Singapore has the foundations to lead tokenised capital markets in Asia, and four building blocks would let it do so
Capital markets have begun to move settlement from electronic book entries to tokenised records, the largest change to their infrastructure this decade. The institutions that move first are setting the standards and the architecture that others will adopt.
Singapore could set the regional standard, export the model and become the settlement and liquidity hub for tokenised markets in ASEAN.
We propose a framework of four building blocks, each a prerequisite for the one above it and each designed for production use:
1. Digital money
2. Tokenised risk-free collateral
3. Digital issuance and registry infrastructure
4. Trading, clearing and settlement infrastructure
Every part of the framework is designed for the whole market. In a market of Singapore's size, separate systems would divide a limited pool of liquidity, so only shared infrastructure can hold that liquidity together.
Digital money and tokenised collateral depend on decisions by the Monetary Authority of Singapore (MAS) and the Ministry of Finance (MOF). Only the state can issue central bank money and Singapore Government Securities (SGS).
Legal and regulatory refinements will also be needed, and we leave them to separate work.
Context
Tokenised settlement is moving into production abroad, and Singapore would gain at three levels by building it now
Capital markets are replacing electronic book entries with tokenised records, and Singapore can lead that change in its region
Capital markets have changed the format of their records once before. Their move from paper certificates to electronic book entries built the infrastructure that has run them for about 40 years.
That infrastructure now limits markets in two ways.
1. Each country runs its own depository and settlement systems. A cross-border holding therefore passes through several intermediaries, each keeping its own record.
2. These systems settle the securities leg and the cash leg separately. They cannot settle atomically, meaning both legs in one indivisible transaction, or move collateral in real time.
Tokenisation issues and records assets directly on programmable ledgers that connected institutions share. An asset can then move between institutions and across borders on a single record, and both legs of a trade can settle in one transaction.
Singapore can define the regional standards and build the region's first production infrastructure before another jurisdiction sets the default.
Market infrastructures abroad are building tokenised settlement now, and their designs point to one architecture
Central banks and market infrastructures in Switzerland, the US and the euro area have committed to tokenised settlement.
The Swiss National Bank (SNB) has run a wholesale CBDC1 pilot in a live production environment since December 2023. The pilot will continue until at least June 2027.2
In March 2026 the US Securities and Exchange Commission (SEC) approved Nasdaq's dual-track framework, under which a security can trade in tokenised or traditional form. Trading in tokenised form takes place within a pilot run by the Depository Trust Company (DTC), part of the Depository Trust & Clearing Corporation (DTCC).3 The approval was the first of its kind at a major exchange.
On 21 September 2026 the European Central Bank (ECB) launched Pontes, which settles wholesale transactions in tokenised assets in central bank money through its TARGET Services. Pontes is the near-term track of a two-track plan that the ECB approved in 2025.4
Together, these initiatives point to one architecture with three parts: (a) government securities provide the risk-free collateral; (b) central bank money in digital form settles the cash leg; and (c) permissioned networks, open only to vetted institutions, keep the records.
Each part corresponds to a building block in the framework proposed here.
Canton Network (Canton) has the strongest institutional backing of any permissioned network built for settlement.
DTCC and Euroclear co-chair the Canton Foundation, which governs the network.5
Investors in Digital Asset, the developer of Canton, include Goldman Sachs, Citadel Securities, DTCC, Nasdaq, S&P Global, BNY, HSBC and BNP Paribas.6
On present evidence, Canton is the most defensible choice of permissioned network for institutional settlement.
Singapore could gain at three levels, from its home market to global capital, and each gain needs infrastructure built now
1. Depth in the local market would grow. Singapore Exchange (SGX) and The Central Depository (CDP) would keep full control of listings, pricing, trading rules and market structure. Only the settlement layer would change. The gains from that change would come early and could be measured.
Shorter settlement and real-time collateral mobility would let institutions reuse capital sooner, reduce the margin that clearing members post and free collateral that is locked today.
Each gain would arrive as soon as its own component works. The first improvement would therefore come before the full framework is in place.
2. Regional liquidity would come within reach. On shared infrastructure, issuers and investors across ASEAN could use cross-border issuance, pooled collateral and common order flow for the first time. The commercial case for these has long existed, and the missing piece has been the infrastructure.
In the near term, a regional company could raise capital from investors in several markets through one tokenised issuance. A Singapore institution could also post tokenised SGS as collateral on a cross-border position without a separate custody arrangement.
Further out, ASEAN markets could pool collateral, net margin across borders and use Singapore as the regional clearing venue. Each step would need commercial and regulatory alignment among the markets involved. The proposed infrastructure would remove the technical barrier that has blocked these steps until now.
3. Global investors would gain simpler access. International institutions could reach SGX-listed and other ASEAN securities through the DTCC or Euroclear infrastructure they already use for other markets. That access would remove most of the operational work that makes ASEAN a difficult allocation today.
DTCC's partnership with Canton offers the near-term route.5 SGX-listed tokenised securities issued on Canton would sit on the same network as DTCC, the world's largest settlement infrastructure, with direct access to international liquidity.
Cross-listing with US and European venues would also become simpler. One tokenised share class would settle in one system, and the two registers would need no reconciliation.
The Framework
Four building blocks would make up the framework, with digital money at the base
Tokenised markets need digital money to settle, so the two must be designed and built together
When a tokenised asset settles against cash paid outside the ledger, the trade carries today's settlement risk in a new format.
Cash and assets must sit on the same ledger before institutions can recycle capital at once or settle trades by delivery versus payment (DvP) in code.7 The ledger then removes settlement risk by moving both legs in one transaction. Today that protection depends on a central counterparty (CCP) taking on each trade through novation.
Each design decision in the asset components must take account of how the money component will work where the two meet.
