Sovereign AI
소버린 AI
A policy concept referring to national strategic efforts to place AI infrastructure (compute, data, models, talent, and regulation) under domestic control. It emerged around 2020 in European digital-sovereignty discourse and was popularized from 2023 by NVIDIA CEO Jensen Huang, who pitched it to governments worldwide. By 2026, over 50 countries were pursuing some 130 sovereign AI projects, yet full-stack sovereignty remains virtually impossible given the transnational entanglement of GPU, cloud, and semiconductor supply chains; the Brookings Institution has proposed 'managed interdependence' as a realistic alternative.
Sources
- Wikipedia (EN) Encyclopedic overview: origins, scope, policy approaches, criticism, and the managed-interdependence debate.
- cambridge.org Robert Dale, 'Sovereign AI in 2025,' Natural Language Processing (Cambridge University Press, 2025): academic analysis of full-stack sovereignty, infrastructure dependence, and the Jensen Huang promotional campaign.
- lawfaremedia.org Pablo Chavez, 'Sovereign AI in a Hybrid World,' Lawfare (2025): policy analysis tracing the concept from digital sovereignty, surveying national model-building and compute-infrastructure strategies.
- brookings.edu Brookings Institution (Feb 2026): 'Is AI Sovereignty Possible? Balancing Autonomy and Interdependence,' the report that introduced 'managed interdependence' as the realistic alternative to full-stack sovereignty.