perspective

The Unlock

Cross-border payments have three constraints: cost, speed, and capital lock-up. Stablecoin transit addressed the first two. Multilateral clearing addresses the third.

Published

The unlock is not stablecoins alone. It is the combination of minutes level settlement and multilateral netting that compresses both payment time and capital requirements.

Reader Brief

Cross-border payments have three constraints: cost, speed, and capital lock-up. Stablecoin transit addressed the first two. Multilateral clearing addresses the third. The unlock is the combination that compresses settlement to minutes and capital requirements by 25x.

What's Inside

Four moves that connect the three locks, the 25x capital multiplier, the capture question, and the strongest counter-arguments.

The perspective starts by separating settlement time, fee cost, and capital efficiency. It then explains why stablecoin transit alone is not enough, why multilateral netting changes institutional treasury economics, and why the dividend is likely to split across bank-led and licensed-operator architectures.

Individual improvements existed since 2017. SWIFT gpi cut time. Bilateral stablecoin cut fees. But capital stays trapped until multilateral netting enters. The combination is the unlock.

At ~96% netting efficiency, $100 of gross flows compresses to $4 of actual capital movement. Combined with minutes-level settlement, this changes treasury economics at institutional scale.

Both sides are building. Historical pattern suggests stratification: wholesale via bank tokens, retail and SME flows via operator clearing. The boundary is transaction size and geography.

Honest projection: 3-5x in year one, 10-15x at 20+ participants. Bank platforms face three structural limits: bank-only membership, permissioned ledgers, and no emerging-market last mile.

The Three Locks

Cost, speed, and capital lock-up are distinct constraints. A rail that solves only one leaves the system structurally expensive.

$10 trillion in prefunded nostro accounts, 6.49% average remittance costs, and 1-5 day settlement times are three locks that have constrained cross-border payments for decades. For the first time, a single architecture can unlock all three simultaneously. Individual improvements have existed since 2017 with SWIFT gpi. The combination has not until now.

The Unlock Has Three Separate Locks


Time, fee, and capital drag each require a different mechanism before the full dividend appears.
  1. key 1

    Speed lock

    Stablecoin transit and gpi compress settlement wait.

    payment moves faster
  2. key 2

    Fee lock

    Direct routing and local liquidity reduce visible costs.

    payment gets cheaper
  3. key 3

    Capital lock

    Multilateral netting compresses gross flow into net positions.

    idle capital releases
common mistakeSpeed without netting still leaves the largest pool trapped.
3
Locks: settlement time, fee cost, and capital efficiency.

Each lock needs its own unlock; solving only one leaves a structural drag behind.

25x
Capital efficiency multiplier at mature multilateral netting [1].

The source frames the multiplier as a CLS-benchmark ceiling, not a launch-state promise.

Lock 1: Settlement Time

Stablecoin transit and SWIFT gpi both compress the time lock where the relevant infrastructure exists.

Correspondent banking settles in 1-5 days. For decades, this was the cross-border norm. SWIFT gpi compressed this to minutes for participating corridors starting in 2017. Stablecoin transit compressed this to minutes for any corridor with a licensed operator on both ends. The settlement time lock is effectively broken where infrastructure exists.

Under correspondent banking, a UK-to-Kenya payment routes through 3-4 intermediary banks, each with its own cut-off times and compliance checks. Actual settlement can take 2-5 business days. A UK-to-US payment through Tier 1 banks can be same day. Settlement time was a property of the relationship chain: which banks, how many hops, and which time zones.

SWIFT gpi compressed this for participating corridors: the source cites 50% of gpi payments credited within 30 minutes. But coverage is uneven. High-volume G7 corridors got fast; thin emerging-market corridors stayed slow. Stablecoin transit removed the variance: USDT on Tron confirms in ~60 seconds whether the corridor is London-New York or Nairobi-Lagos. Settlement time became a rail property, not a relationship property.

Lock 2: Fee Cost

Stablecoin corridors can compress visible payment fees, but fee compression does not automatically unlock trapped capital.

