Enterprise software giant SAP is making a bold bet on structured-data AI, announcing its intention to acquire German startup Prior Labs. While the acquisition price remains undisclosed, SAP plans to invest €1 billion (approximately $1.16 billion) over four years to transform the 18-month-old company into a dedicated AI lab focused on tabular foundation models (TFMs). The move comes amid SAP's stock slump in 2026, partly driven by the so-called 'SaaSpocalypse'—a market correction affecting Software-as-a-Service companies.
Prior Labs was founded by Frank Hutter, Noah Hollmann, and Sauraj Gambhir just 18 months ago. The startup quickly gained traction with its TabPFN model series, which enables AI predictions directly from tables and databases—a more natural fit for enterprise applications than large language models (LLMs). Sources told Pathfounders that the deal was 'almost all cash,' with over half a billion dollars paid upfront to the founders. The startup had previously raised only $9.3 million in a pre-seed round led by Balderton Capital in February 2025.
SAP's interest in Prior Labs is strategic. The company's core products—accounting, HR, procurement, and expense management—rely heavily on structured data stored in relational databases. While LLMs like GPT-4 excel at language, they struggle with tasks involving rows, columns, and foreign keys. Tabular foundation models, by contrast, are designed to understand the relationships within structured data, making them ideal for predicting customer churn, optimizing supply chains, or detecting financial fraud—all critical enterprise use cases.
In parallel, SAP is taking an aggressive stance on agentic AI. The company has updated its API policy to explicitly prohibit AI agents from accessing its products unless they use 'SAP-endorsed architectures.' This effectively blocks OpenClaw and any other agent technology not approved by SAP. However, the policy does endorse Nvidia's Agent Toolkit, which powers NemoClaw—Nvidia's enterprise-ready version of OpenClaw. SAP customers will thus be authorized to deploy NemoClaw agents through the Joule platform, which remains in beta.
SAP's CFO Dominik Asam told CNBC in January that the company views AI as both a threat and an opportunity. 'It’s all about how quickly we as SAP can actually also embark on these technologies in our R&D portfolio to keep the relative economies of scale advantage,' he said. The company has not been idle: it has invested in Anthropic, Aleph Alpha, and Cohere (which intend to merge), and developed its own relational pretrained transformer model, SAP-RPT-1. But the Prior Labs acquisition accelerates SAP's structured data AI roadmap dramatically.
Prior Labs will operate as an independent unit within SAP, maintaining its open-source models—which have already been downloaded over three million times. The lab will continue to research TFMs that combine structured data with language, reasoning, and domain knowledge. Founders have publicly celebrated the deal, with CEO Frank Hutter calling it a 'massive boost' toward building a 'globally-leading frontier AI lab for structured data—in Europe, in the open.'
This acquisition highlights the growing divide among enterprise incumbents. While SAP is locking down its ecosystem, rival Salesforce has adopted a more permissive approach with its Headless 360 architecture, allowing customers to choose their own agents, including OpenClaw. SAP's stricter policy aims to control security and compliance, but also to steer customers toward its own Joule agent platform.
The SaaSpocalypse—a term coined to describe the sharp decline in SaaS valuations and growth rates—has forced companies like SAP and Salesforce to innovate defensively. SAP's stock has been trading slightly upward since the announcement, suggesting investor approval. Balderton partner James Wise called the acquisition 'one of Germany’s biggest ever venture outcomes.'
Looking ahead, the success of SAP's AI lab will depend on its ability to integrate Prior Labs' models into its massive product suite. The lab will work closely with SAP AI Core and SAP Business Data Cloud, as well as the agentic layer offered by Joule. Meanwhile, SAP's ban on unauthorized agents may face pushback from customers who prefer flexibility. But for now, the company is betting that structured data, not language, will unlock the next wave of enterprise productivity—and it's putting more than a billion dollars on that bet.
Source: TechCrunch News