What should an institution evaluate when choosing settlement infrastructure?

Seven things, in rough order: whether the option covers your corridors and currencies; which settlement asset it uses (commercial bank money, central bank money, stablecoins, tokenized deposits); whether it is a network you join or a layer you integrate; where compliance is enforced; who holds the assets and under what protection; what the integration and the exit would cost; and how much concentration risk one provider represents. Most bad choices trace to skipping one of these.

Should an institution pick one settlement network or several?

Increasingly, several. Every live network is partial: each covers certain corridors, settles a specific asset, and admits certain members, and none of them interoperates with the others. Multi-homing is the practical answer, and the real question becomes how to manage several rails without multiplying integrations, policies, and reconciliation. That is the problem a settlement layer above the rails exists to solve.

What is the difference between joining a settlement network and integrating a settlement layer?

A network is a venue: you become a member, adopt its rulebook, and settle with other members in its asset, inside its scope. A layer sits above venues: you integrate once, and it routes each payment to whichever rail fits the corridor, counterparty, and policy. Networks give depth inside their walls; a layer gives breadth across them. Many institutions will end up with both.

Where should compliance sit in settlement infrastructure?

Today it usually sits around settlement: screening before, investigation after, in systems separate from the rail. The design question worth asking any provider is what happens when a payment cannot satisfy its conditions: is it caught before settlement, reversed after, or does it fail to settle at all? Enforcement inside settlement, where a non-compliant transfer simply does not settle, is the strictest of the three and the newest.

How do you compare costs across settlement options?

Look past the per-payment fee to the balance-sheet costs: pre-funding parked in nostro accounts or network wallets, FX spreads, reconciliation and exceptions staffing, and the delay cost of capital in transit. The Financial Stability Board's own progress data shows headline costs are sticky, and industry work puts transaction costs on wholesale cross-border flows above $120 billion a year before counting trapped liquidity.

Explainers

How to choose settlement infrastructure: the institutional buyer's guide

The seven decisions that determine which settlement infrastructure fits: corridors, settlement assets, membership model, compliance, custody, integration, and concentration risk.

Choosing settlement infrastructure comes down to seven decisions: which corridors you need covered, which settlement asset you are willing to hold, whether you join a network or integrate a layer, where compliance is enforced, who holds the assets and under what protection, what integrating and leaving would cost, and how much concentration risk one provider represents. Everything on the market, from SWIFT and its upgrades to stablecoin networks to bank consortium ledgers, is a different set of answers to those seven questions.

This guide walks through each decision and the questions that expose it. The library behind every section goes deeper; this page is the map.

1. Corridors and coverage

Start where the money actually goes. Every alternative to the correspondent chain is partial: instant payment interlinks cover specific country pairs, consortium networks settle a handful of currencies among members, stablecoin corridors are deep where local ramps are strong and thin elsewhere. List your top corridors by value and by pain, then ask each option what percentage it covers today, not on its roadmap. Corridor economics differ enough that the right answer is often different rails for different corridors, which is itself an architectural finding: no single venue will cover you.

2. The settlement asset

What actually changes hands? Commercial bank money on a consortium ledger, balances backed by central bank money, a regulated stablecoin, a tokenized deposit: each is a different claim on a different balance sheet with different risk. The asset decides the counterparty risk you hold, the regulatory treatment you face (Basel’s rules treat stablecoin exposures very differently from deposits), and who will transact with you. The choice between tokenized deposits and stablecoins is the sharpest version of this decision, and it is not one-or-the-other for an institution that can route both.

3. Network membership or neutral layer

A network is a venue you join: a rulebook, a membership process, depth inside the walls, and no connection to rival venues. A layer is something you integrate: one connection that routes across venues as they rise and fall. The tradeoff is real. Networks concentrate liquidity and standards among members; a layer preserves optionality and spans the map. The deciding question is whether you can predict, with confidence, which venue will win your corridors over the life of the integration. If you can, join it. If you cannot, the club-versus-layer analysis argues for the option that does not require you to pick winners.

4. Where compliance is enforced

Every option screens sanctions and enforces limits somewhere. The architectural question is where: before settlement in a separate system, after settlement in investigation and repair, or inside settlement itself, so that a transfer that cannot satisfy its conditions does not settle. Ask each provider what a policy violation does mechanically: does it raise an alert, trigger a reversal, or prevent the settlement event? The further from settlement the check sits, the more screening is repeated across the chain and the more exceptions you staff.

