Startup Introduces 'Large Tabular Model' for Spreadsheet Data

2025-12-15 · Source: aibusiness · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Emerging Technologies & Innovation · Depth: Novice, quick

Summary

Fundamental, a San Francisco-based startup, has launched its "Large Tabular Model" (LTM) named Nexus, designed to process and derive predictions from structured enterprise data, an area where traditional large language models (LLMs) typically struggle. The company recently emerged from stealth with $255 million in funding, comprising a $30 million seed round and a $225 million Series A led by Oak HC/FT. Nexus was developed by Google DeepMind alumni, trained on billions of tabular datasets using Amazon SageMaker HyperPod, enabling it to understand non-linear relationships in spreadsheet-like data. Fundamental claims Nexus integrates easily into existing customer data stacks and automatically learns data patterns without manual training. The company has also partnered with Amazon Web Services, allowing customers to deploy Nexus within their AWS environments.

Key takeaway

For AI Product Managers evaluating solutions for enterprise data analysis, consider Nexus as a specialized alternative to general LLMs for structured, tabular datasets. Your teams can leverage its ability to automatically learn data patterns and integrate into existing data stacks, potentially improving prediction accuracy for tasks like financial fraud detection or energy price forecasting. This could streamline data-driven decision-making across various industries.

Key insights

Nexus is a specialized foundation model for tabular data, addressing LLM limitations in structured data analysis.

Principles

• LLMs excel with unstructured, sequential data.
• Tabular data requires specialized models for non-linear relationships.

Method

Nexus was developed from scratch on billions of tabular datasets and trained on Amazon SageMaker HyperPod to understand non-linear relationships across rows and columns.

In practice

• Integrate Nexus into existing customer data stacks.
• Deploy Nexus within an AWS environment.

Topics

• Large Tabular Models
• Structured Data Processing
• Enterprise AI Solutions
• Amazon SageMaker HyperPod
• AI Funding

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Product Manager, Data Scientist, Investor

Related on AIssential

• Beyond the lakehouse: Fundamental's NEXUS bypasses manual ETL with a native foundation model for tabular data — NEXUS is a Large Tabular Model designed to natively understand and predict from complex, non-sequential business data.
• Towards Structured Data: LLMs from Prototype to Production — Deploy LLMs beyond chatbots for structured data, moving NLP projects from prototype to production.
• ProfiliTable: Profiling-Driven Tabular Data Processing via Agentic Workflows — Robust table processing relies on active, iterative profiling and feedback-driven refinement to bridge linguistic intent with tabular reality.
• [Webinar] TabH2O #2 Webinar Recording — Tabular Foundation Models simplify predictive analytics for structured business data by learning patterns across diverse tables.
• Pinecone Brings AI Agents Directly to Enterprise Data with Microsoft OneLake Integration — Pre-assembling structured knowledge artifacts for AI agents significantly reduces LLM token consumption and accelerates task execution.
• Build AI models that know your enterprise. - Mistral AI — Customizing LLMs with proprietary enterprise data is crucial for domain-specific intelligence and competitive advantage.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by aibusiness.