Cross-border payments are slow because the money, unlike the message, has to pass through a chain of banks, and each one works to its own cut-off, calendar, compliance queue and pre-funded account. The instruction can arrive in minutes. The credit waits for the slowest institution in the path, which is usually the last one.
For the shape of the route first, what a stablecoin corridor is covers it in one page.
Where the days go in an international payment
Five things add waiting time, and most payments meet at least two of them.
The five delays in a cross-border payment, what causes each, and how long it can add.
Delay | What happens | Why it adds time |
Cut-offs and non-business days | A request after the bank's daily deadline, or on a weekend or holiday, waits for the next business day | Until the next business day, which can be a weekend away |
Time zones | One bank's afternoon is the next bank's night | Until the next bank in the chain opens |
Correspondent hops and pre-funded accounts | Each intermediary processes on its own schedule, against money already sitting in its account | Hours or days per intermediary, per the Lumx SLA documentation for SWIFT wires |
Compliance review | Screening at each bank, plus holds for missing or unclear data | Until someone clears it, if the payment is held |
Crediting the beneficiary | The last bank books the funds, sometimes after its own internal review | Until the last bank books the funds |
Only the intermediary line has a source, the Lumx SLA page. The rest is the mechanism, not a measured duration.
How long does a cross-border payment take?
It depends on what you measure. The Financial Stability Board (FSB), which tracks the G20 targets for cross-border payments, publishes independent and dated numbers. Its consolidated progress report of October 9, 2025 measures speed in separate pieces for large payments, retail payments and remittances.
FSB 2025 report, latest data points (wholesale to March 2025).
Measure | Result in the FSB report |
Wholesale payments (USD 100,000 or more) credited within one business day | Above 90% |
Wholesale payments credited within one hour | More than 50% |
Wholesale payments from Central and Latin America credited within one hour | 42.7% |
Retail payment services settling within one business day, global | 67.3% |
Retail B2B payment services crediting within one business day | 39.6% (54% in 2023) |
Retail B2B payment services crediting within one hour | 2.2% |
Average cost of sending USD 200 in remittances | 6.5% |
Three readings matter. Large payments are mostly fine: more than half arrive within an hour. The gap opens at the small end, where B2B payments are the slowest use case, and in the corridors we care about: in Latin America and the Caribbean the FSB found B2B retail payments with no services crediting funds within one hour and only 44.5% within one business day in 2024. And the faster wholesale numbers came largely from the beneficiary leg, the time the receiving bank takes to credit the end customer, which the FSB calls the traditionally slowest part of a cross-border payment.
On Lumx the speed you plan against is published per rail: the documentation lists Pix (Brazil's instant payment system), SPEI (Mexico's interbank transfer system) and Faster Payments (the UK's real-time bank transfer network) as instant and running 24/7/365, ACH as 1 to 2 business days, and a USD wire over SWIFT as 1 to 5 business days. They are processing targets, not contractual commitments.
Cut-offs, weekends and time zones
Every bank sets a daily deadline for outgoing international instructions, and the settlement systems behind them run only on business days. Past the deadline, the request is queued for the next business day in that system's own country. A second country with its own holidays means the calendars stop lining up.
One payment shows the arithmetic. This is our own calculation from Lumx's documented cut-off for a USD wire over SWIFT (3:00 PM ET), not a measured payment.
An illustrative Friday payment, step by step. Assumes the first bank opens at 9:00 AM ET.
When | What happens | Elapsed |
Friday, 5:00 PM ET | The instruction is submitted two hours after the 3:00 PM ET cut-off, so it is queued | 0 hours |
Monday, 9:00 AM ET | The first bank starts processing | 64 hours |
Monday, same day | A correspondent picks it up, on its own schedule | Unknown, depends on the correspondent |
Monday evening in Singapore or Hong Kong | The beneficiary bank is already past its business day | Assumption: credit waits for Tuesday morning |
The payment waited 64 hours and nobody made a mistake. Sent at 2:30 PM, it clears the first cut-off and avoids the weekend wait at that bank, though a later bank can still miss its own. So the first diagnostic on a slow payment is its submit time, in the time zone of the rail it used.
The same logic applies at the edges of a stablecoin payment. Lumx's documentation lists a 2:00 PM ET cut-off for ACH and 3:00 PM ET for FEDWIRE (the Federal Reserve's same-day wire system) and SWIFT. Pix, SPEI and Faster Payments have none. Fuller table: payout cut-offs by rail.