Global average consumer remittance cost is 6.49%. The G20 target is below 3%. The gap has been narrowing slowly. Stablecoin-based B2B corridors now operate at 1.6-4% infrastructure cost with ~20% operator margin, delivering consumer-facing prices below the G20 target in functional corridors. That opens the fee lock, but it does not solve the full cost stack by itself.

  • Consumer remittance: the source cites a 6.49% World Bank average, including fees and FX spread paid by individual senders. Sub-Saharan Africa averages 7.9%; South Asia 4.3%. The G20 3% target has been missed for 15 consecutive years, and the gap narrows by ~0.3% per year.

  • B2B infrastructure cost: the source frames emerging-market corridors at 1.6-4%, covering on-ramp/off-ramp fees, compliance processing, and FX conversion. This is the operator cost floor; below it, a corridor is unprofitable regardless of volume. In mature corridors such as US-Mexico and UK-India, infrastructure costs compress to 0.5-1.5%.

  • Cost of capital: the source uses 12-18% annual emerging-market cost of capital. A $1M prefunded nostro position costs $120-180K per year in opportunity cost with 3-5 day settlement lockup. For institutional flows above $100K, capital cost exceeds transaction fees. It is invisible in consumer data but dominant in the total cost stack.

The fee lock is partially open where licensed operators have built functional corridors. It remains fully locked in thin emerging-market-to-emerging-market corridors, approximately 40% of cross-border payment routes.

Lock 3: Capital Efficiency

The largest lock requires more than faster settlement. It requires netting gross obligations into small net positions.

An estimated $10 trillion sits in prefunded nostro accounts globally to support correspondent settlement. Every dollar of it is idle in the source framing. This is the largest and least-discussed lock in cross-border payments. Solving it requires not just faster settlement, but also multilateral netting. This is the lock that has remained mostly closed until now.

The architecture replaces three traditional capital drains: prefunded corridor balances, Herstatt risk buffers, and bilateral operator float. Minutes-level settlement reduces prefunding, multilateral netting compresses gross flows to net obligations, and atomic PvP settlement removes the need to hold capital against one leg failing while the other clears.

Stablecoin transit alone compresses capital lock-up per transaction but does not eliminate it. An operator still needs float on both sides of every corridor. Multiply that across 20 corridors and the float requirements add up.

Multilateral netting through a shared clearing layer compresses gross flows into net obligations. At CLS-level efficiency of ~96%, for every $100 of gross volume, only $4 of actual capital movement is required. Combined with minutes-level settlement, this is the unlock that changes the economics at scale.

Why All Three Matter

Speed, fees, and netting compound. A partial unlock still leaves the largest capital pool trapped.

SWIFT gpi unlocked settlement time in 2017 but left $10 trillion in prefunding untouched. Bilateral stablecoin transit cut fees to 1-3% but still requires per-corridor float. Only multilateral clearing plus stablecoin transit plus minutes settlement unlocks all three, compressing capital requirements by 25x while cutting fees below 1%.

ConfigurationSettlement timeFee costCapital efficiencyResult

Correspondent banking

1-5 days

3-8%

Low

Legacy state

SWIFT gpi

Minutes for participating corridors

0.5-2%

Low

Time unlocked, capital still trapped

Bilateral stablecoin

Minutes

1-3%

Medium

Time and fee unlocked, capital partially unlocked

Multilateral stablecoin clearing

Minutes

0.3-1%

High through netting

All three unlocks combined

Fast settlement without netting: a bank with $500M in nostro positions saves on per-transaction time but still needs the $500M prefunded. The ROI on faster settlement alone does not justify infrastructure migration because the capital is still trapped.

Netting without fast settlement: multilateral netting compresses gross to net, but if final settlement still takes 2-3 days on traditional rails, the netting cycle must accommodate that delay. Capital is freed per cycle, not continuously, limiting the dividend to ~50% of theoretical.

