Dunia Innovations’ Berlin GigaLab: How an Autonomous AI Lab Could Transform Materials Discovery

Advanced materials—from catalysts and battery components to semiconductor materials—are the foundation of modern industry. But discovering them has traditionally been slow and expensive. Berlin‑based deep‑tech startup Dunia Innovations is working to change that by combining artificial intelligence, robotics, and automated laboratories into a single discovery platform.

One proposed step in that vision is a large autonomous research facility often described as the Berlin “GigaLab.” While publicly available evidence about the exact scale and timeline remains limited, Dunia’s technology platform and partnerships reveal how such a facility would work and why it could matter for Europe’s deep‑tech ecosystem.

The core idea: an autonomous materials discovery engine

Dunia builds autonomous laboratories designed to compress the traditional materials research cycle—design, make, test, and analyze—into a continuous automated loop. Its platform combines physics‑informed machine learning with robotics and experimental data collection to speed up the search for new materials.

Traditional discovery workflows often take years because each step requires manual work and long experimental cycles. Dunia’s approach instead uses AI to generate candidate materials and automated equipment to synthesize and test them rapidly, feeding the results back into the model for the next iteration.

This “closed‑loop” system aims to close the simulation‑to‑reality (Sim2Real) gap—the difference between promising theoretical predictions and materials that actually perform in real‑world conditions.

What the proposed Berlin GigaLab would do

Descriptions of the planned Berlin facility portray it as a large‑scale autonomous laboratory intended to industrialise materials discovery. In that model, the lab would integrate several layers of technology:

1. AI‑driven material design
Machine‑learning systems generate new material candidates, optimize compositions, and predict performance using both physics models and experimental datasets.

2. Automated synthesis and robotics
Robotic lab equipment carries out chemical synthesis, sample handling, and measurement tasks continuously. Industrial robotics systems are increasingly used for automated laboratory workflows across industries such as chemicals and life sciences.

3. High‑throughput testing and characterization
Instruments rapidly measure the physical and chemical properties of each candidate material, producing the experimental data needed to validate—or reject—AI predictions.

4. Simulation and digital twins
Advanced simulation environments can train robotics workflows or test industrial processes virtually before physical deployment. For example, industrial robotics platforms are integrating NVIDIA Omniverse simulation tools to create highly realistic digital environments for training robots and optimizing operations.

Together, these components create a continuous pipeline: AI proposes → robots build → instruments test → data retrains AI.

Partnerships and ecosystem collaborations

While many organizations are mentioned in discussions around the GigaLab concept, only some collaborations are clearly documented.

• Hitachi High‑Tech Europe – Dunia announced a strategic partnership with the company to accelerate the discovery, characterization, and industrial deployment of next‑generation materials for sustainable fuels, chemicals, and clean‑energy technologies.

• ASCEND catalyst initiative – Dunia participates in the European ASCEND program alongside Siemens Energy, BASF, Helmholtz‑Zentrum Berlin (HZB), and the Fritz Haber Institute. The €30 million initiative focuses on accelerating catalyst innovation using AI, automation, and industrial collaboration.

• Industrial robotics and simulation ecosystem – Robotics and simulation technologies from companies such as ABB Robotics and NVIDIA are widely used to enable automated laboratories and industrial robotic workflows. Their integration of robotics with physically accurate simulation platforms highlights how digital environments can accelerate deployment of complex robotic systems.

However, publicly available sources do not confirm that all companies sometimes mentioned in discussions—such as AWS, ABB Robotics, NVIDIA, Siemens, ILS, or Merck—are formal partners in the Berlin GigaLab itself.

Why AI‑generated materials still need massive experimental labs

AI can rapidly propose new materials, but predictions alone are not enough for industrial adoption. Several challenges make large‑scale experimental verification essential:

Real‑world synthesis effects
A predicted material may behave differently depending on how it is synthesized, including impurities, microstructure, or manufacturing conditions.

Performance validation
Industrial materials must be tested under realistic operating conditions—temperature, pressure, chemical exposure, and durability over time.

Reproducibility and scale‑up
Before companies can adopt a new catalyst or battery material, researchers must prove that it can be manufactured reliably and economically.

Autonomous labs help address these challenges by running thousands of controlled experiments, generating the large datasets needed to refine AI models and validate results.

Target industries for AI‑driven materials discovery

Dunia’s work focuses particularly on electrocatalysts and chemical processes linked to the energy transition, including technologies for green hydrogen, ammonia production, and carbon‑dioxide conversion.

Catalysis is especially important because catalysts underpin many industrial processes—from fuel production to chemical manufacturing—and even small efficiency improvements can have enormous economic and environmental impact.

Other sectors frequently associated with advanced materials discovery—such as batteries, semiconductors, and energy‑storage systems—also rely heavily on new materials breakthroughs, though specific programs for these sectors have not been publicly confirmed for the Berlin facility.

Potential impact on Europe’s deep‑tech ecosystem

If built at large scale, an autonomous materials “GigaLab” could reshape how industrial research is conducted in Europe.

Instead of small research teams running experiments sequentially, companies and research institutes could access a factory‑style discovery platform capable of testing vast numbers of materials candidates quickly. That would accelerate innovation in climate‑critical technologies such as clean fuels, sustainable chemicals, and advanced manufacturing materials.

Dunia’s broader strategy—combining AI models, robotic experimentation, and industrial partnerships—signals a shift toward AI‑native research infrastructure designed to industrialize scientific discovery itself.

While details about the proposed Berlin GigaLab’s scale, funding, and launch timeline remain limited in publicly available evidence, the concept illustrates a growing trend: building automated research facilities that operate more like factories than traditional laboratories.

For materials science, that shift could dramatically shorten the path from computational prediction to real‑world industrial deployment.