TabPFN to Bridge AI Gap in Structured Data: Germany’s Prior Labs Secures €9M Investment from Hugging Face, Silo AI, and SAP Founders
Tabular data plays a critical role across various industries, including healthcare, finance, environmental monitoring, and manufacturing. However, the tools currently available for analyzing tabular data have not kept pace with advancements in other AI fields. Freiburg-based AI startup, Prior Labs, is addressing this gap with TabPFN, a cutting-edge AI model designed specifically to unlock the untapped value of spreadsheets and databases used by businesses globally.
Securing €9M in Pre-Seed Funding
Prior Labs has secured €9 million in pre-seed funding, led by Balderton Capital and XTX Ventures, with contributions from SAP founder Hans Werner-Hector’s Hector Foundation, Atlantic Labs, and Galion.exe. The funding round also saw notable AI angel investors joining the effort, including Thomas Wolf (Founder & CSO, Hugging Face), Peter Sarlin (Founder & CEO, Silo AI), and others.
The funding will accelerate the development of TabPFN, expand the team, and broaden the reach of Prior Labs’ pioneering foundation model for tabular data.
Scaling Academic Success into Real-World Impact
Founded in late 2024 by renowned AutoML researcher Professor Dr. Frank Hutter, computer scientist Noah Hollmann, and venture capital expert Sauraj Gambhir, Prior Labs is transforming academic research into practical applications. The team has over 20 years of machine learning experience and has developed TabPFN, a breakthrough model that has already demonstrated its transformative potential in academic settings.
In a recent Nature publication, TabPFN was shown to outperform state-of-the-art models in over 96% of small tabular data use cases. Not only does it require only half the data to match the accuracy of existing models, but it also delivers results in a mere 2.8 seconds, compared to the 4+ hours that conventional models take.
A Game-Changer for Tabular Data
TabPFN is poised to revolutionize industries reliant on tabular data, providing faster and more accurate predictions for business decisions, finance, and analytics. It particularly excels in data-constrained sectors like healthcare and climate science, where data collection is expensive and time-consuming. TabPFN can achieve superior results with only half the data, making it a powerful tool for scientific and commercial breakthroughs.
Prior Labs’ AI model is not limited to task-specific training. With training on 130 million synthetic datasets, TabPFN recognizes and interprets patterns in any dataset, making it highly adaptable to various industries. As a foundation model, it continues to improve through fine-tuning with company-specific data, allowing for enhanced accuracy and scalability.
A Universal Solution for Business Data Workflows
Prior Labs has enhanced TabPFN’s API to integrate seamlessly into business operations at scale. Recent improvements include text feature support, proprietary data fine-tuning, and task-specific contextualization, all contributing to increased model performance.
Frank Hutter, co-founder and CEO of Prior Labs, explained, “The tools for analyzing tabular data are outdated and insufficient. We’re bringing a quantum leap in how businesses make predictions, enabling faster, more accurate insights while empowering them to do more with less.”
Transforming the Future of Tabular Data
James Wise, Partner at Balderton Capital, stated, “While AI has revolutionized text, image, and video analysis, the transformation of tabular data has lagged behind—until now. Prior Labs’ breakthrough allows businesses to leverage machine learning without needing to train models on their own data, providing everyone with supercharged analytics capabilities.”
With its innovative AI model, Prior Labs is set to redefine how businesses and industries unlock the value of their tabular data, enabling smarter decisions, faster results, and more impactful insights.
