Interview with Prof. Dr. Peter Gumbsch, Director of Fraunhofer IWM, Project Coordinator of MaterialsCommons
Why does the world need MaterialsCommons? What motivated the launch of the MaterialsCommons initiative?
High-performance materials are the foundation of European value creation. Approximately 70 percent of all technical innovations are directly or indirectly attributable to them. They are the decisive enablers for the green and digital transformation, for strategic autonomy, and for industrial competitiveness in key markets such as energy, mobility, and communications.
Yet Europe is under massive pressure: dependence on critical raw materials jeopardises security of supply, the OECD forecasts a 40 percent increase in global resource consumption by 2040 and nearly 90 percent by 2060, and fossil fuel dependence is increasingly being replaced by dependence on critical materials. At the same time, competition with the USA and China has intensified dramatically. Europe is falling behind in key technologies such as AI, microelectronics, and energy, all areas in which high-performance materials play a central role.
Particularly alarming is the pace of change relative to the inertia of European innovation processes: the development of new materials today takes 10 to 20 years. This is simply no longer tenable given climate targets and geopolitical upheavals. At the same time, Europe's research and innovation landscape is highly fragmented, poorly networked, and often inaccessible to SMEs. The lack of FAIR data access alone causes estimated losses of over 10 billion euros annually due to inefficient data search and redundant measurements.
From this situation, across three successive strategy documents, the Materials 2030 Manifesto, the AMI2030 Agenda, and the SRIA of IAM4EU, emerged the demand for an integrated, trustworthy, and federated digital infrastructure: the Materials Commons. MaterialsCommons implements exactly that: A platform that connects over 90 percent of European materials data repositories, accelerates materials development by a factor of 4, and is sustainable in the long term through robust governance. It is a strategic necessity for Europe's industrial future.
What problems does MaterialsCommons solve?
Europe's materials research is highly fragmented today. Materials data are still published in classical journals and distributed online repositories. In parallel, numerous platforms have emerged in isolation, such as Materials Cloud, NOMAD, MaterialDigital, DIADEM, or CAPeX. Researchers must query each database individually, struggle with different formats and interfaces, and cannot simply or even automatically combine data from different laboratories.
Common taxonomies, metadata standards, and ontologies that would enable true interoperability are lacking. Moreover, materials science data is frequently poorly documented and not machine-readable, which massively hampers AI-driven innovation. Between academic data ecosystems and industrial systems such as the Asset Administration Shell, there is a gap that is further deepened by insufficient IP protection and lack of data sovereignty.
The consequence of all these deficiencies: the development of new materials and even the development of new components from known materials has so far not been significantly accelerated by digitalisation efforts.
MaterialsCommons aims to solve these problems by creating, for the first time, a federated digital infrastructure with a single access point that connects the relevant European resources and thereby enables acceleration of materials and component development. MaterialsCommons is intended to make relevant materials data findable and usable for academic research and industrial development alike.
Why is access to the various platforms so important?
Even when materials data is openly available, a materials scientist searching for data on a specific alloy or process today must query diverse platforms individually. Each has its own interfaces, its own formats, its own access methods, and its own nomenclature. For experts who know the landscape extremely well, this costs an enormous amount of time. For everyone else, it means that relevant data simply goes unfound, even though it exists and would technically be accessible.
The AMI2030 Agenda clearly identified this problem and called for a network of interconnected infrastructures with a single entry point as the solution. The European Commission followed this agenda, and MaterialsCommons implements exactly that: a single entry point through which researchers from industry and academia can find the relevant European materials data resources. This massively lowers the entry barrier, accelerates data searches many times over, and creates the prerequisite for AI-driven workflows to function across institutional boundaries.
The single entry point is therefore not a convenience feature but the technical precondition for a networked European materials research community.
What does it take for such a large consortium to create a functioning federated infrastructure in four years that can also survive beyond the end of the project?
The platform is technically feasible, the consortium is strong enough, but the challenges are significant. First, what works in our favour: we are not building a new platform from scratch but connecting existing, proven infrastructures i.e. Materials Cloud, NOMAD, MaterialDigital, CAPeX. The leaders of these platforms are directly involved as Principal Investigators in the consortium.