5. Custody and protection

Who holds the value in flight and at rest, and what protects it? Bank deposits carry deposit insurance to scheme limits; e-money sits under safeguarding rules with segregation but no insurance; stablecoin reserves back redemption at par but the holder’s claim is on the issuer. The safeguarding versus deposit insurance comparison lays out the map. For any provider, ask: if you failed tomorrow, what is my claim, on whom, and in what queue?

6. Integration surface and exit cost

The integration you buy is also the exit you will one day pay for. One API that abstracts many rails is faster to adopt and easier to leave than a bespoke connection to one venue; a membership with capital requirements and rulebook obligations binds tighter. Ask what a migration off the platform would involve, because the honest answer prices the lock-in. This is also where orchestration earns its place: routing logic you control beats routing logic embedded in a single provider’s product.

7. Concentration and counterparty risk

Finally, count what one provider represents: settlement asset issuer, network operator, custodian, and software vendor can be four companies or one. The stablecoin market itself carries concentration at the issuer level, and several networks are operated by the issuer of the asset they settle. None of this is disqualifying; all of it belongs in the risk register, with limits sized to it.

Where Frame fits

Frame is the settlement layer for global finance: one integration to orchestrate payments at scale across fiat rails, stablecoins, and tokenized deposits, with compliance enforced on every transaction and settlement in seconds instead of days.

Against the seven decisions, Frame is a specific set of answers: breadth across corridors by routing over whichever rail fits rather than operating one venue; rail-neutrality across settlement assets instead of a commitment to one; a layer you integrate rather than a club you join; compliance enforced inside settlement by Frame’s Rules Engine, so a transfer that cannot satisfy its policies does not settle; white-label operation, so the institution keeps its brand, its client relationships, and its point of record; and one integration surface across the venues, with verifiable evidence produced for every settled transaction. For banks and financial institutions, payment providers and processors, exchanges and trading venues, SaaS and ERP platforms, and enterprises, the buying question this page began with becomes an operating question: what policy should govern each payment, and which rail serves it best today.

See how a rail-neutral settlement layer works: the Frame Blueprint.

Common questions

What should an institution evaluate when choosing settlement infrastructure?: Seven things, in rough order: whether the option covers your corridors and currencies; which settlement asset it uses (commercial bank money, central bank money, stablecoins, tokenized deposits); whether it is a network you join or a layer you integrate; where compliance is enforced; who holds the assets and under what protection; what the integration and the exit would cost; and how much concentration risk one provider represents. Most bad choices trace to skipping one of these.
Should an institution pick one settlement network or several?: Increasingly, several. Every live network is partial: each covers certain corridors, settles a specific asset, and admits certain members, and none of them interoperates with the others. Multi-homing is the practical answer, and the real question becomes how to manage several rails without multiplying integrations, policies, and reconciliation. That is the problem a settlement layer above the rails exists to solve.
What is the difference between joining a settlement network and integrating a settlement layer?: A network is a venue: you become a member, adopt its rulebook, and settle with other members in its asset, inside its scope. A layer sits above venues: you integrate once, and it routes each payment to whichever rail fits the corridor, counterparty, and policy. Networks give depth inside their walls; a layer gives breadth across them. Many institutions will end up with both.
Where should compliance sit in settlement infrastructure?: Today it usually sits around settlement: screening before, investigation after, in systems separate from the rail. The design question worth asking any provider is what happens when a payment cannot satisfy its conditions: is it caught before settlement, reversed after, or does it fail to settle at all? Enforcement inside settlement, where a non-compliant transfer simply does not settle, is the strictest of the three and the newest.
How do you compare costs across settlement options?: Look past the per-payment fee to the balance-sheet costs: pre-funding parked in nostro accounts or network wallets, FX spreads, reconciliation and exceptions staffing, and the delay cost of capital in transit. The Financial Stability Board's own progress data shows headline costs are sticky, and industry work puts transaction costs on wholesale cross-border flows above $120 billion a year before counting trapped liquidity.

Sources

Last reviewed 2026-07-24