Project Guardian is the Monetary Authority of Singapore's industry pilot programme on asset tokenisation, structured as a set of named workstreams across asset-and-wealth management, fixed income, foreign exchange, and trade finance. It is the closest thing in APAC to a regulator-curated commercial sandbox: MAS sets the perimeter, named institutions run the pilots on their own infrastructure under their existing licences, and the workstreams produce design patterns (purpose-bound money, trust anchors, cross-network interoperability for permissioned ledgers) that other jurisdictions then borrow. For a tokenisation product team picking an APAC venue, Guardian is the worked example of how MAS thinks the regulated ecosystem should look once it scales beyond pilots, and it is the place to read the actual operating tradeoffs around interoperability between permissioned bank ledgers and the assets that sit on top of them.

How Guardian is structured

Guardian is not a sandbox in the technical sense. There is no temporary licensing relief, no carve-out from the existing payment, capital markets, or banking perimeters. Participants run the pilots under the licences they already hold. The structure is closer to a coordinating framework: MAS publishes the architectural patterns it wants the industry to converge on, names the participating institutions, and reviews outputs as they ship. The legal authority for any specific pilot still flows through whichever existing regime applies (Securities and Futures Act for tokenised funds, Payment Services Act for payment-related work, Banking Act for bank-issued instruments).

This matters for a couple of reasons. First, the named institutions on Guardian press releases are the institutions doing the work, not a wider universe of "interested parties" sitting outside the perimeter. Second, the regulator is not pre-clearing products through Guardian. A pilot can run inside Guardian and still need to pass through MAS product authorisation channels before it ships to retail or qualified clients. Third, the "graduation" question (when does a pilot become a production product) is governed by the existing regime, not by a Guardian-specific track.

Guardian is organised around four named workstreams. The split is by asset class and use case rather than by technology layer.

Asset and wealth management workstream

The asset-and-wealth-management track covers tokenised investment products and their distribution wrappers: tokenised funds, structured notes, variable capital company (VCC) wrappers issued in tokenised form, and the secondary-market plumbing that lets qualified investors transfer them. Participating asset managers and fund administrators have included global names alongside Singapore-domiciled vehicles, with bank distribution partners sitting on the wealth-management side.

The architectural pattern that has dominated this workstream is the use of a Singapore-domiciled regulated wrapper (typically a VCC or a unit trust) with the on-chain instrument representing the fund interest, and a transfer-agent function performed via permissioned token logic that enforces eligible-investor checks at the contract level. The on-chain side often runs on a permissioned ledger that the manager and bank can both observe. The off-chain registry remains the legal record. This is the same "wrapper plus mirror" construction described in Tokenisation, defined, applied specifically to the Singapore fund regime.

The cross-border angle is what differentiates the Singapore approach from a purely domestic pilot. Several Guardian asset-and-wealth pilots have routed flows through bank distribution partners in other APAC jurisdictions, with the Singapore-issued instrument transferred to a bank-held custody arrangement abroad. The interesting operational question is whether the on-chain transfer is bearer-style among permissioned addresses, or whether each cross-border movement triggers an off-chain instruction set. The pattern in production has been the latter, which is unsurprising once the underlying instrument is a regulated fund interest with KYC-bound holders.

Fixed income workstream

The fixed-income workstream is the most operationally mature of the four because the asset class itself is the easiest to tokenise: large tickets, sophisticated investors, well-understood lifecycle events, and an existing regulated wrapper. Pilots have covered tokenised commercial paper, tokenised bonds (corporate and structured), and tokenised repo. The participating institutions span global investment banks running their tokenisation platforms (the Kinexys lineage is one of the named participants), Singapore banks, and asset managers as buyers.

The November 2025 ICMA Fixed Income workstream deliverables addendum is the most concrete published output of the Guardian fixed-income track to date. Two pieces of operational guidance landed: a tokenised-bond DvP guide that addresses the settlement-asset choice (wholesale CBDC, tokenised bank liabilities, regulated stablecoins) and the atomic-settlement mechanic, and a custody-arrangements guide that addresses how the platform-custody-issuer triangle composes when both the bond leg and the cash leg are tokenised (ICMA news, November 2025). The addendum sits as published guidance on top of the existing Guardian Fixed Income Framework rather than as a standalone framework, and it is the artefact a tokenised-bond product team should read for the operational reference points the Guardian work has converged on.

What makes this workstream interesting beyond the issuance side is the cash leg. A tokenised bond settling against an off-chain payment instruction is not an obviously different product from the same bond issued on a conventional book-entry rail. The differentiator is delivery-versus-payment on a single ledger, or atomic settlement across linked ledgers, with the cash leg also tokenised. Guardian fixed-income pilots have explored both: deposit tokens issued by the bank's tokenisation platform on its own ledger, with the bond and cash legs settling atomically; and bilateral arrangements that bridge between a bond issuer's ledger and a different bank's deposit-token ledger using a hashed time-locked construction or an equivalent.

The institutional adoption signal here is operational rather than symbolic. The participants are settling real flows of meaningful size, not running thought experiments with synthetic test assets. The unit economics still are not public, but the willingness to commit balance sheet to repeated pilots is a non-trivial signal that the participants believe the post-pilot product has a path to production.

