SAP closes its €1B tabular-AI acquisition

SAP completed its acquisition of Berlin-based Prior Labs on July 17, closing a deal first announced in May. Financial terms of the purchase itself weren’t disclosed, but SAP has pledged a four-year, nine-figure-plus euro investment to grow Prior Labs into what it calls a globally leading frontier AI lab, focused specifically on tabular foundation models: AI systems built to work on structured business data like tables, transaction records, and operational logs, rather than free text.

Prior Labs, founded in 2024 by Frank Hutter, Noah Hollmann, and Sauraj Gambhir, had raised only around €9 million before SAP’s commitment. Its technology is aimed at predictions businesses actually run on all the time: late payments, supplier risk, customer churn, sales opportunities, the kind of forecasting that today mostly runs on traditional statistical models or spreadsheet formulas rather than foundation models. Prior Labs will keep operating as a largely independent entity inside SAP, continuing work that builds on SAP’s existing SAP-RPT-1 tabular model.

What it means for operators

Most of the AI-for-business conversation is about text and chat; this deal is a bet that the next real gains sit in the tables underneath your ERP and CRM systems, not the documents next to them. It’s a similar bet to Databricks’ fresh $188B round funding its own governance and agent layers on top of enterprise data, or Zoom folding buyer-intelligence platform Common Room into its AI stack: platform vendors increasingly want to own the data layer, not just the chat interface on top of it. If your team is evaluating AI vendors for forecasting or anomaly detection on structured operational data, tabular foundation models are worth a look as a distinct category from the general-purpose LLM your team probably already uses for text.