The Next Chapter for Prior Labs

When we started Prior Labs, we had a conviction that ran against the grain of where most of the AI world was looking. While the field poured its energy into language, we believed an untapped opportunity in AI sat somewhere else entirely: in the tables, spreadsheets, and structured records that actually run the world's businesses and scientific discoveries.

Today we're announcing the next chapter. Prior Labs signed a definitive agreement to be acquired by SAP to become the next frontier AI lab. The terms are not disclosed and the deal is subject to regulatory approval. We will continue to operate as an independent entity, retaining our brand, our mission, our open-source commitments, our customers and our research-led identity. SAP will invest more than 1 billion EUR over the next four years and support us with long-term investment and a direct path to productization across the SAP portfolio.

This is something truly special and the beginning of Prior Labs’ growth phase.

Why now

Tabular foundation models are at an inflection point. TabPFN, published in Nature last year and cited over 1,000 times, established a new state of the art and has since been validated across hundreds of independent academic studies. TabPFN-2.6 currently sits at the top of TabArena. TabPFN-3 is imminent. Our open-source models have been downloaded over three million times, and the developer community around them keeps growing.

But the frontier we want to reach — models that don't just predict outcomes on tables but reason about causation, integrate domain knowledge, combine tables with language — can be reached much faster with SAP. SAP has multiple customers that span the industries where structured data carries the most value: financial services, healthcare, manufacturing and industrials. This scales the impact of our models to a greater scale and lets us iterate directly on solutions for real problems.

SAP moved early on tabular foundation models with RPT-1, before most of the enterprise software world had recognized the category existed. And critically, they understand what a frontier AI lab needs, which is why Prior Labs will continue as an independent legal entity and operate with autonomy.

What independence means here

Prior Labs will continue to be Prior Labs: same brand, same team, same research mission, same headquarters in Freiburg with our offices in Berlin and New York. Our customer relationships continue, our open-source strategy continues, with SAP's explicit support. The Discord community, the GitHub repository, the academic collaborations — all of it stays.

What changes is the resource envelope and the deployment reach. We get to work on research problems that previously sat outside what was feasible. We get to test models against the messy, high-stakes reality of enterprise data at a scale that sharpens the science. And the broader research community gets a frontier lab in Europe, working in the open, on a category of AI that has been  underinvested relative to its real-world importance.

To our community

To everyone who has used TabPFN, contributed to our open-source repositories, joined our Discord, cited our papers, or built on our models: thank you. None of this happens without you, and none of what comes next changes the relationship. Our models stay open. Our research stays public. Our commitment to the academic community stays exactly where it has been.

To our team: this is a moment to be proud of. This is a testament to what we have built together in this category in 18 months and we are excited for what we get to build together in this next phase of scale.

To our investors and especially, Balderton Capital, thank you for believing in us and your support in building this company together.

To prospective collaborators, customers, and researchers: we're hiring, we're publishing, and we're more ambitious than ever. The mission has not changed. It just got accelerated.

‍

Frank Hutter, Noah Hollmann, Sauraj Gambhir

Co-founders, Prior Labs