Pick a target; the pre-scorer walks its affiliate closure down the documented ownership graph and totals the tracked facilities, compute, and countries it reaches. The estimate is structural only: legal determinations under 15 CFR Part 744 turn on rules the walk does not encode.
The pre-scorer simulates new designations. Already-designated entities are flagged in the results; selecting one switches to inverse-simulation mode.
Inverse simulation starts from a current designation and models the structural reach lost if it were removed. BIS cross-reference →
Simulated designation
Breadth-first walk down the documented ownership graph in the subsidiary / JV / affiliate direction (parent and acquired-by edges are inverted to a parent→child orientation). Max depth 4 hops, matching the offline cascade rule used to backfill historical BIS designations. Each hop names its source and confidence; a closure entry inherits the lowest confidence along its path.
For each closure entry, facilities are matched where the operator, owner, or hardware-owner organization resolves to that entry through the canonical-org map. Direct seed-entity matches are reported as hop 0. The three paths are queried independently, so a facility leased under a different operator from its beneficial owner matches through either; deduplication is on facility id.
Each affected facility is enriched with: (a) current BIS-match status from the cross-reference index, (b) the Allied Coordination Gap classification (NONE / MILD / STRONG / OPAQUE), (c) any sovereign-LP linkage tracing back through fund managers. Each closure entry separately lists: BIS / SDN / EU Sanctions designation status, the entity’s HQ jurisdiction posture (allied-coordinated / allied-partial / non-aligned / restricted) from the regimes table, and the sovereign LPs documented as committing to that entity in the licensed fund/LP database’s fund-LP chains.
SEC-EDGAR Item 21 ingestion produces both real operating subsidiaries (Inc., Corp., Ltd., GmbH) and securities-issuance / cover-page artefacts (rated debt notes, ADS/ADR cover-page text, share-class descriptions, table headers). The closure walk filters the latter by name pattern so they are not counted as subsidiaries. Real operating subsidiaries ingested via SEC EDGAR are unaffected; the count of filtered artefacts is surfaced in the report’s caveats when nonzero.
Transactions not yet ingested from the licensed fund/LP database (weekly refresh), entities outside the canonical-org map, beneficial owners hidden behind unrecorded shell layers, JV partners not disclosed in the licensed fund/LP database, and entity-identity rules BIS uses for designations (alias matching, transliteration, phonetic match) that a structural-graph approach does not encode.
Closure shape and confidence semantics match the offline cascade run that backfills historical designations, so a pre-scorer simulation and an actual cascade agree on what the affiliate graph implies. Each simulation has a stable URL plus a computed timestamp, so a researcher returning to the URL re-runs against current data and can see when the original report was generated.
29fewer entities than BIS named, when this method reconstructs Huawei 2019 — its 51 against the regulator’s 80, a 36.3% shortfall on a case where the answer is known. A comparison of counts, not a roster of missed entities: Federal-Register legal names and ownership-graph entities do not resolve one-to-one in this substrate. It is the reason to read the projection above as a floor.
| Facility | Operator / owner | Power MW | H100-eq |
|---|---|---|---|
| Huawei HoringerCN · AI Training | — | 242 | ~123,799 |
| Huawei Cloud Gui'an Data CenterCN · Hyperscale DC | Huawei Technologies | — | — |
| Huawei Cloud Ulanqab Data CenterCN · Hyperscale DC | Huawei Technologies | — | — |
| Huawei Dongguan Cloud Data Center (Songshan Lake)CN · Hyperscale DC | Huawei Technologies | — | — |
| Hop | Entity | Reached via | Confidence | List status | HQ posture | Sovereign LPs |
|---|---|---|---|---|---|---|
| 0 | Huawei Technologies | seed entity | High | BIS Entity List ×4 | CNRestricted | Kuwait Investment Authority · GIC Private · Temasek Holdings · Mubadala Investment Company4 LPs |
| 1 | Huawei International Co., Limited | Huawei TechnologiessubsidiaryFederal Register (BIS)Huawei International Co.,… | High | — | HKNot classified | — |
| 1 | Huawei International Pte. Ltd. | Huawei TechnologiessubsidiaryFederal Register (BIS)Huawei International Pte.… | High | — | SGNon-aligned | — |
| 1 | Shanghai HiSilicon Technologies Co., Ltd. | Huawei TechnologiessubsidiaryHKEX / SSE public filingShanghai HiSilicon Techno… | High | — | CNRestricted | — |
Auto-picked anchor
This simulation is set beside Huawei 2019 (2019-05-21). Selected as the structural-cascade reference for Chinese-seed designations, where Huawei 2019 remains the most documented analogue. 84 FR 22961.
| Metric | Projected | Anchor · structural cascade(live, computed) | Anchor · BIS-named(initial batch) | Δ vs structural(±15% / ±25% band) |
|---|---|---|---|---|
| Total entities reachedincludes seed for forward simulations | 4 | 51 | 80 | — |
| Affiliates / closure entriesexcludes seed; includes structural-cascade walk | 3 | 51 | 79 | — |
| Facilities (Scrutica-tracked)operator, owner, or hardware owner in the closure | 4 | — | — | — |
| Power capacity (MW)summed across reached facilities | 242 | — | — | — |
| H100-equivalent GPUswhere disclosed | 123,799 | — | — | — |
| Countries reacheddistinct facility-country values | 1 | — | — | — |
The BIS-named column counts the initial batch; across all batches in the Huawei 2019 family, BIS has named 132 entities. The structural-cascade column counts facilities reached by either route: entities BIS named, or affiliates the licensed-database cascade reaches from them.
