Sovereign by necessity? Frontier AI export controls, cyber security, and the limits of national AI capability

2026-08-13

AI Summary

This study addresses the geopolitical and security implications of the extreme concentration of cutting-edge AI models within firms in a few nations, whose export controls threaten the technological sovereignty and cybersecurity of others—particularly smaller states unable to independently develop sovereign AI capabilities. Integrating analyses of training costs, compute concentration, state support mechanisms, and cybersecurity economics, the paper examines the interplay between export restrictions and cyber offense–defense dynamics. It proposes a layered resilience strategy comprising negotiated access guarantees, inference-layer sovereignty, strategic adoption of open-weight models as a hedge, regional capability co-development, and foundational network hardening. The research highlights that open-weight models offer both capability advantages and political risks, and concludes that full AI sovereignty is feasible only for a handful of countries; most must adopt hybrid strategies to mitigate supply disruption risks and enhance defensive resilience.

Abstract

A small number of firms based in two states produce the most capable frontier AI models. The governments of those states have shown both the legal power and the political will to decide which other countries may use these systems. In June 2026 the United States required a leading developer to obtain licences before releasing its most advanced models to any foreign person, including foreign nationals resident in the United States. The affected models were withdrawn worldwide at short notice, partly because the restriction proved impractical to administer. This followed within months of the first documented case of a largely autonomous, AI-run cyber espionage campaign, and coincided with mounting evidence that frontier models alter the economics of both cyber attack and cyber defence. This article examines how these two developments interact, and situates them within the unusual market dynamics now driving large-scale AI development. It argues that access to frontier AI is becoming part of national cyber defence, that such access can be revoked, and that the obvious remedy of sovereign capability remains only partly feasible for all but a handful of states. Drawing on evidence about training costs, the concentration of computing power and the support offered by national AI programmes, it asks what sovereignty can realistically mean for small and middle powers, and for large powers as well. The article proposes a layered strategy: negotiated access guarantees, sovereignty at the level of inference, hedging with open-weight models, pooled regional capability, sustained talent development and continued investment in basic cyber resilience. The open-weight hedge proves at once more capable and more politically exposed than is commonly assumed. Much of the near-term risk lies in how capable models are deployed and contained rather than in their apparent performance.

Problem

• frontier AI
• AI sovereignty
• export controls
• cyber security
• national capability

Innovation

• frontier AI
• AI sovereignty
• export controls
• open-weight models
• inference-level sovereignty