An AI of Europe, by Europe, for Europe: Keynote Speech by Luc Frieden at the European Commission’s Brussels Economic Forum

Artificial intelligence is here to stay. And it is not just another technological cycle. It will drive a fundamental transformation that will redefine our economies and societies in profound ways. That transformation brings immense opportunities. New ways to research and innovate. New ways to develop products and services. New ways to improve wellbeing and drive progress. So, the question is not whether Europe participates in this new technology, but how?

We often frame the global AI race as one between the United States and China. But Europe is not destined to be third. We can lead – if we act together. And I would say, we must lead.

Because we need artificial intelligence of Europe, by Europe, for Europe – if I may borrow a distinctly American expression.

An AI of Europe – because we need an artificial intelligence that reflects our uniquely European values. Because if AI will ultimately come to influence all spheres of life, then it must align with our way of life. Europe is known throughout the world for its values and principles: freedom, democracy, rule of law. And the use of AI must be grounded in those. When journalists use AI tools to write press articles, they must do so in line with the freedom of press and expression. When social media companies use AI tools to define their algorithms, they must do so without harming our democratic debate. When doctors use AI tools to do medical research, they must respect ethical principles. Because progress needs purpose, technological breakthroughs need meaning.

When we speak of AI, we tend to focus on the word "intelligence", and to forget the word "artificial". Yet, what we implicitly say when we speak of artificial intelligence, is that it lacks the human touch.

An AI of Europe does not replace humans, it helps humans – to be more efficient, to make better decisions. But humans still decide. Humans are central to making the AI revolution a success. And that is where Europe's most intrinsic value comes in: human dignity – the conviction that every person possesses inherent worth and value simply by being human.

If we want to build such an AI of Europe, one grounded in European values, it must be an AI by Europe. If not, we risk that the AI solutions we adopt are based on other values – especially in an increasingly competitive international environment.

But what does an AI by Europe mean? First, AI cannot remain an entirely unregulated environment. This does not mean regulating details until there is no room left for innovation.

It means providing general guardrails that keep AI solutions within a space that is aligned with our values. We need to combine regulation with innovation. Second, we need AI to be developed by European universities, implemented by European companies and made possible by European capital. I will get back to how we can achieve this, but the fundamental objective is clear: Europe needs to be a producer of AI solutions for the entire world, not a laggard that adopts solutions developed by others. Because that is the ultimate question of today's AI race: sovereignty.

And that is what I mean by an artificial intelligence for Europe. Sovereignty is not just about borders. To be sovereign, is to be free to make one's own choices. And in the AI age, that freedom comes with control – control over data, infrastructure, and models. Who holds the data? Who operates the cloud? Who builds the models?

If the answers to these questions lie outside Europe, then so will our future. Do not get me wrong. I do not want a Fortress Europe. Pushing for European champions in strategic fields, such as AI, is not protectionist. It is a pro-competition argument, as we need a European alternative that can compete with foreign companies in the Single Market.

But what needs to be done to ensure that our destiny remains in our hands? What actions are needed to ensure Europe shapes the future rather than the other way around? In my view, Europe's success will depend on three things:

• the scale at which citizens and companies adopt AI solutions,
• the speed with which we enable them to do so, and
• the ability to mobilize public and private investments.

First, on AI adoption, let us face the reality: there is a lot of talk, but there are only few companies – or governments for that matter – that have fully integrated AI solutions into their core operations. If this does not change, we may very well build Gigafactories, but they risk remaining ghost factories with no customers. Governments can lead the way by stimulating demand through their own procurement.

That is why I welcome the European Commission's call on EU institutions to procure sovereign cloud services. And I am particularly happy that Luxembourg companies were among those selected to deliver on that call. This is an important milestone, because it turns the public sector's operational needs into an opportunity to lead the way by showing private companies the possibilities, awarding valuable contracts to local champions, and creating demand that in turn drives supply.

And because European ecosystems delivered on the call, this is more than procurement. This is Europe building its capacity to act – with Luxembourg Post Telecom, Clever Cloud and OVHcloud, with StackIT, with Scaleway, with Proximus, Clarence and Mistral AI. But for all those positives, it remains insufficient, by far. We need more Member States and more companies to follow on the same track.

That brings me to my second priority: speed. Progress and innovation are rarely linear, and never predictable. AI does not reward perfectionists, it rewards the fast – those who try, fail, and try again. And while Europe loves plans, our competitors deploy first and refine later.Where we see failure as the end of progress, they see it as part of progress. So, let us dispel the myth of the "perfect" AI strategy, drop the pen, roll up our sleeves and get going.

Because speed matters in more ways than one. First, the faster you move, the faster you gather data. If you fail fast, you can learn fast. If you fail slowly, you become irrelevant before you can improve. Second, if you develop the first AI model in a sector, you set the standards, lock in customers, and shape global norms. And third, speed attracts talent. The best engineers do not join hesitant bureaucracies; they join fast-moving teams. It is essential that we create an environment that encourages speed by setting aggressive deployment targets in the public sector, creating sandbox exemptions for AI startups, and by rewarding execution, not just planning.

And the third priority for Europe to lead in the global AI race is to unlock the large amounts of capital lying dormant in European banks. We must turn yesterday's savings into fuel for tomorrow's innovation. Europe sits on €12 trillion in household savings. But most of that money flows into real estate and government bonds, rather than investments in startups because our financial culture remains risk-averse, because pension funds and banks prefer safe assets, and because our regulations make it harder to invest in deep tech.

All of this starves our innovation ecosystem of the capital it needs – giving way to an unnecessary paradox: Europe produces world-class AI research but has very few global AI companies. We need a savings and investment union that caters to our financing needs, especially in AI – and we need it fast. Let us start by doing less – instead of adding national layers on EU regulations, let us stay true to one European standard and end gold-plating.

Second, let us think about earmarking small amounts of pension funds for deep tech investment and make AI startups easily accessible for retail investors. Third, let us celebrate risk-taking and give founders a second chance when they fail. Those changes will allow AI startups to tap into financial opportunities and allow them to scale much easier and much faster. If Europe can create demand for AI solutions, if Europe can speed up AI adoption across companies and governments, if Europe can unlock savings to finance AI startups, I am convinced that Europe can not just keep up with the US and China – it can compete and outperform them. My government is fully committed to that European effort. Our entire AI strategy is deliberately thought, designed, and executed in a European mindset.

• We are pushing for smart regulation that fosters innovation aligned with our European values.
• We concluded the first whole-of-government contract ever with Mistral AI to show that Europe can build and deploy AI at scale.
• We encourage AI adoption across society and the economy through concrete flagship projects and targeted outreaches to the private sector.
• We emphasize speed by leveraging the shorter decision-making ways of a smaller administration and deliberately making AI a priority at the Prime Minister's level.
• And we are pushing for progress on the Savings and Investment Union, while making strategic public sector investments in high-class infrastructure.

Our entire ecosystem is geared towards building solutions that are open, connected, and scalable across borders – all while ensuring sovereignty at the European level. Our partnerships are designed to strengthen European capabilities, not national silos. Because like capital, data and AI must flow across borders to create value and maximize its impact.

And Luxembourg wants to be an AI platform for all of Europe. We do not think nationally, we think European. We do not build walls, we build connections.

And that is my proposal to you: Think European. Act European. Build European. Help us make sure that we will have an AI of Europe, by Europe, for Europe. Because we are faced with a choice:

• between fragmentation and scale;
• between dependence and sovereignty;
• between hesitation and action.

And most importantly, a choice between thinking small or thinking European. Because in this new age of AI, no country can win alone.

But together, Europe can. And together, Europe will.

Thank you.