Both components can and must be built at the same time, provided their designers agree the design and the interfaces between them at the start.
Singapore's market is too small to carry several incompatible forms of digital money. Separate proofs of concept on such instruments would add to that fragmentation.
Four building blocks would form a complete tokenised capital market, and each requires the one beneath it
All four blocks can and should be built at the same time. The dependency order decides only where design decisions and interfaces sit.
Each block must be ready for production when the block above needs it. The order also shows where design choices are made and where a failure would spread.
Building Block 1, digital money, would combine three instruments: wholesale CBDC in SGD, tokenised deposits and regulated stablecoins.
These instruments would put cash on the ledger, which atomic settlement requires.
Building Block 2, tokenised risk-free collateral, would issue and settle SGS on the ledger to give the market a sovereign risk-free asset.
A market without sovereign collateral on the ledger would have no risk-free asset and could not clear atomically.
Building Block 3, digital issuance and registry infrastructure, would be the ledger on Canton where tokenised securities are issued, recorded and transferred.
It would replace separate national registries with one programmable register designed for regulated institutions.
Building Block 4, trading, clearing and settlement infrastructure, would be the market infrastructure through which members trade tokenised securities and settle them by DvP7 against digital money.
Atomic settlement on a shared ledger would be the end state. The first step would be intraday settlement in batches, with cycles shortening step by step towards real time.
Shorter cycles would reduce risk, free capital sooner and let collateral move in real time at every stage of that path.
Building Blocks 1 & 2
MAS, the Ministry of Finance and the banks would provide digital money and tokenised SGS, the two blocks that the market cannot build alone
Digital money in Singapore would need two tiers, one for sovereign settlement and one for commercial settlement
Tier 1 would provide sovereign and near-sovereign settlement, through wholesale CBDC issued by MAS and regulated SGD stablecoins.
Wholesale CBDC would be the risk-free settlement asset, because it is a direct liability of MAS and settles with finality.
Regulated SGD stablecoins could add a near-sovereign option once licensed issuers operate. MAS finalised its stablecoin framework in August 2023 and published draft legislation for consultation on 1 September 2026.8
Tier 2 would provide the working cash, in tokenised deposits issued by DBS, OCBC and UOB on the same ledger as the securities they settle.
Investors and their brokers settle in commercial bank money today. Tokenised deposits would put that same money on the ledger.
Tokenised deposits would preserve the two-tier monetary system, the credit that banks extend and the intraday liquidity that high-volume settlement needs.
Together, the two tiers would reproduce today's monetary system on a ledger, with central bank money as the anchor and commercial bank money as the working instrument.
Banks can close the digital money gap now, using the licences they already hold
MAS could move the Singapore Dollar Test Network (SGD Testnet)9 into production as its next step towards a deeper local market. Tokenised settlement can begin before that move is complete.
In November 2025 DBS, OCBC and UOB settled interbank overnight loans on the SGD Testnet, in the first live issuance of SGD wholesale CBDC.
DBS, OCBC and UOB can issue tokenised SGD deposits today under their existing licences.
Tokenised deposits are commercial bank money recorded on a ledger, within a class of instrument that banks already issue.
Banks would need no stablecoin authorisation or other new permission to issue them.
On 10 September 2026 the three banks completed their first live interbank SGD payments in tokenised deposits on Swift's ledger, with each bank recording the deposits on its own infrastructure.10
Regulated stablecoins would add a further route once licensed issuers operate under the MAS framework, which still awaits legislation.8
Wholesale CBDC, tokenised deposits and regulated stablecoins would form a complete settlement stack.
Ledger settlement can start now with tokenised deposits. Wholesale CBDC and regulated stablecoins would join the stack as each becomes available, and each can join independently of the other.
09 · Building Block 1: Digital Money (Regional and International)
Singapore could export the framework for digital money, and each ASEAN country would issue its own currency on the shared infrastructure
Singapore could export three things: (a) a policy template; (b) a technical framework; and (c) the connections through which regional jurisdictions join the shared infrastructure.
(a) Regional central banks could adopt MAS's stablecoin framework and digital money standards as a model.
(b) Templates compatible with Global Layer One (GL1)11 and aligned with ISO 2002212 would define how market participants connect to one another to offer digital money on shared infrastructure.
(c) Local market participants would connect to the shared settlement infrastructure through gateway application programming interfaces (APIs) to Canton.
Platforms regulated in Singapore would accept USD digital money where a trade requires it.
USDC could settle cross-border trades with US counterparties under the GENIUS Act. The Act became law in July 2025 and takes effect by 18 January 2027 at the latest.13
Tokenised USD deposits from internationally active banks would be the working settlement asset for USD trades.
USD instruments would operate under the rules of Singapore-regulated platforms. Singapore would therefore keep its monetary sovereignty.
Singapore would export the standard and the technology that applies it, and each jurisdiction would keep control of issuing its own currency.
10 · Building Block 2: Tokenised Risk-Free Collateral (Local)
Tokenised SGS would be the risk-free asset of Singapore's tokenised market and its most consequential piece of infrastructure
Government securities are the main collateral in every capital market: dealers use them in repo, post them as margin and price the benchmark yield curve from them. In a tokenised market, they must sit on the ledger before they can settle atomically.
SGS held off the ledger could support ledger-based repo, margin calls or atomic clearing only by bringing back the settlement risk that the tokenised market exists to remove.
DTCC began its Canton partnership in December 2025 by tokenising a subset of US Treasuries held at DTC, which puts the collateral asset first.5
A tokenised market without sovereign collateral on the ledger would have no risk-free asset.
Collateral on the ledger would then default to private credit instruments.
Private credit carries credit risk, and its value falls under stress, when collateral is most needed.
