Suit Up · Asia's institutional desk on tokenisation

TL;DR

Tokenised MMFs (money-market funds) are the most-mature institutional tokenisation pattern in production. BlackRock's BUIDL, Franklin Templeton's FOBXX, and Ondo Finance's OUSG are the reference programmes a new entrant has to compare design choices against, and the BUIDL reference architecture is the worked example most allocators have already evaluated. The legal-wrapper choice (3(c)(7) for US qualified purchasers, UCITS for EU retail, Singapore VCC plus accredited-investor offering for APAC) is the load-bearing design decision and forces every other choice downstream of it. Pick a partner-versus-build path for issuance and custody early; treating those as separable decisions is one of the most common reasons programmes slip a quarter. Get a single-chain pilot to production before going multi-chain; the operational integration cost of multiple chain identities compounds faster than the distribution benefit at small scale.

Decision frame

The asset manager has an existing MMF franchise (Treasury, Government, prime, or AAA-CLO-style) and is deciding how to ship a tokenised wrapper without breaking the existing fund or its fund-administrator relationships. The tokenisation choice is structurally additive to a live fund-management business rather than a green-field product launch, which means most of the binding constraints are operational integration with what the firm already runs, not technical novelty. The right framing is: the tokenisation programme is a new share class with on-chain mechanics, not a new fund.

Wrong framings to avoid before reading further. "Tokenise everything in the MMF complex at once" loses time and political capital; sequence by jurisdiction and pick one wrapper to ship first. "We will use a public chain for the credibility halo" is the wrong direction of inference; chain choice should follow the distribution channel, not generate it. "We will skip the secondary venue and just do primary subscribe-and-redeem" is the right answer for many MMF programmes (the underlying instrument is primary-market by design, see What to tokenise first for the screening logic) but not for products designed to back DeFi-native collateral demand. "We will tokenise our prime fund first because it is our biggest" is the trap that sinks more pilots than any other single design choice; Treasury and Government MMFs map cleanest to tokenisation, prime adds credit-risk underlying that materially complicates the disclosure cadence and the institutional-buyer diligence pack.

The eight design decisions

These bind in roughly this order. Each downstream choice gets cheaper once the upstream choice is made, and reordering tends to surface as missed deadlines on the operational integration.

1. Legal wrapper

The wrapper choice determines holder eligibility, redemption mechanics, marketing perimeter, and the regulator in the room. Three patterns dominate.

3(c)(7) for US qualified purchasers. The BUIDL and OUSG pattern. Holder set restricted to qualified purchasers (broadly, individuals with USD 5 million-plus in investments, institutions with USD 25 million-plus under management). The whitelist is operationally meaningful, marketing perimeter is restricted, redemption can run on a tighter institutional cadence. See BUIDL reference architecture for the worked example.
1940 Act 2a-7 government MMF. The FOBXX pattern. Both institutional and retail eligibility under specific rules, broader distribution surface, heavier disclosure and conduct overlay. Materially harder to ship a first tokenised product under because the retail-investor protection apparatus has to be in place from day one.
UCITS for EU retail. Less mature on tokenisation; the legal wrapper is well-understood for conventional funds, but the on-chain operative-record question under member-state company law is unresolved in most jurisdictions. Treat as a 2026-2027 path rather than a near-term one.
Singapore VCC plus accredited-investor offering under SFA Sections 274/275. The cleanest APAC route, with Project Guardian's asset-and-wealth workstream as the worked policy precedent (see Project Guardian). Restricts distribution to accredited investors and institutional investors, which collapses the conduct overlay relative to a full retail product.
Cayman SPC for offshore qualified-investor distribution. The default pattern for an APAC manager whose existing fund sits in Cayman. Adds a Reg S overlay for non-US offering and a tokenised share class on top of the conventional structure.

The wrapper is not a marketing decision. It forces operational implications for the transfer agency, the registry, and the on-chain whitelist mechanics. Pick the wrapper before naming the chain.

