MAS and ABS put QR interoperability, AI commerce and larger government payments into PayNow Generation 2 while scams still cost S$913.1 million
By: Christian Brooks – SeaPRwire – PayNow turned nine in 2026 having processed 752.4 million transactions worth S$30.05 billion, with second-half growth of 26.3% year on year and roughly 11 million proxy registrations on the system. Those are the numbers the Monetary Authority of Singapore and the Association of Banks in Singapore chose to lead with, and they deserve the attention. They are also, in a specific and measurable way, beside the point. The same PayNow Generation 2 Phase 1 report that the two institutions published on 25 June 2026 discloses that cross-border payments moving over PayNow amounted to S$371 million, or 0.12% of the total value of domestic and cross-border PayNow payments combined. That share was 0.12% in the prior period as well. It did not move.
A national payment rail that has succeeded at home and barely registers abroad is the defining characteristic of Singapore’s position in 2026, and it is not a failure. It is the arithmetic consequence of a deliberate set of choices, several of which are defensible and at least one of which the regulator has stated openly in Parliament. What follows is an attempt to describe the domestic machine accurately, to show where the cross-border ceiling actually comes from, and to separate the parts of PayNow Generation 2 that will change the number from the parts that will not.
The domestic rail is genuinely excellent
Start with what works, because the domestic story is the strongest in the region and the data is unusually good. MAS publishes semi-annual retail payment statistics with enough granularity to see the shape of the shift, and the shape is unambiguous.
| Instrument | 2021 | 2025 | Direction |
| FAST transactions, millions | 226 | 642 | Up 2.8x |
| FAST value, S$ millions | 296,315 | 797,746 | Up 2.7x |
| FAST average ticket, S$ | 1,323 | 1,242 | Down 6.1% |
| Card payments, millions | 1,982 | 2,528 | Up 27.6% |
| Card average ticket, S$ | 65 | 61 | Down 6.2% |
| SGD cheque clearing, millions | 24 | 5.5 | Down 77% |
| ATM cash withdrawals, millions | 158 | 139 | Down 12% |
| Inter-bank GIRO, millions | 123 | 123 | Flat |
Read the average ticket rows carefully, because they contain the most interesting signal in the table. FAST volumes nearly tripled while the average ticket fell 6.1% to S$1,242. Card volumes grew 27.6% while the average card ticket fell to S$61. E-money average tickets sit between S$1.09 and S$1.27 depending on the period. All three move in the same direction: more transactions, each one smaller. That is what a fast payment rail is supposed to do to a payments market, and Singapore has done it more completely than almost anywhere.
The cheque row is the cleanest illustration. SGD cheque clearing collapsed from 24 million items to 5.5 million, a decline of roughly 77%, and the residual is almost entirely institutional rather than consumer. GIRO, the 1984-vintage system that MAS cites as the beginning of the country’s e-payment history, has gone nowhere at 123 million items in both observation periods. It is the rare case of a legacy rail that is neither dying nor growing, and it is a useful reminder that migration in payments is rarely complete.
Cheques are worth one more sentence. They are the instrument that tells you whether a market has finished a transition, and Singapore has not finished it. A 77% decline is enormous and five years of a flat GIRO line is the counterweight: incumbency in payment rails is sticky in exact proportion to the number of back-office processes built on top of it.
Where the cross-border ceiling comes from
The 0.12% figure needs context before it means anything, and the context is that the ceiling is partly structural and partly a matter of stated policy.
The structural half is simple. PayNow’s cross-border average ticket was S$371 in the latest period against S$286 previously, and both are large relative to the S$132 to S$154 consumer average ticket and the S$109 to S$147 corporate average ticket. Cross-border PayNow is a remittance and high-value transfer instrument, not a point-of-sale one, and the corridors that carry it carry it in small daily amounts. The PayNow and DuitNow link to Malaysia, promised for the fourth quarter of 2022, actually arrived in two stages: QR interoperability on 31 March 2023 and account-to-account connectivity on 17 November 2023, roughly a year late, with transaction limits on the three corridors that have stayed in the S$1,000-per-day range for years. A rail with a S$1,000 daily ceiling will not produce meaningful value share no matter how good the underlying plumbing is.
| Corridor | Promised | Actual | Gap |
| PayNow to DuitNow, QR | Q4 2022 | 2023-03-31 | About 1 quarter |
| PayNow to DuitNow, account-to-account | Q4 2022 | 2023-11-17 | About 1 year |
| Thailand to Malaysia, phase 2 | Q4 2021 | Last week of January 2022 | Minor |
| Thailand to Malaysia, phase 3 | Q4 2022 | No go-live confirmation | Open |
The policy half is more interesting, because MAS has stated it. In a written Parliamentary reply of 5 March 2026 on SGQR, the authority said it does not target full QR interoperability, on the grounds that doing so would weaken competition and affect the service and cost outcomes for merchants and consumers. That is a real position with a real rationale, and it deserves to be quoted rather than paraphrased away. It also means the 0.12% number is not primarily a technical shortfall. It is the expected output of a regulator who has decided that a single interoperable QR layer is worse for the market than several competing ones.