Nostro and vostro accounts: why the chain needs pre-funded money
Most banks do not hold an account with every other bank. To pay into a country where it has no presence, a bank asks a correspondent that does, and sometimes two. The correspondent keeps accounts for other banks in the currencies it serves.
A nostro account ("ours") is the account a bank holds in a foreign currency at a foreign bank. The same account, seen from the other side, is a vostro account ("yours"). One account, two names, depending on whose books you read.
This slows payments because the account is pre-funded. A correspondent can only pay out what is already there, so the sending bank must have moved money in beforehand, in the right currency, and the correspondent spends it down one instruction at a time, inside its own business hours and screening. Every nostro-backed hop is one more institution that must act, in its own time zone, before the next can start. The longer comparison is in stablecoins versus SWIFT for B2B payments.
On Lumx, the pricing model is "transparent FX on locked quotes, no pre-funding". That means the customer does not leave capital parked in settlement (nostro) accounts before a transaction: conversion and settlement happen when it starts, and a USDC balance in a Lumx wallet does not contradict that. On the SWIFT rail, which in Lumx is USD only, the documentation is straight about what is outside its control: international wires can pass through correspondent banks, each can add hours or days and may deduct fees. Why fees get deducted along the way is in cross-border payment costs in 2026. Every SWIFT payout carries UETR and MT103 confirmations, so the customer can see where a wire sits.
Compliance holds and repairs
Each bank in the chain screens the payment against sanctions lists and its own anti-money-laundering rules. Some payments stop: a name that partly matches a listed party, a missing beneficiary address, or a purpose too vague to classify. Someone has to read it, and the clock keeps running.
Repairs are the quiet version of the same delay: a payment with missing or unclear data needs a person before it can move. According to Swift, as reported by the FSB, the faster wholesale payments may be attributed to greater bank automation, and the migration to ISO 20022, the structured message standard on the Swift network, may leave further room to improve. See Swift ledger vs stablecoins.
A stablecoin payment does not escape this. The checks run at the on-ramp and the off-ramp instead of at each correspondent. On Lumx, compliance sits inside the flow: a transaction flagged by monitoring is held in an RFI (request for information, a compliance hold that asks for a document before a transaction clears). The documentation says compliance reaches out by email or Slack, usually asking about the relationship between sender and receiver, the purpose and the expected outcome, and that a transaction with no answer in 24 hours may be refunded to the sender. Every on-ramp and off-ramp carries a purpose field, and a mismatch with the real use of the funds can hold the transaction or raise an RFI. The common causes are in why stablecoin payouts get rejected.
Why tracking shows the delay but does not remove it
A wire can carry a unique end-to-end reference in its Swift message, the UETR, so each bank in the chain can report status against one identifier. That ends "we sent it, we do not know where it is". It does not remove the cut-off, the hop or the manual check, because tracking reports the queue and does not shorten it.
On Lumx, every SWIFT payout carries UETR and MT103 confirmations, every transfer has a per-rail reference, and webhooks fire on each status change, so a late payment shows where it sits.
I would sort delayed payments by cause before sorting them by corridor. A payment that missed a Friday cut-off and one in a compliance hold look identical to the person waiting, but they need opposite fixes: a different rail or send time for the first, better data at onboarding for the second. I recommend logging the submit time and the status shown for every late payment.
What changes when the middle leg is a stablecoin
A stablecoin changes the middle of the payment. Instead of hopping across correspondent banks, value moves on a blockchain between two parties, and the chain does not close on weekends. The payment ends on a local rail, so the last leg is domestic.
Which delays a stablecoin leg removes, and which stay.
Delay | Does a stablecoin leg remove it? | What stays |
Correspondent hops | Yes, the middle leg has no intermediary bank | Nothing in the middle |
Time zones and weekends in the middle | Yes, the blockchain runs every day | The fiat legs keep their own calendars |
Pre-funded accounts in each country | Yes, nothing is parked in settlement accounts ahead of the payment | Someone still sources local liquidity |
Currency conversion | No | It happens at the on-ramp or the off-ramp, so it needs a priced quote |
Compliance review | No | KYC, KYB and monitoring still run |
Cut-offs on the fiat legs | No | ACH, FEDWIRE and SWIFT keep their deadlines |
Crediting the beneficiary | Depends on the local rail | Pix, SPEI and Faster Payments run around the clock |
So the fastest flows pair a stablecoin in the middle with a local rail that has no cut-off at the end. Over a wire the weekend comes back. The final leg is explained in what a stablecoin off-ramp is, and the settlement side in real-time cross-border settlement.