2. Chain choice

Single-chain at launch. The institutional gravity argument points to Ethereum mainnet (BUIDL, FOBXX, OUSG, ACRED all originated there), the regulatory-perimeter argument points to a permissioned ledger (Canton for the Goldman / BNY tokenised-deposit work, Provenance for some private-fund issuance). Both are defensible. What is not defensible is going multi-chain at launch; BUIDL and FOBXX both ran single-chain for their first 12 to 18 months before extending, and the engineering cost of cross-chain whitelist consistency is non-trivial.

Multi-chain as expansion. BUIDL operates across nine chain identities, FOBXX across eight, OUSG across four. The pattern is to launch on one chain, accumulate operational track record, then extend on the back of confirmed institutional or DeFi distribution demand on the destination chain. The destination should be downstream of where investors actually settle, not where the operator wants to be visible.

The chain choice is an integration question dressed up as a brand question. Where does your distribution channel actually settle.

3. Issuance platform partner

The choice is structurally similar across MMF programmes; the differentiator is which set of regulatory wrappers the platform brings. See Evaluating issuance platforms for the full evaluation rubric.

Securitize. SEC-registered transfer agent plus broker-dealer plus ATS in one counterparty. The US-perimeter default. Centre of gravity is qualified-purchaser MMFs and private-fund interests under the 3(c)(7) wrapper.
Centrifuge. Public-chain native, SPV-plus-tokenisation pattern. The fit when DeFi-native distribution is the primary channel (the Aave Horizon institutional-collateral lane is the worked example).
Ondo Finance. Post-Oasis-Pro acquisition (October 2025), brings broker-dealer plus ATS plus transfer-agent registrations in-house. Issues its own products plus is a platform layer.
Marketnode. Singapore market-infrastructure venture, SGX-aligned, recurring Project Guardian participant. The natural counterparty for a Singapore-issued tokenised product needing local market-infrastructure-aligned positioning.
In-house. Franklin Templeton's pattern (own SEC-registered transfer agency on the FOBXX programme). Structurally available to any asset manager willing to build the registered-transfer-agent capacity in-house, but materially harder to justify at single-product scale.

4. Custody partner for the underlying

The off-chain T-bill, repo, and cash sleeve has to sit at a bank-grade conventional custodian. BNY and State Street are the default counterparties for US-domiciled funds; the same firm typically already custodies the conventional fund and the tokenised wrapper inherits the relationship. This is rarely a green-field decision; the underlying-custody counterparty is set at fund formation. See Evaluating custody providers for the role-by-role evaluation rubric and the asymmetry between the four custody functions. Skipping the bank-grade underlying custodian is a red flag your distributor's risk team will reject; it surfaces in 100 percent of institutional diligence packs.

5. Custody partner for the tokens

The on-chain token-level custody is a separate function from the underlying custody. Three patterns.

Federally chartered US trust. Anchorage Digital is the only OCC-chartered digital-asset bank, plus the conditionally approved cohort that includes BitGo, Fidelity Digital Assets, Paxos, and Circle (see OCC trust bank charter).
State-chartered qualified custodians. Coinbase Custody Trust (NYDFS limited-purpose trust), BitGo Trust (South Dakota state trust). Common counterparty for institutional and ETF-like flows.
Bank-incubated APAC specialists. Sygnum (FINMA banking licence plus Singapore MAS perimeter), Zodia Custody (Standard Chartered-incubated, FCA-registered with Singapore presence). The natural fit for APAC-distributed flows where the distributor wants a regulated regional counterparty.
Infrastructure tech model. Fireblocks sits behind a large share of regulated-counterparty flows but operates as infrastructure rather than as the regulated custody perimeter; the regulated perimeter sits with the customer.

Match the token custodian to the holder set. Institutional holders default to a federally chartered or bank-incubated counterparty; DeFi-native holders to a self-custody or MPC-backed wallet posture.

6. Distribution channel

The channel determines almost everything about the operational requirements. Three families.