Digital money has an interim answer, because tokenised deposits can settle trades while wholesale CBDC reaches production. Collateral has no equivalent, so tokenised SGS is the only route to a risk-free asset on the ledger.
MAS could close that gap by issuing tokenised SGS alongside the existing SGS auctions and settling them in SGD digital money.
Tokenised SGS would be the risk-free asset on which Singapore's tokenised collateral arrangements rest.
11 · Building Block 2: Tokenised Risk-Free Collateral (Local)
Moving collateral in real time would bring the largest early gain, and one day of shorter settlement released USD 3 billion of margin in the US
Members of SGX Derivatives Clearing (SGX-DC) could move tokenised SGS between clearing accounts in real time. That ability would reduce the haircuts and liquidity buffers they carry today.
US experience with T+1 shows the size of the gain from shorter settlement alone.
USD 3bn
The clearing fund of the National Securities Clearing Corporation (NSCC)14 fell by an average of USD 3.0 billion (23%) after the US moved from T+2 to T+1 in May 2024, a one-day reduction made with no tokenisation.15
SIFMA, ICI and DTCC · T+1 After Action Report, September 2024
Atomic settlement with real-time collateral mobility would produce a larger gain still.
SGX-DC members can calculate the size of the gain in Singapore directly from their clearing volumes and the haircuts that SGX-DC applies.
12 · Building Block 2: Tokenised Risk-Free Collateral (Regional and International)
Tokenised SGS would also be Singapore's most exportable infrastructure, because ASEAN sovereigns could copy it and international collateral could connect to it
Together, tokenised SGS and tokenised US Treasuries would give the ledger a complete set of risk-free collateral, in SGD for the home market and in USD for international business. Other ASEAN sovereigns could copy the SGS model.
For ASEAN, tokenised SGS would be the proof of concept that matters most.
A sovereign bond issued on Canton, settled in SGD digital money and supported by open-source issuance templates would give other sovereigns a complete model to copy.
Singapore would export the architecture, standards and technology that let any ASEAN sovereign tokenise its own government debt on the shared infrastructure. The SGS themselves would stay in Singapore.
For global capital access, tokenised US Treasuries would provide risk-free collateral in USD.
US Treasuries are the global collateral benchmark. With both on the ledger, institutions working across SGD and USD markets would have risk-free collateral in each currency.
Building Blocks 3 & 4
We propose building the registry and settlement infrastructure on Canton, while keeping the option to move to another network
13 · Building Block 3: Digital Issuance and Registry Infrastructure (Local)
Canton is the most defensible network for Singapore's digital securities infrastructure, because the institutions Singapore needs to reach already use it
Adoption decides the choice of network. Canton has gathered an institutional base that no other permissioned network matches.
DTCC and Euroclear co-chair the Canton Foundation.5
Goldman Sachs, Citadel Securities, DTCC, Nasdaq, S&P Global, BNY, HSBC and BNP Paribas have invested in Digital Asset, the developer of Canton.6
Canton's design suits regulated capital markets infrastructure.
Digital Asset designed Canton for permissioned institutional networks. On Canton, each party sees only the parts of a transaction that concern it.
Contracts on Canton run in DAML,16 and the ledger settles both legs of a DvP7 trade atomically.
Digital Asset built Canton for regulated financial market infrastructure from its first design.
Singapore would meet its interoperability objective most directly by joining the network that the institutions it needs to reach already use.
In network infrastructure, each new participant adds value for all the others. A market that joins late must therefore catch up with those already connected.
Other permissioned networks exist, and on present evidence none combines Canton's institutional adoption with direct links to US and European settlement infrastructure. Canton is therefore the most logical choice at this stage.
14 · Building Block 3: Digital Issuance and Registry Infrastructure (Local)
We propose committing to Canton while keeping the framework portable in case network conditions change
Legal rights and business logic would sit apart from the smart contracts on any one network. That separation would keep the framework portable.
A legal framework held outside the smart contracts would define the rights each token carries, such as ownership, transfer entitlements and economic claims.
Business logic, such as matching engines, settlement sequencing and position management, would connect to the network through adaptors and could be maintained separately from it.
Market operators could move both to another network without rebuilding them. The commitment to Canton would therefore be considered and reversible.
Other permissioned networks exist, and MAS has used them in its own projects.
MAS's Project Ubin built prototypes on R3's Corda, among other platforms, in 2017.17
Separation of rights and logic from the network would keep the option to move open while Singapore commits to Canton for production.
Hydra X already builds on Canton in this way, with business logic connected to the network through a separate interface layer.
Hydra X has tokenised over USD 100 million of assets for institutional clients on Canton. In July 2026 it launched the HX Gateway API, a standard interface that lets institutions build on Canton without specialist DAML skills.18
An interface layer of this kind would give market operators a technical basis for production testing.
Regional trading links have stalled for want of shared post-trade infrastructure, and the framework would build it
From 2012 until its closure in 2017, the ASEAN Trading Link routed orders between Singapore, Malaysia and Thailand, and each trade cleared and settled in its home market.19 Its commercial aim was sound, and its weakness lay in post-trade, where it had no central clearing or settlement.
Earlier links joined incompatible national systems after the fact. Shared settlement infrastructure would build interoperability into its foundations from the start.
Where commercial interest and political will exist, markets could act on them, because the shared infrastructure would remove the technical obstacle.
Each ASEAN jurisdiction that modernises its market infrastructure must choose between an architecture built for tokenisation and an upgrade of its existing systems. Building for tokenisation would add little cost. A jurisdiction that chooses it early would also gain time on those that wait.
Singapore is the only ASEAN jurisdiction with the institutional depth to build this infrastructure today.
Hong Kong's Project Ensemble is the nearest competing effort in the region. The Hong Kong Monetary Authority (HKMA) moved it to live transactions in tokenised deposits in November 2025 and plans 24/7 settlement in wholesale CBDC by the end of 2026.20 Singapore's time to move first is therefore shortening.
