From ancient Greeks to applied AI
The ancient Greeks laid the foundation of Western civilization. I wonder what figures like Socrates and his students Plato and Xenophon would have thought about the concept of artificial intelligence. Might they have interpreted AI as a tool of practical wisdom, rather than the harbinger of a paradigm shift in human history? If that wild speculation holds any water, then modern Europe’s approach to AI would align quite well with the perspectives of those 4th century B.C. philosophers. While Europe does not play a major role in developing foundation models (Mistral, Aleph Alpha, and iGenius are among the few exceptions) and thus does not drive this fundamental transformation, it is exceptionally good by global comparison at making practical sense of AI and putting it to work.
Adoption growth rates
Looking at AI patent registrations, Germany, Europe’s economic powerhouse, accounts for only 6% of the global AI patents. This places it well behind leading countries such as China and the US (with 29% and 27% respectively). Furthermore, these two countries significantly outpace any European nation in terms of overall patent applications. China reportedly applied for 70% of AI patents globally in 2024, followed by the US in a strong, though distant, second place. Obviously, these sheer numbers reveal little about the quality underlying the technology.
However, the picture of the continent is much brighter on the AI adoption front. Here, we can observe significant momentum across Europe. A recent Eurostat report (2025) provides deeper insights: companies ranging from small (10+ employees) to larger enterprises (250+ employees) show a sizable year-over-year growth in AI adoption. Not surprisingly, large enterprises not only demonstrate the highest growth rate but also the highest percentage of AI usage. This is no different from the situation in the U.S. More exposure to competition in global markets is certainly one of the triggers for this.
This picture changes when we look at the specific use cases for AI. Interestingly, the AI adoption gap between large enterprises and SMBs narrows significantly in the Sales and Marketing space, while it remains high in areas like production processes or ICT Security. Several reasons explain this:
- Better availability of Sales and Marketing data,
- faster and more immediate ROI,
- lower entry barriers due to the readily available AI-driven tools in these areas.
Deconstructing the paradox
The gap between AI adoption and foundation model development appears paradoxical. However, there are a number of reasons that explain this phenomenon. They are primarily structural and financial.
While the tech giants in the US are predominantly consumer-facing, Europe maintains a strong B2B focus, deeply rooted in its economic history. For Europe, applying AI rather than prioritizing the development of massive foundation models does make sense, even if this approach carries inherent risks in today’s rapidly evolving global landscape. Below is a brief selective list illustrating Europe’s industrial DNA.
Germany: Automotive and Industrial Engineering, Pharmaceuticals.
France: Luxury Goods, Aerospace and Defense, Nuclear Energy.
Italy: Industrial Robotics, Fashion and Design, Machinery and Equipment.
Netherlands: Semiconductor Equipment, Agrifood Technology.
Sweden: Industrial Engineering and Manufacturing, Green Technology.
Denmark: Life Sciences and Pharmaceuticals, Wind Energy Technology.
Austria: Mechanical Engineering and Plant Construction.
UK: Aerospace and Defense, Life Sciences and Pharmaceuticals, Financial Services.
What hinders European foundation model development?
A strong AI adoption should not preclude a robust foundation model industry. This isn’t an "either/or" question. In fact, it’s easy to imagine that widespread AI adoption would actually accelerate the development of models rather than hinder it. Three main reasons for Europe’s ‚model gap‘ can be identified:
- Venture Capital: The most significant barrier is a staggering gap in investment. In 2024, U.S. private AI investment grew to $109.1 billion—nearly 12 times China’s $9.3 billion and 24 times the U.K.’s $4.5 billion. The overall AI investment in the EU was roughly one-eighth of that in the US.
- Regulatory "Superpower": The EU Commission, speaking for 27 countries, follows a ‚regulation first‘ approach. While creating a safe, trustworthy, and ethical AI continent is a meaningful goal, it carries the risk of being left behind and fosters a more costly and complex compliance environment.
- Market Fragmentation: Europe officially represents a single market, which in many ways facilitates business across its combined 450+ million people economy. Still, it remains fragmented by languages, cultures, and national regulations.
What lies ahead of Europe?
Combining smart AI applications with a solid industrial base is an excellent starting point and makes for a strong value proposition in a global economy. Industry-specific, "vertical AI" and application knowledge – rather than generic AI – is a better way to compete for Europe’s industries.
The EU AI Act has a potential to help by setting an international standard for trustworthy AI. The Act applies not only to AI systems developed and used within the EU but also to those developed by companies outside the EU if the output produced by the system is intended to be used in the EU. This forces global companies to understand and potentially adapt their AI systems to comply with EU standards if they want to access the European market.
And last but not least: The EU Commission and governments across Europe are investing a lot of money to bolster the continent’s AI sovereignty. This includes e.g. direct investments in AI development ("GenAI4EU"), so-called AI factories to provide supercomputing power and infrastructure and the support of homegrown champions.
And not to forget: All of this makes Europe an attractive destination for companies that want to benefit from high adoption rates.
