
• The acquisition strengthens SAP's Business Data Cloud and Joule AI agent strategy
• Apache Iceberg has emerged as the dominant open table format, backed by companies like Apple, Netflix, and Airbnb
• SAP joins a growing list of major vendors — including Snowflake and Confluent — making aggressive Iceberg bets

Why Dremio Matters to SAP's Data Strategy

Dremio has carved out a distinctive position in the data infrastructure market by building a lakehouse engine that queries data in place. Unlike traditional data warehouses that require costly ETL (extract, transform, load) pipelines, Dremio lets analysts and AI systems run SQL queries directly against data stored in formats like Apache Parquet on cloud object storage — all governed by Iceberg's table format.

This 'query-in-place' architecture is precisely what SAP needs. The company's core strength lies in the transactional data flowing through its ERP, supply chain, and HR systems. But enterprise AI agents — like those SAP is building with its Joule platform — need access to far more than just ERP records. They need customer interaction logs, IoT sensor data, third-party market data, and unstructured documents scattered across cloud data lakes.

Previously, SAP relied on a partnership with Databricks to bridge this gap. That partnership, announced with considerable fanfare, allowed SAP's Business Data Cloud to federate queries across SAP and non-SAP data sources. But relying on a partner — one that competes in adjacent markets — introduced strategic risk. Owning Dremio gives SAP direct control over the integration layer.

The Apache Iceberg Factor

Apache Iceberg has rapidly become the de facto standard for open table formats in the data lakehouse world. Originally developed at Netflix, Iceberg provides a metadata layer that makes data lakes behave more like traditional data warehouses — with features like ACID transactions, schema evolution, time travel, and partition pruning.

The format's rise has been nothing short of meteoric:

• Snowflake launched Iceberg Tables as a core feature and acquired related startups
• Databricks added Iceberg compatibility to its Unity Catalog after initially championing its own Delta Lake format
• Confluent integrated Iceberg support into its Kafka-based streaming platform via its Tableflow product
• AWS built Iceberg support directly into services like Athena, Glue, and EMR
• Apple runs one of the world's largest Iceberg deployments, managing exabytes of data

By acquiring Dremio — one of the most prominent Iceberg-native platforms — SAP is planting its flag firmly in this ecosystem. It is a clear signal that SAP views open formats as the future of enterprise data, not proprietary silos.

A Strategic Pivot Away from Databricks Dependence

The Databricks relationship is the subtext that makes this acquisition particularly interesting. SAP and Databricks announced a deep partnership in 2023, positioning Databricks' lakehouse platform as the connective tissue between SAP's structured data and the broader data ecosystem. The integration was real and technically functional, but it created an uncomfortable dynamic.

Databricks is not just a data integration company. It competes aggressively in AI and machine learning with its Mosaic ML infrastructure and Dolly language models. It sells its own analytics tools. And it has its own vision for how enterprise AI should work — one that does not necessarily center SAP.

By bringing Dremio in-house, SAP can now:

• Control the roadmap for how external data integrates with SAP systems
• Eliminate licensing dependencies on a potential competitor
• Accelerate Iceberg-native features in SAP Business Data Cloud without waiting for a partner's release cycle
• Differentiate its AI agents by offering seamless access to lakehouse data that competitors cannot match
• Reduce total cost of ownership for customers who previously needed separate Databricks contracts

This does not necessarily mean SAP will sever ties with Databricks entirely. Large enterprises often maintain multiple data platform relationships. But the center of gravity has clearly shifted.

Implications for SAP's AI Agent Ambitions

SAP has been aggressively building out its Joule AI assistant and a broader ecosystem of domain-specific AI agents designed for procurement, finance, HR, and supply chain workflows. These agents need data — lots of it — and they need it from sources that extend well beyond traditional SAP databases.

Consider a procurement AI agent that needs to evaluate supplier risk. It might need SAP purchase order data, but also third-party credit ratings stored in a Snowflake warehouse, news sentiment data in an S3 data lake, and IoT quality metrics from a manufacturing partner's systems. Dremio's federated query engine can pull all of this together without requiring massive data movement or duplication.

This 'data gravity' problem is one of the biggest obstacles to enterprise AI adoption. Models and agents are only as good as the data they can access. By owning Dremio, SAP can position its AI agents as uniquely capable of reasoning across the full breadth of enterprise data — not just the data that happens to live in SAP systems.

The timing aligns with broader industry trends. Gartner has predicted that by 2026, more than 75% of organizations will adopt a data fabric or data mesh architecture. Dremio's technology maps directly to this vision, providing the semantic and query layer that makes distributed data accessible without centralization.

What This Means for Enterprises and Developers

For existing Dremio customers, the acquisition raises both opportunities and questions. On the positive side, deeper SAP investment could accelerate Dremio's product development, expand its global support infrastructure, and improve integration with one of the world's largest enterprise software ecosystems.

However, some Dremio users — particularly those who chose the platform specifically because it was vendor-neutral — may worry about SAP-centric roadmap priorities. SAP will need to tread carefully to maintain Dremio's appeal to non-SAP shops.

For SAP customers, the benefits are more clear-cut:

• Simplified architecture for accessing non-SAP data from within SAP analytics tools
• Potential cost savings by eliminating redundant data integration middleware
• Faster time-to-value for AI agent deployments that require external data
• A clearer path to adopting open data formats without abandoning SAP investments

For developers and data engineers, the acquisition validates the Iceberg ecosystem as the winning bet in the open table format wars. Skills in Iceberg, Parquet, and federated query engines are increasingly valuable across the enterprise stack.

Looking Ahead: The Data Platform Wars Intensify

SAP's acquisition of Dremio is the latest salvo in an intensifying battle among enterprise software giants to control the data layer that feeds AI. Microsoft has its Fabric platform. Google is pushing BigQuery with Iceberg support. Oracle is expanding its Autonomous Database capabilities. And Snowflake continues to evolve from a pure data warehouse into a broader application platform.

The common thread across all of these moves is a recognition that data access is the bottleneck for enterprise AI. The models are increasingly commoditized. The real competitive advantage lies in connecting AI systems to the right data, at the right time, with the right governance.

SAP's bet on Dremio and Iceberg suggests it sees the future of enterprise AI not in building the biggest language model, but in building the most comprehensive data access layer. If SAP can make its AI agents the easiest path to unified enterprise intelligence — spanning ERP, data lakes, and third-party sources — it could lock in its position as the indispensable platform for large-scale business operations.

The integration timeline remains to be seen, but expect SAP to showcase Dremio-powered capabilities at its next SAP Sapphire conference. For the broader data industry, the message is clear: the era of open table formats is here, and the biggest enterprise vendors are racing to own the stack that sits on top of them.