17 · Building Block 3: Digital Issuance and Registry Infrastructure (Regional)
Every ASEAN participant would keep local control, because no single party would own the network
Sovereignty is usually the first objection in bilateral discussions across ASEAN, and it deserves a direct answer.
A distributed ledger has no single owner, which is what makes shared infrastructure acceptable to sovereign markets.
Each participating jurisdiction would issue on its own terms, within its own regulatory perimeter.
Each would keep full control of the rules for its own market and participants.
Singapore could help build the shared infrastructure and connect ASEAN markets to it. Every participant would be an equal validator, meaning an operator that confirms transactions on the network.
Local control would apply across the network, with Singapore as one participant among many.
Every participant, Singapore included, would gain local control, regional liquidity and global capital access.
Canton Foundation governance supports local control by design.
Set up with the Linux Foundation as an independent non-profit, the Canton Foundation has a decentralised structure designed to stop any single entity from controlling the network.21
No participant, Singapore included, can change the network's technical standard on its own.
18 · Building Block 3: Digital Issuance and Registry Infrastructure (Regional and International)
The framework would start in asset classes where ASEAN has yet to build registry infrastructure, with links to international settlement from launch
Entry would be easiest in asset classes where ASEAN exchanges have yet to build registry infrastructure.
Digital bonds, sukuk, carbon credits and private credit offer the most immediate openings, because no incumbent holds a position across them in ASEAN.
Once the framework has proved itself in these asset classes, it could extend to others.
DTCC is putting its tokenisation service into production on networks that include Canton. That service would give Singapore a route to global capital.
DTCC began limited production trades in July 2026 and plans to launch the full service in October 2026.22
Euroclear co-chairs the Canton Foundation and has taken part in collateral mobility trials on Canton since February 2025.23
SGX-listed tokenised securities issued on Canton would sit on the same network as DTCC and Euroclear. They could reach international liquidity with no separate integration layer.
19 · Building Block 4: Trading, Clearing and Settlement Infrastructure (Local)
A dual-track market would keep liquidity together during the move to atomic settlement, and off-ledger business logic would keep the system portable
In a dual-track market, tokenised and traditional versions of a security would trade under the same identifier, rules and investor rights, and each trade would settle by the route its format requires. A dual-track market is one sensible way to keep liquidity together during the transition.
In March 2026 the SEC approved Nasdaq's rules for this design, within DTC's tokenisation pilot.3 SGX equities settle at T+2 and SGS at T+1.24 Those cycles would be the starting point for shortening settlement step by step.
Shorter settlement would bring the gains shown under Building Block 2 at every stage, beginning with intraday settlement in batches and moving towards real time.
Market-facing business logic would sit off the ledger and connect to the settlement layer through adaptors. It covers the matching engine, pre-trade risk checks, collateral calculation, position management and post-trade reconciliation.
Operators could integrate their existing systems for each function with contained adaptation work.
Each function could then be updated or moved separately as the settlement layer changes.
Issuers, investors and other market participants could therefore adopt tokenised settlement step by step, using adapted market logic during the transition, until enough of the market has moved for full migration to follow.
The Opportunity
Singapore could build the shared infrastructure that ASEAN markets lack and own it, as Nasdaq and Intercontinental Exchange (ICE) own theirs
Singapore could fill the gap in ASEAN by exporting a complete stack that it would first prove in its own market
No ASEAN jurisdiction yet has the standards, infrastructure and expertise to build regional tokenisation infrastructure.
ASEAN regulators and exchanges are following developments and want to take part. They lack the capability to build the infrastructure themselves.
Singapore would face no established rival in this work. It brings MAS's regulatory standing, SGX's institutional relationships and existing links with DTCC, Euroclear and the Bank for International Settlements (BIS).
Singapore would export a complete stack ready for production, covering digital money, tokenised collateral, registry, trading and settlement. It would prove each component at home before offering it to the region.
Any regional exchange or central securities depository (CSD) could connect its own market logic through adaptors and use the shared settlement infrastructure without rebuilding its architecture.
Integration work would be kept small by design, so each market could adopt the framework in stages.
Edge markets would be the first entry point: digital bonds, sukuk, carbon credits and private credit, where ASEAN exchanges have yet to build registry infrastructure and no incumbent holds a position.
These asset classes offer the easiest route into ASEAN markets, and the framework could extend to others as it proves itself.
Singapore's aim should be to own the infrastructure, which is how Nasdaq and ICE built revenue that outlasts listing cycles
Nasdaq and ICE grew largely through acquisition, and the precedent that matters for Singapore is what they became.
Both became infrastructure and technology companies, and most of their revenue now comes from services other than their own listings.
Nasdaq also sells market technology to other exchanges, including markets it competes with. SGX launched the matching engine for its own securities market on Nasdaq technology.25
A Singapore operator could take the equivalent position in tokenised markets. Canton itself would stay neutral, because the Canton Foundation governs it and no single party owns it.
That operator could own the connections that participants use to reach Canton: (a) the adaptors that link each participant's business logic to the ledger; and (b) the validators of participants that choose to have them hosted.
Those services would let a regional exchange or bank join the network sooner and at lower cost, because it could connect through the operator's adaptors and hosted validators without building its own. They would also give SGX a practical contribution to offer in its discussions with regional exchanges on trading links and market access.
Other markets would then depend on that operator, as exchanges, SGX among them, depend on Nasdaq's matching engines. That position would bring durable revenue and a central place in the region's market infrastructure.
Assets under management and listing rankings rise and fall with market cycles. Revenue from infrastructure that other markets depend on recurs year after year, whatever the cycle.
A Singapore operator would have to build this position from within, which is harder and slower than acquisition. It would also need the full architecture in production.