Foreign exchange workstream

The FX workstream is where the purpose-bound money (PBM) construction has been most visible. PBM is a Singapore-coined design pattern: a wrapper around a payment instrument that carries programmable conditions on use, expiry, recipient, or purpose. The Singapore literature treats PBM as a generalisation that can sit on top of tokenised deposits, regulated stablecoins under the SCS framework, or future wholesale CBDC. PBM is not itself a settlement asset; it is a contract layer that conditions how a settlement asset can be used.

The Guardian FX work has used PBM-style constructions to test atomic FX swap settlement with conditional release based on receipt confirmation, and conditional payment release tied to delivery in a goods context. Participating institutions have included Singapore banks running their own deposit-token rails alongside global banks bringing their G10 currency platforms. The cross-jurisdictional dimension is explicit: an FX swap between SGD-denominated bank money and USD-denominated bank money, settling atomically across two banks' ledgers, is a worked example of how a tiered-ledger model survives a multi-currency, multi-jurisdiction transaction.

The 3 July 2025 ISDA + Ant International industry report on FX with tokenised bank liabilities is the most consequential FX-workstream output to date (ISDA). The report quantified a USD 50bn cross-border cost saving by 2030 if interoperability standardises, with contributors including BNY, HSBC, OCBC, GFXD, and the Ant International Whale platform liquidity-provider role surfaced as part of the workstream output. The December 2025 Standard Chartered tokenised-deposit production launch on Whale (see the SC + Ant Whale theme) is the operational follow-through from the Guardian FX work, with HKD, CNH, USD, SGD live as the multi-currency tokenised-deposit substrate.

The interoperability layer is the load-bearing piece here. Two banks running different permissioned ledgers, with potentially different tokenisation stacks underneath, need a shared settlement protocol or a trusted bridging arrangement. The Guardian work has tested several patterns, including hashed time-locked atomic swaps and a notary-style "trust anchor" model where a regulated coordinator confirms both legs before either releases. None of these patterns has emerged as the dominant standard, which is itself useful information: the design space is still open in 2026, and a product team committing to one pattern is implicitly betting on its adoption.

Trade finance workstream

The trade-finance workstream is the youngest and the smallest by named participation. The pilots have focused on tokenisation of trade documents (electronic bills of lading, letters of credit) alongside the cash leg, with a programmable money construction conditioning payment release on document presentation or delivery confirmation. Several Singapore banks have participated, with a smaller set of corporate counterparties on the other side of the flow.

The operational tradeoff in this workstream is the integration with the off-chain trade-document ecosystem. Tokenising the cash leg does not solve the problem if the underlying trade documents are still moving via conventional channels. The pilots that have shipped meaningful flow have tended to be those where the document layer was already digitised and could be hooked into the on-chain payment release logic. This is a harder shape to scale than tokenised funds or tokenised bonds, and the Guardian trade-finance work should be read as an early-stage exploration rather than a near-term production pipeline.

Symbolic versus operational participation

A useful filter when reading Guardian announcements is to separate two kinds of named participation. Symbolic participation is "logo on the press release": the institution lent its name to the workstream, possibly seconded a researcher, and produced a co-authored design paper. Operational participation is "balance sheet committed": the institution issued the actual tokenised instrument, custodied real assets, settled real flows of meaningful size on its own ledger or someone else's, and integrated the pilot with its production compliance and operations stack.

Most Guardian press releases mix both. The asset-and-wealth and fixed-income workstreams have produced more operational participation by named institutions than the FX and trade-finance workstreams to date, partly because the asset classes are more mature on the tokenisation side and partly because the Singapore bank participants have committed deposit-token rails to the work. Reading the participation list with this filter in mind is a faster way to assess where Guardian work is actually moving production thinking, versus where it is still architectural.

Cross-border and standard-setting reach

Guardian has a deliberate outward face. MAS has signed cooperation arrangements with peer regulators in Japan, Switzerland, the UK, and other jurisdictions to align Guardian-style pilots with comparable workstreams elsewhere, including PIP in Japan and the BIS Innovation Hub experimental work. The architectural patterns produced by Guardian, particularly PBM and the trust-anchor model for cross-network settlement, have been picked up in BIS and BIS Innovation Hub work and discussed at CPMI and IOSCO level.

The consequence is that Guardian outputs reach further than the Singapore perimeter alone. A design pattern that lands in a Guardian paper has a non-trivial chance of appearing in a BIS report two cycles later, which is in turn the basis for international standard-setting work. For a tokenisation product team, this means Guardian is worth tracking even if the team is not Singapore-domiciled, because the patterns it surfaces have a habit of becoming the regional default.

Open questions

• Which Guardian pilots have crossed from named-pilot status into production-grade product authorisation, and which are still in pilot mode as of late 2025. MAS has not published a consistent graduation tracker.
• The unit economics of any Guardian pilot. None of the participants have published settlement-cost or balance-sheet-mobility data publicly.
• Whether the FX workstream's PBM patterns will survive contact with retail-facing distribution, or whether they remain confined to wholesale and bank-to-bank flows.
• Whether the trade-finance workstream produces a workable end-to-end pattern or stays as document-tokenisation-plus-payment-release plumbing without broader uptake.

Related

• singapore mas scs framework.
• singapore dtsp extraterritoriality.
• Project Ensemble (Hong Kong counterpart).
• Payment Innovation Project (Japan counterpart).
• hong kong, japan.
• Tokenisation, defined, Stablecoin types, Permissioned blockchains.
• Tokenised collateral.
• BIS, BIS Innovation Hub, CPMI, IOSCO.