Sovereignty is treated as durable control over critical outcomes and dependencies, not as a requirement for total technological self-sufficiency.
Core definition
The paper defines sovereign AI as the capacity to exercise durable, lawful and practical control over critical AI capabilities, data and operational choices. The test includes authority, continuity, substitutability, assurance and strategic choice.
Strategic interdependence
Domestic hosting alone does not guarantee sovereignty, and foreign infrastructure does not automatically remove it. The key question is whether dependencies are visible, governed, portable and replaceable within an acceptable time and risk envelope.
System layers
The research examines semiconductors, advanced compute, cloud concentration, electricity, data governance, model access, connectivity, cybersecurity, talent, public-sector capability and the legal conditions that shape operational control.
Decision discipline
The monograph argues against prestige-driven self-sufficiency. Institutions should identify the workloads that matter, map the dependencies that could stop them, test supplier substitution and recovery, and invest first where bottlenecks create the greatest continuity risk.
Research boundary
The publication includes proposed measures, hypotheses and research methods for future empirical work. Analytical indices and sovereignty measures are presented as research constructs that require validation and local calibration.