A network standard on its own can be proposed. Only working infrastructure, however, can be exported into production.
All five layers (money, collateral, registry, trading and settlement) would need to operate before other markets could adopt them and depend on them.
Each regional market would connect to infrastructure that Singapore builds and maintains, and no participant would control it alone
SGX · CDP · MAS (illustrative)First and primary user of the infrastructure
Singapore Infrastructure Layer
Digital Money
Tokenised Collateral
Issuance & Registry
Trading & Settlement
▶
DTCC
Euroclear
International
Indonesia
Malaysia
Thailand
Philippines
Vietnam
Hong Kong
Each market connects to the shared infrastructure as an equal participant.
Each connected market would issue on its own terms, clear on the shared infrastructure and reach international liquidity through infrastructure that Singapore builds and maintains. Governance would be decentralised, so no participant, Singapore included, could control the infrastructure alone.
Implementation
MAS, SGX, CDP and the banks could build all four blocks at once, starting from capabilities that exist today
Banks would keep their core functions in tokenised markets and could earn from several new services
Tokenisation would change the infrastructure beneath the banks. Banks would continue to provide credit, make liquidity, transform risk and distribute to clients.
Banks would therefore keep the intermediation that runs on the new infrastructure.
Collateral management across tokenised platforms would be a new business for banks, and clearing members would pay for it.
A bank could manage tokenised SGS for clearing members, moving it to wherever it is needed. It could set its own price, because no existing service does the same work.
Today clearing members hold collateral as a static buffer. Under the framework, a bank could manage that pool actively. The bank that manages it best would earn fees that grow with the volume it handles, and rivals would find that business hard to win away.
Banks could earn from four further services, each of them new and each growing with market volume:
1. Issuance agency and trusteeship
2. Custody of institutional digital assets
3. Market-making in tokenised secondary markets
4. Tokenised deposit infrastructure for settlement
Each service would add revenue, because it serves activity that does not yet exist. Its market would grow with the infrastructure.
DBS, OCBC and UOB are well placed to lead because of their client relationships, regulatory standing and local knowledge.
International banks entering the region cannot quickly replicate those advantages.
Goldman Sachs's Digital Asset Platform already runs on Canton and is building tokenised settlement and custody services.26 Singapore's banks hold the home advantage, provided they move before international banks establish themselves.
The building blocks depend on one another in a fixed order, and all four can still be built at the same time
Each building block depends on the one before it, through three links:
1. Digital money must exist before collateral can sit on the ledger.
2. Collateral must be on the ledger before the registry can settle atomically.
3. The registry must be live before the settlement infrastructure can clear trades.
That order governs design, because each component must be built with the one above it in view. It also shows where a problem would spread if a component falls short.
All four blocks can be built at the same time, provided the teams agree the hand-off points in advance.
Each component should reach production at the point when the block above it needs it.
A sequential build would add years to the programme and leave other jurisdictions to set the standards.
Switzerland shows how concurrent work can succeed.
SIX, the SNB and commercial banks worked at the same time on the wholesale CBDC1 pilot, the SIX Digital Exchange (SDX)27 registry and tokenised bond issuance.
All three were in live use within a short period, because no component waited for another.
Four decisions would move the framework forward, and progress on any one would help the others
Three of the four already have a technical base in Singapore: the SGD Testnet, live tokenised deposits at DBS, OCBC and UOB, and MAS's planned trial of tokenised MAS Bills.9, 10
1. MAS could move wholesale CBDC on the SGD Testnet from pilot to limited production for approved institutions.
2. MAS could issue a tokenised SGS tranche alongside the existing auction programme.
3. DBS, OCBC and UOB could put their tokenised SGD deposits into use for securities settlement.
4. SGX and CDP could connect new or updated market logic for matching, settlement, risk and collateral management to their existing infrastructure through Canton.
28
Singapore could lead the next era of Asian capital markets, and this framework sets out one considered way to get there.
1Wholesale CBDC. Wholesale central bank digital currency: central bank money in digital form, issued only to regulated financial institutions for settlement between them. A retail CBDC would be issued to the public.
2SNB, Project Helvetia. The SNB has issued wholesale CBDC on the SIX Digital Exchange platform since 1 December 2023. In June 2025 it extended the pilot to at least June 2027, and it states that the pilot is no commitment to issue wholesale CBDC permanently. Source: SNB, Project Helvetia, snb.ch.
3Nasdaq. The SEC approved rule change SR-NASDAQ-2025-072 on 18 March 2026. Tokenised and traditional shares would share a CUSIP, symbol, order book and shareholder rights. Trading in tokenised form is limited to DTC's tokenisation pilot, and Nasdaq will give 30 days' notice before it begins. Source: SEC Release No. 34-105047.
4ECB, Pontes. The Eurosystem launched Pontes on 21 September 2026. It links distributed ledger technology (DLT) platforms to TARGET Services so that wholesale transactions in tokenised assets settle in central bank money, with further services to follow in stages. The ECB Governing Council approved the two-track strategy, Pontes and the longer-term Appia, on 1 July 2025. Pontes serves wholesale markets and is separate from the retail digital euro. Source: ECB press release, 21 September 2026.
5DTCC and Canton. On 17 December 2025 DTCC and Digital Asset announced a partnership to tokenise a subset of US Treasuries held at DTC on Canton. DTCC also joined Euroclear as co-chair of the Canton Foundation. Source: DTCC press release, 17 December 2025.
6Digital Asset investors. Digital Asset raised USD 135 million in June 2025, from investors including Goldman Sachs, Citadel Securities and DTCC, and USD 50 million in December 2025 from BNY, Nasdaq, S&P Global and iCapital. In June 2026 it raised USD 355 million in a round led by a16z crypto, with HSBC, BNP Paribas, Apollo, Citadel Securities, S&P Global and Tradeweb among the participants. Sources: Digital Asset press release, 11 June 2026; Bloomberg, 4 December 2025.
