The GENMAT Project: Artificial intelligence accelerating the discovery of new materials
Cutting the time it takes to bring a new material from the laboratory to market from decades to a few years: this is the ambitious goal of GENMAT, the new European project coordinated by Politecnico that aims to revolutionize materials development through artificial intelligence. Funded under the European call "AI Foundation Models in Science", GENMAT brings together a consortium of 14 partners spanning universities, companies, and organizations with expertise in AI ethics. The project will run for three years and aims to develop an innovative foundation model dedicated to materials science.
Unlike traditional AI applications, which are often designed for specific tasks, GENMAT will develop a single system capable of working across multiple fronts. This model will integrate scientific data of different kinds – from atomic chemistry to microscopic images and sensor data – to support the entire lifecycle of materials, from design to production and use. The innovative element is its ability to connect information across multiple scales, from atoms to molecules, from polymers to composite materials, delivering scientifically consistent and physically plausible results and overcoming the limitations of current generative AI tools.
The development of new materials is crucial for the competitiveness and economic security of the European Union, with strategic applications in sectors such as energy, mobility, construction, healthcare, and electronics. In this context, GENMAT will focus on four application areas, ranging from advanced composite materials with specific mechanical and electrical properties, such as those used in aerospace, to innovative materials that support the circular economy, such as next-generation recyclable polymers. The project will also address the development of more sustainable coatings free from PFAS – substances known as "forever chemicals" due to their persistence in the environment – and intelligent systems for monitoring the structural health of composite materials, improving their safety and lifespan.
One of the main expected benefits is a dramatic reduction in development time and costs. Today, designing a new material often requires years of testing and hundreds of attempts. Instead, GENMAT’s AI will enable the identification of a limited number of promising solutions from the earliest stages, significantly accelerating the innovation process. The project will not immediately produce a commercial product but represents a fundamental step towards a new generation of digital tools for the materials industry, destined to be adopted by chemical, mechanical, and construction companies.
GENMAT also aims to build an open, transparent, and trustworthy AI infrastructure, in line with European values. The model will be trained on certified scientific data and will incorporate physical constraints, interpretability tools, and systems for estimating the uncertainty of results, ensuring greater confidence in decisions. The resources developed will be shared through European open science platforms, contributing to strengthening Europe's technological sovereignty in a strategic sector.
"As coordinator of the GENMAT project, I will be able to draw on the experience built up within the Mechanics of Advanced Materials research group, coordinated by Professor Davide Paolino, which has been working for years in the field of advanced and intelligent composite materials – comments Raffaele Ciardiello, professor at the Department of Mechanical and Aerospace Engineering-DIMEAS – The expertise developed in the mechanics of composite materials, in the production and integration of sensors into advanced materials, in the development of multifunctional materials for structural monitoring, in multiscale and AI-based numerical modelling, and in advanced testing allows us to contribute to the development of a foundation model capable of bringing together data, physical knowledge, years of experimental experience and real application requirements."
In addition to coordinating the project, Politecnico di Torino will be involved in developing applications related to composite materials and will participate in designing the AI model, with the contribution of DIMEAS researcher Alberto Ciampaglia.
Source: Politecnico di Torino