A stablecoin corridor is not magic, and it is not a single rail. It is a sandwich: fiat in your sending country converts to a dollar token, the token moves on a public blockchain, then it converts back to fiat in the destination. The middle leg is fast and cheap. The two ends, the on-ramp and the off-ramp, are where the money is actually made or lost. If you only price the middle, you will misjudge every corridor you build.
We treat stablecoins the way we treat any rail in this module: as plumbing with specific failure modes, not as a thesis about the future of money. The question is never "are stablecoins good." It is "does a stablecoin leg beat the correspondent chain on this specific lane, this week, at this size."
What you are actually replacing
Correspondent banking moves money by passing it through a chain of banks that hold accounts with each other. A payment from a smaller bank in one country to a smaller bank in another can hop through two or three intermediaries, each taking a fee and an FX spread, each adding a day. Settlement through that chain often takes several business days, and the all-in cost of a traditional wire commonly runs 2 to 7 percent once fees, FX spread, and intermediary deductions are counted. The World Bank's global average cost of sending $200 across borders sat near 6.4 percent in late 2025, with some corridors well above 12 percent.
The reason this matters now is supply-side. Active correspondent banking relationships have fallen by roughly a quarter globally since 2011, and by 30 to 40 percent in the hardest-hit emerging-market regions, as compliance costs pushed banks out of marginal corridors, and the thinning is worst in exactly the emerging markets where remittance and B2B demand is growing fastest. Where the correspondent chain has already left, a stablecoin leg is not competing with a cheap incumbent. It is competing with nothing, or with informal channels.
Where the corridor wins
Three conditions have to line up for a stablecoin leg to beat the alternative.
Deep liquidity on both ends
The token moves in seconds. Settlement finality is roughly 15 seconds on Ethereum and under two seconds on faster chains, which collapses the multi-day correspondent chain. But the win only survives if you can convert into and out of local fiat at a tight spread. Over 60 percent of global stablecoin liquidity sits in a handful of major pairs. In a Tier 1 corridor with deep markets, conversion back to local currency can clear in 10 to 45 minutes. In thinner markets the off-ramp can take hours and the spread widens enough to eat the headline savings.
Real cost savings after the ends are priced
Independent surveys, including BVNK's 2026 stablecoin report across 15 countries, put the saving from stablecoin transfers around 40 percent below traditional remittance channels, and B2B implementations land consistently in the 30 to 50 percent range once you include fees, FX spread, float, and intermediary deductions. That is a strong number, but note what it includes: the off-ramp. A 0.1 to 0.5 percent on-chain cost is real, and irrelevant on its own if your off-ramp partner charges 3 percent to hand over local currency.
Speed that the receiver can actually use
Faster settlement only helps if the recipient can spend or bank the money. In corridors with mature local payment rail or mobile money integration, the destination fiat lands in a usable account quickly. Where it cannot, you have moved the float problem rather than solved it.
A worked example
Take a payments company sending supplier payouts from the US into a Latin American market on a $50,000 transaction.
The correspondent route: a wire at roughly 3 percent all-in costs about $1,500 and settles in three to four business days, with the supplier's bank applying its own FX markup on receipt.
The stablecoin route: on-ramp $50,000 to USDC at about 0.2 percent ($100), move it on-chain for cents and seconds, then off-ramp into local fiat through a partner with deep local liquidity at, say, 0.8 percent ($400). All-in cost lands near $500, roughly a third of the wire, with funds usable the same day. The saving is real, and it is the off-ramp spread, not the on-chain fee, that determines whether it materializes.
Now change one variable. Move the same payout into a thin-liquidity currency where the off-ramp spread is 3.5 percent. All-in cost climbs back toward $1,800, the corridor loses, and you would have been better off on the wire. Same token, same chain, opposite answer. That is the whole lesson.
What does not change
The token leg is the easy part. The hard parts are the parts this course has already covered.
The on-ramp and off-ramp are regulated money services. In the US, the GENIUS Act, signed into law in July 2025, created the federal framework for payment stablecoins: 1:1 reserve backing in cash and short-term Treasuries, monthly public reserve disclosures, Bank Secrecy Act treatment for issuers, and a hard requirement that issuers be able to freeze or burn tokens under lawful order. None of that exempts you, the operator moving customer funds across the corridor, from your own licensing and safeguarding obligations. If you touch customer money on either end, you are running a money services business, and the license question and safeguarding rules from the earlier modules apply in full.
Reserve quality is a live risk, not a settled one. A payment stablecoin is only as good as its redemption, and a corridor built on a thinly disclosed token inherits that token's counterparty risk on every transaction.
The takeaway
Price the sandwich, not the middle. A stablecoin corridor beats correspondent banking when both ends have deep local liquidity, when the off-ramp spread plus on-ramp cost still clears comfortably under the 2 to 7 percent wire alternative, and when the destination fiat is usable on arrival. It loses, quietly and expensively, in thin currencies where the off-ramp spread swallows the on-chain saving. Run the all-in number per corridor, re-run it as liquidity shifts, and never let a 0.2 percent on-chain quote stand in for the cost of the whole route.