International AI competition aims to speed up the development of materials for the green transition
The Pioneer Center CAPeX at DTU has announced the winners of the first phase (Stage 1) an international competition in partnership with the Novo Nordisk Foundation, and the Danish Centre for AI Innovation (DCAI), on using machine learning models to predict synthesis recipes for novel nanoparticles.
Denmark’s first AI supercomputer, Gefion, which the winners in the Stage 2 of the competition will get access to, allowing them to further develop their machine learning models. Photo: DCAI
Friday 06 February 2026
Katrine Damkjær
The green transition critically relies on new advanced materials for Power-to-X technologies such as catalysts, energy storage materials, and sustainable fuel production. However, discovering and optimizing materials, in the form of nanoparticles, remains slow and resource intensive.
To attract leading machine learning and materials science researchers, CAPeX partnered with the Novo Nordisk Foundation and DCAI to launch the ‘AIS25 AI4Materials NP Challenge’ at the European Commission’s AI in Science (AIS25) summit in Copenhagen. The competition invited international teams to train machine learning models to predict the conditions required to synthesize nanoparticles in real-world laboratory settings.
The winning teams have been found among two of the world’s leading research environments in the field: Filippo Bigi and Cesare Malosso from the COSMO group of Prof. Michele Ceriotti, École Polytechnique Fédérale de Lausanne (EPFL) and Ilyes Batatia from the group of Prof. Gábor Csányi, Cambridge University, and the Max Planck Institute for Polymer Research (MPIP).
They will now be onboarded on the Danish DCAI NVIDIA GPU supercomputer Gefion for second phase of the competition (Stage 2). There, they will gain access to unique computational and experimental datasets to further develop and train their models to predict experimentally observable outcomes in nanoparticle synthesis from synchrotron experiments. This will be the first time machine learning models are optimized to directly predict experimental nanoparticle synthesis conditions.
“To achieve CAPeX’s goals of delivering scalable solutions for Power‑to‑X, we need to develop much faster and more accurate machine learning models, and for this, we need to collaborate with the very best researchers worldwide. A competition like the AIS25 AI4Materials NP Challenge is an opportunity for doing just that. By combining an AI‑driven approach with large FAIR dataset and research infrastructure across borders, we can shorten the time from discovery to industrial uptake and reduce the cost of discovering new advanced materials. In doing so, we strengthen both Europe’s green transition and competitiveness,” says Professor Tejs Vegge, Director of CAPeX.
Fakta
About the winners
PET-MAD from École Polytechnique Fédérale de Lausanne (EPFL): The winning models are developed by researchers from Prof. Michele Ceriotti’s group: A model from Filippo Bigi (model 1: pet-bc619ed-mammoth) and a model from Cesare Malosso (model 3: pet-7a655c5-lynx)
MACE from Cambridge University and the Max Planck Institute for Polymer Research: The winning model is developed by a researcher from Prof. Gábor Csányi’s group: Ilyes Batatia (model 2: mace-mh-1-ft)
Figure by Emil Frost, DTU Energy. The model performance (test scores) in terms of RMSE (root mean square error) on atomic energies and forces, measured in four separate challenges.
About CAPeX
CAPeX is short for ’Pioneer Center for Accelerating P2X Materials Discovery’ and the centere is run in a partnership between DTU and Aalborg University (AAU) and is located in a new interdisciplinary "Climate Challenge Laboratory" at DTU. Here, researchers from DTU, AAU, the University of Copenhagen, Aarhus University, and the University of Southern Denmark will create a ground-breaking interdisciplinary environment for Power-to-X technology in collaboration with international partners from Stanford University, Utrecht University, and University of Toronto.
Five Danish foundations – the Danish Ministry of Higher Education and Science, the Danish National Research Foundation, the Carlsberg Foundation, the Lundbeck Foundation, the Novo Nordisk Foundation, and VILLUM FONDEN – have granted a total of DKK 300 million to the establishment of CAPeX.
