While the AI summit is taking place in Paris as part of a week dedicated to this technology (which we will discuss later), a group of businesses united under the name UE AI Champions Initiative has published a report entitled “An Ambitious Agenda for European AI”. It is interesting to look at what they say and then sift through the results of the summit, which may validate or invalidate this ambition.
Before going any further, it should be remembered that this is not a political initiative but an entrepreneurial initiative.
Moreover, the report emphasizes that Europe must urgently take concerted action in the field of AI to take advantage of its strengths and overcome its structural weaknesses, which is not a compliment to politicians.
Then I couldn’t help but notice a juicy detail. While the initiative claims to bring together 70 major European businesses(AXA, BNP Paribas, EDF, Total Energies, Lufthansa, SAP, etc.), it also includes pure players such as Mistral. More surprisingly, the report was written under the leadership of General Catalyst, which also signed it and claims authorship.
Now Digital Catalyst (a member of the initiative) happens to be an American investment fund. Of course, they have a presence in Europe with offices in London and, more recently, in Germany (only since the summer of 2024) with the acquisition of the La Famiglia fund, and they have invested in Mistral, but this is a detail that I could only note.
Europe: between ambition and ambiguity
The report reiterates why AI is a challenge for Europe (but who would doubt it?), but also points the finger at the European contradictions in this area.
First of all, because Europe is under economic pressure, with productivity lagging behind the United States, and what’s more, with more limited investments and an ageing workforce.
All this is part of a global AI race with all the major powers investing heavily in the subject. We will talk about it again, but it is no longer a technological issue or even a business issue; we are talking about sovereignty and geopolitics.
But Europe has an ambiguous position on AI, as we might add on innovation in general. It has trouble choosing between its desire to innovate and an irrepressible need to regulate (for reasons that in my opinion are as good as they are bad).
Of particular note are strict data protection requirements such as the GDPR or the recent AI Act (The European AI Act for dummies), which can hamper innovation and adoption compared to countries with more permissive regulations.
Add to this the processes for approving technology that are often longer and more complex than elsewhere, and it is understandable that businesses, whether they are simple users or producers of technology, are worried about entering a race against over-trained opponents with a millstone around their necks.
These businesses therefore suggest reconciling security and agility with a more adaptive regulatory approach that favors experimentation. Or a way to stimulate innovation while reassuring those who swear by governance.
AI, a driver of transformation
The report proposes concrete courses of action for implementing AI to give it a real industrial impact.
Unsurprisingly, we start with the optimization of productivity through the automation of cognitive tasks, which could unlock growth opportunities.
There is also talk of structural or even infrastructural impacts in sectors such as networks, energy and semiconductors. Beyond the efficiency gains, I sense the underlying discourse that AI will help optimize these sectors to make digital technology more sustainable and environmentally acceptable, using technology to solve the harms of technology. I will let you be the judge of whether or not this idea is realistic, while many voices are now saying that the only solution for environmentally friendly digital technology is not sustainability but frugality, i.e. a reduction in both the use and manufacture of equipment ([FR]AI and ecological disruption: getting away from the techno-solutionist illusion).
Finally, sectors have been identified as particularly conducive to the deployment of AI because of the competitive advantage they could gain from it, namely aerospace, automotive, finance, healthcare and logistics.
Europe is not starting from scratch
Yes, Europe likes to beat itself up, and often rightly so when it comes to innovation and investment, but it is not starting from scratch and has advantages to offer.
The report refers to leadership in research and open source, areas in which European universities produce a wealth of talent (provided they do not leave the continent). In addition, the open source approach encourages innovation and, beyond the highly questionable figures put forward, what we should remember from the recent DeepSeek episode is perhaps that open source models have won out over proprietary models (The Biggest Winner In The DeepSeek Disruption Story Is Open Source AI and IBM finally has an AI strategy the market loves).
Furthermore, Europe has cutting-edge industries in highly regulated sectors. This can indeed be seen as an advantage, but with a limitation that has been pointed out so many times from abroad: Europe regulates so much that creating business models that exploit regulation is an easy niche, but creates products that are not easily exportable because they are built on local specificities.