7DvP. Delivery versus payment: the securities leg and the cash leg settle at the same moment, so neither party is exposed to the other leg failing after its own has settled.
8MAS stablecoin framework. MAS finalised its framework for single-currency stablecoins in August 2023. On 1 September 2026 it published draft amendments to the Payment Services Act 2019 to give the framework legal effect. The consultation closes on 16 October 2026, and no commencement date has been set. Source: MAS Consultation Paper P015-2026.
9SGD Testnet. The Singapore Dollar Test Network, MAS's shared ledger for live trials of settling tokenised assets in wholesale CBDC. On 13 November 2025 MAS announced that DBS, OCBC and UOB had settled interbank overnight loans on it, in the first live issuance of SGD wholesale CBDC, recorded in the banks' official books. MAS also announced a future trial of tokenised MAS Bills issued to primary dealers and settled in wholesale CBDC, with details due in 2026. Moving the SGD Testnet into production would make wholesale CBDC available for live institutional settlement. Source: MAS media release, 13 November 2025.
10Tokenised deposits. On 10 September 2026 DBS, OCBC and UOB completed their first live interbank SGD transactions in tokenised deposits on Swift's ledger. Each bank recorded the deposits on its own tokenised-deposit infrastructure. Source: joint release by DBS, OCBC and UOB, 10 September 2026.
11GL1. Global Layer One, an initiative supported by MAS with the Bank of England, Banque de France, the ECB and global financial institutions. It develops common principles, standards and toolkits for interoperable networks of regulated institutions. Source: MAS speech, Singapore FinTech Festival, November 2025.
12ISO 20022. The global messaging standard for payments and securities. Swift ended coexistence with its older MT format on 22 November 2025, and ISO 20022 is now required for all cross-border payment instructions. An aligned implementation exchanges messages in the same format as the major payment and settlement systems. Source: Swift press release, 25 November 2025.
13GENIUS Act. Signed into US law on 18 July 2025. It takes effect on the earlier of 18 January 2027 or 120 days after the federal regulators issue final rules. The Office of the Comptroller of the Currency (OCC) expects to issue its final rule by November 2026. Sources: OCC, remarks of 19 August 2026; Anchorage Digital, The GENIUS Act at One, July 2026.
14NSCC. National Securities Clearing Corporation, the US central counterparty for equities and corporate and municipal bonds, and a subsidiary of DTCC.
15T+1 and the NSCC clearing fund. After the US moved to T+1 on 28 May 2024, the NSCC clearing fund fell on average by USD 3.0 billion (23%), from a prior three-month average of USD 12.8 billion to USD 9.8 billion. Source: SIFMA, ICI and DTCC, T+1 After Action Report, 12 September 2024.
16DAML. The programming language that Digital Asset developed for smart contracts on Canton, designed so that each party sees only the data it is entitled to see.
17Project Ubin. In Phase 2 of Project Ubin, MAS and the Association of Banks in Singapore built interbank payment prototypes on Corda, Hyperledger Fabric and Quorum. Source: MAS and ABS, Project Ubin Phase 2 report, November 2017.
18Hydra X on Canton. Hydra X became the first licensed custodian in APAC to hold Canton Coin in November 2024 and launched the Sigma Value Token, a tokenised structured product, on Canton in April 2025. It launched the HX Gateway API on 23 July 2026 and reported then that it had tokenised over USD 100 million of assets for institutional clients on Canton. Sources: Canton Network, 20 November 2024; Hydra X press releases, 8 April 2025 and 23 July 2026.
19ASEAN Trading Link. SGX and Bursa Malaysia launched the link in September 2012, and the Stock Exchange of Thailand joined in October 2012. The link ceased operation on 6 October 2017. Sources: ASEAN Exchanges; Bloomberg, 13 October 2017.
20Project Ensemble. The HKMA launched EnsembleTX, the pilot phase for real-value transactions in tokenised deposits and digital assets, in November 2025, to run through 2026. Hong Kong's 2026 Policy Address sets a target of 24/7 settlement in wholesale CBDC under EnsembleTX by the end of 2026. Sources: HKMA, 13 November 2025; Hong Kong 2026 Policy Address.
21Canton Foundation. An independent non-profit, created in partnership with the Linux Foundation, that governs the Global Synchronizer, Canton's shared interoperability infrastructure. It was called the Global Synchronizer Foundation until September 2025. Sources: Canton Foundation, 22 September 2025; Canton Network documentation.
22DTC tokenisation service. DTC received an SEC no-action letter in December 2025 allowing it to offer a tokenisation service for three years. DTCC planned limited production trades for July 2026 and a full launch in October 2026. On 1 July 2026 Tradeweb completed a real-time trade of tokenised US Treasuries against tokenised cash on Canton. Sources: DTCC press release, 4 May 2026; Tradeweb press release, 1 July 2026.
23Euroclear on Canton. Euroclear and Digital Asset launched the first phase of a tokenised collateral mobility initiative on Canton in February 2025. Euroclear has since taken part in industry working-group transactions on Canton, including in January and February 2026. Euroclear has announced no production date on Canton. Sources: Digital Asset and Euroclear, 25 February 2025; Digital Asset, 15 January and 24 February 2026.
24Settlement cycles. SGX equities have settled at T+2 since 10 December 2018, and SGX is consulting the industry on a possible move to T+1 with no decision yet made. SGS settle at T+1 on a DvP basis over MEPS+. Sources: SGX; MAS, Market Features: SGS, Money Market, Repo and Swaps Markets.