The consequence is visible in the domestic landscape too. Singapore’s own two national QR schemes do not interoperate with each other; the PayNow Generation 2 report says so directly, and the pilot that would connect PayNow to the NETS QR network is scheduled for the end of 2026. Fourteen payment providers are usable at NETS merchants, and seven country QR schemes are accepted for cross-border payments at those merchants. That is a functioning arrangement, and it is also a fragmented one by design.
What PayNow Generation 2 actually contains
The Generation 2 study published on 25 June 2026 was the product of consultation with 37 institutions and benchmarking against 11 jurisdictions. It set out four enhancement directions, and they are worth listing because the distinction between them is where the realistic expectations live.
- PayNow and NETS QR interoperability, with a pilot targeted for the end of 2026. This is the item that would remove the most visible domestic inconsistency, and it is also the one with the most institutional friction behind it.
- Deep-linking for PayNow QR codes to improve online checkout. Technically modest, commercially significant, and the least controversial item on the list.
- A sandbox for larger-value government PayNow transactions, starting in 2027. This is the quiet one that could matter most, because government disbursement and collection is the single largest untapped pool of value in the system.
- Expanded payment capabilities, including support for AI-driven commercial activity, which is the phrase that has attracted the most commentary and the least specificity.
Only the third item has an obvious path to moving the cross-border share, and it does so indirectly. Larger government flows create a domestic value base that makes bilateral corridor agreements more attractive to counterparties, but they do not by themselves create a corridor.
The multilateral option and its missing date
Project Nexus is the BIS-backed answer to the corridor-by-corridor problem, and its status in 2026 is the clearest example of a pattern that runs through all of Singapore’s cross-border work. Nexus Global Payments was established in 2025 to take the project forward, BIS handed the platform over to its participants in October 2024, and the official launch year has not been published. India and four ASEAN central banks have signed on. The blueprint is sound. The date is absent.
The same pattern produced the PayNow and DuitNow timetable above, and it is worth stating plainly: in cross-border payments, Singapore publishes architecture years before it publishes availability. Anyone building a business plan on the architecture rather than the availability will be wrong for longer than their runway.
There is a global ceiling on all of this that is not Singapore’s fault and not within its control. BIS and the Financial Stability Board report that only 35% of cross-border retail payments settle within one hour against a G20 target of 75%; the retail figure sits at 33.5% and the wholesale figure at 50.6%, and both declined year on year. Around 65% of payment messages still contain unstructured addresses, and roughly 65 market infrastructures have no alignment plan for the November 2026 milestone, according to SWIFT’s own disclosures. Singapore can build the best rail in the world and still hand its customers to a counterparty that has not finished migrating.
The number that grew fastest is the one nobody wanted
While the cross-border share sat flat, scam losses moved. Singapore Police Force data for 2025 records 37,308 scam reports and S$913.1 million in losses. Both fell: reports down 27.6% year on year and losses down 17.9%, the first double decline since 2021. That is a genuine achievement and it should be recorded as one.
The composition is where the concern sits.
| Scam category | Losses, S$ millions | Note |
| Investment scams | 336.2 | Largest single category |
| Government official impersonation | 242.9 | Cases roughly doubled |
| Government official impersonation, cases | 1,504 | Doubled year on year |
| Total reported losses | 913.1 | Down 17.9% year on year |
Government official impersonation doubled in case count while total reports fell more than a quarter. A scam type that grows while the aggregate shrinks is a scam type that has found a working method, and the method here is specific: it exploits the trust that a well-run state accrues. That is an uncomfortable externality of the same institutional quality that makes Singapore’s payment system work, and there is no technical fix for it inside the rail.
What SGQR+ showed, and what it did not
The SGQR+ proof of concept is the most encouraging piece of evidence in the 2026 record, and it is also small. Transaction volume grew 105% and transaction value grew 212% during the pilot. About 75% of participating merchants said they would continue, and 82% of financial institutions said they would support expansion.
Those are strong numbers from a small base, and the honest reading is that they establish viability rather than impact. The MAS position quoted earlier still governs what happens next. A proof of concept that most participants want to continue is not the same thing as a mandate, and the authority has said it does not intend to issue one.
