At Rise of AI 2025 in Berlin, Axel Voss, Member of the European Parliament and a lead architect of the EU AI Act, argues that Europe faces fragmented digital markets, insufficient investment and eroding sovereignty, and that survival, not leadership, is now the realistic goal unless the bloc changes radically. He contrasts the speed and flexibility of big tech and autocratic states with the EU's slow legislative cycle of drafting, consultation and compromise, and argues national politicians still think in outdated national terms instead of pooling forces. He calls for suspending some competition law for years to let European companies merge freely, for a sovereign 'Euro Stack' digital infrastructure, and for concentrating investment in robotics, biotech and quantum computing. He is openly critical of his own AI Act, saying an internal evaluation found as many points against it as in favor, and pushes for harmonized standards, sandboxes, a risk-based approach by company size, and resolving overlaps between the AI Act, GDPR, the Digital Services Act and related laws. He also argues Europe has overreached on data protection and should shift toward a narrower concept of privacy protection instead.
Key takeaways
- Voss argues Europe's realistic goal now is survival, not leadership, unless it fundamentally rethinks its digital strategy.
- He proposes suspending competition restrictions for seven years or more to let European companies merge into joint digital projects.
- He is openly critical of the AI Act he helped negotiate, citing an internal evaluation of 32 points in favor against 32 points opposed.
- He wants GDPR reformed to explicitly allow personal data for training algorithms and to shift from data protection toward a narrower concept of privacy protection.
- He calls for concentrating European AI investment in robotics, biotech and quantum computing, backed by a sovereign 'Euro Stack' digital infrastructure.
Detailed notes
Europe's shrinking window
Voss says that six years ago he would have described a five-year window to become more independent digitally; now the challenges are more critical still. He points to fragmented digital markets, insufficient investment, and undermined sovereignty and competitiveness across the EU, and argues Europe needs a full rethink of strategy rather than incremental fixes. He frames big tech companies as receiving political backing without limits, operating in dimensions of money and power rather than values, and says Europe's opportunity lies in connecting values to technology, but only if it can move as fast as its competitors.
A legislative process too slow to compete
He contrasts the EU's process, paper, discussion, consultation, further discussion, voting, and slow execution, with the speed and flexibility of big tech and autocratic states. He proposes that Europe answer new problems within three months with a broad regulatory frame rather than detailed rules, says national politicians are still 'sticking to the thinking of the last century,' and argues no single member state can act alone any longer. He proposes suspending competition restrictions for seven years or more so European companies can merge into joint projects without market-dominance objections, alongside reformed subsidies, public procurement, infrastructure, funding and cybersecurity as preconditions.
A sovereign 'third way' and investment priorities
Voss argues Europe needs a 'third way' built on a sovereign digital infrastructure, open source and privacy-preserving technology, referencing a European 'Euro Stack' concept modeled on what India has done. He warns Europe could otherwise be 'held hostage' by overseas partners geopolitically, and names robotics, biotech (where he says Europe is still leading but may not remain so) and quantum computing as the minimum areas requiring concentrated investment. He also calls for eliminating regulatory fragmentation to build a genuine digital single market.
A self-critical view of the AI Act and GDPR
Voss says he is 'not the biggest fan' of the AI Act he helped negotiate, citing an internal evaluation that found 32 points in favor and 32 against, and argues it risks hampering the competitiveness of the European AI ecosystem. He describes the final compromise as shaped by seven political groups with no government or opposition at EU level, and says Germany and France pushed through elements at the end that he personally considered 'totally idiotic' for face-saving reasons. He calls for harmonized technical standards, legal certainty, a functioning code of practice, resolving overlaps between the AI Act, GDPR, the Digital Services Act, Digital Markets Act and Data Act, and a more centralized governance structure. On data, he argues GDPR as implemented, especially in Germany, is restricting access to the quality data an AI-driven economy needs; he wanted a clause explicitly allowing personal data for training algorithms and did not get a majority for it. He argues for shifting the underlying principle from 'everything is forbidden unless allowed' to 'everything is allowed, but do not touch citizens' privacy,' and for moving from broad data protection toward a narrower concept of privacy protection, citing as an example that tire pressure data is treated as personal data in Germany.
Sandboxes, talent and less punitive regulation
He proposes regulatory sandboxes, including sector-specific ones, so member states stop competing with incompatible national approaches, and a risk-based approach that spares smaller companies from rules designed for big tech. He argues education systems, hampered by Europe's federal structures, need to do more to build digital talent, and that constantly threatening companies, especially smaller ones, with penalties discourages entrepreneurship rather than motivating it.
Q&A highlights
Asked about the Paris AI Summit pledge of 200 billion euros in investment and a later claim that 100 billion of it would come from national states, Voss says he doubts national leaders were consulted before that figure was announced, and calls for changing rules so private capital, including insurers, can invest in AI the way they already can in the US. Asked whether the political will to move faster reaches down to the directorates-general actually running EU projects, such as the 24-month development phase of the IPCEI-AI project, Voss agrees the courage to take risks is often missing at the administrative level, and suggests the European Investment Bank could do more.
Quotes
I'm representative of the CDU, but I'm asking for revolution and anarchy in this direction, otherwise I can't see how we will survive at the end.
Axel Voss
When we evaluated the result, we find out 32 points in favor and 32 points against. So that's why I would say it's not the best way forward.
Axel Voss
Questions answered in this talk
What does Axel Voss think of the EU AI Act he helped negotiate?
At Rise of AI 2025, Axel Voss says he is not its biggest fan, citing an internal evaluation with 32 points in favor and 32 against, and argues it risks hampering the competitiveness of the European AI ecosystem.
What is the 'Euro Stack' Axel Voss refers to for European digital sovereignty?
According to Axel Voss at Rise of AI 2025, it is a proposed sovereign European digital infrastructure built with open source and privacy-preserving technology, modeled on the kind of national digital stack India has already built.
How does Axel Voss think GDPR should change to support AI development?
Axel Voss argues at Rise of AI 2025 that GDPR should explicitly allow personal data to be used for training algorithms, and that Europe should shift its underlying principle from 'everything is forbidden unless allowed' to 'everything is allowed, but do not touch citizens' privacy'.