NTT DATA Uses AI To Accelerate CO₂ Capture and Conversion Research

• Research in Italy demonstrates how AI accelerates catalyst discovery by up to 20 times with the University of Palermo and the University of Catanzaro.
• Screening millions of molecular candidates turns complex materials research into a fast, predictive and repeatable, reducing R&D time and cost.
• The research identified new manganese‑based catalysts for CO₂ capture and conversion using AI framework from NTT DATA.

MILAN — May 21, 2026 — NTT DATA, a global leader in AI, digital business and technology services, together with the University of Palermo and the University of Catanzaro, has developed an artificial intelligence framework that accelerates the discovery of catalysts for CO₂ conversion by 20 times, addressing one of the key challenges for the decarbonization of industrial processes.

In materials informatics, an AI framework is a structured pipeline that combines data, feature engineering and advanced AI models to turn materials discovery into a fast, predictive and repeatable process. In this project, the framework enabled the systematic exploration of vast chemical spaces that would be impractical using traditional approaches alone.

Funded by Spoke 7 of the Italian Research Center in High Performance Computing, Big Data and Quantum Computing (ICSC), the project shows how materials informatics can transform R&D into a predictive, data‑driven workflow. By combining machine learning with advanced computational chemistry, the system can screen millions of molecular candidates and highlight the most promising ones for industrial use.

The research identified a new class of homogeneous metal–salt catalysts, with manganese-based systems showing strong potential for efficient CO₂ conversion and stable product release.

Beyond CO₂ conversion, the methodology serves as a replicable model for any R&D intensive sector — from energy and chemistry to pharmaceuticals and advanced materials — helping companies:

• Speed up time‑to‑market
• Reduce development costs and risk
• Link sustainability goals with industrial performance

NTT DATA designed and implemented the underlying AI framework, developing a dedicated database of chemical properties, algorithms to convert molecular structures into machine readable descriptors, and fine tuned deep learning and generative AI models able to learn structure–property relationships.By shifting materials discovery from intuition to large‑scale predictive exploration, NTT DATA and its partners are paving the way for a new generation of sustainable industrial innovation.

“We developed an AI framework to support materials informatics workflows at an industrial level,” said Antonio Policicchio, Quantum Computing and Materials Informatics Lead at NTT DATA Italia. “This approach accelerates sustainable innovation and strengthens industrial competitiveness.”

Academic partners provided scientific leadership in computational chemistry, ensuring rigorous methods, meaningful chemical insights and reliable model validation.

“By integrating machine learning with atomistic simulations, we identified the key parameters needed for more efficient catalysts,” said Adriana Pietropaolo, University of Catanzaro.

“This collaboration bridges academic excellence and industrial needs, boosting global innovation from southern Italy,” added Professor Giampaolo Barone, University of Palermo.

The initiative also aligns with the mission of ICSC’s Spoke 7 Materials & Molecular Sciences, which promotes advanced materials research for strategic national applications. “The project shows how the ICSC can bring together public and private expertise to develop deep tech solutions with real industrial impact. It perfectly reflects the mission of Spoke 7, applying advanced materials research to strategic challenges such as CO₂ capture and conversion,” commented Stefano Fabris, leader of Spoke 7 of the ICSC National Center and Director of the Department of Physical Sciences of the National Research Council.

Visit our website to find out more about NTT DATA Materials Informatics.