SAP announces two acquisitions: AI research lab Prior Labs and data lakehouse platform Dremio. With an investment of over one billion euros, SAP is building a leading position in AI for structured business data. Both deals are still subject to regulatory approval.
Prior Labs specializes in Tabular Foundation Models (TFMs). It will continue to operate as an independent unit following the acquisition. SAP is investing over one billion euros over the next four years to develop it into a world-leading AI research lab for structured business data. Financial details regarding the acquisition prices themselves have not been disclosed.
SAP previously launched the Business Data Cloud, a data platform that combines SAP and non-SAP data as a foundation for AI applications. Prior Labs and Dremio each add a new building block to their respective platforms.
Tabular Foundation Models: AI for Structured Data
TFMs are fundamentally different from large language models. In SAP’s view, LLMs are less suited to accurate predictions from structured business data due to their limited understanding of tables, numbers, and statistics. TFMs are built for this type of data and can predict payment delinquencies, supplier risks, customer churn, and upsell opportunities.
SAP has already taken steps in this domain. Earlier this year, the company launched SAP-RPT-1, a table-native foundation model for relational business data from systems such as SAP S/4HANA. That model supports in-context learning for predictions and classifications without retraining.
Prior Labs builds on the same foundation but has established a strong position of its own. Its flagship TabPFN-2.6 ranks first in the TabArena benchmark for TFMs. The model’s growth is remarkable: in less than a year, TabPFN’s support expanded from 10,000 to 10 million rows of data. The open-source model now boasts over three million downloads.
By combining Prior Labs and SAP-RPT-1, SAP aims to deploy TFMs via SAP AI Core and the SAP Business Data Cloud. The agentic AI layer Joule plays a coordinating role in this. Users can ask questions, select datasets, and run what-if scenarios via a conversational interface in plain language, without requiring data science expertise. The model adapts immediately to new use cases, without the need for separate model training.
“Early on, SAP recognized that the greatest untapped opportunity in enterprise AI wasn’t large language models; it was AI built for the structured data that runs the world’s businesses,” said SAP CTO Philipp Herzig.
Dremio Makes SAP Business Data Cloud Apache Iceberg-Native
In addition to Prior Labs, SAP is also acquiring Dremio. This is an open, serverless data lakehouse platform designed to accelerate analytics and AI workloads. While Prior Labs strengthens the model side, Dremio addresses data fragmentation.
Many enterprise AI projects fail not because of the models, but because of the underlying data. That data is fragmented, locked in proprietary formats, and lacks the business context that makes predictions meaningful. Dremio is specifically designed to remove those barriers.
Dremio focuses on the open-source projects Apache Iceberg, Apache Polaris, and Apache Arrow. In April 2026, Dremio released Apache Iceberg V3 support, featuring improved JSON support, change data capture, and schema evolution.
With the acquisition, the SAP Business Data Cloud becomes Apache Iceberg-native. SAP and non-SAP data can now coexist on the same open foundation without data migration or format conversion. Dremio is serverless and automatically scales up during peaks in demand and back down when the load decreases.
“Enterprise AI doesn’t stall because the models aren’t good enough; it stalls because the data isn’t ready for AI agents,” said Herzig.
Through Dremio, SAP is also building a universal, open catalog based on Apache Polaris and the Apache Iceberg REST Catalog API. This catalog will serve as the basis for the SAP Knowledge Graph, incorporating business relationships, organizational structures, regulatory classifications, and cross-system lineage as native properties.
SAP expects to complete both deals in 2026. The Prior Labs transaction is expected to close in the second or third quarter of 2026. Dremio is expected to follow in the third quarter. SAP confirms that Prior Labs’ open-source strategy around TabPFN and Dremio’s contributions to Apache Iceberg, Polaris, and Arrow will continue.