Where allied control binds
Where allied export control actually binds: per facility, whether compute inside an allied regime traces to a parent beyond it — the structural arbitrage two regimes could be played against. Per facility, the analyzer walks the documented ownership chain (operator to ultimate parent, ten hops), classifies each hop against a four-class posture taxonomy (allied-coordinated, allied-partial, non-aligned, restricted), and emits NONE / MILD / STRONG / OPAQUE — with the per-hop provenance to reproduce it by hand. A documented non-aligned hop dominates downstream opacity; OPAQUE names a chain that terminates before reaching a public-filing or government endpoint, never a chain the record simply confirms is clean.
Methodology
How the analyzer classifies each facility
Four-step deterministic pipeline; every step is reproducible against the substrate plus the versioned regime snapshot linked below.
1. Chain walk
Per facility, the classifier starts at COALESCE(operator_org_id, owner_org_id) and walks ownership_chain_edges upward, picking the highest-authority documented parent at each hop. Termination conditions: a publicly-listed entity (CIK present), a government entity, a public-filing-grade ancestor edge, a cycle, or 10 hops, whichever lands first. The walker is the same one the Facility Intelligence Card uses, so the card view and the analyzer never disagree about what the documented chain says.
2. Posture lookup
Each hop's country_hq normalizes to ISO 3166-1 alpha-2 and resolves through jurisdictional_regimes.json (v2026-04-25, built from the Wassenaar participant list, EU 2021/821, EU Delegated Regulation 2025/2189, national export-control authorities, and CSIS Allen & Goldston, March 2025). Hops whose country does not resolve surface as 'unrecognized' rather than getting silently bucketed.
3. Severity rule
STRONG when the chain reaches a non-aligned or restricted jurisdiction. MILD when the chain reaches at least one allied-partial jurisdiction without reaching a non-aligned/restricted one. OPAQUE when the chain terminates short of a public-filing or government endpoint AND no non-aligned/restricted jurisdiction has been documented; the gap may exist upstream but the public record does not say so. NONE when every documented jurisdiction is allied-coordinated AND the chain terminates at a documented endpoint.
4. Conservative defaults
If a chain terminates opaquely AND a non-aligned hop is already documented, the row is STRONG (the documented signal carries). If a chain has no documented jurisdictions at all, the row is OPAQUE (no fabrication, period). Operators with no organization linked at all surface as a separate data-quality bucket inside OPAQUE so they can be filtered out without contaminating the gap analysis.
Severity reference
STRONGChain reaches a non-aligned or restricted jurisdiction. Compute physically sitting in an allied posture has parent-chain control extending to a jurisdiction outside the coordinated framework; this is the coordination-arbitrage signal Feature 15 is built to surface.
OPAQUEChain terminates short of a public-filing or government endpoint. A coordination gap may exist upstream of the documented chain but the public record cannot confirm it. Surfaced explicitly rather than defaulted to NONE.
MILDChain reaches at least one allied-partial jurisdiction (Wassenaar member or EU dual-use applicable, but no AI-compute-specific instruments); coordination on AI-compute items is not end-to-end.
NONEEvery documented jurisdiction in the operator-to-ultimate-parent chain runs coordinated AI-compute export controls aligned with the US-led trilateral framework.
Regime classifications are versioned; snapshots live under data/processed/jurisdictional_regimes_versions/ so a researcher citing a Scrutica snapshot can re-run the classification against the historical regime data even after a later session updates the live file. Current snapshot: v2026-04-25 JSON snapshot.
Peer cross-reference — CNAS Sovereign AI Index
The closest peer index is CNAS’s Sovereign AI Index (Chavez, Chilukuri, Scanlon, Center for a New American Security, April 2026). The two surfaces overlap on the question of jurisdictional control over AI compute, but the unit of analysis differs; the panel below states that relationship explicitly so a reader citing one in a paper that already cites the other has no need to triangulate from scratch.
Convergence
Both frameworks treat jurisdictional control as a primary signal for AI-compute governance, and both publish stage-aware indicators. CNAS uses announced → signed → appropriated → deployed for sovereign-program disbursement; Scrutica's severity escalates monotonically as the documented chain reaches further into non-coordinated jurisdictions (NONE → MILD → STRONG), with chain coverage as the closest analogue.
Divergence
CNAS publishes a country-level Sovereign AI Index ranking; Scrutica's classifier deliberately bans country-level coordination-failure rankings (per the data-access module's docstring), and every classification is per-facility. The two are not directly comparable as rankings. Scrutica's contribution is the per-facility decomposition that isolates structural arbitrage cases — compute physically inside an allied regime whose ultimate parent extends to a non-coordinated jurisdiction.
Calibration
When CNAS lists a country whose Scrutica-classified facilities carry STRONG-severity gaps, the two should triangulate. When they don't, the divergence sits in the unit of analysis (program-level vs facility-level), not the underlying assessment. A reader reproducing one from the other should expect the facility-level and program-level decompositions to surface different cases, and should not expect program-level ranks to fall out of a facility-count ratio.
Cross-page calibration
The stage-aware framing here mirrors the Sovereign-Dashboard validation work (May 2026): commitments are reported at the stage the source actually documents (announced / committed / disbursed), never collapsed into a single "deployed" aggregate. The Sovereign Execution Tracker carries the same convention, so a reader cross-referencing the two surfaces sees consistent stage attribution rather than an artefact of which page they entered through.
Policy context
Regime coverage, scenarios, and the export-control timeline
Technology-by-regime coverage matrix, multilateral-regime structure, named coordination scenarios, and the export-control policy timeline. Reference material for the analyzer above; not a country-level ranking, and not intended as one.