⚡ Instant payments around the world

What an instant rail really is

An instant payment rail is defined by three properties that must all hold, and only three. The service runs 24 hours a day, 7 days a week, 365 days a year. Funds reach the payee within seconds. The payment is irrevocable once accepted. Everything else is an overlay built on top of the rail, not part of the rail itself: QR codes, phone-number addressing, request to pay, recurring mandates. The line between the two layers determines which infrastructure a provider joins, how risk is split among the parties, and how much development work to plan for.

The overlay is often the brand the public knows, while the underlying rail stays invisible. In Australia, customers see Osko, never NPP. In Sweden, they see Swish, not RIX-INST. In Spain, they see Bizum, not SNCE. In Brazil, they see Pix, while the infrastructure is called SPI and the directory DICT. Providers join SPI; the product the public sees is called Pix. Confusing the two layers means talking to the wrong counterpart, signing the wrong contract, and working to the wrong timeline.

The global landscape: how the rails compare

An instant rail’s weight can be measured in two different ways: the number of transactions it processes and the value they carry. By transaction count, instant payments are now the leading retail rail in several of the world’s largest markets. The Prime Time for Real-Time report from ACI Worldwide and GlobalData counted 266.2 billion real-time transactions in 2023, up 42.2% year over year. It projected 511.7 billion by 2027. Two systems account for most of that volume: UPI in India and Pix in Brazil.

Behind this top tier comes a second wave of fast-growing rails. For a business that needs to collect payments locally, they often matter more than the market leaders. They include InstaPay in the Philippines (2018, PPMI/BancNet), NAPAS 247 in Vietnam (2016; 8.9 billion transactions in 2024, according to NAPAS), and Raast in Pakistan (2021, State Bank of Pakistan). Next come sarie in Saudi Arabia (2021; 750 million transactions worth US$621 billion in 2025), Aani in the UAE (2023, Al Etihad Payments), Fawran in Qatar (2024, Qatar Central Bank), and PayShap in South Africa (2023, PayInc). Then Bre-B in Colombia (2025, Banco de la República), SINPE Móvil in Costa Rica (2015, BCCR), Transfer365 in El Salvador (2020, BCR), and Bakong in Cambodia (2020, National Bank of Cambodia). The list ends with GhIPSS Instant Pay in Ghana (2016), KWiK in Angola (2022), and eKash in Rwanda (2025, built on Mojaloop).

These figures call for one methodological caveat. Volume and value measure two different things. In the US, RTP claims about 98% of interbank instant payments by transaction count. FedNow processed $853.4 billion in 2025 with an average payment of $101,435, or about 40% of total value. RTP therefore carries mostly high-volume, low-value retail payments, while FedNow carries infrequent, large treasury transfers. Ranking the rails on either measure alone produces an order that the other measure reverses.

Three governance models, three distributions of power

Governance of an instant rail means who owns the infrastructure and who has the power to write its rules. It determines who sets prices, who controls access, and how fast the system evolves. Three models dominate, each with concrete consequences for a PSP or merchant looking to connect.

Opening up to nonbank institutions is the most active governance issue of the moment. In the euro area, EU Regulation 2024/886 requires payment institutions and e-money institutions to receive and send instant credit transfers by April 9, 2027. In Canada, the RPAA (Retail Payment Activities Act) framework opened Interac e-Transfer to nonbank PSPs. In Japan, the Zengin System, in operation since 1973, has been open to funds transfer operators since 2022. In Hong Kong, e-money issuers have participated in FPS from the start, which in practice made Hong Kong’s wallets interoperable with one another.

Who pays what: the real economics of an instant rail

The economics of an instant rail come down to three lines: the unit cost charged by the infrastructure, the price the PSP charges its customer, and the value of the services built on the rail. One structural feature sets them apart from card economics. Instant payments carry no interchange: no fee flows from the payee’s PSP to the payer’s PSP. As a result, acceptance costs a merchant structurally less than card acceptance, and an issuer earns structurally less.

“Free” describes the price the user pays. The cost of production doesn’t go away; it shifts to other parties. India is the clearest example. Zero MDR on UPI generated enormous volume, yet every transaction consumes real bank resources. NPCI responded with UPI Lite, an on-device wallet inside the app that is debited without a PIN below a low limit. Its main purpose is to take micropayments off bank servers. Funding comes from the government budget: the incentive pays industry participants 0.15% of value on merchant payments under ₹2,000 at small merchants, and nothing above that.

For merchant acceptance, the cost gap with cards remains the instant rail’s main selling point. That shortcut is wrong, because the price the merchant pays is set by its PSP, not by the rail. In Brazil, the central bank requires Pix to be free for consumers but does not cap what business payees are charged; that is left to competition among acquirers and payment institutions. A business case that pits “free Pix” against “cards at 2%” compares a central bank rule with a negotiated market price, two figures at different points in the acceptance chain. Such a business case is flawed by design.

What makes a rail take off

A rail takes off when it moves from marginal to everyday use in a country’s retail payments. Forty years of attempts point to a consistent set of factors that separate widely adopted rails from those that remain niche. None of them is technical: all come down to decisions on governance, pricing, and distribution.

