☁️AWS Machine Learning Blog•Stalecollected in 5m
Fundamental’s NEXUS Large Tabular Model joins SageMaker JumpStart
💡Easily deploy specialized large tabular models on AWS with the new NEXUS integration in SageMaker JumpStart.
⚡ 30-Second TL;DR
What Changed
NEXUS is now integrated into the Amazon SageMaker JumpStart ecosystem.
Why It Matters
This launch simplifies the adoption of specialized tabular AI for enterprises, reducing the infrastructure overhead typically associated with deploying large-scale models.
What To Do Next
Deploy the NEXUS model via SageMaker JumpStart to benchmark its performance against your existing tabular data processing pipelines.
Who should care:Enterprise & Security Teams
Key Points
- •NEXUS is now integrated into the Amazon SageMaker JumpStart ecosystem.
- •Users can streamline the deployment process for large tabular models.
- •The platform supports running predictions directly against enterprise-grade datasets.
🧠 Deep Insight
Web-grounded analysis with 10 cited sources.
🔑 Enhanced Key Takeaways
- •Fundamental AI, the company behind NEXUS, was founded in October 2024 and emerged from stealth in February 2026 with $255 million in funding, achieving a $1.2 billion valuation.
- •NEXUS is a Large Tabular Model (LTM) specifically designed with a non-transformer, tabular deep learning architecture to overcome limitations of traditional Large Language Models (LLMs) and classical machine learning in handling structured, non-sequential tabular data.
- •The model was pre-trained on billions of tabular datasets using Amazon SageMaker HyperPod and is engineered for direct ingestion of raw tables, minimizing the need for manual feature engineering or extensive data science pipelines.
- •Fundamental has already secured multiple "seven-figure contracts" with Fortune 100 companies across various industries, utilizing NEXUS for critical applications like fraud detection, predictive maintenance, and demand forecasting.
- •As of early 2026, Fundamental has not published third-party benchmarks or peer-reviewed papers validating NEXUS's accuracy claims, a contrast to some other tabular foundation models like TabPFN and KumoRFM.
🛠️ Technical Deep Dive
- NEXUS is categorized as a Large Tabular Model (LTM) and a foundation model specifically for enterprise prediction.
- It employs a proprietary tabular deep learning architecture, explicitly noted as a non-transformer architecture, to process structured data.
- The model was pre-trained on billions of diverse tabular datasets, leveraging Amazon SageMaker HyperPod for its training infrastructure.
- NEXUS is designed to ingest raw tabular data directly, automatically learning underlying structures, patterns, and dependencies across rows and columns without requiring extensive manual feature engineering or bespoke model design.
- It natively handles various data types including numbers, categories, dates, and free text.
- The architecture addresses inherent challenges of applying LLMs to tabular data, such as the tokenization of numbers (which can lead to a loss of understanding of numerical distributions) and the order-invariance of tabular data, which LLMs struggle with due to their sequential nature.
🔮 Future ImplicationsAI analysis grounded in cited sources
Increased adoption of specialized foundation models for structured data.
NEXUS's integration into SageMaker JumpStart highlights a growing market need for AI solutions tailored to enterprise tabular data, where general-purpose LLMs are less effective.
Reduced reliance on extensive data science teams for predictive analytics.
By automating feature engineering and model design, NEXUS could enable enterprises to deploy predictive models faster and with fewer specialized resources.
Enhanced accuracy and efficiency in critical enterprise decision-making.
The model's ability to uncover subtle signals and non-linear relationships in tabular data promises more precise predictions for use cases like fraud detection and demand forecasting.
⏳ Timeline
2020-12
Amazon SageMaker JumpStart announced, offering a curated catalog of pretrained foundation models.
2022-01
Amazon SageMaker JumpStart adds popular tabular models and algorithms including LightGBM, CatBoost, XGBoost, and Linear Learner.
2022-06
Amazon SageMaker introduces four new built-in tabular data modeling algorithms: LightGBM, CatBoost, AutoGluon-Tabular, and TabTransformer.
2024-10
Fundamental (AI company, developer of NEXUS) founded.
2026-02-05
Fundamental emerges from stealth with $255M funding at a $1.2B valuation and publicly launches its NEXUS Large Tabular Model, announcing a strategic partnership with AWS.
2026-06-03
Fundamental’s NEXUS Large Tabular Model becomes available for deployment via Amazon SageMaker JumpStart.
📎 Sources (10)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
📰
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: AWS Machine Learning Blog ↗