A Market in Definition
In recent months, the landscape of materials research and development has begun to take shape more precisely around a key concept: Materials Informatics (MI). It is no longer an abstract term or a long-term vision but a well-recognized market area with dedicated reports, specialized companies, and international players already in the field.
Intellico is a pioneering AI company in this field, both in Italy and in Europe.
According to analyses by IDTechEx and other sources1, Materials Informatics represents one of the areas with the highest growth potential for the next ten years. This field applies data infrastructure and artificial intelligence techniques to accelerate the design and optimization of new materials, as well as formulation processes across various industries, such as chemicals, cosmetics, and food.
The Materials Informatics market is described as solutions and technologies that apply data science, machine learning, and artificial intelligence to accelerate the discovery, design, and optimization of new materials2.
The sector includes software platforms for large-scale laboratory data analysis, specific cloud infrastructures, proprietary databases, predictive algorithms, and companies that provide consulting or specialized services in materials innovation through data-driven approaches.
A Transition That Cannot Be Ignored. The USA and Japan Have Already Embraced the Challenge
According to the IDTechEx report¹, neglecting this shift in research and development puts organisations at a disadvantage. More companies are realising that overlooking these changes means missing out on launching competitive products, improving supply chains, spotting future opportunities, and expanding their business or materials range.
Several companies in the USA and Japan have already begun this adoption following three main approaches:
- Collaborating with an external company;
- Operating entirely in-house;
- Joining a consortium.
Each of these approaches has pros and cons and requires careful evaluation based on three key factors:
- The expected robustness and scalability of the solution;
- The availability of internal resources dedicated to managing the tools;
- The volumes of data required for the type of prediction: for example, in the rubber sector, it is essential to reach a critical mass by aggregating data from other companies in the sector, while for other predictive types, a smaller volume of data may be sufficient.
The report notes that Japanese companies are key end-users of this technology, with most new external entities originating from the United States. Major consortia and academic labs are primarily based in Japan and the US.
The Two Main Market Challenges
Materials Informatics solutions do not just speed up existing processes but introduce two fundamental directions:
- The forward direction, where starting from a material, its properties are analyzed and optimized;
- The inverse direction, considered the ultimate goal: designing new materials starting from the desired properties.
The latter direction promises a revolutionary impact but is also the most complex: it requires a large and structured data base, as well as machine learning models capable of handling sparse and multidimensional information.
The Data Maturity Problem
The primary challenge arises from the fact that data maturity within this sector remains limited at present. Companies often work with:
- Fragmented data, distributed among legacy systems, spreadsheets, or even paper archives;
- Small and heterogeneous datasets;
- Bias and irrelevant data that make it difficult to train advanced algorithms.
Based on our experience at Intellico, the challenges associated with Materials Informatics solutions differ significantly from those encountered in other AI-driven domains, such as autonomous driving or social networks. The prediction targets in this field are governed by principles of physics and chemistry, which means that the probabilistic methods underlying neural networks may not always be sufficient. Achieving predictive accuracy requires alignment with the expected behavior dictated by the relevant physical or chemical laws. Consequently, it is crucial to recognize that, besides ensuring high-quality and ample data, integrating neural networks with physics-informed models can substantially improve outcomes in this domain.
Intellico’s Approach for Empowering R&D Teams
In this context, Intellico, through the Matilde platform, has positioned itself as an effective resource for R&D managers. Even when available data is incomplete or limited in quantity, the platform enables users to gain initial insights promptly and initiate a structured process of data collection and value enhancement, all without delaying or interrupting the adoption of these tools.
Intellico’s approach differs in two main aspects:
- Don’t know what data governance is? No problem.
Thanks to targeted solutions, Matilde allows creating value even when data is not well organized and only partially tracked. This means:
- Access to tools capable of integrating heterogeneous and fragmented sources;
- Possibility to perform comparative analyses;
- Obtaining AI-driven suggestions that guide the researcher even with partial initial information.
In this way, companies can start experimenting with the benefits of Materials Informatics without waiting for the data to be complete and numerically relevant.
One of the advantages of this approach is that the R&D team, together with IT colleagues, gains direct experience in structuring the collection and archiving of laboratory data. At the same time, data collection is organized in a targeted manner concerning actual archiving and traceability needs, with the aim of optimizing and accelerating product development.
- Don’t Know How and How Much to Trust These Tools? Intellico Proposes Only “Explained” and “Transparent” AI solutions
One of the key points highlighted by the report analysts is that the most sophisticated tools have little impact if they are not accessible and understandable to formulators and, more generally, to R&D teams. For this reason, Matilde integrates algorithmic logics and a transparent and intuitive user experience, which together allow:
- Visualizing data and analyses through graph techniques that facilitate the identification of relationships and similarities;
- Easily understanding the origin of the results, always providing contextually what the input was, for example, the mix of ingredients, and how each ingredient affected the result, for example, the prediction of viscosity;
- Tracing all stages of trials and experiments, from the trial phase to testing to approval, ensuring traceability.
The progressive integration of analysis dashboards and visualization tools is also an integral part of the adoption process as the R&D team is ready to use more comparative and analysis tools.
Structuring, Innovating
The foundational elements of Intellico’s approach support not only the organisation of laboratory data but also the enhancement and optimisation of the formulation process where necessary. As with any credible innovation initiative, it is critical to evaluate whether a newly introduced tool merely automates existing tasks or constitutes a fundamentally different method of operation. In practice, Materials Informatics solutions and technologies frequently facilitate the adoption of novel workflows. This capability is often where genuine organisational innovation emerges.
For example, similarity searches compared to previous projects are no longer necessary: the tool directly proposes the most similar formulations to start from on the digital desk. Checking if a particular ingredient has been used, when, and in which product category becomes an action executable with a simple click. The choice of the mix of ingredients to bring to the laboratory can initially be addressed in simulation and only subsequently validated with real tests, which at that point focus solely on the most promising combinations.
Accelerating Product Development Also for Regulatory and Technical Documentation Consultation
The Generative AI component plays an increasingly important role in expediting the review of technical and regulatory documentation. As a result, Intellico’s solution includes an intelligent assistant tool that provides actionable recommendations throughout the workflow. This strategy reduces barriers for new users, optimises onboarding for the R&D team, and supports faster decision-making.
By integrating the capabilities of AI with the expertise of domain specialists, this approach fosters a collaborative environment that values both technological innovation and subject matter knowledge. This enables teams to continuously improve their efficiency, adaptability, and creativity.
Conclusions and Future Outlook
Materials Informatics is now a growing market with active players globally. Although data maturity is currently limited, the sector already requires experimentation, even using imperfect datasets, to gain useful experience for future development.
With the Matilde platform, Intellico supports R&D managers in this transition, providing tools for immediate value creation, promoting transparency and trust criteria, and collaborating in building a functional data base for the sector’s evolution over the next decade.
Authors:
- Marco Rossetto – Product Owner Intellico
- Francesca Saraceni – CEO Intellico
References:
- Materials Informatics 2025-2035: Markets, Strategies, Players, available here
- Zivic et al., 2025, Materials informatics: A review of AI and machine learning tools, platforms, data repositories, and applications to architectured porous materials,([https://doi.org/10.1016/j.mtcomm.2025.113525](https://doi.org/10.1016/j.mtcomm.2025.113525)).
- To read a real case study in the food sector: Manuel Dileo, Raffaele Olmeda, Margherita Pindaro, and Matteo Zignani. Graph machine learning for fast product development from formulation trials. Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track, pages 303–318, 2024. doi:10.1007/978-3-031-70378-2_19.