BIS designated Huawei Technologies Co., Ltd. and 68 non-U.S. affiliates on 2019-05-16, effective 2019-05-21 (FR 84 FR 22961). Subsequent batches on 2019-08-21 (FR 84 FR 43495, +46 affiliates) and 2020-08-20 (FR 85 FR 51603, +38 affiliates) raised the cumulative Huawei-family Entity List count to 132 active designations by 2026-05-03. Scrutica's substrate carries all 132 in export_control_designations. The 80 initial-batch designations anchor the comparable structural reach: a hypothetical Huawei designation announced today would equate structurally to the May 2019 batch, not the cumulative.
Structural-vs-legal divergence: BIS's Entity List determinations apply 15 CFR Part 744 entity-identity rules (alias matching, transliteration, phonetic match), end-user analysis, license-review presumptions, and foreign-policy criteria the Pre-Scorer's licensed-database affiliate-graph cascade does not encode. For state-linked Chinese conglomerates whose subsidiary networks are partially undisclosed in public records, BIS's named-entity reach typically EXCEEDS the structural cascade reach (Pre-Scorer's affiliate graph is missing entities BIS named via manual research). For Western publicly-listed corporations with subsidiaries disclosed in the licensed corporate-ownership database, the two typically agree within ±15% facility / ±25% capacity bands. The Huawei 2019 anchor is a high-divergence case — BIS reach > structural reach by 4-5x for the initial batch.
Live numbers (computed for this report): the structural cascade reaches 4 Scrutica-tracked facilities via the affiliate-graph walk.
Follow-on batches:
Huawei Technologies’s HQ jurisdiction sits in BIS’s Presumption-of-friction band (23.9% denial and RWA share). Denial + RWA share ≥15%. Notable proportion of applications denied or returned-without-action; the licensing regime itself is a binding constraint at the application layer.
24,992 approved1,890 denied5,964 returned without actionCY2018–CY2022
Aggregate across all ECCNs; specific-ECCN posture may diverge (for instance 3A090 advanced-compute integrated circuits after the October 2022 rule). Source: BIS OTE per-country publications under Section 1765 of ECRA / 50 USC §4824. Authority tier 1.
Sovereign LPs (state-linked investors) with documented fund-position commitments to entities in the affiliate closure. Stage-aware caveat: the LP→fund→target chain does not resolve whether the fund manager has actually deployed capital into each named target — the substrate has commitment status (Current / Recommended / Former), not target-level disbursement. Treat “Current” as evidence of an active fund position, not as evidence of deployed capital at the target.
Closure entitiesAll 4 rows: 1.
Affected facilitiesAll 4 rows: 4.
| Sovereign LP | Type | Disclosed $M(across positions) | Fund mgrs. | Stage breakdown | Cross-link |
|---|---|---|---|---|---|
| Kuwait Investment Authority | Sovereign investor | $300M2/2 disclosed | 1 | Current 2 | Sovereign AI dashboard → |
| GIC Private | Sovereign wealth fund | —0/2 disclosed | 1 | Current 2 | Sovereign AI dashboard → |
| Mubadala Investment Companystage qualification required | Sovereign entity | —0/3 disclosed | 2 | Current 3 | Sovereign AI dashboard → |
| Temasek Holdings | Sovereign wealth fund | —0/1 disclosed | 1 | Current 1 | Sovereign AI dashboard → |
The pre-scorer models the structural reach of an affiliate graph built from licensed corporate-ownership databases (affiliate disclosures and ownership edges). The output stops short of legal interpretation: BIS designations under 15 CFR Part 744 (and the parallel SDN and EU Sanctions frameworks) turn on entity-identity rules, end-user analysis, license-review presumptions, and foreign-policy criteria the pre-scorer does not encode.
Stable URL: https://scrutica.com/export-controls/impact?entity=org-huawei&type=BIS_ENTITY_LIST
@misc{scrutica_prescorer_org_huawei_20260903,
title = {Scrutica Regulatory Impact Pre-Scorer: Huawei Technologies},
author = {{Scrutica}},
year = {2026},
url = {https://scrutica.com/export-controls/impact?entity=org-huawei&type=BIS_ENTITY_LIST},
note = {Computed 2026-09-03; 3 affiliates, 4 facilities, 1 country.},
}