For most of financial history, sending money across borders has felt like mailing a letter through a postal system designed in the 1950s — slow, expensive, and surprisingly opaque. You hand your money to a bank, it disappears into a chain of intermediaries, and somewhere between two and five days later it reappears in another country, diminished by fees that nobody fully explained. The miracle isn’t that this cross-border payments system has worked for so long. The miracle is that we’ve tolerated it. That tolerance is running out.
A convergence of technological innovation, regulatory ambition, and genuine competitive pressure is forcing a reckoning with the way the world moves money. What’s emerging isn’t a single replacement for the old correspondent banking model but a plurality of new rails, protocols, and institutions that together resemble less a system than an ecosystem. Understanding what’s driving this shift requires looking not just at the technology, but at the incentives and power dynamics underneath it.
The Correspondent Banking Problem
The traditional cross-border payment model relies on a network of bilateral relationships between banks. Your bank knows a bank in Frankfurt; that bank knows one in Singapore; that one maintains an account at a local institution in Jakarta. Each link in this chain adds time, cost, and points of potential failure. Compliance checks multiply. Liquidity gets trapped in pre-funded accounts. Fees accumulate with little transparency for the end customer.
The G20 has put hard numbers on the ambition to fix this: retail cross-border payments costing under 1% and completing in under one hour by 2027. Today, the global average cost for sending remittances hovers around 6.5%, and in some corridors, particularly those serving lower-income countries, it remains in double digits. That gap between aspiration and reality represents hundreds of billions of dollars extracted annually from people who can least afford it.
New Architectures, New Politics
The response has come from multiple directions simultaneously, and not all of them are aligned.
On one end of the spectrum sits ISO 20022, a messaging standard that may sound mundane but carries real significance. By enabling payments to travel with richer, more structured data like full counterparty details, payment purpose, reconciliation references, etc., it dramatically reduces the manual intervention and false compliance positives that slow transactions down. Swift’s migration to ISO 20022 is effectively a modernization of the plumbing that most of global finance still depends on.
On the other end sits something more politically charged: CIPS, China’s Cross-border Interbank Payment System. Launched in 2015 and expanded substantially since, CIPS provides an alternative clearing and settlement infrastructure that operates outside the Western-dominated Swift network. It processes yuan-denominated transactions between a growing number of participant banks globally. CIPS isn’t yet a peer competitor to Swift in volume, but its existence signals something important — that the architecture of global payments is increasingly contested geopolitical terrain, not neutral infrastructure. As more bilateral trade is settled in non-dollar currencies, CIPS offers an alternative corridor that some countries find strategically attractive.
Between these two poles, a quieter but potentially more disruptive transformation is underway: the linking of domestic real-time payment systems. India’s UPI and Singapore’s PayNow are already interconnected. Brazil’s Pix, the EU’s SEPA Instant, Thailand’s PromptPay… All these are being woven together through bilateral agreements and multilateral frameworks. The vision is a mesh of fast payment networks that allows a payer in São Paulo to reach a recipient in Mumbai at low cost, in seconds, without touching the correspondent banking chain at all. This isn’t science fiction; partial versions of it already exist.
The Digital Money Question
The emergence of stablecoins and central bank digital currencies (CBDCs) adds another dimension entirely — one that’s simultaneously more promising and more uncertain than the infrastructure conversation.
Stablecoins like USDC have demonstrated something important: you can move dollar-denominated value across borders in minutes, at minimal cost, on infrastructure that runs around the clock. For businesses operating in multiple currencies, this is genuinely transformative. The question isn’t whether stablecoins work technically — it’s whether regulators will allow them to scale, and under what conditions. The EU’s MiCA framework and similar efforts in other jurisdictions are beginning to answer that, but the picture is still forming.
CBDCs represent the state’s version of the same idea. Projects like mBridge, which links central banks in China, the UAE, Thailand, and Saudi Arabia, are exploring whether wholesale digital currencies can replace some of the pre-funded liquidity that currently makes international settlement so capital-intensive. The technical results have been encouraging. The governance questions, e.g. who controls the rules, who gets access, what data gets shared — are considerably harder.
AI as Infrastructure
Cutting across all of these developments is the role of artificial intelligence, which is becoming less of a feature and more of a requirement. The compliance burden associated with cross-border payments: sanctions screening, AML monitoring, fraud detection, and more, has grown faster than the transaction volumes it’s meant to protect. Legacy rule-based systems generate enormous numbers of false positives, creating delays and costs that contradict every efficiency gain the industry is trying to make.
Machine learning is changing this calculus. Real-time risk scoring, behavioral anomaly detection, adaptive fraud models — applied thoughtfully, these tools can dramatically reduce the friction associated with compliance without reducing its effectiveness. For payment corridors into markets with weaker banking infrastructure, AI-assisted identity verification is also beginning to extend access to people who previously couldn’t participate in the formal financial system at all.
The Inclusion Imperative
That last point deserves more weight than it typically receives in the infrastructure debate. Cross-border payment reform isn’t just a story about corporate efficiency or geopolitical competition. Remittances to low- and middle-income countries represent roughly $800 billion annually. It’s more than foreign direct investment and more than official development aid. Every percentage point shaved off the cost of those transfers has direct, measurable impact on household incomes in some of the world’s most economically vulnerable communities.
The future of cross-border payments, then, isn’t only a question of which technology wins or which standard dominates. It’s a question of who gets to participate in the global economy and on what terms.
The old geography of money, defined by correspondent relationships and dollar dominance, is being redrawn. What replaces it will depend on choices being made right now: by central banks piloting digital currencies, by regulators drafting stablecoin frameworks, by engineers designing message formats, and by policymakers deciding whether financial inclusion belongs at the center of the agenda or at its margins.
The plumbing is being rebuilt. The question is what values get built into it.