Both at once: $500M in nostro compresses to ~$20M in net positions and settles in minutes instead of days. The freed $480M earns 4-5% annually, or $19-24M per year in recovered yield per institution. At this scale, treasury strategy changes. When Tier 1 treasury strategy changes, market structure follows within 3-5 years.

Who Captures The Unlock

The dividend can stay inside banks, move to licensed operators, or split across layers.

JPMorgan's Kinexys, Partior (DBS plus JPMorgan plus Standard Chartered), and BIS Agora bet on banks capturing the dividend through tokenized deposits. Licensed operators building on commercial stablecoins bet on a different outcome. The likely end state is stratification: wholesale via bank tokens, retail and SME flows via operator clearing networks.

Projects like JPMorgan's Kinexys, Partior, BIS Agora, and mBridge consolidate the unlock inside the regulated banking system. Banks' existing nostro capital gets redeployed more efficiently. Their balance sheets improve. Their role in cross-border payments is preserved or strengthened.

This is the incumbent-wins-the-unlock scenario.

Licensed operators in emerging-market corridors already handle significant cross-border volume through bilateral stablecoin settlement. Their advantage is that they have built on-ramp/off-ramp infrastructure in markets that banks abandoned when exiting correspondent relationships. The source cites the CBR exodus as hitting 34% of African banking links.

If these operators connect through shared multilateral clearing, they capture the capital efficiency dividend previously available only to Tier 1 banks. A network of 20 licensed operators across 15 corridors could achieve 10-15x netting efficiency, enough to compete with bank-led platforms on capital cost while serving markets those platforms cannot reach.

The constraint: regulatory frameworks must tolerate non-bank operators at settlement scale. GENIUS Act, MiCA, and UAE PTSR all create pathways. But in jurisdictions without clear frameworks, operators remain bilateral, and bilateral operators cannot access the netting dividend.

Every major payment infrastructure shift has produced stratification, not replacement. CLS handles $6.6T per day in FX settlement alongside Fedwire's $4.7T. Visa processes 700M+ daily transactions alongside ACH batch settlement. SWIFT routes 45M+ messages per day alongside banks' internal systems. None replaced the other; each found its layer.

The likely stratification: wholesale settlement, meaning $1M+ between Tier 1 institutions, uses tokenized deposits or CBDCs. Kinexys, Partior, and Agora are building this layer. Retail and SME flows, meaning $100 to $100K, use licensed operator clearing on commercial stablecoins: faster to deploy, lower integration bar, broader emerging-market coverage.

The boundary is not just transaction size. It is also geography. Bank-led platforms dominate G7-to-G7 corridors where relationships are strong. Operator-led clearing dominates emerging-market corridors where banks have retreated. Both layers benefit from the unlock; neither captures it entirely.

The Transition Window

The unlock is technically available now, but adoption depends on regulation, integration cost, and network depth.

The unlock is technically available now. The economics favor it. But transition is slow because the institutions with the most to gain are the ones with the highest operational friction to adopt. This is why the market has not already moved decisively.

Three frictions slow adoption:

  1. Regulatory clarity: institutions need supervisory comfort before committing treasury to new infrastructure. Frameworks like GENIUS Act, MiCA, and UAE PTSR reduce this friction but do not eliminate it in every jurisdiction.

  2. Integration cost: joining a clearing network is a multi-quarter project. Legal, technical, operational, and compliance lift all matter.

  3. Network effects: the dividend compounds with participant count. Early joiners invest before the network is deep enough to deliver full value.

The institutions building now are those that see the 3-5 year horizon and prefer being in the first cohort. Late adopters will get the same dividend at lower integration cost, but lose competitive positioning during the transition.

Counter-Arguments & Limitations

The strongest objections challenge the maturity of the 25x benchmark and the role of commercial stablecoins.

Every perspective has boundaries. The 25x capital multiplier assumes CLS-level netting efficiency, which took CLS a decade to reach with the world's largest FX banks. Stablecoin clearing networks are starting from zero participants.

