SAP’s EUR 1bn bet on AI: Understanding Spreadsheets, Not Chatbots
While everyone focuses on building chatbots, SAP has just made a significant investment of over €1 billion in artificial intelligence (AI) that delves into spreadsheets. The enterprise software giant has completed the acquisition of Prior Labs, a German lab specializing in ‘tabular’ foundation models. This rare European success story in frontier AI stands out as most major AI names tend to merge or dissolve rather than foster independent champions.
The Acquisition and Prior Labs’ Focus
The deal, with regulatory approvals secured, marks a new phase for Prior Labs now operating within SAP. Over four years, SAP will invest more than €1 billion in this venture.
Prior Labs builds foundation models tailored for structured data, often found in business spreadsheets and databases. Their TabPFN model, published in Nature, sets the standard across various independent studies for tabular prediction. Unlike AI models focused on text, images, or audio, Prior Labs‘ expertise lies in forecasting and classifying data within spreadsheets and databases.
Benefits and Applications
This capability offers vast potential as it has already been applied to financial risk assessment, loan approvals, predictive maintenance for railways, and even cancer diagnostics.
SAP, with its existing business focused on structured enterprise data, finds a perfect fit. The acquisition addresses a gap in enterprise AI, where most companies struggle with structured-data handling, despite significant investments in chatbots.
A European Success Story
The strategic significance of this move is notable as it showcases a European software giant supporting a local frontier lab. Prior Labs maintains its brand, location, open-source models, and an advisory board including Meta‘s Yann LeCun. This outcome aligns with EU policymakers’ efforts to foster academic-to-commercial transitions in AI.
The Larger Impact
The acquisition signifies the arrival of tabular AI as a significant category, with tech giants like Microsoft, Google, and AWS also exploring structured-data models. SAP’s investment corrects the focus towards data prediction, which might prove more valuable than chatbot capabilities.