Your tariff exposure just moved from the HS code to the origin file

The White House Office of Trade and Manufacturing Policy released a report on August 13, 2026 describing an analytics system, called the "detective border," built to identify and intercept goods rerouted through third countries to avoid US tariffs on Chinese products, and naming more than 40 countries as participants ^[1][2]. For importers, that relocates the audit: the exposure is no longer only the tariff code you selected, it is the country-of-origin claim you attested to and whatever file you kept to support it.

The scale figure comes from a vendor, not the government. Exiger, an AI supply chain firm, estimates that roughly $75 billion of goods may have been illegally transshipped between February 2025 and February 2026, costing the Treasury between $19 billion and $34 billion in tariff revenue ^[3][4]. Take those two numbers together and the implied effective duty on the suspect flow is 25 to 45 percent ^[5], which is the arithmetic that makes an enforcement build worth funding. Treat the estimate with the caution any number deserves when the firm producing it sells software for the problem it describes ^[3].

The mechanics matter more than the branding. According to the report, the system combines shipment data, routing histories, production capacities, ownership ties and packaging patterns, and adds X-ray imaging at ports to catch gaps between declared paperwork and container contents ^[6][7]. Two of those inputs are the ones importers cannot see from their own ERP. Production capacity analysis asks whether a supplier's factory could physically have made what it invoiced. Ownership mapping asks who actually controls that factory. An importer can be entirely honest about its documents and still fail a capacity test it never ran on its own vendor.

The legal hinge is substantial transformation: origin changes when a product undergoes meaningful processing in another country ^[8]. A Chinese steel coil stamped into auto parts in Thailand becomes Thai-origin goods; the same coil repacked in a Thai warehouse with new paperwork does not ^[8]. That distinction has always been the rule. What changes is that the government now proposes to test it statistically rather than read it off a certificate. The practical consequence is that bills of materials, process records, energy and labour inputs, and supplier ownership disclosures become the evidence base, and thin files convert into assessments.

Note who was named. Alongside the expected routes, the report lists Mexico, Canada, the European Union, India, Japan and South Korea among the top enablers ^[9], which means an origin file for a Canadian or German supplier is not automatically lower risk than one for a Vietnamese supplier. The China+1 build-out of the first term produced both real second plants and label-deep ones in places like Vietnam, Thailand and India ^[10]; the analytics do not distinguish by intent, only by pattern.

The cost side is where operators should plan. The report's own framing points to higher compliance spend, shipment delays and retroactive penalties on past shipments the system flags ^[11], with technology, manufacturing and consumer goods most exposed ^[12]. Retroactivity is the item to price: goods already sold carry duty exposure that no future sourcing change removes.

Watch three things. Whether the named allied jurisdictions ^[9] draw the same scrutiny as Southeast Asia, or whether the list is leverage. Whether X-ray discrepancies at the port ^[7] become the standard trigger for a full origin audit, which would make detention rates the leading indicator. And whether any penalty actions actually reach back across the February 2025 to February 2026 window ^[3][11], because that is the difference between a compliance-cost story and a balance-sheet one.