National University of Singapore Builds ‘AI Foundry’ to Transform Materials Discovery via Continuous Prediction, Experimentation, and Learning Cycle

September 13, 2026

EurekAlert! Singapore

Overview

Professor Ong is leading the Materialyze.AI Lab at NUS to establish an ‘AI Foundry’ for transforming materials discovery, aiming to connect various AI capabilities to create a continuous cycle of prediction, experimentation, and learning. As a founding developer of the Materials Project and creator of pymatgen, Prof. Ong previously developed a foundation potential AI model that enabled the large-scale screening of millions of hypothetical crystal structures, far exceeding conventional quantum mechanical methods.

In Depth

Key Findings

Professor Ong at the National University of Singapore (NUS) is leading the Materialyze.AI Lab to establish an ‘AI Foundry’ aimed at fundamentally transforming materials discovery. This initiative seeks to dramatically increase the speed and efficiency of new material discovery by linking diverse AI capabilities to create a continuous, integrated cycle of prediction, experimentation, and learning.

Technical / Clinical Details

The ‘AI Foundry’ at Materialyze.AI Lab fuses data-driven AI models with first-principles simulations, incorporating the physical laws of materials science. Specifically, AI models first learn from existing data and academic knowledge to theoretically predict millions of novel material candidates with potential functional properties. This prediction phase leverages technologies like the foundation potential AI model previously developed by Prof. Ong, which can screen hypothetical crystal structures at a scale and speed far exceeding conventional quantum mechanical calculations. Subsequently, promising predicted candidates are actually synthesized and evaluated by automated experimental systems (autonomous laboratories). The experimental data obtained is then fed back into the AI model, improving its accuracy and forming the basis for new predictions. This closed-loop learning cycle accelerates and optimizes the entire materials discovery process. Prof. Ong, as a founding developer of the Materials Project and creator of pymatgen (a Python library for materials science), brings deep expertise and a proven track record in data-driven materials science to this foundry.

Background & Context

The discovery of new materials forms the cornerstone of advancements in all cutting-edge technologies, including energy, electronics, and medicine. However, traditional materials research has historically required extensive time and cost for trial-and-error experimentation and computations, becoming a bottleneck for innovation. The concept of an ‘AI Foundry’ addresses this challenge by combining AI’s predictive capabilities, automated experimentation, and a continuous learning loop. Prof. Ong’s initiative embodies the ‘fourth paradigm’ of materials science (data-driven science) and reinforces Singapore’s position as a hub for materials informatics in the Asian region.

Strategic Significance & Outlook

The AI Foundry at Materialyze.AI Lab is expected to accelerate materials discovery across a wide range of fields, including new-generation battery materials, high-efficiency catalysts, high-performance semiconductors, and biomaterials. This platform will dramatically shorten the research and development cycle, enabling faster innovation and market entry. The ultimate goal is to achieve ‘fully autonomous scientists,’ where AI independently designs, synthesizes, and evaluates materials. While challenges remain, such as managing large datasets, automating complex experiments, and ensuring the reliability and interpretability of AI models, this endeavor represents a significant step in shaping the future of materials science research.

Source: https://www.eurekalert.org/news-releases/1143109

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.