This article announces the availability of Fundamental's NEXUS, a large tabular model now deployable on Amazon SageMaker JumpStart for structured data prediction.

• NEXUS is a foundation model pre-trained on billions of tabular datasets for deterministic predictions
• Produces consistent, reproducible results unlike probabilistic LLMs on structured data
• Handles feature engineering automatically without manual data pipeline setup
• Processes multi-dimensional relationships across enterprise tables and databases
• Deploys on SageMaker with scikit-learn compatible Python SDK and single-tenant GPU instances
• Supports industry-specific variants for finance, healthcare, and manufacturing
• Enterprise use cases include fraud detection, churn prediction, and predictive maintenance
• Data remains in AWS environment with network-isolated, compliant infrastructure

NEXUS enables enterprises to generate accurate predictions from structured data in days rather than months, reducing time-to-value for tabular data analysis workloads.