Private-bank wealth platforms. DBS Private Bank, HSBC Private Bank, JPMorgan Private Bank, Standard Chartered Wealth Management. The cleanest fit for a Treasury MMF tokenised wrapper because the channel already buys conventional MMFs into qualified-investor accounts. The integration is the on-chain holder-account reconciliation against the bank's internal client-account ledger.
Institutional treasury management. Corporate cash desks, large-corporate treasurers, family offices. The buyer is allocating to the tokenised MMF as a cash-equivalent slot in the existing investment-policy statement (see allocator psychology for the comfort-pattern dynamics).
DeFi liquidity venues. The Aave Horizon institutional-collateral lane, Sky / Grove for stablecoin-issuer collateral demand. Different operational requirements (whitelist consistency at the venue level, on-chain accounting cadence) and a different buyer profile (DeFi-native treasuries, on-chain stablecoin reserves seeking yield-bearing dollar collateral).

The distribution channel determines the chain choice, the redemption mechanics, and the disclosure cadence. Pick the channel before the chain.

7. Subscribe and redeem mechanics

Two binary choices and one trade-off.

T+0 versus T+1 redemption to fiat. T+0 is the marketing line; T+1 is more honest. Banking-rail integration on weekends and holidays is the binding constraint. Promising T+0 to fiat without confirmed weekend coverage is a red flag.
On-chain redemption to USDC versus fiat. The Circle USDC swap adjacent to BUIDL is the cleanest published worked example of an on-chain liquidity primitive that does not require waiting for the off-chain redemption cycle. See Circle for the structural design. Restricting on-chain redemption to institutional holders is a defensible design and the typical pattern.
Primary-only versus secondary-trading-supported. MMFs are primary-market by structure. The on-chain transferability is a secondary-market substitute for many use cases but the secondary-trading venue (Securitize Markets, Oasis Pro) is rarely load-bearing for a tokenised MMF programme. Build it if the channel demands it; do not over-engineer it.

8. Reporting and disclosure cadence

Tokenised MMFs are typically held to a higher disclosure cadence than non-tokenised peers. Daily NAV publication is the institutional-holder default. Weekly portfolio holdings is the institutional-buyer expectation; monthly is the floor that some 2a-7 funds operate at, but a tokenised wrapper distributed into qualified-purchaser channels usually commits to weekly. Monthly attestation by the platform partner is the BUIDL pattern (Securitize-side attestation on share-register integrity). Annual audited financial statements at the fund level on the conventional cadence.

Asset-class scorecards by underlying

Different MMF subtypes have different tokenisation maturity. Treat the subtype as a binding constraint on the platform shortlist, not a soft preference.

Treasury MMFs. Cleanest. BUIDL (BlackRock), FOBXX (Franklin Templeton), and OUSG (Ondo) are the reference programmes. US Treasury underlying gives Basel SCO60 Group 1a treatment a clean path under the targeted review. The institutional-buyer diligence pack is well-understood.
Government MMFs. Similar to Treasury, with a slightly broader collateral pool (agency paper, repo against Treasuries). The 2a-7 government MMF route is the FOBXX pattern; institutional and retail eligibility under specific rules. Disclosure cadence inherits from the underlying fund.
Prime MMFs. Less common to tokenise. The credit-risk underlying (commercial paper, certificates of deposit) adds material complexity to the disclosure cadence and the institutional-buyer diligence pack. Holders need exposure-by-issuer reporting, not just NAV. The structural reason to hold off on tokenising the prime franchise first is that the credit-risk underlying compounds the operational integration risk on top of the tokenisation-specific operational risk; both are non-trivial in isolation, harder to debug when they overlap.
AAA CLO and private-credit MMFs. Janus Henderson's JAAA on Centrifuge is the worked example, sitting at the boundary between MMF and private credit and inheriting some of each's complexity. The Sky / Grove allocator integration is the worked DeFi-distribution channel. Treat as a phase-2 product rather than a first-tokenisation candidate unless the firm's existing credit franchise already runs CLO-shaped strategies.

Worked example

An APAC asset manager with a USD 5 billion Treasury MMF franchise (Cayman SPC, USD-denominated, distributed across Singapore, Hong Kong, and selective other regional accredited-investor channels) is deciding how to ship a tokenised version into Singapore and Hong Kong wealth distribution.

Walk through the eight decisions.