• The participation mandate. No voluntary rail has achieved universal coverage. Pix requires participation above 500,000 accounts, Tanzania for all transactions between providers, and the European Union by regulation. Greece went further still: through tax legislation (2024–2025), it requires merchants and professionals to accept IRIS, the only case in Europe where the obligation falls on the merchant rather than the PSP.
• Alias addressing. Nobody reads out an IBAN or an 18-digit CLABE. Mexico learned this the hard way: DiMo was created in 2023 specifically to add a phone-number alias overlay to SPEI and fix the failure of CoDi.
• Free for the payer. PromptPay is free below a limit, Pix is free for consumers, UPI has a zero MDR by law, and Raast and Transfer365 are free. No successful rail charges consumer payers.
• A single standardized QR code. Bank Indonesia’s QR mandate (QRIS, 2019) ended wallet fragmentation: 50.50 million users and 32.71 million merchants enrolled, with transaction growth of 226.5% year over year (Bank Indonesia, 2024 data). DuitNow QR applies the same logic in Malaysia.
• A killer use case that builds the habit. P2P in Nigeria and Sweden, in-store merchant payments in Thailand and Brazil, G2P transfers in Pakistan, corporate treasury for RTP and FedNow. A rail without an everyday use case never becomes a habit.

The corollary applies to markets that rely on organic adoption, with no participation mandate and no pricing rules. Convenience alone does not change a payment habit, as PayShap shows: in South Africa, it spent three years growing slowly, held back by bank pricing. The rail then went from about 14 million transactions a month over its first 33 months to about 89 million a month over the first five months of 2026 (ClearingPost / PayInc, 2026). That inflection did not coincide with any technical change to the rail.

What fails, and why

An instant rail project counts as a failure when it is shut down before reaching meaningful coverage, or when it remains a minority player in its own market for the long term. The cases below come down to three causes: no mandate, competition from an entrenched incumbent, and governance that cannot make decisions. None of these projects failed for technical reasons.

One last cause of failure lies in market structure rather than in the rail itself. The rail exists, but the country’s payments economy was built around a different instrument. In Ghana, GhIPSS Instant Pay has operated since 2016 but structurally carries less weight than mobile money interoperability, in a country built around the mobile wallet rather than the bank account. In Kenya, PesaLink was launched in response to M-Pesa and never reversed the trend. An interbank rail assumes that the population has bank accounts. Where account ownership remains low, the real issue is interoperability between mobile wallets and banks, not the choice of an instant rail.

Irrevocability, fraud, and liability

Irrevocability means that neither the payer nor the payer’s provider can cancel a payment once it has been accepted. An instant payment is irrevocable by design: there is no equivalent of the card chargeback, no dispute window, and no unilateral reversal. On every such rail, from day one, this shifts the target of fraud from counterfeiting the instrument to manipulating the payer, who authorizes the payment. The industry calls this authorized push payment (APP) fraud.

• Real-time sanctions screening: EU Regulation 2024/886 replaces transaction-by-transaction screening with a daily check of the customer base, precisely because per-transaction screening cannot fit into a few seconds. Check the rules that apply in each jurisdiction before sizing your screening process.
• Configurable limits and time-of-day restrictions: limits are set by the scheme in some cases and by each bank in others. Aani caps transfers at AED 50,000; RTP and FedNow cap them at $10 million; Cotra in Japan caps them at ¥100,000 per transaction. These limits are the leading cause of rejections in production.
• Detection on the payee side: under a shared-liability regime like the UK’s, the fraudster’s bank pays half. That makes fighting money mule accounts a P&L issue, not just a compliance one.
• A recovery mechanism is not a guarantee: Brazil’s MED recovers only funds that are still in the account. Marketing “protection” without stating its limits invites complaints.

Accepting payments on an instant rail: a practitioner’s checklist

Acceptance-side integration of an instant rail covers the technical and contractual connections a merchant needs to collect payments on it. It differs structurally from a card integration in several ways: there is no authorization and capture, no D+1 clearing cycle, and no liability framework standardized by an international network. The points below must be settled before signing an acceptance contract in a given market.

Reconciliation is matching the payments received in the bank account against the sales the merchant has recorded. It is the most commonly underestimated cost. The rail credits funds continuously, 24/7, including nights and weekends, with no end-of-day cutoff to close the books against. Matching data therefore has to become a continuous feed rather than daily files. Native ISO 20022 rails make this easier by carrying structured references. Branded overlays make it harder, because the identifier the customer sees (a Pix key, a PayID, a CliQ alias) differs from the one in the interbank message.

Cross-border interconnection: the next frontier

Cross-border interconnection means linking two national instant rails directly, so a payment can flow from one to the other without going through a correspondent bank. All of these rails are domestic by design: they settle in one currency, under one jurisdiction, with a local directory. Since 2021, work has focused on connecting them. Two approaches coexist. Bilateral links are live but do not scale (n × n agreements). Multilateral schemes reduce each system to a single connection but are not yet in production.

Nexus is a multilateral interconnection scheme designed at the BIS Innovation Hub and later transferred to Nexus Global Payments (NGP). This Singapore-incorporated nonprofit was set up on March 26, 2025, by the central banks of India, Malaysia, the Philippines, Singapore, and Thailand, later joined by Bank Indonesia. Instead of bilateral agreements, each instant payment system connects only once to the network to reach all the others. The stated goal is a cross-border payment in under 60 seconds. The blueprint was completed in July 2024, but nothing has gone into production since, and legal and technical milestones remain.

Three distinct arrangements are routinely sold under the same label, especially around UPI. The first is acceptance of a foreign QR code by a traveler: UPI has been accepted in Bhutan since 2021, in the UAE via NeoPay, in Singapore, Nepal, Sri Lanka, and Mauritius, and in Qatar via QNB. The second is interconnection of two national systems (UPI–PayNow). The third is export of the software stack, with NPCI International selling the architecture to other countries. Peru’s central bank, the Banco Central de Reserva del Perú, signed such an agreement with NPCI International to build a public retail payments platform. The three arrangements involve different contracts, flows, and risk exposures.