The EU, artificial intelligence and the politics of more

The more domains Europe declares to be strategic, the less strategic each becomes. Three EU announcements within a year; a new chips package, a strategy pushing artificial intelligence (AI) into industry and a delay to the AI Act all illustrate a Union mistaking motion for direction. Draghi’s estimate of an €800 billion annual investment gap and Letta’s telecoms figures expose ambitions that outrun delivery, while the shortfall of the first Chips Act compared to Asian and American investment confirms the pattern. The next phase of European autonomy requires prioritisation rather than proliferation. This means a short list of true non-negotiables such as digital infrastructure and defence; a contested middle ground where European firms can still lead, including industrial AI, semiconductors and pharmaceuticals; and a third category of managed interdependence, where dependence is negotiated rather than denied. Europe’s AI debate has also neglected societal consequences beyond regulation, a gap Pope Leo XIV’s recent encyclical exposes and one that should be fixed. The state should build the scaffolding; firms should execute. Europe’s credibility will ultimately rest on the courage to subtract, not add.