Tabular foundation models are powerful, but we're still in early days. Here's what we don't understand, where they fail, and the fascinating collision course with large language models.

The technology is impressive, but when should you actually use it? This practical guide helps enterprises understand where tabular foundation models deliver value—and where the classics still reign.

Tabular Models are trained on million of datasets, but they were all fake. This isn't a bug; it's the entire point. Here's why synthetic data is the secret weapon of tabular foundation models.

If you've ever applied for a credit card, had your insurance premium calculated, or seen a "customers also bought" recommendation, you've been touched by gradient boosted decision trees. These algorithms—with names like XGBoost, LightGBM, and CatBoost—have been the undisputed rulers of tabular machine learning for over a decade. Before we get into the advantages of tabular foundation models, let's understand why these traditional methods have worked so well.

Foundation models—the technology behind ChatGPT and image generators—are now being built specifically for the rows and columns of your business data. This series explores how they work, their value for enterprises, limitations, and future developments in the field.

This piece examines why structured data represents the next $600bn opportunity for AI adoption. It shows how Predictive AI and tabular foundation models can transform how enterprises deploy predictive capabilities on their data, with clearer ROI and production-ready performance.

A joint interview with Business Times: "Created a year ago by Antoine Moissenot and Alexandre Pasquiou, Neuralk-AI develops AI models that help data science teams improve their predictions across a range of concrete use cases"

Neuralk AI joins the 2025 French Tech 2030 cohort! This recognition will help us accelerate our mission to bring Frontier AI to Enterprise data, with a foundation model purpose-built for tabular business data.

Discover how retailers can transform product categorization from a tedious task into a strategic driver of sales.

The article covers the LSA IA Tech For Business Awards 2025, highlighting Neuralk AI’s Jury’s Favorite Award for its tabular foundation model.

Introducing TabBench, an open-source benchmark built by Neuralk-AI to evaluate tabular ML models on practical, real-world industry tasks, starting with commerce-related use cases.

Discover how Tabular AI can outperform classical ML with instant predictions and minimal manual effort.

LLMs are good general-purpose tools. But when the stakes are high and the data is complex, Tabular AI can truly make the difference.

We put our Tabular Foundation Model head-to-head against the recently released LLM-based classifier from Mistral AI, focusing on Product Categorization.

The article explores tabular foundation models and structured data as a new AI frontier, and reports on our $4M raise to accelerate R&D in Predictive AI.

This TechCrunch article explains why structured, tabular data remains central to enterprise value, and how Neuralk-AI is building Predictive AI models purpose-built for production use cases on enterprise data.