The 6,000-square-metre autonomous R&D facility, backed by Siemens, ABB Robotics, NVIDIA, AWS and ILS, is positioned as Europe’s answer to the materials-verification bottleneck that AI-led design has opened up.
Dunia Innovations, the Berlin-based deeptech company building autonomous infrastructure for materials R&D, announced plans on Wednesday for a 6,000-square-metre, €280m facility in Berlin called GigaLab, designed to discover and develop advanced materials at industrial scale.
Siemens, ABB Robotics, NVIDIA, AWS and ILS will provide core technology. The facility is expected to create over 200 direct jobs and begin operations in 2028.
Founded in 2022, Dunia operates an integrated platform that combines AI, lab automation and simulation into a closed-loop system serving customers in catalysts, batteries and semiconductors. The first-generation platform launched in 2023; the second-generation IRIS platform went live in May 2025.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
GigaLab is the company’s bet that the next bottleneck in frontier materials discovery is not algorithmic but physical: the experimental-verification capacity needed to validate the millions of candidate materials AI models are now generating.
Dunia’s core argument, as laid out by chief executive and co-founder Dr Alex Hammer in the announcement, is that the published scientific record is too fragmented to train the large-scale models reshaping other domains, and that simulation alone has failed to predict how materials behave under real-world conditions of temperature, pressure and contamination. ‘
With AI already dreaming up millions of new materials, the demand for experimental verification is exploding,’ Hammer said. ‘We need factories that do science at industrial scale. GigaLab will be the first facility of its kind to do exactly that.’
The industrial consortium Dunia is assembling around the facility carries most of the strategic weight. Siemens provides digital-twin and process-simulation technology; ABB Robotics provides lab automation for fully autonomous experimentation; AWS handles cloud infrastructure and large-scale analytics; NVIDIA supplies high-performance compute for AI model training via its Inception programme; ILS contributes advanced high-throughput parallel testing equipment.
Merck has expressed industry interest in GigaLab’s capabilities to accelerate next-generation semiconductor materials. The Siemens-and-NVIDIA combination matches the same partnership architecture the two companies have been deploying in industrial robotics, with the digital-twin-plus-edge-compute stack now extending into materials science.
The European competitive context matters here. Dr Dirk Demuth, Head of Corporate Development and co-founder of hte GmbH, said in the announcement that the seriousness of the integration is what separates the Berlin GigaLab from prior generations of materials-AI infrastructure.
‘We’re building AI, automation, and industrial-grade workflow design all together from the ground up, not bolting them onto each other,’ he said.
Dunia is positioning the project as strategically relevant to European competitiveness, sustainability and sovereignty, and expects the GigaLab to attract significant public co-investment alongside venture capital and industrial partners.
The company has separately made the case for a €500m EU-funded materials-testing facility, a longer-cycle initiative running in parallel.
On the funding-history line, Dunia has raised approximately $11.5m (€10.6m) in October 2024 co-led by French VC Elaia and Swiss VC redalpine, and is currently raising additional capital to fund GigaLab. The European Commission’s Horizon programme has separately supported Dunia’s electrocatalyst-discovery work under a dedicated CORDIS grant.
The €280m GigaLab is materially larger than anything Dunia has previously announced; the company has not disclosed the funding mix that will finance the build, though the announcement signals a combination of venture capital, industrial-partner investment and expected European public co-investment.
The wider European deeptech context is the part this story sits inside. European deeptech VC structures have struggled to underwrite the multi-decade, capital-intensive infrastructure bets that frontier materials work requires, and successful first funds like Berlin-based World Fund’s record-breaking €300m climate-tech vehicle have begun to address that gap from the climate-tech end.
A €280m single-facility commitment with named Siemens, ABB, NVIDIA and AWS partners is the largest visible European materials-infrastructure announcement of the year.
Whether the project’s announced 2028 opening timeline holds will depend on how quickly the public-co-investment piece materialises alongside the existing industrial commitments.