The retail mix, instrument by instrument
The aggregate numbers hide a set of instrument-level shifts that matter more to anyone actually operating in the market. MAS’s own caveat should be attached to all of them: the authority states that its retail payment statistics are sample-based industry statistics, that sample sizes and aggregation methods differ across instruments, and that cross-instrument comparisons may therefore not be conclusive. That is an unusually candid disclaimer and it should be repeated whenever these numbers are quoted side by side.
| Instrument | Latest volume | Average ticket | Trend |
| Card payments, total | 2,528 million | S$61 | Growing, ticket shrinking |
| POS credit and charge | 726 million | Not separately published | Growing slowly |
| POS debit | 960 million | Not separately published | Growing faster than credit |
| E-money | Not published | S$1.09 to S$1.27 | Stable, very small ticket |
| ATM cash withdrawals | 139 million | S$51,692 million total | Declining |
| Inter-bank GIRO | 123 million | Not separately published | Flat |
Debit is growing faster than credit at the point of sale, which is consistent with a market where account-to-account rails have taken the low-value, high-frequency segment outright. E-money average tickets of around one dollar describe stored-value transit and small retail behaviour rather than a wallet economy competing with cards. And ATM withdrawals of 139 million items against S$51,692 million in value imply an average withdrawal above S$370, which is what you would expect once the small daily cash withdrawer has migrated to a QR code.
The instrument mix also explains why the cross-border share looks so small. Everything that has grown fast in Singapore is small-ticket and domestic. Everything that is cross-border is large-ticket and structurally capped. Those two facts are not in tension, and reading them as if they were is the most common error in commentary on this market.
What the November 2026 milestone actually requires
The ISO 20022 migration is the least visible item on the 2026 agenda and the one with the widest blast radius. MT and MX coexistence ended on 22 November 2025, which means the two-phase timetable is now the operative one rather than the single November 2026 date that circulated in earlier coverage. SWIFT’s own published position is that roughly 65% of payment messages still contain unstructured addresses and that about 65 market infrastructures have no plan to align by the November 2026 milestone.
For a Singapore-based institution, the practical consequence is counterparty risk rather than domestic compliance. A Singapore bank can finish its own migration perfectly and still find that a receiving institution in a non-aligned market rejects, repairs or manually handles the message. Repair rates are the operational metric that will show where the pain lands, and they will be concentrated on exactly the corridors where the 0.12% figure already suggests volume is thin.
How the region’s corridors actually compare
Singapore’s cross-border figure looks better and worse depending on which comparison you pick, and both comparisons are worth making.
| Market | Cross-border QR volume | Value | Note |
| ASEAN total, 2025 | 36.2 million | US$716 million | P2P average ticket US$191 |
| Malaysia, H1 2025 | 11.8 million | RM967 million | Exceeded full-year 2024 |
| Indonesia QRIS cross-border | Not published | Not published | Growth of about 225% |
| Singapore PayNow cross-border | Not published | S$371 million | 0.12% of PayNow value |
ASEAN moved 36.2 million cross-border QR transactions worth about US$716 million in 2025, with an average peer-to-peer ticket of US$191. Malaysia’s first half alone, at 11.8 million transactions worth RM967 million, exceeded its full-year 2024 total. Indonesia’s QRIS cross-border usage grew roughly 225%. Against that regional picture, Singapore is not behind on capability; it is behind on volume, and the reason is that Singapore’s corridors carry higher-value transfers with lower daily limits rather than the tourist-scale QR payments that drive the ASEAN totals.
The cost comparison is the one that should worry policymakers. The World Bank’s Remittance Prices Worldwide data puts the global average cost of sending remittances at 6.36% in the third quarter of 2025, with banks at 14.99% and still the most expensive channel type. Every corridor that ASEAN has opened competes against that number, and the corridors that will win are the ones that beat it on price rather than the ones that open first.
A working checklist for treasury teams
Anyone deciding whether to route through PayNow’s cross-border links in 2026 should work through the following in order, because the answers change the decision.
- Check the daily limit before the launch announcement. Singapore’s bilateral corridors have sat near S$1,000 per day for years. A corridor capped at S$1,000 cannot carry a payroll or a supplier payment.
- Separate QR interoperability from account-to-account connectivity. These arrived a year apart on the Malaysia corridor and they are not substitutes. QR solves the tourist and small retail case; account-to-account solves the remittance case.
- Model the repair rate, not the go-live date. With roughly 65% of messages still carrying unstructured addresses, the cost sits in exception handling on the receiving side.
- Price the alternative. Stablecoin settlement via the card-network platforms now covers more than 200 countries. For corridors where no bilateral agreement exists, that is the live comparison, and it is often cheaper at the high-value end.
- Read the MAS position on interoperability as settled. Full QR interoperability is not coming, and planning that assumes otherwise will not survive contact with the regulator.
