Looking for European winners of the AI boom? These 6 stocks deserve attention
European Central Bank President Christine Lagarde recently reminded that Europe largely missed the first digital revolution and cannot afford the same mistake with artificial intelligence. The European Commission had earlier signaled a similar direction when at the end of July it opened a tender for up to seven so-called AI gigafactories with 10 billion euros of public money and a promise of at least twenty billion from private investors. That sounds like a giant bet. But it is a fraction of what the rest of the world is pouring into AI. For context: major American hyperscalers are expected to invest roughly 725 billion dollars combined this year.
Key points
Brussels versus global capex. Europe is giving 10 billion euros to AI, but major American hyperscalers will spend around 725 billion dollars this year. And that disparity determines which European companies are actually making money from AI.
Six names, two groups. At four companies, AI is already visible in the numbers, at two it is not yet, and two big, well-known names do not belong in the clean selection at all today.
The safest bet has a catch. One stock is the most readable exposure to data center construction, but today you pay a premium for that growth that leaves little room for error.
+380% and still a loss. One European stock has added hundreds of percent this year even though the company itself remains loss-making and its entire price rests on a single hot segment.
Cheap stock, dangerous paradox. The cheapest name in the selection sells exactly the AI that could undercut its own business, and therein lies why it is so cheap.
That is the whole problem in one comparison. The money that actually drives AI construction today is global, not Brussels money. The growth of the most interesting European suppliers today does not rest on Brussels money; they mainly profit from global investments by hyperscalers and data center operators, regardless of whether the new data center is in France, the US, or Asia.
More important than European policy is therefore another question. At which European names is AI already really visible in revenue, orders, and margins, and at which is it still a side topic for now? We went through six companies. At four, the impact of AI can already be traced directly in orders or results; at two, AI does not yet form a significant part of the investment thesis. And the differences are enormous: from a company with triple-digit growth in data center demand to a stock that has added hundreds of percent this year even though it remains loss-making.
AI is not just about chips. Europe has strong companies right here
When you say artificial intelligence, most investors think of chips from Nvidia. But the processors themselves are only the core of a much broader bill. For an AI data center to work, it needs electricity in volumes that whole factories used to consume, power distribution and protection for that electricity, cooling that can keep thousands of graphics cards packed in one hall cool, optical and data networks through which chips communicate at speeds a normal server room never needed, and finally software and people who connect all that with the operations of a specific company.
In the last two years, it has also become clear that the bottleneck in construction is not so much the chips as available electricity and the ability to remove heat from the hall. That is crucial because power and cooling suppliers see demand earlier than anyone else. Cooling is shifting from air to water: card density in AI halls is so high that air can no longer remove the heat, and the switch to liquid cooling means more expensive and complex equipment, thus higher revenue for its suppliers. Similarly, investments in the internal data center network have exploded: training large models moves huge volumes of data between chips, so optical connectivity has gone from a supporting technology to a critical part of every AI cluster.
And it is precisely in these layers outside the processors themselves that European firms have strong positions. Power and cooling, optical connectivity, materials for optical components, and AI integration into enterprises are all currently occupied by large European suppliers. But demand for them is not driven by Brussels, but by the global capex of hyperscalers and foundries, regardless of where the data center ultimately stands. Europe is weak mainly in the most advanced AI processors and large foundation models, which are dominated by Nvidia, American clouds, and Asian foundries. It is strong, on the other hand, in what is built around them.