Sustainable Materials Discovery

For our AI Circular Economy Conference, we are delighted to have had the opportunity to speak with Josua Viethen, CEO of ExoMatter GmbH. In this insightful interview, he shares his perspective on how AI-driven data intelligence is accelerating sustainable materials discovery and enabling more informed, holistic decision-making in R&D.

We sincerely thank our speaker partner for this inspiring conversation and the valuable insights into how structured data and AI can reduce uncertainty and drive more sustainable material pathways. We are also pleased to announce that Xaiza Aniban (ExoMatter GmbH) will present at the AI Circular Economy Conference, showcasing how AI-powered approaches are transforming materials research and innovation.

Join the dialogue on how AI and data are shaping the circular economy of tomorrow.

Register to join the conversation: https://ai-circulareconomy.eu/registration/
Become a sponsor: https://ai-circulareconomy.eu/sponsoring/
Exhibit on-site: https://ai-circulareconomy.eu/exhibition-booking/

1. Accelerating Sustainable Innovation Through AI Materials Discovery

ExoMatter focuses on leveraging artificial intelligence to accelerate materials research and innovation. From your perspective, how can AI-driven discovery platforms help industries transition toward circular and low-carbon material systems?

At ExoMatter, we start from a simple but often overlooked fact: materials are one of the largest levers for technological progress and sustainability. In almost every industry that produces physical goods, materials account for at least 20–30% of technical innovation.

AI-driven discovery platforms allow us to explore this leverage systematically. Instead of relying on a small set of familiar or standardized materials, AI enables companies to screen millions of known materials and simulate entirely new combinations to find alternatives that are lower-carbon, recyclable, or better suited for circular use.

This is particularly powerful for substitution problems: replacing scarce, toxic, or carbon-intensive materials. With AI-based screening and simulation, industries can identify viable substitutes early, reduce dependency on critical raw materials, and design products with circularity in mind from the outset rather than as an afterthought.

2. From Data Scarcity to Knowledge Acceleration

In materials science, reliable datasets are often fragmented or incomplete. What innovations in data management or model training are helping ExoMatter overcome these limitations and ensure that AI models deliver trustworthy, scalable insights?

Data scarcity is a real challenge in materials science, whether it is about volume, structure and usability. At ExoMatter, we’ve built our own proprietary database with hundreds of thousands of materials, each described by more than 40 curated material properties.

What makes the difference is not just collecting data, but standardizing, validating, and contextualizing it so that AI models can learn meaningful physical relationships rather than statistical noise. We combine curated experimental data with physics-based simulations to fill gaps and increase robustness.

On the model side, we focus on approaches that work well with sparse or heterogeneous data: transfer learning, active learning, and uncertainty-aware models. The goal is not to claim perfect predictions, but to dramatically accelerate early-stage decision-making and guide researchers toward the most promising candidates with confidence.

3. Building Transparent and Responsible AI Pipelines

Transparency and explainability are critical for building trust in AI-driven innovation. How do you ensure that AI outputs are interpretable for researchers, and how do you address the challenge of “black-box” decision-making in materials development?

Trust is critical in materials development. No engineer or scientist will act on a result they don’t understand. That’s why we place strong emphasis on interpretability and traceability.

Our platform doesn’t just output a ranked list of materials. It shows which parameters matter, how trade-offs emerge, and where uncertainties lie. Researchers can inspect why a material is suggested, which properties drive its performance, and how sensitive the result is to assumptions.

We see AI not as a replacement for scientific reasoning, but as a decision support system that makes complex search spaces navigable. Responsible AI, in our view, means keeping humans firmly in the loop and designing systems that support scientific judgment.

4. Collaboration Between Human Expertise and Machine Intelligence

AI can accelerate discovery, but human expertise remains essential for context and creativity. How does ExoMatter’s approach combine data-driven insights with scientific intuition to amplify innovation capability across industries?

AI is extremely good at exploring vast combinatorial spaces. Humans are extremely good at asking the right questions and understanding context. ExoMatter is built around combining these strengths.

Our platform allows researchers to interact with the system using both precise parameter definitions and natural language via a trained LLM. This lowers the barrier between scientific intuition and computational power. A materials expert can translate experience, constraints, and creative hypotheses directly into the search process.

The result is a workflow where AI accelerates screening and simulation, often saving up to 90% of early laboratory experiments, while human experts focus on interpretation, validation, and innovation. That combination is where real breakthroughs happen.

5. The Next Decade of AI-Driven Circularity

Looking ahead, how do you see AI and digital infrastructures shaping the materials industry over the next decade? What kind of global collaboration and policy frameworks do we need to ensure that AI truly supports a regenerative, circular economy?

Over the next decade, AI and digital materials infrastructures will fundamentally change how materials are developed. Instead of trial-and-error dominated processes, we will see digitally guided, simulation-first pipelines that shorten development cycles and reduce waste dramatically.

For this transformation to support a truly circular economy, we need more than technology. We need cross-industry collaboration, and policy frameworks that incentivize sustainable materials choices early in product development.

Materials innovation sits at the intersection of industry, climate policy, and global supply chains. If we get this right, combining AI, open collaboration, and responsible governance, materials can become a key enabler of a regenerative, low-carbon industrial system rather than a constraint.

Source

nova-Institute, original text, 2026-01-29.

Supplier

ExoMatter GmbH
nova-Institut GmbH

Share