As artificial intelligence rapidly expands its role in scientific research, one major bottleneck is becoming increasingly clear: validating the enormous number of new material candidates generated by AI systems. While modern AI models can now propose millions of theoretical materials for applications ranging from semiconductors to batteries, testing whether those materials actually work under real world industrial conditions remains slow, fragmented, and difficult to scale. Berlin based company Dunia Innovations aims to solve that challenge through a new industrial scale research facility designed specifically for autonomous materials discovery and validation.
The company has announced plans for Berlin GigaLab, a 6,000 square metre facility expected to require approximately €280 million in investment. The project is designed to combine artificial intelligence, robotics, automation, simulation, and industrial testing into a unified infrastructure platform capable of discovering and validating advanced materials at industrial scale.
The facility is expected to begin operations in 2028 and create more than 200 direct jobs.
Building Infrastructure for AI Driven Science
Founded in 2022, Dunia Innovations develops AI powered infrastructure for advanced materials research across industries including batteries, catalysts, semiconductors, clean manufacturing, and critical raw material substitution.
The company’s platform combines AI systems, laboratory automation, simulation technology, and industrial experimentation into a closed loop research environment. Dunia launched its first generation platform in 2023, while its second generation IRIS platform became operational in May 2025.
With GigaLab, the company aims to scale this approach significantly by creating a facility purpose built for autonomous experimentation and industrial materials validation.
The platform will integrate AI guided materials design, digital simulations, autonomous laboratory workflows, and industrial grade characterisation systems capable of generating highly structured scientific data at a scale difficult for traditional academic laboratories to achieve.
Solving the Verification Bottleneck
As AI models increasingly generate vast numbers of theoretical material candidates, Dunia argues that the key challenge is no longer discovery itself but experimental verification.
The company says existing scientific data remains too fragmented and incomplete to support the large scale AI systems now emerging in materials science. Additionally, simulations alone often fail to accurately predict how materials behave under real industrial conditions involving heat, pressure, contamination, and operational stress.
According to Dr Alex Hammer, AI systems are already capable of generating millions of potential materials, creating growing demand for infrastructure capable of testing and validating those discoveries rapidly.
Hammer described GigaLab as a new type of scientific infrastructure designed to perform experimentation at industrial scale and remove critical bottlenecks slowing frontier technologies.
Industrial Consortium Behind the Project
To develop GigaLab, Dunia is assembling a consortium of major industrial and technology partners contributing expertise across automation, cloud computing, robotics, simulation, and laboratory systems.
Siemens will contribute digital twin and process simulation technologies, while ABB Robotics will support autonomous laboratory automation systems.
Amazon Web Services will provide cloud data infrastructure and analytics capabilities, while NVIDIA will support high performance AI model training through its Inception programme.
Additional partners include ILS for high throughput testing systems and Merck, which has expressed interest in using GigaLab’s capabilities for semiconductor materials development.
Strengthening European Technology Sovereignty
Dunia believes the project carries broader strategic importance for Europe’s industrial competitiveness and technological sovereignty, particularly in areas such as semiconductors, clean energy technologies, and advanced manufacturing.
The company expects the project to attract significant public sector co investment alongside venture capital and industrial funding as Europe increasingly seeks to strengthen domestic research and manufacturing capabilities.
Dr Dirk Demuth, who previously co founded and led high throughput experimentation company hte GmbH, said earlier attempts to combine AI and materials science often failed because technologies were developed separately.
According to Demuth, Dunia’s approach differs because AI, automation, and industrial workflow systems are being built together from the beginning as one integrated infrastructure platform.
As industries increasingly depend on advanced materials for energy systems, semiconductors, defence technologies, and AI hardware, facilities capable of accelerating scientific discovery at industrial scale are becoming strategically important infrastructure for the next generation of technological innovation.