Nums AI, which develops a Tabular Foundation Model (TFM) for structured data, said on July 20 it has raised its first investment worth 4 billion won.

Stonebridge Ventures led the investment, with participation from SBVA, KT Investment and Base Ventures.

Structured data refers to numerical data in table form organised in rows and columns, such as sales records, transaction histories and customer information. TFM is a foundation model for structured data that supports predicting missing values with a single inference, much like a large language model predicts the next word.

In the past, the company said, task-specific models for problems such as demand forecasting, anomaly detection and credit assessment had to be developed over several months. It said TFM can cover all of these tasks with one model. The company stressed this can reduce dedicated staff and costs previously spent on developing and maintaining task-specific models.

Nums AI is an AI company founded by Chief Executive Jae-min Yoo (유재민), an assistant professor in the Department of Computer Science and Engineering at Seoul National University, together with fellow researchers.

Nums AI plans to use the investment to expand GPU infrastructure and its research organisation and speed up TFM development. It will provide an API service and an on-premise solution for companies that find it difficult to export data, and plans to expand its business centred on industries where using structured data is key, including manufacturing, finance, commerce and healthcare.

Yoo said, "Going forward, how effectively structured data is used will become a key factor that determines corporate competitiveness." He added, "Based on our research capabilities and expertise, we will grow into a core AI company that leads global model competition."