Discovery to Design: The Applications Driving Materials Informatics

 

 

 

For decades, discovering a new material relied on years of trial and error, with researchers working through countless experimental iterations. But that process is being rewritten. Materials informatics replaces that guesswork with machine learning, compressing years of laboratory work into a fraction of the time, and the world's largest technology companies have taken notice. If artificial intelligence can already narrow down which material to build next, how long before it starts designing entirely new ones on its own?

 

IDTechEx has covered materials informatics since 2020, following its transition from an emerging technology to an integral part of modern materials R&D. The latest edition of the report, "Materials Informatics 2025-2035: Markets, Strategies, Players," is now in its fifth update. It provides detailed breakdowns of algorithmic approaches, data infrastructure requirements, and strategic options for adoption, alongside a ten-year market forecast to 2035 and independent and unbiased analysis of over 35 key players, based on primary research.

 

 

 

 

 

Materials discovery often requires searching through tens of millions of possible candidates, far beyond what can realistically be evaluated through conventional simulations or laboratory experiments. Materials informatics therefore addresses this challenge by using machine learning as an initial screening stage. This involves rapidly filtering the search space before the most promising ones progress to detailed physics-based simulations and experimental validation. Microsoft's Azure Quantum Elements platform recently demonstrated this approach for Pacific Northwest National Laboratory, reducing 32 million potential battery electrolytes to 500,000, from which 18 were selected for laboratory testing. This process ultimately led to the discovery of a solid-state electrolyte using 70% less lithium than conventional alternatives. Furthermore, at the Advanced Materials Show 2026 in Birmingham, IDTechEx spoke with Glass Futures, a UK glass industry research organization, which has developed a physics-informed neural network that cut a traditional multiphysics furnace simulation from six days down to milliseconds.

 

 

Such approaches require extensive training datasets and significant computational infrastructure, giving major technology companies a clear advantage over specialist materials informatics vendors. IDTechEx has observed several strategies emerging in recent years. For example, Microsoft provides commercial screening platforms directly to industrial customers, while Meta has focused on releasing open datasets and pretrained models, such as OMat24, enabling organizations across the sector to accelerate materials discovery.

 

 

Developing formulations presents a different challenge from discovering a single material. Polymers and liquid electrolytes are both created by combining multiple ingredients under specific processing conditions, creating a vast number of possible combinations. Industrial datasets add a further challenge, often being incomplete, inconsistent and spread across decades of laboratory records. Together, these factors led formulation companies to adopt data-driven methods earlier than much of the wider materials industry, making formulations and soft materials the most established application for materials informatics according to IDTechEx.

 

One company addressing this challenge is Intellegens, whose Alchemite platform estimates missing experimental values while providing confidence levels for each prediction, allowing researchers to prioritize experiments despite incomplete datasets. Intellegens has continued to expand the platform through the launch of Alchemite Suite, reporting reductions of up to 80% in physical experiments. IDTechEx also spoke with Lucideon at the Advanced Materials Show. This UK-based materials science consultancy has used its IMPACT™ platform to predict sensory panel results for an oral healthcare product directly from analytical data, lessening the need for resource-intensive physical panel testing. Such commercial activity suggests the sector remains not only well established, but among the most commercially active in materials informatics.

 

 

 

Generative and inverse design shifts the focus from evaluating existing materials to creating entirely new ones. In this approach, researchers begin with a target set of properties and use the model to generate candidate structures. Kebotix was one of the earliest companies to pursue this, using generative models to explore chemical space and identify promising candidate molecules. However, IDTechEx has extensively researched generative materials design and identified synthesizability as one of the key barriers to wider adoption. In practice, a material predicted to have desirable properties may still prove impossible, or impractical, to produce experimentally.

 

Researchers at KAIST have begun addressing this directly, introducing a "Crystal Likeness Score" to estimate whether a generated inorganic material is likely to be experimentally realistic. When tested against materials deliberately withheld during training, the model successfully rediscovered 25 of 31 known compounds. Progress such as this suggests the gap between computational design and experimental reality is beginning to narrow. If synthesizability can be predicted with greater confidence, generative design has the potential to become one of the most transformative applications of materials informatics.

 

 

For more details on materials informatics, including profiles of over 35 leading companies and a full breakdown of application areas, see the IDTechEx market report "Materials Informatics 2025-2035: Markets, Strategies, Players."

 

You can get full report details and downloadable sample pages at www.IDTechEx.com/MaterialsInformatics. For more information on IDTechEx's other reports and market intelligence offerings, please visit www.IDTechEx.com/Research.

 

IDTechEx guides your strategic business decisions through its Research, Subscription and Consultancy products, helping you profit from emerging technologies. For more information, contact research@IDTechEx.com, or visit www.IDTechEx.com.