The European Commission unveiled a new technology sovereignty package in Brussels intended to reduce the EU’s dependence on third‑country suppliers for semiconductors, artificial intelligence (AI), cloud services and open‑source technologies. The package contains two legislative initiatives — Chips Act 2.0 and the Cloud and AI Development Act — together with an EU open‑source strategy and a strategic energy‑digitalisation roadmap focused on data centres and AI integration.
Objective: not closure but manageable dependencies
The Commission stresses the aim is not to close the European market but to ensure the EU is not vulnerable to external actors for technologies essential to hospitals, energy grids, public services and industrial systems. Officials describe the objective as creating "manageable dependencies": mechanisms to lower the risk that third countries could cut access to critical technology in a crisis.
Chips Act 2.0: targeted capacity, faster approvals and demand‑side measures
The Commission warns that AI‑related components could account for more than 70% of the semiconductor market by 2030, motivating a revision of chip policy that goes beyond building capacity to speed up permitting, designate strategic projects, introduce a semiconductor excellence region label and increase demand‑side measures.
According to the Commission’s assessment, the first Chips Act mobilised about €52 billion in investment for the European semiconductor industry and created some 16,000 direct and 30,000 indirect jobs. Despite these gains, the EU’s global share of semiconductor production remains roughly 10.5%. The updated regulation aims to complement supply‑side measures with incentives to create European buyers, industrial adoption and public procurement demand for advanced chips.
Measures proposed include continuing design centres, pilot production lines and competence hubs, elevating photonics as a strategic priority, and creating lab‑to‑fab accelerators to bring research closer to manufacturing and industrial use. The package would also establish a semiconductor supply‑chain forum to help spot bottlenecks early and enable authorities to request market information pre‑crisis if needed.
Cloud and AI Development Act: triple data‑centre capacity and a sovereignty framework
The proposed cloud and AI legislation targets a threefold increase in European data‑centre capacity over the next five to seven years and would set a unified EU framework to determine which sensitivity level of public‑sector data may be hosted in what type of cloud.
The cloud sovereignty framework differentiates four levels:
- Level 1: public‑sector data remain in Europe — the Commission estimates ~70% of public cloud needs can be handled at this level by current providers.
- Level 2: third‑country access is restricted and continuity of service cannot be jeopardised without EU jurisdictional control — an additional ~20% of needs.
- Level 3: providers owned and controlled in Europe, though some third‑country providers could be recognised as equivalent if they meet legal and data‑protection guarantees (about 9%).
- Level 4: full control over the entire software technology stack — a narrow category (about 1%) for highly sensitive defence, law‑enforcement or justice functions.
Member states would be required to perform risk assessments of public systems and map appropriate sovereignty levels to identified risks. The Commission does not intend to mandate a single sectoral list; sensitivity can vary inside organisations and across applications. The procurement rules would not bind the private sector but the Commission recommends that critical private operators incorporate the four levels into their risk assessments.
Open‑source strategy: mobilising European contributors for industry and public demand
The Commission estimates around 3 million open‑source contributors are active in Europe. The EU open‑source strategy aims to better align this community with industrial and public‑sector demand — particularly relevant for the highest sovereignty level where control over the full software chain matters.
Energy‑data roadmap: expanding capacity sustainably
The Commission reports roughly 12 gigawatts (GW) of data‑centre capacity in the EU today. The plan foresees tripling capacity over the next five to seven years and reaching around 60 GW by 2036. An industry estimate mentioned in the discussions puts the required investment at about €200 billion (for buildings, cooling, servers, chips and related infrastructure — excluding operational energy costs).
The roadmap rests on three pillars:
- "Energy for AI": manage data‑centre expansion so it does not overload grids or undermine climate goals; develop three‑party agreements between data‑centre operators, energy actors and authorities; introduce a new data‑centre rating covering energy efficiency, water use, flexibility, waste heat recovery and environmental performance.
- AI and digitalisation in the energy system: shift away from automatic grid expansion toward more smart‑grid, AI‑based and demand‑side solutions; examples include vehicle‑to‑grid schemes where shared EV fleets store solar surplus and return power at peak times.
- Opening and cross‑border use of energy data: create an EU framework for cross‑border energy‑data exchange to support demand‑response, bidirectional EV charging and other smart‑energy services.
The Commission projects demand‑side flexibility could yield about €71 billion in direct annual consumer savings, while smart bidirectional EV charging could save consumers anywhere from several hundred to nearly €3,000 per year.
Two concrete initiatives were proposed alongside the roadmap: a memorandum of intent among 14 energy and data‑centre industry associations to cooperate on sustainable integration agreements, and the ai.grid project bringing together 48 European organisations to develop AI base models for network management. The Commission would allocate €50 million from Horizon Europe to ai.grid and nearly €300 million across this and next year’s work programmes for digital energy R&D and innovation.
Financing approach: public seed capital, private majority funding
Financing for the package would not be purely public. The Commission expects private capital to supply most investment, with public funds acting as seed, de‑risking and coordinating instruments. Examples cited include AI gigafactories where public financing may cover about 30% while companies provide the remaining 60–70%.
Potential instruments include the next multiannual EU budget, member‑state support, InvestEU‑style loan and quasi‑equity tools, and a possible new European technology equity vehicle that is still under political and market testing.
Regional participation and political trade‑offs
A practical question for member states, especially in Central and Eastern Europe, is how industrial regions can join the chip and data‑centre investment wave. The semiconductor excellence region label would signal to investors that a region has investment plans, an industrial ecosystem and basic infrastructure — it is a visibility tool rather than direct funding.
Political debates will focus on how to reduce critical technology dependencies while keeping markets open, cooperating with like‑minded partners and avoiding renewed fragmentation among member states. The Commission maintains the package seeks greater choice, capacity and control in technologies underpinning industry, public administration and energy systems — not decoupling from global value chains.
Conclusion
The Commission’s package combines regulatory measures, targeted public funding logic, industry engagement and a strategic energy dimension to address EU vulnerabilities in semiconductors, cloud, AI and open‑source technology. Public money is positioned as a catalyst; most investment is expected to come from private investors. The proposals will now enter detailed legislative and member‑state negotiations that will shape how the EU balances sovereignty, openness and competitiveness.