25SGX and Nasdaq. SGX launched a multi-matching trading engine for its securities market on the Nasdaq Financial Framework, Nasdaq's market technology platform. Source: Finance Magnates, reporting on Nasdaq's technology agreements.
26GS DAP. Goldman Sachs's Digital Asset Platform runs natively on Canton. Goldman Sachs has announced that it is exploring a spin-out of the platform into industry ownership. Sources: Canton Network ecosystem directory; Goldman Sachs announcement on GS DAP.
27SDX. SIX Digital Exchange, Switzerland's DLT-based financial market infrastructure. In 2021 the Swiss Financial Market Supervisory Authority (FINMA) licensed SIX Digital Exchange AG as a central securities depository and SDX Trading AG as a stock exchange. SIX announced in March 2025 that SDX would join its Securities Services unit, and on 5 May 2026 it reported FINMA approval to merge the digital CSD into SIX SIS AG. Sources: FINMA, 2021; SIX press release, 5 May 2026.
28US Treasury market. Marketable US Treasuries outstanding were USD 31.8 trillion as of August 2026. SIFMA put the average daily amount of repo outstanding at USD 4.4 trillion in 2024. Sources: SIFMA, US Treasury Securities Statistics, 5 October 2026; SIFMA, Fixed Income Market Structure Compendium, 2025.
29SGS bonds. SGS bonds outstanding rose from SGD 172.6 billion at the end of 2024 to SGD 178.6 billion on 3 November 2025, and MAS expects similar growth in 2026. Source: MAS, 2026 SGS Issuance Calendar, 31 October 2025.
30Enhanced Repo Facility. MAS gives SGS primary dealers exclusive access to its Enhanced Repo Facility, through which they can borrow specific SGS. Source: MAS, A Guide to SGS Primary Dealer Operations, 11 September 2025.
31Broadridge DLR. Broadridge's Distributed Ledger Repo processed USD 7.5 trillion of repo in September 2026, an average of USD 359 billion a day. Source: Broadridge, 8 October 2026.
32BIS Project Agorá. A project convened by the BIS and the Institute of International Finance, with seven central banks and more than 40 regulated financial institutions. It combined tokenised commercial bank deposits and tokenised central bank reserves on a shared platform, and its prototype showed atomic, multi-currency settlement of wholesale cross-border payments. Source: BIS, Project Agorá report, 27 May 2026.
33Project Meridian. A project of the BIS Innovation Hub London Centre and the Bank of England, concluded in April 2023. A synchronisation operator linked an RTGS system to other ledgers so that funds and assets changed hands together. The Bank of England has since tested the same approach for tokenised securities in Project Meridian Securities. Sources: BIS, Project Meridian, April 2023; Bank of England, Project Meridian Securities.
Food for Thought
Atomic settlement raises three questions about Singapore's post-trade system that are easier to resolve early
CDP's role under atomic settlement · The practical case for tokenised SGS · How interbank settlement could work
Question 1
CDP's role under atomic settlement
29 · Post-Trade in a Tokenised Market
Atomic settlement would gradually take over both of CDP's clearing functions, so Singapore should decide early where netting will sit
CDP clears trades through two operations: (a) novation, in which it becomes the counterparty to each side and takes on the settlement risk; and (b) multilateral netting, in which it reduces gross obligations to net amounts before settlement.
Atomic settlement would replace novation directly.
When securities and cash settle together on a shared ledger, the settlement itself removes counterparty risk.
Each step from T+2 towards real time shortens the period in which novation protects the parties. Continuous atomic settlement would remove that period entirely.
As an operation, netting would continue, and it would move towards the banks over time.
In batched settlement, CDP could still net the trades in each batch before the batch settles atomically. It could keep this function for as long as batching continues.
As batch windows shorten, fewer trades fall within each batch, so netting releases less capital.
In continuous atomic settlement, the securities leg would settle gross, trade by trade. The banks would then be the parties that accumulate and net the remaining cash obligations.
CDP could take on netting for the banks. Even then, the function would follow the cash, which sits with the banks.
Once both functions change, Singapore will need to decide what operational role supports CDP's designation as the national depository. That question is easier to settle before the move to atomic settlement forces it.
Both questions, where netting sits and what role underpins CDP, are easier to answer early, while Singapore still controls the timing.
Question 2
The practical case for tokenised SGS
30 · SGS and US Treasuries
Singapore must judge tokenised SGS on the facts of its own market, which differ sharply from the US Treasury market
US Treasuries outstanding reached USD 31.8 trillion in August 2026, and the US repo market provides over USD 4 trillion of financing on an average day.28 Thousands of participants trade Treasuries continuously. That depth reflects their supply, the dollar's reserve status and decades of infrastructure built for the market.
SGS bonds outstanding stood at about SGD 179 billion in November 2025.29 The government issues SGS to build the yield curve and to meet banks' liquidity requirements. It does not issue them to finance deficits.
At SGX-DC, US Treasuries and Japanese government bonds make up most non-cash collateral because of settlement currency and funding costs. SGS are eligible and play a smaller part.
MAS runs an Enhanced Repo Facility for primary dealers within fixed hours.30 Singapore has no deep, continuous repo market to match that of the US.
Tokenisation would leave SGS supply, the currency mix of derivatives exposures and the reserve status of the SGD unchanged. Its value would lie in removing the operational limits on how existing SGS can be used.
31 · Collateral Velocity
Tokenised SGS could turn a static liquidity buffer into working collateral, and each extra use of a limited supply would be worth more
Banks hold most of their SGS as static positions, either in custody to meet the High Quality Liquid Asset (HQLA) requirement under Basel III or as long-duration investments.
Each movement between accounts needs a matched pair of Swift instructions into the MAS Electronic Payment System (MEPS+), settled within fixed operating hours. That overhead is acceptable for a static buffer. It becomes a binding constraint, however, once banks use SGS actively as collateral.