Legal wrapper. Keep the existing Cayman SPC. Add a tokenised share class under Reg S for the non-US offering, plus the existing accredited-investor and institutional-investor exemptions under SFA Sections 274/275 for Singapore distribution and the SFC professional-investor regime for Hong Kong distribution. The conventional fund continues to operate alongside; the tokenised share class is structurally additive.
Chain. Ethereum mainnet at launch (institutional gravity, Securitize and Anchorage operationally proven on the chain, the institutional-buyer set already evaluating Ethereum-resident tokenised MMFs). Plan multi-chain expansion to a permissioned APAC-aligned chain (Canton or a Quorum-derived ledger) for HK and Singapore wealth distributors that prefer a permissioned settlement venue, plus potentially Solana for DeFi-curious institutional buyers. Do not commit to multi-chain at launch.
Issuance. Securitize for US-perimeter compliance and the SEC-registered transfer-agent stack on the Reg S-accessible share class, plus Marketnode for the Singapore market-infrastructure positioning into Project Guardian-aligned distribution. The dual-platform pattern adds operational complexity but matches the dual-jurisdictional distribution thesis.
Custody (underlying). Existing BNY relationship continues. The off-chain T-bill, repo, and cash custody is set at fund formation and the tokenised wrapper inherits it. No change.
Custody (tokens). Anchorage for institutional US-perimeter holders (federally chartered, the cleanest counterparty profile for a US-domiciled buyer holding the Reg S share class through an offshore vehicle), plus Coinbase Custody for HK and Singapore distributor integration where the distributor's existing custody relationships favour a state-trust-chartered counterparty.
Distribution. DBS Private Bank, HSBC Private Bank, Standard Chartered Wealth Management for APAC private-bank channels. Defer DeFi distribution to phase 2; the institutional channel commitment is the gating decision.
Subscribe and redeem. T+0 to fiat for institutional holders with confirmed banking-rail coverage in both jurisdictions; on-chain redemption to USDC for institutional holders only (the Circle swap pattern adjacent to the BUIDL share class, see Circle). Primary-only secondary; no dedicated secondary venue.
Reporting. Daily NAV publication; weekly portfolio holdings; monthly Securitize attestation on share-register integrity; annual audited financial statements on the conventional cadence.

The 6-9 month rollout sequence: months 1-2 (wrapper documentation, Reg S overlay, tokenised share-class prospectus addendum, internal stakeholder alignment across compliance, risk, legal, distribution, fund administration); months 3-4 (Securitize and Marketnode contracting, Anchorage and Coinbase Custody onboarding, Ethereum mainnet contract deployment and audit); months 4-5 (private-bank distributor integration, KYC and whitelist provisioning, banking-rail confirmation for T+0 redemption); months 5-6 (controlled launch with anchor institutional buyer, daily NAV and weekly holdings reporting cadence); months 7-9 (broader distributor onboarding, evaluation of phase-2 chain expansion). The named gating decisions: the Reg S overlay sign-off (compliance), the Securitize-and-Marketnode contracting decision (commercial), the anchor-buyer commitment (distribution), the banking-rail confirmation (operations).

Red flags

Tokenising the prime MMF first because it is the firm's biggest. The credit-risk underlying complexity is structural, not marketing; it compounds the tokenisation-specific operational risk on top of the credit-risk diligence pack. Sequence Treasury or Government first.
Going multi-chain at launch. The cross-chain whitelist administration is non-trivial engineering and the operational reconciliation surface compounds with each additional chain. BUIDL and FOBXX both ran single-chain for 12 to 18 months before extending. Pick one chain first.
Skipping the bank-grade underlying custodian to simplify the stack. The distributor's risk team will reject in 100 percent of institutional diligence packs. The off-chain underlying custody is non-negotiable.
Promising T+0 redemption to fiat without confirmed banking-rail integration on weekends and holidays. The marketing line is honest only with continuous banking-rail coverage; without it, the published cadence is T+1 dressed up as T+0 and the first weekend-redemption request will surface the mismatch publicly.
Letting the platform partner own the customer relationship. The platform should provide infrastructure; the asset manager owns the customer. Pitches that propose platform-side white-labelling of the customer relationship turn the asset manager into a vendor of its own product.

[Suit Up]

Tokenising money-market funds in practice