Materials Informatics and AI Drive New Design Innovations

The materials informatics market is forecast to reach $725 million with a CAGR of 9.0%, according to a new report from IDTechEx, "Materials Informatics 2025-2035: Markets, Strategies, Players". In a webinar covering the research findings, Senior Technology Analyst Sam Dale explained some of the technology innovations and challenges facing the market.

According to IDTechEx, materials informatics (MI) uses data-centric approaches for the advancement of materials science. Primarily, the technology is “based on using data infrastructures and leveraging machine learning solutions for the design of new materials, discovery of materials for a given application, and optimization of how they are processed.”

MI can help reduce the number of experiments needed for materials development and selection, which can accelerate time to market or help take research in new directions. 

“This bridges the gap between data scientists and materials scientists,” Dale said.

Dale also noted that materials informatics allows users to reverse the direction of innovation – in addition to simulating material structures to predict material properties, scientists can select desired properties to suggest material candidates. However, material science data is multi-dimensional and “noisy.”

That data usually comes from some combination of experiments, computational modeling, material data repositories, or data mining. The latter is being improved through the use of artificial intelligence and large language models (LLMs) that can help pull contextual data despite its complexity.

Dale noted that there are opportunities in every stage of material design and development for AI to improve things, including within data analysis, data handling, creating a hypothesis, and knowledge extractions. 

“The natural endpoint of this is self-driving labs,” he said. “Using robotics to perform physical experimentation, along with AI and computational modeling. However, it’s very hard to do this because of the variability of process and conditions.”

Dale said that industry-agnostic AI vendors are moving into the space – for example, Meta released a large materials data set last year. Other trends he noted were:

• SaaS models in materials informatics have had trouble reaching profitability

• The MI market is still a highly competitive and crowded space

• There is still a lot of funding available, although there have been fewer funding rounds overall

• MI is becoming widely established as an essential tool for the materials industry

You can watch the webinar and download the report here.