# Sovereign AI Program Index — Dataset v2026.3 Scrutica # Data & Citation The packaged Scrutica datasets and how to cite them. Each bundle ships as a CSV with a paired JSON Schema, methodology Markdown, SHA-256 checksum, and Croissant JSON-LD descriptor; the Citation view carries the reuse forms for the datasets and for any number rendered on the site. # Sovereign AI Program Index Cite Sovereign AI investment programs across roughly thirty jurisdictions, with announced / committed / disbursed spend stages and qualitative interdependence ratings (NVIDIA / TSMC / U.S. / BIS reach). The three funding stages are tracked separately because conflating them is the single most common source of inflation in sovereign-AI reporting. Version **2026.3**·34 rows·Released 2026-07-17·License [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)·SHA-256 19a19e3f91fa… Download bundle [CSVFlat data, UTF-8](/api/datasets/sovereign-ai-index/csv)[Data DictionaryColumn definitions](/api/datasets/sovereign-ai-index/data-dictionary.md)[JSON SchemaColumn types](/api/datasets/sovereign-ai-index/schema.json)[MethodologyMarkdown](/api/datasets/sovereign-ai-index/methodology.md)[Croissant JSON-LDMLCommons 1.0](/api/datasets/sovereign-ai-index/croissant.json)[SHA-256Integrity](/api/datasets/sovereign-ai-index/sha256) Hosted mirrors ZenodoDeposition pending — a versioned Zenodo record with a resolvable DOI is in preparation. Until it lands, cite by version, URL, SHA-256, and retrieval date. Pending Hugging FaceMirror pending — a Hugging Face dataset mirror is in preparation. The canonical download above is complete and hash-verifiable in the meantime. Pending Citation Gringras, D., & Scrutica. (2026). Scrutica Sovereign AI Program Index (Version 2026.3) \[Data set\]. Zenodo. (DOI pending deposition) BibTeXCopy BibTeX @dataset{scrutica\_sovereign\_ai\_index\_2026\_3, author = {Gringras, David and {Scrutica}}, title = {Scrutica Sovereign AI Program Index}, year = {2026}, version = {2026.3}, publisher = {Zenodo}, note = {DOI pending Zenodo deposition; cite by version + SHA-256 hash from /datasets/sovereign-ai-index until assigned}, url = {https://scrutica.com/datasets/sovereign-ai-index} } Methodology \# Scrutica Sovereign AI Program Index — v2026.3 \*\*Release date:\*\* 2026-07-17 \*\*License:\*\* CC-BY-4.0 \*\*Suggested citation:\*\* see \`CITATION\` block at the bottom. ## Coverage 34 sovereign AI investment programs across 34 jurisdictions. Each row carries three distinct funding stages — \`announced\_usd\`, \`committed\_usd\`, \`disbursed\_usd\` — because conflating them is the single most common source of the "$X billion sovereign AI" inflation pattern in trade press. \`announced\_govt\_only\_usd\` separates the government-pledged component from stacked private FDI pledges (Stargate, Cumulative Microsoft, etc.). ## Funding-stage definitions - \*\*Announced\*\*: total publicly-attributed pledge, including private FDI partners. Inflation-prone. - \*\*Committed\*\*: subset that has cleared legislative / board sign-off, where verifiable. - \*\*Disbursed\*\*: subset that has actually been contracted / spent, where verifiable. \`\*\_is\_estimated = true\` flags when the figure is analyst-assigned (e.g., a backed-out fraction of a stacked pledge). ## Interdependence scoring Four qualitative axes (low / medium / high) describe each program's structural exposure: - \`nvidia\_dependency\`: reliance on NVIDIA accelerators - \`tsmc\_dependency\`: reliance on TSMC foundry capacity - \`us\_dependency\`: reliance on U.S.-jurisdiction inputs (chips, tools, IP) - \`bis\_jurisdiction\_reach\`: whether the program sits inside BIS Entity-List / FDPR reach These are not numeric scores; they are analyst calls anchored on the published primary chips, partners, and supply-chain edges of the program. The qualitative ladder is preserved to surface uncertainty. ## Currency handling \`currency\_of\_record\` records the native announcement currency. USD figures use the FX rate in \`fx\_rate\_used\` on \`fx\_rate\_date\` (with the source in \`fx\_rate\_source\`); we do not re-baseline historical announcements to a single year's USD purchasing-power index — the rate is the rate that turns the literal announcement into USD at announcement time. ## Refresh path The upstream JSON (\`data/processed/sovereign\_programs.json\`) is curated against the program-page primary-source layer in the Scrutica platform; this dataset is regenerated by \`scripts/package-datasets.ts\`. ## Known limitations - Some programs (e.g., Chinese provincial / military-aligned) have \`structural\_non\_disclosure = true\` — the announced figures should be read as floors, not actuals. - Stacked pledges are flagged in \`key\_notes\` and broken out in the platform's \`/sovereign-ai\` decomposition view; the flat CSV preserves the stacked total in \`announced\_usd\`. - The qualitative interdependence ratings are intended for triage, not for cross-program quantitative ranking. ## Citation \`\`\`bibtex @dataset{scrutica\_sovereign\_ai\_index\_2026\_3, author = {Gringras, David and {Scrutica}}, title = {Scrutica Sovereign AI Program Index}, year = {2026}, version = {2026.3}, publisher = {Zenodo}, note = {DOI pending Zenodo deposition; cite by version + SHA-256 hash from /datasets/sovereign-ai-index until assigned}, url = {https://scrutica.com/datasets/sovereign-ai-index} } \`\`\` See also The interactive surface that consumes this dataset lives at [/sovereign-ai](/sovereign-ai); for the platform-wide data-provenance overview, see [/methodology](/methodology). Notify me about New dataMajor featuresRegulatoryResearch Email addressSubscribe [CC BY-SA 4.0·© 2026 Scrutica](https://creativecommons.org/licenses/by-sa/4.0/ "Creative Commons Attribution-ShareAlike 4.0") Built for the AI governance research community