German software group SAP has reached an agreement to acquire the Freiburg-based AI startup Prior Labs. The company, which specializes in so-called Tabular Foundation Models, is to continue operating as an indepconcludeent unit following the completion of the transaction. SAP plans to invest more than one billion euros in Prior Labs over the next four years in order to develop the startup into a globally leading AI research laboratory.
From Early-Stage Investment to Billion-Dollar Investment
Prior Labs was founded in 2024 by Frank Hutter, Noah Hollmann, and Sauraj Gambhir and has so far completed only a single funding round. In February 2024, the startup raised nine million euros. Investors included venture capital firms Balderton Capital, Atlantic Labs, and XTX Ventures, as well as several angel investors, among them Robin Rombach, co-founder of Black Forest Labs, and Thomas Wolf, co-founder of Hugging Face.
The acquisition by SAP thus comes less than 15 months after the first and only funding round. Balderton partner James Wise commented on the development as a remarkably rapid outcome for a European tech company. The purchase price was not publicly disclosed. The closing of the transaction is expected in the second or third quarter of 2026 and remains subject to regulatory approvals.
“The largegest untapped opportunity in enterprise AI was not large language models, but AI built for the structured data on which the world’s businesses run,” declared SAP CTO Philipp Herzig.
Prior Labs’ Specialty: AI for Tabular Data
Prior Labs develops Tabular Foundation Models (TFMs), i.e. AI models specifically designed for the analysis of structured data in tables and spreadsheets. In doing so, the startup addresses a gap that conventional Large Language Models (LLMs) cannot close: LLMs have only a rudimentary understanding of tables, numbers, and statistical relationships, and are therefore only partially suited for precise predictions based on structured business data.
The startup’s core product, TabPFN, is a widely utilized open-source tool with more than three million downloads. The current version TabPFN-2.6 holds first place on TabArena, the leading benchmark for TFMs. The model can instantly achieve the accuracy of a four-hour automated machine learning pipeline without any model training. Use cases include, among others, the prediction of payment delays, supplier risks, upselling opportunities, and customer churn risks.
The company’s scientific advisors include Yann LeCun, former Chief AI Scientist at Meta and recipient of the ACM A.M. Turing Award, as well as Bernhard Schölkopf, Director of the Max Planck Institute for Ininformigent Systems and President of ELLIS.
SAP’s Plans for Prior Labs
SAP intconcludes to continue operating Prior Labs as an indepconcludeent unit in order to maintain its research velocity. At the same time, the startup is to be closely integrated into the SAP ecosystem. The developed models are to be integrated into the entire SAP portfolio via SAP AI Core and SAP Business Data Cloud, as well as through the agentic layer with the AI assistant Joule.
- Research indepconcludeence: Prior Labs operates as an autonomous laboratory to ensure speed of innovation.
- Open-source strategy: SAP explicitly commits to continuing the open-source strategy around TabPFN.
- Product integration: The TFM technology is to flow directly into SAP products and benefit enterprise customers.
- Causality over correlation: In the long term, the models are intconcludeed not only to create predictions, but also to explain cautilize-and-effect relationships in business data.
- Global scaling: With the planned investment of more than one billion euros, Prior Labs is to be developed into a globally leading frontier AI laboratory for structured data.
With this relocate, SAP underscores its conviction that structured enterprise data represents the greatest untapped potential in the enterprise AI space. The company had already taken initial steps toward TFMs with its own model SAP-RPT-1 and sees the acquisition of Prior Labs as the logical continuation of this strategy.
Leave a Reply