Why cross-border payments are still slow — and what's changing.

Guide · 6 min read

The correspondent banking bottleneck

Despite decades of upgrades — SWIFT gpi, ISO 20022, instant domestic schemes — the legacy model for moving money between countries is still structurally slow for high-frequency, high-value business flows. The reason is architectural: a traditional wire passes through a sending bank, one or more correspondent banks, and a beneficiary bank, with each hop adding fees, an FX spread, and a settlement window.

Two forces compound the problem. Cross-border processing carries high fixed costs, which pushes activity through a concentrated set of intermediaries; and the number of correspondent banking relationships has been declining for years as institutions de-risk. Fewer relationships means longer routing chains and more corridors where settlement is measured in days rather than minutes.

The performance gap, in numbers

Industry analysis through 2026 puts traditional correspondent-bank settlement at roughly three to five business days, with all-in costs (fees plus FX) in the low single-digit percentages. Blockchain-settled stablecoin rails, by contrast, can reach finality in seconds to minutes at a fraction of the cost — which is why a growing slice of business payouts has migrated onto them.

The Financial Stability Board has noted that the value moved across the SWIFT network every few days approaches the size of world GDP, yet the underlying rails still miss the G20's own targets for speed and cost on a meaningful share of business corridors. The gap is not theoretical; it is felt every time a payout sits in transit over a weekend.

What's actually changing

Three shifts are reshaping the landscape at once. First, fintech infrastructure layers sit on top of banking networks and use their APIs to add pre-funding, automatic rerouting, real-time status, and built-in compliance handling — solving the visibility and predictability problems without replacing the underlying rails. Second, stablecoin settlement has graduated from a crypto-native curiosity into a substantial cross-border rail, with institutional adoption and regulatory frameworks now in place in both the US and EU. Third, the incumbent networks are adapting: SWIFT has moved a shared-ledger design into active development to enable interoperability between banks' tokenised deposits, and instant-payment schemes have announced cross-border ambitions.

Where stablecoins fit — and where they don't

Stablecoins have meaningfully displaced legacy rails in a narrow but expanding band: small-to-mid-value payouts into corridors where licensed local off-ramps exist. They have not replaced correspondent banking for large bank-to-bank settlement, regulated trade finance, or flows that depend on documentary credit. The realistic picture is hybrid — the right rail depends on the corridor, the value, the counterparties, and the regulatory route available in each market.

The real challenge: orchestration, not a single rail

For a platform launching cross-border products, the decisive problem is not picking one rail. It is orchestrating across many: choosing the right route per corridor, enforcing eligibility and compliance before value moves, keeping a clean ledger, and reconciling provider confirmations back to a single source of truth. That coordination layer — eligibility, compliance, routing, ledger, reconciliation — is exactly where most of the operational risk lives.

This is the layer Axora provides. Rather than integrating every bank, scheme, and regulated route directly, platforms connect once and route across regulated fiat partners and, where authorised, regulated digital-asset infrastructure — with compliance and reconciliation handled as part of the workflow rather than bolted on afterward.