Fiat rails, classified by how the money actually moves

Ask which funding method is cheapest and the answer is unstable. Ask what route the money takes and the answer holds for years, because clearing systems change far more slowly than the products built on top of them.

Why the route is the useful classification

A funding page presents you with a list of names: bank transfer, card, some payment brand, peer-to-peer. Those names describe the interface, not the mechanism, and the mechanism is what has consequences.

Three things follow directly from the route and from nothing else:

• Where the cost sits. Routes that settle through a clearing system charge a visible fee. Routes that involve a currency conversion bury the cost in the rate. This is a property of the route, not a policy the platform chose.
• How fast it settles, and why. Business-hours dependency, weekend gaps and value-dating are properties of clearing systems. The exchange cannot make its bank clear faster.
• What has to fail for it to stop working. Each additional business on the route is an additional party that can independently decide it no longer wants this category of customer.

Once you see a method in terms of its route, its behaviour stops being surprising.

Route one — direct through the banking system

Money moves from your bank account to a bank account the platform controls, through the local clearing system or, for cross-border transfers, through a chain of correspondent banks.

Cost: normally a visible flat fee, which makes this the only one of the four where the published number is close to the whole story. Being flat, it is expensive as a share on small amounts and very cheap on large ones.

The exception is cross-border. A wire routed through correspondent banks can have deductions taken by intermediaries you never see and cannot enumerate in advance. The amount that arrives is then less than the amount you sent minus the fee you were quoted, and reconstructing why is difficult after the fact.

Speed: hours to business days, governed by the clearing system rather than by either institution's willingness. Instant local schemes settle in seconds; older batch systems do not settle at all outside business hours.

Failure mode: the platform's banking relationship ends. This is the most common single cause of a fiat rail disappearing, and it is usually a decision by the bank about a category of business rather than about the platform specifically. There is rarely notice.

The detail that catches people: most platforms require the sending account to be in the same name as the verified account. A transfer from a spouse's account, a joint account with a different primary name, or a business account is routinely rejected and returned — a process that can take days and sometimes costs a return fee.

Route two — the card networks

The transaction runs through card authorisation and settlement, with your issuing bank, the card network and the platform's acquirer all involved.

Cost: typically the highest of the four, and typically not shown as a fee. What you are quoted is a rate, and the cost is the distance between that rate and the mid-market rate at the same moment. A funding page can accurately say "no fee" while this route is the most expensive one on it.

Speed: near instant, which is the entire reason it exists.

Failure mode: your card issuer's policy, not the platform's. Issuers can and do decline this whole merchant category, and when yours does, the exchange cannot do anything about it. The transaction fails at your bank.

The detail that catches people: some issuers classify this category as a cash advance rather than a purchase. When they do, the cost changes character entirely — a cash advance fee applies, and interest usually accrues from the transaction date with no grace period. This is decided by your issuer and is worth one phone call before the first attempt rather than one statement afterwards.

Route three — a payment institution in the middle

A licensed payment provider handles the fiat leg; the exchange handles only the crypto leg. From your side it looks like part of the platform. Structurally it is a separate regulated business.

Cost: the provider's markup, usually inside the rate rather than shown as a fee.

Speed: minutes to hours.

Failure mode: this is the least stable of the four, and the reason is structural rather than any comment on the businesses involved. There is one more independent company on this route, with its own risk appetite, its own regulatory constraints and its own commercial priorities. Any of those changing ends the arrangement, and none of them have anything to do with you or with the exchange.

The detail that catches people: when something goes wrong mid-transaction, the question of who is responsible is genuinely ambiguous. The exchange will point at the provider and the provider will point at the exchange, and both may be describing their situation accurately. Knowing which business actually took your money — it is on your bank statement — is what lets you address the right one.

Route four — peer to peer

Fiat moves directly between two users' accounts by ordinary bank transfer. The platform never touches it; it holds the crypto in escrow and releases it when the seller confirms receipt.

Cost: usually appears as a spread rather than a fee. Sometimes the spread is negative — the rate is better than the market — and that is worth pausing over rather than celebrating.

Speed: as fast as the underlying bank transfer, plus however long the counterparty takes to respond.

