Beware of GeeksBearing Gifts: Building True EU Frontier AI Sovereignty

EU researchers propose a comprehensive framework for achieving frontier AI sovereignty across five pillars and five technology stack layers, addressing Europe's structural dependence on US and Chinese AI models. The analysis reveals fragmentation in existing EU policy and demonstrates how the proposed sovereignty-centered approach could guide strategic interventions across 92 Commission initiatives.

The European Union faces a critical strategic vulnerability in frontier artificial intelligence. The bloc depends almost entirely on US and Chinese models while controlling only 15% of global hyperscale data center capacity and possessing just 6% of US AI supercomputing resources. This asymmetry threatens economic competitiveness, defense capabilities, and democratic sovereignty in an AI-driven world. The research addresses this gap by constructing a decomposed framework linking frontier AI infrastructure to broader strategic objectives rather than treating autonomy as a purely technological problem.

The sovereignty framework integrates five pillars—economic competitiveness, resilience, security and defense, European values, and foreign relations—with a detailed 26-component AI stack analysis. This structured approach reveals how existing EU initiatives operate in silos without coherent prioritization or conflict resolution. The AI Gigafactory Initiative case study demonstrates how narrow economic framings obscure deeper trade-offs between different sovereignty dimensions, suggesting current policy optimizes for visibility rather than strategic outcomes.

For the European technology sector and policymakers, this framework provides actionable guidance for allocating resources across competing initiatives. Rather than incremental adjustments to fragmented programs, the analysis suggests a coordinated approach spanning semiconductor manufacturing, data center infrastructure, model development, and regulatory capability. The identification of critical gaps across 29 sub-components enables targeted intervention in areas where Europe currently lacks redundancy or indigenous capability.

Looking forward, implementation success depends on political coordination across EU member states and sustained investment in non-competitive infrastructure. The framework's explicit treatment of inter-pillar trade-offs creates space for honest policy debates about priorities—whether European AI development emphasizes values-alignment, military capability, or commercial competition with US incumbents.

• →EU controls only 15% of global hyperscale data center capacity and possesses roughly 6% of US AI supercomputing resources, creating structural dependence on external frontier AI models.
• →Existing European Commission initiatives lack coherent prioritization mechanism, with 92 programs operating without explicit sovereignty framework across five strategic pillars.
• →The proposed five-layer AI stack decomposition identifies 26 components and 29 sub-components enabling systematic gap analysis currently obscured by fragmented policy approaches.
• →Economic framing alone misses critical trade-offs between competitiveness, security, defense, values-alignment, and foreign relations objectives in frontier AI development.
• →Implementation requires coordinated member state investment in indigenous semiconductor, data center, and model development capabilities rather than incremental program adjustments.