Europe can still win with AI. The key is focusing on physical AI

In the race to keep up with the dizzying pace of artificial intelligence (AI), Europe can “win” by focusing on physical AI. An industry-heavy AI play makes sense for Europe and capitalizes on Europe’s history as an engineering and manufacturing hub and its current manufacturing leadership in key industries such as chemicals, pharmaceuticals and aerospace.

European companies are built on strong capabilities that drive efficiency and effectiveness, ensure quality, and achieve operational excellence. European business leaders have a clear opportunity to build on these strengths by applying AI to the supply chain, logistics, operations, machinery and robotics.

In addition, a focus on physical AI also aligns with the European Union’s (EU) AI regulation, since all major regulations concern consumer/citizen data privacy, not confidential corporate intellectual property.

Europe’s founding engineering heritage positions it well for the future. In physical AI, high-labour-cost economies benefit most, as returns are faster and more predictable.

European companies can harness AI’s potential by investing in “hidden champions.” While the United States is full of industry giants and entrepreneurial upstarts, Europe’s strength lies in the middle.

Hidden champions are mid-sized companies that might not make headlines but they drive excellence and lead markets in their niche and many can be found in Europe. They hold patents, shape the market and are the perfect use case for new AI-powered improvements in robotics and machinery.

Europe can double down on its strengths by innovating physical AI for these “hidden champions.”

While the US startup ecosystem is optimized for software, Europe’s corporate ecosystem is optimized for asset-intensive industries: e.g. automotive in Germany, France, Italy and Sweden; industrial machinery in Germany, Austria and Italy; logistics and manufacturing in the Netherlands, Belgium, Czechia and Poland; and healthcare/pharma in the Nordics, Germany, Switzerland and Italy.

Europe’s environment of industry conglomerates, hidden champions with engineering heritage and public-private partnerships favours long-horizon, capital-intensive innovation if played well.

Europe has strong research institutions but lacks the scale to commercialize AI and robotics. To compete, Europe will need to develop AI-friendly regulation, nurture AI expertise and embrace data sharing.

Regulatory changes would accelerate Europe's entry into the AI era. Countries could build joint European programmes for embodied AI models, similar to past regional initiatives for semiconductor and aerospace.

Governments could develop harmonized safety standards so that a robot certified in one EU country can operate across the bloc.

Additionally, governments should incentivize shared infrastructure (compute clusters, robotics testbeds and simulation environments). Leaders should also consider fast-track regulatory pathways for testing autonomous physical systems such as drones, industrial robots and humanoids.

Overall, adopting a flexible regulatory model would allow companies to experiment in controlled environments while preserving public trust.

Most executives agree that traditional process optimization and automation have reached their limits, so the move to physical AI will require leaders who are ready to think AI-first.

We’ve seen that an AI-first mindset can reshape companies with traditional, linear operations into AI-driven transformation engines that execute at scale.

More broadly, Europe needs a skills and labour strategy that includes workforce development for robotics technicians, AI engineers, maintenance specialists and roles in human-robot interaction.

Europe won’t succeed if it continues on a path of disintegration; instead, organizations must collaborate and share non-confidential data to build “world simulations” and “digital twins” and train AI models.

Physical AI requires massive, heterogeneous, real-world data. No single European company, however large, has enough coverage across environments, industries and operational conditions to train robust models.

Thus, a shared dataspace and collaboration approach would enable broader statistical diversity, reduce model brittleness and accelerate the development of foundational models for the physical world.

Data sharing would accelerate innovation across every industry. For example, if car companies shared supply chain data and collaborated directly, they could build an autonomous driving system that rivals Tesla’s.

In the pharmaceutical industry, companies need to collaborate throughout the supply chain. To build faster automotive innovation, car companies need to share data.

The world simulations (digital twins, physics engines, generative 3D) that will power physical AI are only as good as the real-world distributions they approximate.

With shared data spaces, companies could create high-fidelity digital twins of manufacturing, logistics, mobility, energy and healthcare environments; training loops where simulated data and real data continuously improve each other; and scenarios that would be unsafe or impractical to generate in the real world.

Europe has an enormous advantage because its industries already produce structured operational data. The missing piece is interoperability and pooling, not data availability. The silver bullet isn’t new tech; it’s collaboration.

Companies are already applying an AI-first mindset to reshape their industrial operations. For example:

A global industrials company was targeting ambitious growth over the next decade but needed a stronger supply chain operation to get there. Years of mergers and acquisitions had left their IP landscape fragmented and constant global supply chain disruption was impacting their ability to reliably deliver.

Supply chain leaders deployed an AI-powered inventory model. The models were based on enterprise resource planning (ERP) data, which they validated and calibrated with local inventory managers in plants. It identified how to reduce stock levels by 17% across SKUs (stock keeping unit codes) and plants, saving millions of euros.

A leading global automotive original equipment manufacturer needed to transform its inefficient, error-prone manual procurement processes.

Leaders used AI to transform their procurement process: they automated and streamlined the procurement function, leveraging out-of-the-box automation, AI applications and point solutions, resulting in efficiency improvements of over 20% and created a seamless buying experience.

To fully realize the potential of physical AI, businesses must embrace data sharing. Europe’s AI regulation protects consumer and citizen data, not corporate IP. Non-consumer data falls outside government regulation, leaving its use to business leaders.

Many leaders instinctively hoard data, but isolating it limits learning and value creation. The smarter question is no longer, “How do we build a moat around our data?” but “What new moat can we build with data and AI?”

It’s time to collaborate. Post-generative AI Europe will be shaped by companies that openly share, learn from their own data, and draw insights from across their industries.

When we read about AI, we encounter a spectrum of scenarios, ranging from scary to sanguine to sceptical. One thing is certain: something momentous is happening when we look at all the investments in AI but we still cannot agree on how the technology will change humanity.

Political and business leaders must make active decisions to ensure that humans will not
become obsolete in the AI ​​era. We need more debates about how we use AI for societal
progress. In Europe, our AI-powered change should focus on growth for prosperity and social cohesion, grounded in our industrial strength.

The authors wish to thank Roland Scharrer, partner, Kearney, for his valuable contributions to this article.