SAP's EUR 1 billion investment in AI focuses on tables rather than chatbots.

      The AI sector is highly focused on models that can generate text and speech. However, SAP has recently made a significant investment in a different type of model designed for structured data, completing its acquisition of Prior Labs.

      The acquisition is finalized with all necessary regulatory approvals in place, and the Freiburg-based lab is now operating under SAP. The German software giant is committing over €1 billion to this initiative over a four-year period.

      What Prior Labs does

      While most foundational AI models are geared toward text, images, or audio, Prior Labs specializes in developing foundational models for tables—meaning the structured data in rows and columns that organizations rely on.

      Its TabPFN model, published in Nature, has set a new standard in tabular prediction according to numerous independent studies. This model analyzes and categorizes data from spreadsheets and databases instead of traditional prose.

      Despite its seemingly mundane functionality, this capability has vast applications. It has already been utilized in areas such as financial risk assessment, loan approval processes, predictive maintenance for railways, and cancer diagnostics.

      Why SAP is interested

      The alignment is almost perfect. SAP’s core business revolves around structured enterprise data found in various company systems, which is precisely the type of data that Prior Labs’ models are designed to process.

      Additionally, it addresses a significant gap in the market. Many businesses struggle with the structured-data aspect of enterprise AI even while investing heavily in chatbots.

      SAP has been making substantial investments to catch up in this area, including an autonomous enterprise formed by over 200 AI agents and integrating new automation features into its AI studio. The company even implemented a hiring and travel freeze to support this initiative.

      A notable European success

      The strategic significance of this acquisition is particularly noteworthy. A European software leader has successfully acquired a cutting-edge European lab and retained its independence.

      This outcome aligns with the goals that EU policymakers have pursued for years. Prior Labs retains its name, maintains its presence in Freiburg, continues to offer open-source models, and has an advisory board that includes Meta's Yann LeCun.

      This acquisition is remarkable given Europe’s historically thin track record in this field. Other prominent AI companies in the region have typically merged or been absorbed by larger corporations rather than cultivating a leader at home.

      The broader implication

      This deal signifies the emergence of tabular AI as a legitimate sector. Major players like Microsoft, Google, and AWS are also entering the market for structured-data models, and SAP has acquired one of the leading independent ventures in this area.

      Moreover, this acquisition serves as a subtle correction to the prevailing obsession with chatbots. The most valuable AI within a business may not necessarily be the one that composes emails but rather the one that can anticipate loan defaults or predict which train is likely to fail next.

      SAP is wagering that the next frontier in enterprise AI lies within the data that organizations already possess. The models remain open, the lab continues its work in Freiburg, and SAP's investment now exceeds €1 billion.