Europe also reassures through its approach based on trust. European data security and protection standards would thus constitute a competitive advantage in a global context that is increasingly sensitive to these issues. Except when faced with competitors capable of developing by freeing themselves from them, even if it means adapting afterwards. Which brings us back to the eternal debate, which boils down to asking whether it is possible to innovate if you regulate before innovation arrives….
Finally, Europe, thanks to its linguistic diversity, is better placed to create AIs that can operate effectively in different markets. Unless, that is, this richness becomes a hindrance, once again, to the speed at which solutions are developed.
But Europe must not flatter itself.
These are real assets, but are they enough to compete with the massive investments in the United States and China? Some of the announcements made at the summit may reassure us, but we must admit that we are entering the race a little late, and it remains to be seen how all this will materialize and, above all, how quickly.
Then there is the European approach, strongly focused on responsibility and ethics, in contrast to the strategy of the major world powers, which prioritize performance and industrialization above all else. There would be a lot to say about Lucie, the “ethical and sustainable” French AI project that didn’t work and drew the wrath and jeers of the general public. Yes, the objective of the project had been poorly communicated, but in 2025 it seems that before an AI is “ethical and sustainable” it must already work.
On this point, I also note that the report somewhat distances itself from European dogma. Yes, it talks about “strengthen Europe’s position in the global tech landscape, while preserving commitment to ethical innovation and democratic principles” but on closer inspection, the word ethics is only mentioned once in the 56 pages of the document, in the introduction. A way of saying “we said the word and now let’s move on to something else”?
Finally, European research still struggles to convert its innovations into industrial applications and to attract the capital necessary for their large-scale deployment.
I hear people talking about the “Airbus of AI” in reference to the only European industrial project that has really worked and produced a world leader. But before Airbus, there were national technological champions that were made to work together, whereas we lack these champions in AI.
What action plan for European AI?
The report therefore suggests structuring European efforts in a number of directions.
Harmonization of regulations between countries is already essential if European businesses are to grow by easily expanding beyond their national market and have less uncertainty. If we start putting obstacles in each other’s way, there is little chance that we will be able to unite in the face of international competition.
The public sector is also called upon to act as a catalyst by adopting AI in administrations to stimulate the ecosystem and validate technological options. Three thoughts come to mind here: the first is that I am not sure that the public sector has the means, the second is that it wants to be a guinea pig, the third is that in my view, when a sector lives from public procurement, it is often disconnected from the market and ends up dying when the perfusion stops (because it always stops one day).
Europe should then encourage partnerships between start-ups, industries and the education system. An excellent idea, but if it hasn’t been done or has been done badly so far, there must be a reason. And when we see that something as obvious as a European Small Business Act has been adopted with a half-hearted commitment and no binding effect.
Then there are roses on infrastructure, security of data sharing, skills development, etc.
Bottom line
Yes, European businesses have an ambition, but will they or rather can they turn it into reality? As is often the case, excessive caution and regulatory bulimia can cause them to fall fatally behind.
There is no denying the lucidity of businesses in the face of the potential of AI and what they need to express their full potential and exist in global competition against the backdrop of the geopolitical issues that accompany leadership on the subject.
The question is whether their ambition will not come up against the wall of politics and European bureaucracy and the continent’s inability to carry out an industrial project.
Gaia-X, the sovereign cloud project? It has suffered from governance problems, internal tensions and, from a commercial point of view, has not found its market.
EuroHPC, which aims to develop European supercomputers? It is experiencing industrialization problems and remains confined to limited fields, with no impact on commercial AI.
Past experiences leave little room for optimism, and if we look at reality and history, the proposed avenues of progress may be nothing more than wishful thinking.
Unless, following the summit, Europe decides for once to move in the right direction and to do so quickly? To go into “commando” mode, putting urgency before certain principles?
We will see.
In any case, the future of AI in Europe is being decided now.
Image : AI by Gumbariya. via Shutterstock.