Tokenised SGS on a shared ledger would remove most of that overhead.
Collateral would move as an operation on the ledger, with no separate settlement event for each transfer. It could also move outside MEPS+ operating hours.
Counterparties could trigger margin calls, repurchase returns and substitutions on the ledger. The ledger would enforce atomic and conditional execution at each step.
Each unit of SGS serves one function at a time today. A bank could lend tokenised SGS under repo during the day and take them back before the reporting snapshot. The same bonds would then meet the HQLA requirement and raise secured funding within 24 hours.
Because the supply of risk-free assets is limited, each additional use of the same bond is worth proportionally more.
32 · SGS Repo Market
Singapore's SGS repo market remains shallow mainly because each trade costs too much at current volumes, and tokenisation could cut that cost
In the US, decades of investment have cut the cost of each Treasury repo to a minimum. That low cost lets the market provide over USD 4 trillion of financing a day.28
In Singapore, each SGS repo needs separate matched instructions through MEPS+ for delivery, for return on the repurchase date and for any margin or substitution during the term. That overhead is too large for the frequent, continuous trading that builds a liquid funding market.
Repo on a shared ledger would shorten the whole lifecycle, in three steps:
1. At inception, SGS tokens and digital cash would move together on the ledger.
2. During the term, either counterparty could trigger margin calls, substitutions and early termination on the ledger, and the ledger would enforce atomic and conditional execution at each step.
3. On the repurchase date, the reverse leg would settle as agreed.
Ledger-based repo already works at institutional scale: Broadridge's Distributed Ledger Repo platform processed USD 7.5 trillion of tokenised repo in September 2026, an average of USD 359 billion a day.31
MAS is moving in the same direction. In November 2025 it announced a trial of tokenised MAS Bills issued to primary dealers and settled in wholesale CBDC.9 MAS Bills are short-dated, and the trial would still test the mechanics that a wider SGS repo programme needs.
33 · Three Stages of Value
Tokenised SGS could deliver value in three stages, each building on the one before
These stages describe what could happen, and each would follow only if participants find enough commercial value in the one before.
1. In the first stage, SGX-DC could accept tokenised SGS as margin collateral that moves in real time.
This stage needs the fewest counterparties and offers the most measurable benefit. Clearing members could calculate the capital released when static SGS buffers become working collateral that meets both the HQLA requirement and margin calls.
It could begin once a tokenised SGS tranche is issued, possibly through CDP.
2. In the second stage, tokenised infrastructure could support an SGS repo market.
MAS's Enhanced Repo Facility could form the core of that market if MAS opened it beyond primary dealers and used ledger-based repo.
Such a market would give Singapore the secured SGD funding it lacks at scale.
3. In the third stage, tokenised SGS could become the proven model for ASEAN sovereign debt.
Singapore would tokenise a sovereign bond, settle it in SGD digital money and prove the standards and architecture at home before offering them to the region.
Singapore's value to ASEAN counterparts would be the working precedent and the expertise to guide adoption. That value depends on the first two stages being in production.
Question 3
How interbank settlement could work
34 · Settlement Scenarios
The seller would receive funds through its bank, by a path that depends on whether buyer and seller use the same bank
In scenario 1, buyer and seller would use the same bank, DBS in this illustration.
Under atomic DvP,7 the tokenised securities and deposits would transfer together on the ledger. DBS would credit the seller at once, with no movement between banks.
In scenario 2, the buyer would bank with DBS and the seller with OCBC, again as an illustration.
1. Atomic DvP would lock the trade: the securities would pass to the buyer, and the buyer's bank would commit an irrevocable payment obligation to the seller's bank.
2. The seller's bank would verify that the securities leg has settled and credit the seller in advance, relying on that verification and the irrevocable obligation.
3. The intraday credit exposure between the two banks would clear at the next interbank settlement. Banks carry the same kind of exposure today on PayNow and FAST.
In scenario 3, which is theoretical, MAS's wholesale CBDC would be in production.32
Atomic DvP would transfer the securities and the wholesale CBDC together, so no interbank credit exposure would arise.
That configuration is one long-term possibility, and it would remove the interbank credit step entirely.
35 · Netting Migration
Interbank netting would take over from CDP's multilateral netting, so the operation would continue in a new place
Today, CDP nets both legs. It compresses all gross trades multilaterally, so only net obligations move on settlement day.
In a tokenised model, the securities leg would settle atomically. CDP's securities netting would then have no remaining role.
Atomic settlement would itself deliver the capital efficiency that securities netting provides today.
On the cash leg, banks would take over netting in three steps:
1. Banks would accumulate bilateral obligations through the trading day.
2. They would calculate multilateral net positions at a set cycle.
3. They would settle only the net amount, through MEPS+33 or wholesale CBDC.
Singapore could choose among three institutional forms for interbank netting:
(a) Bilateral arrangements between banks
(b) A multilateral utility run by the banks or MAS
(c) CDP in a new role
36 · Batch to Continuous Settlement
Intraday batches are the realistic near-term step, and each shorter batch window would move more netting from CDP to the banks
Stage
Settlement model
Where netting sits
Stage 1 (near term)
Trades would settle in one end-of-day batch, netted and settled atomically at the close.
CDP would keep its netting role, and banks would settle net interbank positions once a day.
Stage 2 (medium term)
Trades would settle in two to four intraday batches, with less netting in each smaller batch.
CDP's netting would release less capital as batches shrink.
Stage 3 (potential end state)
Each trade would settle atomically as it matches, and securities would move gross.
Banks would net interbank cash on their own cycle, and CDP's netting would end in its current form.
At every stage, the seller would receive funds promptly, either by book entry within one bank or by its bank's advance credit against a verified commitment. The interbank cycle would decide only when the credit exposure between banks clears.
