SAP is making a bigger push into the data layer behind enterprise AI. SiliconANGLE reported that SAP plans to acquire Dremio and Prior Labs, two companies focused on helping organizations manage and analyze tabular business data. Financial terms were not disclosed.
Dremio is best known for a data-management platform built around open table formats and metadata. SiliconANGLE noted that Dremio’s technology uses Apache Iceberg and Apache Polaris, projects that help companies store, govern and query large datasets across cloud environments. SAP plans to use the acquisition to strengthen Business Data Cloud, its service for combining enterprise datasets and making them usable by AI systems.
Prior Labs brings a different but related capability. The Berlin startup has developed TabPFN-2.5, an AI model optimized for data stored in rows and columns. That kind of model is aimed at everyday enterprise problems such as detecting bad entries in inventory data, analyzing structured records or supporting decisions where the source material is not text but tables.
Why it matters
The acquisitions underline a practical truth about enterprise AI: models are only as useful as the data they can trust. Many companies have strong business applications, but their information is split across warehouses, SaaS tools, spreadsheets and legacy systems. If AI agents cannot understand permissions, metadata, lineage and table structure, they are likely to produce brittle or risky results.
For SAP, Dremio and Prior Labs could help turn Business Data Cloud into a more credible foundation for agentic workflows. For customers, the deal points to a broader market shift. The next wave of AI spending may move from model access toward data readiness, governance and specialized models that understand business records. In other words, the most valuable AI infrastructure may be the layer that makes ordinary enterprise data clean, governed and usable.
Source: SiliconANGLE, published May 4, 2026.