Legacy vs. Tokenised

Why This Approach

Legacy vs. tokenised — the honest trade-off

What tokenised settlement actually buys you versus today's rails, what it costs to build, and where it isn't worth it yet.

New to the wording on this page? 7 terms explainedShow

Correspondent banking

The traditional way cross-border payments work: banks hold accounts with each other, and a payment hops through a chain of these relationships to reach its destination. Each hop adds time, cost and a place it can stall.

For example: Paying a supplier in Brazil from Hong Kong might route through a US bank, then a regional bank, then the supplier's bank — three hops, three sets of fees, and three places where a compliance check can hold it for days.

RTGSReal-Time Gross Settlement

Central-bank systems that settle large payments one at a time, immediately and irreversibly, using central-bank money. 'Gross' means each payment settles individually rather than being bundled and netted off.

For example: CHAPS in the United Kingdom is how the money moves when you buy a house — a single large payment settled on its own, same day, and impossible to reverse once done.

Netting

Adding up everything two parties owe each other and moving only the difference, instead of settling every transaction separately.

For example: Over a day, Bank A owes Bank B US$100m across many trades, and Bank B owes Bank A US$97m. Rather than moving US$197m in total, they move US$3m once. Far less cash is needed.

ISO 20022

The modern international standard for the messages banks send each other about payments. It carries far more structured detail than the older format it replaces — including what a payment is actually for.

For example: The old format might carry a cramped free-text line like 'INV 4471 PYMT'. The new one carries separate, properly labelled fields for invoice number, purpose code and full party addresses — so a computer can read it rather than a human guessing.

Cut-off time

The daily deadline after which a payment will not be processed until the next business day.

For example: A treasurer realises at 6pm Hong Kong time that the Singapore office needs cash. The payment window closed at 5pm, so the money cannot arrive until Monday — leaving Singapore short all weekend while the cash sits unused in Hong Kong.

Nostro account

An account one bank holds at another bank, usually abroad and in that country's currency, so it can make payments there. Latin for 'ours' — our money, held at your bank.

For example: For HSBC to pay someone in Brazilian reais, it keeps a pot of reais in an account at a Brazilian bank. That cash sits there earning little, purely so payments can be made — money that is parked rather than working.

Settlement

The moment value actually changes hands and the deal is genuinely done — not just agreed, promised, or sitting as pending. Before settlement, someone is still owed something.

For example: You tap your card at 9am and the shop says 'approved'. That is authorisation, not settlement. The money may only actually leave your account two days later — that later moment is settlement.

See every term used across this site

Side-by-side comparison

DimensionLegacy railsTokenised settlement
AvailabilityBound by cut-off times and banking hours per corridor; weekend and holiday gaps24/7 in supported corridors, no cut-offs
Settlement speedSame-day to T+2 depending on corridor and rail (SWIFT, RTGS, correspondent chain)Seconds — atomic settlement on-ledger
VisibilityStatus often opaque mid-transit through correspondent chains; SWIFT gpi improved but didn't solve this fullyReal-time, per-transaction status on a shared ledger
Reconciliation effortManual or batch reconciliation across multiple parties' independent recordsContinuous automated reconciliation against a shared source of truth
Settlement riskSequential legs create a window of exposure (e.g. Herstatt risk in FX)Atomic DvP/PvP removes the window — both legs or neither
Legal & regulatory maturityDecades of case law, established finality, well-understood by every counterpartyGenuinely new — legal finality has to be established corridor by corridor
InteroperabilityUniversal — every bank already speaks SWIFTImmature — requires counterparties on compatible infrastructure or a bridge
Operational complexityWell-worn playbooks, mature tooling, deep institutional muscle memoryNew failure modes (key management, smart-contract bugs, network outages) without decades of institutional experience to draw on
ProgrammabilityLargely manual; conditional/automated payment logic is bolted on, not nativeNative — conditional and automated settlement logic is part of the platform

Advantages

Removes principal and settlement risk

Atomic DvP/PvP makes the classic 'I paid and they didn't deliver' failure mode structurally impossible, not just less likely.

Collapses reconciliation effort

A shared, continuously-reconciled ledger removes most of the manual matching work that legacy multi-party settlement requires.

Removes the cut-off constraint entirely

Not 'faster within business hours' — genuinely always-on, which changes what's possible for corporate treasury, not just how fast today's process runs.

Programmable settlement

Conditional logic (release funds only when X is true) becomes a platform capability instead of a manual workaround or a bespoke integration project.

Disadvantages

Legal finality isn't settled yet

Every corridor needs its own legal opinion before production use — this isn't a technology problem, and it can't be engineered away.

Real integration cost

Core banking, treasury systems, screening and reconciliation all need new integration points — this isn't a drop-in replacement for existing rails.

Requires counterparty readiness

The value compounds with adoption — a single bank's tokenised ledger is far less useful until clients and counterparties are actually on it too.

New operational risk surface

Key management, smart-contract logic and network resilience are new failure modes without decades of institutional playbook behind them yet.

Genuinely new skills required

Engineering, risk, legal and operations teams all need capability they may not have today — this is a change-management cost, not just a build cost.

Cost

What it actually takes to build

Using this case study's own delivery backlog as a real effort proxy, in story points — not a made-up number.

Effort by area

137 points total across 7 epics — see the full backlog for story-level detail

Client onboarding & entitlements

New KYC-to-entitlement workflow, role-based approval limits

18 pts

Token issuance, transfer, redemption

Core ledger lifecycle, 24/7 availability engineering

31 pts

Ledger integration & reconciliation

Real-time core-banking posting, continuous reconciliation

19 pts

Financial-crime & wallet controls

Pre-settlement screening, wallet risk scoring, AML monitoring tuned for tokenised movement

21 pts

Operations, exceptions & reporting

New exception-handling and incident playbooks for a system with no institutional history

18 pts

Pilot rollout & client support

New commercial and support model — nothing to reuse from the legacy playbook

11 pts

Commercialisation & go-to-market

Pricing, RM enablement, client segmentation for a genuinely new proposition

19 pts

See the full delivery backlog for every story and acceptance criterion behind these totals.

The real question

Is it worth it?

Not a yes/no answer — it depends on the client, the corridor and the counterparty. That's exactly why the roadmap is phased rather than a single big-bang launch.

High-volume, multi-entity corporate treasury clients with real cut-off pain

Worth it — this is exactly where 24/7 movement and reconciliation savings show up fastest and most visibly.

A single, occasional cross-border payment need

Probably not worth it yet — legacy rails already handle this adequately, and the integration cost isn't justified by the volume.

A trade that requires DvP/PvP with an external counterparty on incompatible infrastructure

Not worth it until the counterparty is reachable — the value of atomic settlement depends on both sides actually being on compatible rails.

A corridor where legal finality hasn't been established yet

Not worth it for production — this is a hard blocker, not a trade-off to weigh, regardless of how compelling the technology case is.