The stablecoin and card-network flank
One reason the 0.12% figure may matter less in three years than it does today is that the competitive set has changed underneath it. Mastercard completed its acquisition of BVNK on 3 August 2026, bringing a stablecoin infrastructure business with roughly US$30 billion in annualized payment volume and coverage of more than 200 countries inside a card network. The two launched a Global Money Platform on 19 October 2026 that lets a single integration settle in both fiat and stablecoin. Stripe’s 2025 annual letter reported US$1.9 trillion in platform volume, up 34%, with stablecoin payment volume roughly doubling to about US$400 billion.
None of that runs on PayNow, and all of it competes for the same cross-border flows. If the multilateral rail route stays slow, the market will route around it through networks that were never designed as national infrastructure, and the 0.12% share will look less like a ceiling and more like the share of a market that has partly migrated elsewhere.
Judge the corridor by its limit, not its launch
The operational conclusion for anyone building on Singapore’s rails in 2026 is to read the fine print before the press release. A corridor announcement tells you that two institutions have agreed to connect. The transaction limit tells you what the corridor will actually carry, and the limit is the number that has stayed flat across every Singapore corridor to date.
Three things would change the picture, and only one of them is in Singapore’s hands. Publishing a firm launch date for Nexus would convert an architecture into a schedule, and MAS and BIS could do that unilaterally. Raising the daily limits on the existing bilateral corridors would move more value in a quarter than any new agreement, and it requires only the two counterparties. And closing the government sandbox into production in 2027 would put the largest single pool of domestic value onto the rail, which is the precondition for every other item on the list. Until at least two of those happen, the domestic numbers will keep compounding at 26% and the cross-border share will stay where it is.
Sources
- MAS, Semi-Annual Retail Payment Statistics, H2 2025 — https://www.mas.gov.sg/-/media/mas-media-library/statistics/payment-statistics/semi-annual-retail-payment-statistics/retail-payment-statistics-h2-2025.pdf
- MAS, payment statistics portal — https://www.mas.gov.sg/statistics/payment-statistics/semi-annual-retail-payment-statistics
- MAS and ABS, PayNow Generation 2 Phase 1 Report, 2026-06-25 — https://www.mas.gov.sg/-/media/PNG2-Phase-1-Collaterals_PDDC/PayNow-Gen2-Phase-1-Report_25-June-2026.pdf
- MAS, written Parliamentary reply on SGQR, 2026-03-05 — https://www.mas.gov.sg/news/parliamentary-replies/2026/written-reply-to-parliamentary-question-on-sgqr
- MAS and ABS, PayNow Generation 2 media release, 2026-06-25 — https://www.mas.gov.sg/news/media-releases/2026/mas-and-abs-chart-instant-payments-enhancements-with-paynow-generation-2-study
- MAS, e-payments development page — https://www.mas.gov.sg/development/e-payments
- MAS, PayNow and DuitNow linkage media release, 2021-09-27 — https://www.mas.gov.sg/news/media-releases/2021/singapore-paynow-and-malaysia-duitnow-to-link-in-2022
- Bank Negara Malaysia, PayNow and DuitNow linkage page — https://www.bnm.gov.my/-/mysg-pmt-en
- Singapore Police Force, Scams and Cybercrime Briefs — https://www.scamshield.gov.sg/resources/scams-and-cybercrime-briefs/
- BIS and FSB, cross-border payments targets — https://www.fsb.org/uploads/P211024-3.pdf
- BIS, cross-border payments report 2025 — https://www.bis.org/publ/biscbt2025.htm
- SWIFT, ISO 20022 removal of unstructured addresses — https://www.swift.com/standards/iso-20022/iso-20022-faqs/iso-20022-removal-unstructured-address
- SWIFT, ISO 20022 news, 2025-11-22 — https://www.swift.com/insights/news/iso-20022-new-era-global-payments
- Reserve Bank of India, Project Nexus press release — https://rbi.org.in/scripts/FS_PressRelease.aspx?prid=58197
- BIS, Project Nexus — https://www.bis.org/about/bisih/topics/fmis/nexus.htm
- World Bank, Remittance Prices Worldwide Issue 54 — https://remittanceprices.worldbank.org/sites/default/files/2026-04/RPW_main_report_and_annex_Q325.pdf
- Mastercard, completion of BVNK acquisition, 2026-08-03 — https://bvnk.com/press/mastercard-completes-acquisition-of-bvnk-to-advance-global-stablecoin-capabilities
- Stripe, 2025 annual letter — https://stripe.com/annual-letter/2025
- BIS CPMI, red book statistics — https://data.bis.org/topics/CPMI_CT
- IMF, ASEAN digital payment revolution selected issues paper, 2026-02-20 — https://www.imf.org/en/publications/selected-issues-papers/issues/2026/02/20/aseans-digital-payment-revolution-a-new-frontier-for-regional-integration-thailand-574175