Materials Cloud and NOMAD represent the most widely used materials data repositories in Europe. They offer resilient and inherently scalable structures, but also tools to keep data sovereignty with the provider. This is complemented by a proven multi-tier access model for private, restricted, and open data, as well as workflow management systems such as AiiDA or pyiron. MaterialDigital excels in ontology work: its PMDco 3.0 is a domain-wide mid-level ontology for materials science and engineering. OPTIMADE, a predecessor project involving several project partners, has already demonstrated that a unified interface across 25 repositories is feasible.
Nevertheless, there are significant development tasks: the semantic quality and thus the intelligibility of data descriptions for humans and machines is still insufficient. Data is often stored in a technically correct manner but not annotated in a way that allows automatic combination with other data. This is precisely the gap that MaterialsCommons closes with its semantic foundation, shared schemas, and federated architecture.
On top of this, there is enormous political tailwind from the Advanced Materials Act and the IAM4EU partnership. And we are deliberately targeting TRL6, a functional prototype, not a finished product. That is realistic for four years.
At the same time, I am very aware of the risks. Twenty-six partners plus over 30 associated partners mean an enormous coordination burden. Researchers are reluctant to change their habits even if the technology works, adoption is not guaranteed. Companies will only share their most valuable data if IP protection and data sovereignty are proven. And agreeing on common ontologies is extremely demanding both politically and technically.
That is why we have defined four clear conditions for success: First, researchers need a visible benefit within the first 18 months, a »wow effect« that shows the effort is worthwhile. Second, we rely on pragmatism over perfection. Better 80 percent interoperability for five platforms than 100 percent for none. Third, we need external incentives for adoption, for example if funding agencies require MaterialsCommons-compatible data depositories. And fourth, financial sustainability beyond 2030 must be secured which is why we are already establishing the MaterialsCommons Foundation during the project and working on a viable business model with membership fees, usage charges, and national co-financing, similar to the ELIXIR model in the life sciences.
The prospects are much better than for comparable undertakings. Because the motto »build it and they will come« has rarely worked in the European research landscape. What will be decisive is acceptance by the relevant academic community and industrial utility. And here, the size and reach of the consortium is our greatest strength.
What lies behind the use cases? Aren't the use cases already huge research projects in their own right? Who implements the use cases?
The use cases are deliberately not standalone materials research projects. They are demonstrators whose task is to validate the MaterialsCommons infrastructure, not to solve the complete research problem. Each use case must show that our digital tools work in practice: semantic interoperability, interoperable workflows, and federated data spaces. They are implemented not by individual partners alone but as a cross-cutting effort.
Each use case involves between 6 and 10 partners working together with several industrial partners. In total, we have written commitments from 30 external partners. Industry contributes real data, laboratories, and requirements, while academic partners develop and integrate the infrastructure components. The incentives are concrete for both sides: industrial partners expect time savings of 25 to 60 percent and cost reductions of up to 80 percent. Academic partners gain access to industrial data that would normally be out of reach, publish in high-ranking journals, and use the developed workflows for follow-up projects. Crucially, each use case plays a triple role: it delivers requirements to the technical work packages, validates their results, and demonstrates industrial added value. This tight integration ensures that we do not develop infrastructure that misses actual needs.
THE project promise is the acceleration of materials development by a factor of 4. What are the success hypotheses, and what are the hurdles?
The acceleration factor rests on several success hypotheses:
Semantic interoperability, that is, linking materials data across organisational and disciplinary boundaries, enables researchers to no longer have to manually search for scattered datasets.
Second, the consortium expects that interoperable workflows, i.e. the ability to seamlessly combine work processes from different systems, will drastically reduce »friction losses« between simulation, experiment, and machine learning.
Third, Self-Driving Labs are expected to execute more than five steps autonomously without manual intervention, allowing far more research experiments to be conducted in the same amount of time. AI models are intended to suggest the most informative next experiments or material compositions and thereby narrow the search space much more quickly.
In addition, surrogate models are expected to partially replace time-consuming and costly physical tests or simulations. And finally, by connecting the many relevant EU repositories, significantly more high-quality data will actually be available.
The hurdles are the insufficient FAIR maturity of real-world data, partners' willingness to share data, and of course the high complexity of technically integrating platforms and workflow systems.
But the greatest leverage lies in the data-rich industrial use cases, where we will test the performance of the MaterialsCommons concept. And against the backdrop of the global challenges we face, EU-wide integration of know-how in key technologies is without alternative.
to top
Fraunhofer Institute for Mechanics of Materials IWM