The argument: CLS achieves ~96% netting efficiency because it has 70+ direct participants and thousands of third-party members generating massive bilateral offset opportunities. A new stablecoin clearing network with 5-10 members will see netting efficiency of 30-50% at best. The 25x figure is aspirational, not operational.

Valid. Early-stage netting efficiency will be far lower. The honest projection is 3-5x in year one, scaling toward 10-15x with 20+ participants, and approaching CLS-level only at 50+ participants with diversified corridor coverage. The 25x is the theoretical ceiling, not the launch capability. We use it to show what the architecture enables at maturity, not what it delivers on day one.

The argument: JPMorgan's Kinexys processes $2B+ per day in intrabank transfers. Partior connects three of Asia's largest banks. BIS Agora is exploring wholesale tokenized settlement. If banks build their own multilateral clearing on tokenized deposits, the capital efficiency unlock happens inside the banking system with no role for commercial stablecoins or non-bank operators.

Partially valid. Bank-led platforms will capture some of the wholesale dividend. But they face three structural limitations: (1) they require all participants to be banks, excluding the licensed non-bank operators that dominate emerging-market corridors; (2) they operate on permissioned ledgers with limited interoperability; (3) they have not solved the last-mile on-ramp/off-ramp in markets where banking infrastructure is thin. The more likely outcome is stratification: banks use their platforms for wholesale, operators use stablecoin clearing for retail and SME flows. Both layers benefit from the unlock.

Anton Titov

Author of The Unlock. Building a stablecoin clearing network, solving interoperability between licensed financial institutions across stablecoins, chains, and jurisdictions. He focuses on connecting payment infrastructure between emerging and developed markets. Speaker at Money20/20 Asia 2025, Stablecoin Summit Africa (Johannesburg, 2025), Stablecoin & Blockchain Conference Kenya (2026), and Fintech Week Central Europe (2026).

About This Perspective

Scope, disclosure, and method.

Author
Anton Titov, CEO, Plexo
Published by
Plexo Institute
Data vintage
2023-2025

Plexo builds a Stablecoin Clearing Network and marketplace for licensed financial institutions - structuring multi-party cross-border settlement, compliance packaging, and liquidity coordination across complex corridors. Plexo Institute publishes our perspective on the market we operate in. This analysis reflects the vantage point of an infrastructure builder, not a neutral observer.

Capital efficiency benchmarks come from CLS Bank operational data and BIS CPMI multilateral platform analysis. Nostro prefunding estimates come from McKinsey Global Payments Report (2025). Cost layer taxonomy comes from World Bank Remittance Prices Worldwide, McKinsey B2B payments research, and operator disclosures. The piece takes a position that the three unlocks only transform market structure in combination; counter-views that any single unlock is sufficient are acknowledged but not treated as the base case.

Continue Reading

Where the capital is trapped - $10T prefunded capital and how multilateral netting frees it.

The settlement architecture - fiat in, stablecoin transit, fiat out in minutes.

The clearing layer - matching, netting, and compliance coordination.

The stratified outcome - six competing architectures and which layer each wins.

References

BIS CPMI, Exploring Multilateral Platforms (2023); CLS Bank operational data

SWIFT, gpi Performance Data

World Bank, Remittance Prices Worldwide (March 2025)

McKinsey, Global Payments Report (2025)

Three-locks capital unlock diagram showing speed lock, fee lock, and capital lock, with stablecoin transit opening speed and fee while multilateral netting opens idle capital release.
Stablecoin transit opens speed and fee locks; multilateral clearing is needed for the capital lock.

Evidence And Sources

This raw HTML export preserves source visibility for crawler and contractor review. Indexing decision: index, follow.

  1. Exploring Multilateral Platforms for Cross-Border Payments - BIS CPMI / CLS Bank operational data
  2. SWIFT gpi Performance Data - SWIFT
  3. Remittance Prices Worldwide - World Bank
  4. Global Payments Report - McKinsey

Internal Graph