Lead or Lag: Europe’s AI Materials Race
How AI and Robotics are reshaping the race for materials discovery.
Artificial Intelligence (AI), autonomous labs and robotics are shifting the materials science field from predictive to practical. While China and the US pour capital into this “physical AI” revolution, Europe, with a limited window to lead, shows unique strengths in engineering, process discipline, and industrial infrastructure.
From Prediction to Physical AI
Materials science has long been Europe’s hidden industrial powerhouse. Catalysts, polymers, coatings, and specialty chemicals have quietly fueled its competitiveness for decades, anchoring its position as a global innovation hub in this space.
Now, a new chapter is being written: the fusion of artificial intelligence with robotics to create what many call physical AI. Algorithms can propose molecules and materials at unprecedented speed. But a design on paper (or in silico) is not enough. Models only improve when their predictions are tested, verified, and fed back as new training data. Even better: doing so on an industrially relevant scale. That requires experiments that are reproducible, highly scalable, and fast. In practice, this means robots, automation, and high-throughput labs that can generate the kind of trusted datasets on which both science and industry rely.
The scale of progress is already astonishing. DeepMind and collaborators recently mapped millions of stable crystal structures, expanding the known “materials space” at a scale that would have been inconceivable a decade ago.
Without experimental proof, predictions remain hypothesis, and industry does not operate on hypotheticals.
A Science article last year also highlighted how AI is no longer just predictive but generative, designing entirely new functional compounds. Even Google, historically a software company, is building its own physical lab. Google has understood that no matter how elegant an AI model may be, simulated compounds don’t necessarily translate into the physical world well, especially not at scale. Without experimental proof, predictions remain hypothesis, and industry does not operate on hypotheticals.
This is physical AI in action: robots carrying out chemistry, generating reproducible datasets, and feeding the machine-learning systems, in a constant feedback loop, guiding the next iteration. Chemists remain in control, but the repetitive, precision work is increasingly automated.
Europe is not standing still. Berlin-based Dunia, for example, developed IRIS, the continent’s first commercial self-driving lab for materials discovery in under a year. The platform combines machine learning with modular chemical robotics to test catalysts and electrochemical systems under industrially relevant conditions. IRIS' throughput is equivalent to more than a hundred chemists. This might illustrate what happens when vision meets engineering execution.
A Global Race—And Europe at Risk
The race is not hypothetical. Global competition is already fierce. In China, GCL has teamed up with XtalPi to work on materials discovery, while ByteDance (of TikTok fame) is collaborating with BYD to build better batteries. On the opposite side of the world, American startups such as Periodic Labs, Lila Sciences, and Radical AI have raised close to $700 million between them.
Meanwhile, Europe, as it often does, advances more cautiously. Pioneers like Dunia, Entalpic, and PhaseTree together raised just €22 million. The disparity is not a reflection of Europe’s scientists or entrepreneurs. European talent is world-class, as is the boldness of its ideas. What holds Europe back is capital: investors here remain more cautious, with less appetite for the kind of high-risk, high-reward bets that American backers routinely make. The result is underfunded pioneers trying to compete against rivals with twenty times the firepower.
Yet Europe’s foundation is stronger than it may appear. Chemical leaders such as BASF, Syensco, and Covestro have invested heavily in digital R&D and global innovation hubs. Flagship initiatives like BIG-MAP and AMANDA at HI-ERN established data-driven materials science long before it became fashionable. Dunia’s early decision to bet on this model, when few thought it viable, now looks prescient.
Winning means more than clever code: it’s about building robotic systems that actually work.
Why Europe Can Lead
Execution over hype is where Europe shines. Hype can sell apps and SaaS, but deeptech plays by different rules. Here, substance beats story. Winning means more than clever code: it’s about building robotic systems that actually work, that generate trusted data at scale, plug into industrial processes, and prove performance in the real world.
This is Europe’s comfort zone. Precision engineering, chemical process design, regulatory discipline, and the foundations are all here. What’s missing is supporting capital and urgency.
What Must Change
- For policymakers, AI-driven materials science warrants the same strategic priority as semiconductors and batteries. That means targeted capital, cross-border consortia, and support for ambitious, high-risk ventures; not just incremental projects.
- For industry leaders, the message is equally clear: AI for materials is not a distant curiosity. It will determine who owns the next generation of catalysts, coatings, and electrolytes. Early adoption, joint labs, and partnerships with startups rapidly innovating in this space will shape tomorrow’s market leadership.
A Narrow Window
The stakes are clear and they are high. If Europe embraces this opportunity, it can own the platforms that design and validate the materials underpinning clean energy, digital technologies, and industrial decarbonization. If not, it risks licensing breakthroughs from abroad, ceding advantage at the very core of its industrial base.
Europe has faced such crossroads before. Time and again, improbable visions became industrial realities. AI for materials science is the next such leap. The window to lead is open, but it will not stay open for long.
The choice is stark: lead, or be leapfrogged.
Alex Hammer, Co-founder and CEO, Dunia Innovations
Alex Hammer is the CEO and co-founder of Dunia Innovations, a deeptech startup tackling climate challenges through AI, robotics, and advanced materials. He holds an M.Phil. in Chemistry from the University of Cambridge and a Ph.D. from the University of Glasgow, where he worked on autonomous systems for chemical discovery. Before founding Dunia, Alex held roles in management consulting and R&D digitalization at Siemens and BASF, where he bridged scientific innovation with industrial application.