Conversion is the delay people forget. On a wire the rate is applied somewhere in the chain, often after the fact. On Lumx the rate is a step the customer can price first: a locked quote holds the rate for 30 seconds at no fee, or for one or five minutes with a fee that covers volatility, and an expired quote is rejected so the customer re-quotes. An off-ramp then moves through TRANSFERRING_STABLECOIN, TRADING and TRANSFERRING_FIAT to SUCCESS, with a webhook on each change.
When a stablecoin leg is the wrong fix
It is the wrong fix when the delay is not in the middle. A payment in a compliance hold stays in one with a stablecoin leg, and so does one funded by a transfer that misses its cut-off. If your late payments are mostly compliance or data problems, changing the rail changes little.
It is also the wrong fix when the payout ends in USD by SWIFT, because that payout keeps the rail's 1 to 5 business days whatever happens in the middle. The coverage page shows which rail serves each country, so check it before promising a faster route.
What the slow legs look like on a Lumx payout
For a company paying a supplier or a seller in Brazil or Mexico, the old chain becomes two legs you can plan around: funding, with the cut-off of the rail you fund on, and payout, with the cut-off of the rail you pay out on. Lumx is stablecoin payments infrastructure for businesses that move money between Latin America and the rest of the world: one API to collect, hold, convert, and pay out in BRL, MXN, COP, USD, EUR, and GBP or in USDC and USDT, over local rails such as PIX, SPEI, PSE, Bre-B, ACH, FEDWIRE, SEPA, and Faster Payments, with SWIFT and on-behalf-of payments and collections (POBO and COBO) in USD, EUR, and GBP, plus named virtual accounts, custodial wallets, and KYB/KYC built in.
On the funding side, the stablecoin is credited about five minutes after a Pix or SPEI deposit. A USDC payout to a Brazilian bank account goes out over Pix with no cut-off, and one to a Mexican account goes out over SPEI with none either, so the weekend wait disappears on the last leg (Banxico non-business days queue SPEI to the next business day). The pairs are on the USDC to BRL and USDC to MXN pages, and the route is described on global payments. A payout that needs a wire stays on the wire clock, USD only. The comparison with banking APIs is in stablecoin API versus a banking API, and the regional picture is in cross-border payments in Latin America.
Methodology and sources
All figures were read on November 1, 2026. The speed and cost figures come from the FSB's G20 Roadmap for Cross-border Payments: Consolidated progress report for 2025, published October 9, 2025, read in full from the report PDF. The Lumx statements come from its documentation on cut-off times, transactions, request for information, exchange rates and coverage, read the same day.
Verified on November 1, 2026. Operational context, not legal, tax, or investment advice.
Cover photo: Michael Jasmund on Unsplash.
Why are cross-border payments slow?
Because the money passes through several banks that each work to their own cut-offs, calendars, time zones and compliance queues, often against pre-funded accounts abroad. The message moves in minutes, but the credit waits at each stop, and the last bank is often the slowest. On Lumx, a payout over Pix, SPEI or Faster Payments has no cut-off and runs every day of the year, while a USD wire over SWIFT stays on its 1 to 5 business day target.
How long does a cross-border payment take?
From minutes to several business days. The FSB's 2025 report found more than half of large wholesale payments credited within an hour, but only 67.3% of retail payment services settling within one business day. On Lumx, the documented targets are instant for Pix, SPEI and Faster Payments, 1 to 2 business days for ACH, the same business day for FEDWIRE and 1 to 5 business days for a USD wire over SWIFT.
What are nostro and vostro accounts?
A nostro account is the account a bank holds in a foreign currency at a foreign bank. The same account, seen from the other bank, is a vostro account. It must be funded before a payment can be paid out of it, which is why correspondent banking needs pre-funded money in each currency. On Lumx, the pricing model is "transparent FX on locked quotes, no pre-funding", and SWIFT payouts are in USD with UETR and MT103 confirmations.
Why do compliance checks delay an international payment?
Each bank screens the payment against sanctions lists and its own rules, and a partial name match, a missing address or an unclear purpose can stop it until someone reviews it. On Lumx, a flagged transaction is held in an RFI, compliance asks by email or Slack for the relationship, the purpose and the expected outcome, and a transaction with no answer within 24 hours may be refunded to the sender.
Do cross-border payments stop on weekends?
Bank rails with cut-offs do. A request after the deadline, on a weekend or on a bank holiday is processed on the next business day of that rail's country. On Lumx, Pix, SPEI and Faster Payments run every day of the year with no cut-off, while ACH, FEDWIRE and SWIFT requests wait for the next business day.
