The technology sovereignty package of 3 June 2026: is Europe really changing course?

On 3 June 2026, the European Commission presented its "technology sovereignty package", a set of texts and initiatives covering four pillars: semiconductors, artificial intelligence, cloud computing and open source. The chosen definition is progress in itself: technological sovereignty is described as Europe's capacity to act independently in the digital world, by developing and controlling key technologies, data and infrastructure, while reducing dependence on third-country suppliers. In other words, neither autarky nor incantation: a question of capacity to act.

Two regulations, a strategy, a roadmap

The legislative core of the package rests on two proposals: a "semiconductors 2.0" regulation, which takes stock of the first Chips Act's limits, and a regulation on cloud and AI development, which aims to build European compute and hosting capacity at scale. Added to these are a Union open source strategy — which we come back to in Open source: illusion or pillar of sovereignty? — and a roadmap for digitalisation and AI in the energy sector.

The momentum carried through the summer: in late July, the Commission launched a call for projects for "AI gigafactories", aiming to mobilise more than 30 billion euros of investment in very-large-scale compute capacity. In early August, it announced the accelerated deployment of the sovereign IRIS² constellation. The pace, unusual for Brussels, reflects a realisation: the window for existing in frontier AI is closing fast.

Parliament prepared the ground

This package does not come out of nowhere. The European Parliament adopted by a large majority a resolution on technological sovereignty and digital infrastructure, noting that digital power is concentrating in the hands of non-European companies, to the detriment of Europe's capacity to innovate, to compete and to keep control of its economy, its society and its democracy. The diagnosis is now consensual; it is execution that divides.

What can change — and what will not

Three elements argue for a real effect. First, the capability-based approach (compute, manufacturing, orbit) rather than regulation alone: a change of nature after a decade centred on law. Second, the articulation with public procurement: several provisions open the way to preference criteria for European-controlled offerings on critical segments. Third, the enshrining of open source as a sovereignty instrument in its own right, not a peripheral topic.

Three limits, on the other side. The amounts, first: 30 billion for the gigafactories remains modest against the annual capex of the American hyperscalers. The timeline, next: two regulations to negotiate in trilogue means two to three years before application. Demand, finally: without purchase commitments from Europe's large buyers — states included — the capacities created risk running empty. Sovereignty cannot be decreed on the supply side; it is built on the demand side.

For European technology companies, the message is twofold: funding instruments and a preference framework are being put in place, but the credibility of the pivot will play out over the next eighteen months, in the first tenders and the first gigafactories actually contracted. That is where to look — not in the press releases.