A country’s ability to build and run AI without depending on another country’s infrastructure, models or data centres. Motivated by security, resilience, language and cultural fit, and industrial policy. It is expensive, and the alternative is dependence on decisions made elsewhere.
Technical view
Components include domestic compute capacity, national or regional models trained with local language and cultural data, data residency guarantees, and skills. National programmes have been announced across many countries and blocs, with varying scope from procurement of compute through to full-stack ambitions. The economics are demanding: matching frontier capability independently is costly, so most strategies focus on sovereign infrastructure and adapted open-weight models rather than frontier pre-training.
Related
Part of AI Terminology Index, section: Governance, Law and Regulation.