Today, we’re thrilled to announce that Dremio has agreed to join forces with SAP, pending regulatory approval. Together, we will be able to deliver one open platform where agents reason over all enterprise data, decide, and act. This acquisition will give us the scale and backing to accelerate our agentic vision, while also incorporating innovative SAP technologies.
For SAP, it will bring the open, agentic lakehouse natively into the SAP Business Data Cloud. SAP customers will be able to combine SAP application data with every other source of enterprise data, and run their analytics, their agents, and their business on a single, open, governed data platform.
We are excited, and we are just getting started.
Explore this interactive demo and see how Dremio's Intelligent Lakehouse enables Agentic AI
When we started out, we saw the promise of the data lake, a place where all data could come together. At that time data lakes were mainly the domain of data scientists, not analysts. The existing query engines were slow, so while there was a lot of "potential" with data lakes, it largely went unrealized. This was the initial mission of Dremio, we set out to create the fastest data lake engine to enable business analytics on the lake. We co-created Apache Arrow, built our SQL engine on the technology, and delivered against our mission. The business started taking off.
As we gained traction with customers, it was clear that enterprises wanted to bring all of their data together, but they didn’t always want to set up pipelines and copy the data into their data lake. Query federation was the solution, but it was often too slow, so we set off to build lightning-fast query federation. We innovated, building a broad set of performance optimized connectors, advanced caching, and a technology called Reflections, which materializes data to automatically accelerate queries. Analysts can build a view of data that combines lake and federated data, query the view, and since we designed Reflections to be transparent to the analyst, they never had to rewrite the SQL. This innovation enabled sub-second performance on both large datasets and federated queries. With the fastest data lake engine and the best query federation, we kept growing fast.
It was clear that table formats such as Apache Iceberg and Delta Lake would elevate the data lake architecture to a new level by simplifying data management and enabling data warehouse workloads on the lake. But to achieve these objectives, it was critical to ensure that the industry would embrace an open format with a diverse ecosystem that was not controlled by a single vendor.
Recognizing Iceberg’s potential to revolutionize data management, Dremio became the first technology provider to embrace it in early 2021, and we set off on a mission to evangelize the project and build the leading Iceberg-native lakehouse. We rearchitected Dremio’s internals to use Iceberg throughout the engine. For example, Reflections mentioned above are Iceberg tables. We implemented full DML, built easy to use ingestion capabilities, and co-created Apache Polaris, the leading open-source catalog, with Snowflake. Iceberg did become the lakehouse standard, benefiting our customers and driving our continued growth.
Three years ago, we placed a bet others weren't yet making. Most data platforms had been built around the same assumption: a human is at the helm. A data engineer would orchestrate pipelines and curate data. A BI analyst would build semantic layers and dashboards. An application developer would build applications that leverage enterprise data. Generating a new dashboard would take 6-8 weeks, with the knowledge worker going back and forth with an analyst and the analyst going back and forth with the engineer.
We could see the world was about to change. Generative AI was going to put data directly in the hands of knowledge workers without specialized training, and also put data “in the hands” of autonomous agents that act on their own. We asked ourselves, how do we make it so anyone, human or agent, can move from question to answer to action without a queue of specialists in between?
So we pioneered the agentic lakehouse, the first analytical data platform designed from the ground up for agents, and managed by agents. We shipped the first MCP interface in the industry, built an AI-centric CLI for modern agent access, and created a unified and autonomous knowledge graph to provide agents the context they need to deliver accurate answers. When coupled with our query federation innovations, agents could rapidly find the right data, analyze it, return insights, and take action.
We also wanted it to make platform operations easier and easier for engineers and admins. We took our Reflections technology and made it autonomous, so materializations are automatically created based on query patterns, and queries are transparently accelerated. We also delivered a set of automatic Iceberg table optimization capabilities, including Iceberg clustering, variant shredding, compaction, and vacuum
We believe it is still early days for the agentic enterprise. Our vision is to enable knowledge workers working with agents and agents operating autonomously, to find, understand, and use enterprise data to answer any question at any time, and to determine the best actions to take and then take those actions.
SAPs's early success with tabular foundation models, anchored by RPT-1, can accurately predict business outcomes from structured business data. Joule Agents use these models to predict what will happen and then autonomously kick off application workflows or build new agents to proactively take action. We expect that SAP’s data assets and Knowledge Graph will complement Dremio’s capabilities.
SAP is one of the one of the world's largest providers of enterprise application software, with its customers generating 84% of total global commerce.
Enterprises want a complete view of their enterprise: SAP application data combined with every other system that runs alongside it, databases, data lakes, third-party SaaS, all of it. They want to run their analytics on that combined data, deploy AI agents on top of it, and govern the actions all the way through. They want to run their businesses on a single, governed, open data platform.
With Dremio, SAP will be able to deliver an open, governed Iceberg-native, agent-native lakehouse that unifies a customer's entire data estate. This is what every enterprise has been asking for. We will soon be able to deliver it together, faster, and at a scale neither of us could reach alone.
Enterprises want to run on data platforms that provide maximum interoperability and broad ecosystems. They want the flexibility to use a variety of engines, tools, and applications on their data.
As the only Iceberg-native lakehouse, Dremio provides the background services that customers need to continuously and automatically optimize their Iceberg tables, providing fast read and write performance through any client. Dremio’s Polaris-based metadata catalog enables Iceberg-compatible engines and clients to access the data with end-to-end security and governance.
To make this flexible world a reality, Dremio has spearheaded open source projects such as Apache Iceberg, Apache Polaris, and Apache Arrow. SAP is fully committed to these projects and will continue to invest in them.
Dremio isn’t going anywhere; we are doubling down on our agentic vision. Together with SAP, we will, once the transaction has closed, create a leading data platform and the foundation for an enterprise’s agentic future. Dremio customers will benefit from our increased scale and use of SAP’s innovative technologies.
To every customer and community member who believed in Dremio and what we are building, thank you. You made this possible.
The best part? We are still just getting started.
Deploy agentic analytics directly on Apache Iceberg data with no pipelines and no added overhead.
Agentic AI is an artificial intelligence system that is designed to operate autonomously. With minimal human supervision it can be expected to make decisions and perform tasks with specifically trained agents. This is thanks in large part to Large Language Models (LLMs) which provide agentic AI with enhanced reasoning and the ability to understand context. […]
Dremio is a commercial platform, and we’re straightforward about that. But the standards and projects that power it are genuinely open, and Dremio has been an active contributor to building them, not just consuming them. Apache Arrow, Apache Iceberg, and Apache Polaris all have Dremio fingerprints on their design, specification, and governance. That work matters […]
Building an agentic lakehouse with Dremio moves your organization beyond the "phone book" catalog of the past and into a future where the catalog is a dynamic knowledge base.
