SAP acquires Prior Labs in €1B to build next-gen AI for business data

SAP is making a bold call on where enterprise AI is headed—and it’s not betting on chatbots. The German software giant has agreed to acquire Prior Labs, a young startup focused on a niche many in AI have overlooked: structured business data.

Alongside the deal, SAP plans to invest more than €1 billion to scale the company over the next four years, making it a central part of its AI strategy. Terms weren’t disclosed, and the transaction still needs regulatory approval.

At a time when much of the industry is fixated on large language models, SAP is leaning into a different reality inside enterprises. Most business decisions don’t come from text—they come from tables. Spreadsheets, financial records, supply chain logs, and customer databases still drive how companies operate. And that’s exactly where Prior Labs has been building.

Founded in late 2024 by machine learning professor Frank Hutter, Noah Hollmann, and Sauraj Gambhir, the Freiburg-based startup specializes in tabular foundation models, or TFMs. These models are built to work directly with structured data, making predictions about things like payment delays, churn risk, supplier issues, and upsell opportunities. It’s a category that tackles a known limitation in AI today: language models struggle with numbers, statistics, and relational data.

“Early on, SAP recognized that the greatest untapped opportunity in enterprise AI wasn’t large language models; it was AI built for the structured data that runs the world’s businesses,” SAP CTO Philipp Herzig said. “We built SAP-RPT-1 to prove that conviction for enterprise data. Prior Labs has built a leading TFM on public benchmarks and built one of the leading research teams in this category. Combining their frontier model work with enterprise data and customer reach is how we intend to lead this category globally.”

Prior Labs will continue operating as an independent unit out of Freiburg, keeping its team intact and maintaining its open-source direction. SAP plans to integrate the startup’s technology into its own stack, including AI Core, Business Data Cloud, and its agentic AI layer, Joule.

SAP’s €1B AI Move: Why Prior Labs Could Change How Businesses Predict the Future

The appeal is speed and simplicity. Prior Labs’ models can take raw business data and generate predictions instantly, without the long training cycles typically required in machine learning. Its flagship system, TabPFN, has already gained traction in the research and developer community, with millions of downloads and strong benchmark performance. The latest version, TabPFN-2.6, delivers results comparable to hours-long automated machine learning pipelines in a single pass.

“Over the last 18 months, Prior Labs has built an incredible team, increasing the velocity in tabular foundation models,” Prior Labs CEO Frank Hutter said. “Joining the SAP family gives us the resources, data environment, and customer reach to take this category to its full potential.”

For SAP, the acquisition signals a broader shift in how enterprise AI will be built and used. The company is pushing toward systems that go beyond answering questions and start making decisions—tools that can interpret data, run scenarios, and explain outcomes. With a conversational layer on top, business users could query datasets, test assumptions, and get predictions without needing deep technical expertise.

That vision leans into a deeper challenge in AI: moving from correlation to causation. Predicting what might happen is useful. Explaining why it might happen is where real value starts to show up inside companies.

Prior Labs brings more than just technology. Its team includes researchers with backgrounds at Google, Apple, Amazon, Microsoft, and leading financial firms, as well as ties to top academic institutions. Advisors such as Yann LeCun and Bernhard Schölkopf add further weight as the company scales its research efforts.

If the deal closes as expected in the second half of 2026, SAP will gain a dedicated AI research hub focused on one of the least glamorous but most critical parts of enterprise computing: the data that runs the business itself.