SAP said on May 4 it will acquire AI research lab Prior Labs and data lakehouse platform Dremio at the same time. Both deals require regulatory approval.

According to a Techzine report, SAP plans to acquire Prior Labs and invest more than 1 billion euros over the next 4 years to develop it into a global AI research institute in structured business data. It did not separately disclose the size of the acquisition. Prior Labs will continue to operate as an independent organization after the acquisition.

Prior Labs specializes in table foundation models (TFM), which it says are fundamentally different from large language models (LLM).

SAP believes LLMs have limited understanding of tables, numbers and statistics, making it difficult for them to produce accurate predictions in structured business data. It says TFMs are specialized for this type of data and can be used to predict payment delinquency, assess supplier risk, predict customer churn and identify cross-selling opportunities.

SAP plans to combine Prior Labs with its in-house SAP-RPT-1 model and deploy TFMs through SAP AI Core and SAP Business Data Cloud. Its agentic AI layer Joule will coordinate this. Users can ask questions through a conversational interface, choose datasets and run virtual scenarios without data science expertise, the company said.

Dremio is an open, serverless data lakehouse platform that supports analytics and AI workloads. It is based on open-source projects such as Apache Iceberg, Apache Polaris and Apache Arrow.

With the Dremio acquisition, SAP plans to convert its Business Data Cloud to an Apache Iceberg-based architecture so SAP and non-SAP data can be used together without data migration or format conversion. SAP CTO Philipp Herzig said, "The biggest untapped opportunity in enterprise AI is not LLMs but AI for structured data that runs the world's businesses." He said, "The reason enterprise AI is stalling is not because there are not enough models, but because the data is not ready for AI agents."