Fundamental | Bringing Foundation Models to Enterprise Tables

Forecasting demand, anticipating equipment failures, and assessing credit risk often require separate models, each with its own development cycle. Large Tabular Models offer a shared foundation for these prediction tasks. Fundamental’s NEXUS is pretrained across billions of rows and columns to learn patterns that can transfer to new enterprise datasets. In this keynote, Neil Leiser, Senior Forward Deployed Engineer at Fundamental, explores how pretraining changes the economics of predictive AI, where specialized models still earn their place, and what it takes to turn the information in enterprise tables into useful predictions.