The Hidden Cost of Going Global Isn't Shipping

Cross-border card payments carry a cost stack that a blended processing rate usually hides. Cards issued outside the acquirer's home market draw a higher interchange rate: Visa's own published fee schedule puts foreign-issued card interchange at 1.10% to 2.00% for U.S. merchants. The EU shows the same gap from the other side: Regulation (EU) 2015/751 caps domestic interchange at 0.2% for debit and 0.3% for credit cards, while transactions originating outside the EEA aren't covered by that cap and can run 1.15% or higher. And when currency conversion is involved, the acquiring bank applies its own spread on top of interchange, commonly 0.5% to 3%, and wider on less liquid currency pairs.

None of this shows up as a single line item on a standard statement; it's baked into a blended rate that looks competitive until someone breaks it apart. Local acquiring, routing a transaction through an acquirer domiciled in the shopper's own country, exists specifically to avoid the interchange premium and reduce FX exposure.

US de minimis has been suspended since August 2025, and it just survived its first real legal test: a US Court of International Trade ruling in August 2026 upheld the administration's authority to keep it suspended. The EU took a different path: it eliminated its own €150 duty exemption effective July 1, 2026, replacing it with a temporary flat €3 charge per HS-code line in a shipment, a stopgap that itself expires in 2028 and gets replaced by full tariff treatment. And the UK hasn't moved at all. Its £135 threshold is untouched, with abolition not scheduled until March 2029.

That's three different postures on the same underlying pressure: one market actively restructuring, one that just changed and is already scheduled to change again, one standing still by choice. A brand that extrapolates one market's trajectory onto another will get the timing wrong in at least two of the three.

Add to that Section 321—the general US statute permitting duty-free informal entry for qualifying low-value shipments regardless of origin—and the specific practice some brands built on top of it: staging inventory in Mexico to route the last-mile leg of a US-bound shipment through Section 321. That specific fulfillment play is finished. The underlying statute isn't, which is its own trap for anyone conflating the two. Sources cite as many as 75 separate regulatory changes on the US side alone since de minimis went to zero— a pace no static pricing model, in any of these three markets, can keep up with.

Recoverable duty and tax on international returns is a second, quieter margin opportunity most brands leave on the table. When an international shopper returns a product, the brand refunds the product, the shipping, the duty, and the tax, but the duty and tax portion can be claimed back through drawback, the formal process for reclaiming duties paid on goods later exported, destroyed, or returned. Filing means matching the return against the original import entry, confirming the HS code and duty paid, and submitting inside CBP's filing window: a few lookups and a form, repeated at the scale of a brand's return volume, which is exactly why almost nobody does it.

But it's a rules-based, repetitive task with a clear right answer, which makes it one of the more obvious candidates for an agent to handle systematically rather than something a finance team gets to eventually. Matching records, tracking a filing deadline, and submitting a claim doesn't require judgment, it requires persistence, which is what agentic systems are built for. It's recovered margin with no acquisition cost attached, sitting uncollected because the process is tedious, not because the money isn't real.

A composable, MACH-based platform already makes the merchant-of-record, tax engine, or acquiring provider a configuration decision instead of a re-platform. And increasingly, an agent can be the thing that decides which configuration applies, classifying a SKU, flagging a basket about to trip a threshold, or matching a return against its original import entry, faster than a team can do it manually.

But the platform decision underneath that is what actually matters: not whether it supports multiple markets, but whether it can see acquiring, duty, tax, and shipping as separate lines instead of one blended cost, and adapt to the next rule change before that rule change costs real margin. That's the capability international expansion actually requires now.