Tabular data underpins key operations in healthcare, finance, environmental monitoring, and manufacturing. While data analysis is paramount in these fields, the tools for analysing tabular data have fallen behind. Freiburg-based Prior Labs, an AI startup pioneering foundation models for spreadsheets and databases, aims to solve this problem with TabPFN, a groundbreaking AI model for data analysis that unlocks the untapped value in spreadsheets and databases used by businesses worldwide.
The company has just grabbed €9M in pre-seed funding, led by Balderton Capital and XTX Ventures, with participation from SAP founder Hans Werner-Hector’s Hector Foundation, Atlantic Labs, and Galion.exe. Notable AI angel investors also joined the round, including Thomas Wolf (Founder & CSO, Hugging Face), Peter Sarlin (Founder & CEO, Silo AI), Guy Podjarny (Founder, Snyk & Tessl), Ed Grefenstette (Director, DeepMind) Robin Rombach (Founder & CEO, Black Forest Labs), Chris Lynch (Founding Investor Data Robot & CEO, AtScale), Ash Kulkarni (CEO, Elastic), and other business leaders.
The funding will accelerate product development, expand the team, and bring Prior Labs’ groundbreaking foundation model to more users.
Prior Labs: scaling the academic success to deliver real-world impact
Prior Labs was founded in late 2024 by Professor Dr. Frank Hutter, the world’s most cited AutoML researcher; Noah Hollmann, a computer scientist with Google and BCG experience and an X-Prize finalist; and Sauraj Gambhir, an expert in venture capital, M&A, and enterprise growth. With over 20 years of machine learning expertise, the team has developed a state-of-the-art foundation model for tabular data.
Their groundbreaking work, detailed in a paper recently published in Nature, showcases TabPFN’s transformative potential. Prior Labs is now bringing this academic success to the business world by integrating its API into business data workflows, helping companies unlock the hidden potential of their tabular data.
In the recent Nature paper, TabPFN demonstrated remarkable results, outperforming state-of-the-art models in over 96% of small tabular data use cases. The model requires only half the data to match competitors’ accuracy and achieves better results in 2.8 seconds compared to the 4+ hours needed by existing models. Its intuitive design ensures quick application to any dataset with minimal coding, providing a user-friendly experience.
Frank Hutter, co-founder and CEO of Prior Labs, said: “Most of the world’s critical decisions rely on tabular data, yet the tools to analyse it are outdated and insufficient. We’re bringing a quantum leap to how businesses make predictions from their most valuable data, creating a future where working with tables is as seamless as using AI for text or images. We can deliver faster, more accurate predictions that empower businesses to do more with less.”
A universal solution for tabular data analysis
TabPFN has the potential to drive profitability through faster, more accurate predictions for critical decisions in trading, finance, and business analytics. The system particularly shines in data-constrained fields like healthcare, medicine, and climate science — where data collection is challenging and costly. TabPFN achieves superior results with half the data, enabling scientific breakthroughs while competing with established players like Neuralk-AI, Alteryx, Microsoft Power BI, Tableau, Rows, and Databricks.
Prior Labs’ TabPFN model represents a universal approach to tabular data analysis. Having trained on 130 million synthetic datasets, it can instantly recognise and interpret patterns in any dataset without task-specific training. As a foundation model, it improves through fine-tuning with company-specific data, enhancing its accuracy and real-world adaptability.
Prior Labs has built on this foundation with an enhanced API, allowing businesses to integrate TabPFN seamlessly into their operations at scale. The company’s commitment to continuous improvement is evident in its steady enhancements to the baseline model’s performance across all metrics. Recent updates, such as text feature support, proprietary data fine-tuning, and contextual task information, have significantly contributed to greater accuracy and usability.
James Wise, Partner at Balderton Capital, said: “Tabular data is the backbone of science and business, yet the AI revolution transforming text, images, and video has had a marginal impact on tabular data—until now. Prior Labs’ breakthrough gives everyone the superpowers of machine learning without needing to train their models on their data.”
The post TabPFN to bridge AI gap in structured data: Germany’s Prior Labs lands €9M from Hugging Face, Silo AI, and SAP founders appeared first on Tech Funding News.