National compute capacity across 7 supply chain layers (materials, EDA/IP, chip design, fabrication, packaging, systems integration, cloud/deployment). Capacity scores derived from facility-level FLOP estimates, market share data, and production node assessments. Export control constraint overlay from EAR/Dutch/Japanese regimes. For ownership-transparency tier distribution and Hosted-vs-Controlled FLOP attribution per country, see Compute Visibility. Government AI spending for each country: Sovereign AI. Export control constraints by country: Export Controls.
National capacity at each step of the AI chip supply chain, from raw materials to cloud deployment, with export control constraints highlighted. The matrix exposes a structural reality: very few countries are self-sufficient across all layers, most depend on allies for at least one critical step, and export controls actively constrain the rest. For where computing power physically sits versus who controls it (the hosted–controlled gap), and for the per-country ownership-transparency tier distribution, see Compute Visibility. See government AI spending commitments in Sovereign AI and access constraints in Export Controls.