Customs agencies inspect a fraction of one percent of global cargo — the rest clears on the strength of paperwork. Fraud exploits that gap: misdeclared weights, false ports of origin, phantom transshipments and timing anomalies that evade tariff schedules or sanctions screens. A sovereign satellite stack changes the odds by generating an independent ground-truth layer that no shipper can falsify after the fact.
The detection stack fuses three data streams. Synthetic-aperture radar measures the waterline displacement of vessels at anchorage and compares it to the declared manifest weight — a 10,000-tonne discrepancy is visible from orbit. RF survey payloads track AIS continuity gaps that signal deliberate transponder suppression during transshipment. Optical revisit at ports-of-interest records container dwell times and yard configurations, feeding a behavioural baseline model that surfaces statistical outliers for human review.
The operational outcome is a prioritised targeting list delivered to customs risk-management officers before a vessel berths, not after it clears. Interception rates rise without proportionate increases in inspection staffing. Equally important, the data is owned nationally: no foreign commercial provider can throttle access, demand audit rights over the ML model, or comply with a foreign court order that exposes a live investigation.