SWIFT connects more than 11,000 financial institutions in over 200 countries and facilitates trillions of dollars in cross-border payment messages annually. It is the closest thing the world has to a universal language for moving money between banks. It is also slow, expensive in aggregate, and built on an architecture — the correspondent banking chain — that adds days and hidden costs to payments that could, on modern rails, settle in seconds.
Stablecoin B2B payments are not replacing SWIFT yet. But they are replacing the most painful part of it: the multi-day, multi-hop settlement chain that slows down payroll, supplier invoices, and treasury flows in exactly the corridors where businesses need speed most.
What SWIFT actually does — and does not do
SWIFT (Society for Worldwide Interbank Financial Telecommunication) is a messaging network. It does not move money; it sends standardized instructions between member banks. When your business initiates an international wire, your bank sends a SWIFT message to the recipient's bank — or, more often, to a correspondent bank that has a relationship with both.
The correspondent banking model works like a chain of postal offices. If your bank in New York needs to send funds to a supplier in Jakarta, and there is no direct relationship, the message may route through a US correspondent → a regional hub in Singapore → the Indonesian correspondent → the final bank. Each link in the chain:
- Must receive and validate the message
- Holds nostro/vostro accounts (pre-funded pools of the other currency)
- May deduct fees from the principal
- Settles through its national RTGS system, which operates on a schedule
The result: a payment that carries information instantaneously (SWIFT messages travel in seconds) can take 1–5 business days to settle because the actual fund movement clears through national systems at each hop. A three-hop payment may experience overnight settlement windows at each correspondent, compounding delay.
Correspondent banking cost anatomy for a $100,000 B2B transfer:
| Component | Typical cost |
|---|---|
| Sending bank wire fee | $25–$50 |
| Correspondent bank deductions ($10–$100 per hop) | $20–$300 |
| Receiving bank fee | $10–$30 |
| FX spread vs. mid-market (1–3% for major pairs) | $1,000–$3,000 |
| Float cost (3–5 days of capital tied up) | Significant for high-volume businesses |
| Total on a $100,000 transfer | $1,055–$3,380+ |
The FX spread is often the largest single cost and the least visible — it appears as an exchange rate rather than a line-item fee.
What stablecoin B2B rails offer instead
A stablecoin payment moves a dollar-pegged token directly between blockchain addresses. For USD-to-USD flows — which cover the majority of international B2B invoicing, since most global commerce is priced in dollars — there is no FX step at all. The stablecoin is already a dollar. No correspondent chain, no overnight clearing, no nostro account.
The settlement chain comparison:
| Step | SWIFT wire | Stablecoin on-chain |
|---|---|---|
| Instruction sent | Milliseconds (SWIFT message) | Milliseconds (transaction broadcast) |
| Correspondent routing | 1–3 hops, each adding hours–overnight | None |
| Settlement | 1–5 business days | < 1 second (Tempo, Simplex BFT) |
| Reversal possible? | Yes, within limits | No — deterministic finality |
| Principal deductions en route | Yes (correspondent fees) | No |
On a payments-first chain like Tempo, the on-chain step is deterministic: once a block is finalized under Simplex BFT consensus, the payment is irreversible. There are no re-orgs. For a corporate treasury officer, "settled" means settled — not "probably settled barring some bank's end-of-day reconciliation."
The float problem
One cost that does not appear in any fee schedule is float — the capital tied up in a payment during the 1–5 days it is in transit. For a business making $10 million per month in cross-border supplier payments, average float of 3 days means roughly $1 million is perpetually in-flight and unavailable. At a cost of capital of 5%, that is approximately $50,000 per year in opportunity cost, before any explicit fees.
Stablecoin settlement eliminates float on the on-chain leg entirely. The payment is received within seconds; the capital is available to the recipient immediately.
The reconciliation problem — and how payment memos solve it
SWIFT's MT103 message format carries a reference field, but businesses frequently find that payment references are truncated, modified, or lost in transit through correspondent banks. Accounts-payable teams spend significant time matching incoming wires to invoices.
Purpose-built payment chains address this directly. Tempo's TIP-20 token standard includes a 32-byte memo field on every transfer, and the protocol is designed for ISO 20022-compatible payment data. DoorDash, which processes stablecoin payouts to merchants and delivery workers in 40+ countries on Tempo, specifically cited the structured memo fields as enabling automated reconciliation.
Who is building B2B stablecoin infrastructure
The institutional adoption stack has assembled rapidly since Tempo's mainnet launched in March 2026:
| Partner | Role |
|---|---|
| Stripe | Stablecoin settlement for businesses in 100+ countries; co-incubator and validator |
| Visa | Corporate validator; exploring card rail and stablecoin rail interoperability |
| Zodia Custody (Standard Chartered) | Institutional custody; corporate validator |
| Coastal Bank | Sponsor bank for fintech stablecoin-settled cross-border payments |
| DoorDash | Global merchant and worker payouts in 40+ countries |
| Deel | Payroll for 40,000+ businesses across 150 countries via DLUSD stablecoin |
| ARQ | $10B+ annualized cross-border payment volume across LatAm |
Stripe's pricing for its stablecoin settlement product is a flat $0.001 per transaction — a hundred-fold reduction from the per-transaction cost of a wire. For a business processing a thousand supplier payments per month, that is $1 in transaction fees versus thousands in wire costs, before the FX savings.
Where SWIFT retains its role
Stablecoin rails are not yet appropriate for every institutional use case:
- Regulated capital flows — Some jurisdictions require SWIFT MT103 references or correspondent-bank audit trails for regulatory reporting. A blockchain transaction hash is not yet accepted as an equivalent.
- FX settlement on non-USD pairs — For flows between two non-dollar currencies (EUR→JPY, for example), stablecoin infrastructure is thinner. SWIFT and FX markets are more mature for these flows.
- Credit and trade finance — Letters of credit, guarantees, and trade finance instruments are embedded in correspondent banking relationships and have no stablecoin equivalent at scale.
- Very large single transactions — Eight- and nine-figure treasury movements benefit from institutional credit lines and contractual settlement guarantees that stablecoin networks are still building.
The practical path for businesses
The cleanest B2B adoption path today is corridor-by-corridor substitution, not wholesale replacement:
- Identify your highest-volume, highest-cost cross-border corridors (often to emerging markets with multi-hop SWIFT routing).
- Set up a stablecoin account through a regulated provider (Stripe, Kraken, or a bank like Coastal offering stablecoin settlement).
- Negotiate supplier acceptance in USDC or USDT — increasingly straightforward in LatAm, Southeast Asia, and parts of Africa.
- Keep SWIFT for jurisdictions without off-ramp infrastructure and for FX-heavy flows.
The hybrid model captures most of the cost and speed benefit without requiring a complete infrastructure transition.
The bottom line
SWIFT's correspondent banking chain is not going away — it is too embedded in global compliance and too universal in coverage. But its role is narrowing. For USD-denominated B2B payments in corridors with stablecoin infrastructure, the case for routing through 2–3 correspondent banks over 3 days is getting harder to make. Settlement that used to take business days now takes seconds, the cost that used to be measured in percentage points is now measured in fractions of a cent, and the institutional infrastructure — validators that are Stripe, Visa, and Standard Chartered — is no longer experimental. The question for most treasury teams is not whether to adopt stablecoin rails; it is which corridors to start with.