SAP Acquires Berlin AI Lab Prior Labs in €1B+ Deal 18 Months After Launch

Prior Labs developed tabular foundation models (TFMs), a category of AI built for structured enterprise data. Tech.eu reported that its TabPFN system uses a single pre-trained foundation model to handle prediction tasks such as payment delays, churn, supplier risk and demand forecasting, rather than requiring a separate model for each dataset. The technology has been applied with Hitachi to help prevent train failures, with TD for financial forecasting, and across research projects spanning pancreatic cancer diagnosis, wildfire prediction and battery materials. The company's latest model is TabPFN-3-Thinking.

Co-founder and CEO Frank Hutter said that eighteen months ago Prior Labs was a research project, and that the company is now beginning its next chapter with the resources to take on problems it could not previously address, as quoted by EU-Startups. He added that advancing tabular foundation models requires better data environments, deployment surfaces and long-term research investment, which he said SAP is positioned to provide.

EU-Startups noted the deal follows broader investment in European infrastructure for training, deploying and governing enterprise AI, citing its 2026 coverage of roughly €1.93 billion across comparable and adjacent companies, including Nscale's €1.7 billion AI-compute round and Verda's €100 million financing.

Grace Investment Machine raised $20 million in its third round within a year, as agentic AI investment tools gain momentum globally.