Failure mode: it does not really have one in the sense the others do, which is why it is often the last route standing when the others have gone. Its risks are of a different kind.

The detail that catches people, and it is the serious one: the escrow protects the crypto side of the trade. It does not and cannot protect the fiat side. If the funds you receive turn out to have come from a fraud, the bank on the other end may reverse or freeze the transfer after the crypto has been released, and the account that received them can be frozen while it is investigated. You will have completed a technically flawless trade and still be dealing with your bank about where the money came from. Unusually favourable rates are worth treating as a reason for caution rather than an opportunity.

The property that separates the four

There is one characteristic that explains most of the otherwise puzzling behaviour of fiat rails, and it is almost never discussed on a funding page: how long the incoming payment can still be taken back.

Crypto transfers are final. Fiat transfers are not, and the window in which they can be reversed differs enormously by route. A card payment can be disputed by the cardholder for a period measured in months. A bank transfer through most clearing systems is difficult to reverse once settled, though not impossible where fraud is alleged. A payment-provider transaction inherits whatever the underlying instrument allows.

The exchange is on the wrong end of that asymmetry. It hands you an irreversible asset in exchange for a reversible payment, and the gap between the two is a real exposure that someone has to carry.

Which explains three things that otherwise look like obstruction:

• Post-deposit holds. Funds arriving by a reversible rail may be unable to leave for several days. This is the exchange holding the position open until its own exposure closes, not a liquidity problem.
• Why the most reversible rail is the most expensive. Card pricing carries the cost of dispute risk. That is not the whole markup, but it is a real component and it is why the fastest route is also the priciest.
• Why name matching is enforced so rigidly. A payment from an account that is not yours is the exact shape of a disputed transaction, so platforms refuse it categorically rather than case by case.

Reading a rail through this lens makes its rules predictable. Fast and reversible costs more and is held longer; slow and final costs less and clears sooner once it arrives.

When a rail is withdrawn while you are using it

Rails do not usually give notice, and the awkward case is a transfer already in flight when one closes.

What normally happens is that the money is returned, because the receiving institution rejects it and the funds retrace their route. That process is slower than the original transfer, sometimes considerably, and cross-border returns can lose money to intermediary fees on the way back.

What makes it worse than it needs to be is that the exchange often cannot tell you anything useful. If its partner has stopped accepting the transfer, the transaction is with institutions the exchange has no visibility into, and "we have not received it" is an accurate rather than evasive answer.

Two practical consequences. Keep your own record of the transfer — reference number, date, amount, sending and receiving details — because your bank can trace it and the exchange cannot. And do not send a second transfer because the first has not arrived. The failure mode where both eventually land, or both eventually return, is common and takes far longer to unpick than waiting would have.

Why we publish no country table

The obvious thing to publish alongside this is a matrix of countries against available methods. We are not going to, and the reason is that it would be wrong in the direction that costs readers money.

Country availability is the fastest-moving data any exchange holds. It changes when a partner bank revises its risk policy, when a provider ends a relationship, when a rule changes in one jurisdiction — and none of those arrive with an announcement. Within weeks of publication some rows would be stale, with nothing to indicate which ones. A reader acting on a stale row attempts a rail that no longer exists, or plans around a cost structure that has changed.

What holds is the route-level description above, and the method for finding out what applies to you. The rail lookup covers the same four routes in a form you can query directly.

Finding out what is actually available to you

1. Check inside a verified account. Available rails depend on your verification level and registered region, so public help pages show a superset of what you can use. The deposit screen after verification is authoritative; nothing else is.
2. Check the withdrawal side at the same time. The two sets are not always symmetrical, and a rail that funded your account may not be able to empty it. Learning this after you have a balance is the expensive order to learn it in.
3. Run a small round trip. One small deposit and one small withdrawal give you the real cost, the real settlement time and proof that both directions work. No documentation provides any of the three.
4. Identify a second route before you need one. Given how these rails fail — abruptly, because of a decision made by a business you have no relationship with — having a fallback identified is worth more than optimising the fee on your primary one.

If you are setting up an account from scratch, the order matters: verification gates which rails you can even see, which is why our walkthrough completes verification before funding rather than after.