Suit Up · Asia's institutional desk on tokenisation

TL;DR

Tokenised private credit sits where institutional credit asset management meets DeFi-native distribution, and as of early 2026 it is the dominant on-chain RWA (real-world asset) category by AUM (per the onchain credit market theme). The architectural question is mostly settled: the SPV-plus-tokenisation pattern, where an off-chain SPV (special-purpose vehicle) holds the underlying credit and an on-chain wrapper exposes the SPV's beneficial interest as a tradeable token, has become the dominant production model. The Centrifuge V3 + Aave Horizon + Janus Henderson JAAA stack is the canonical worked example on the public-chain side; Apollo's ACRED on Securitize is the canonical analogue on the US-perimeter regulated side. The decision tree turns on three questions: distribution intent (DeFi-native versus institutional-only), the underlying credit mandate (AAA CLO tranche, direct lending, asset-backed, distressed, mezz, each with different liquidity and valuation profiles), and how much DeFi composability you are willing to absorb on the asset's behalf.

Decision frame

The asset manager has a private credit franchise (direct lending, asset-backed finance, mezzanine, distressed, AAA CLO tranches, structured credit, or some combination) and is choosing how to ship a tokenised version. Before any platform conversation, the framing has to be right.

Several wrong framings show up routinely in tokenisation conversations and are worth eliminating up front. "DeFi cannot handle credit" is wrong: the JAAA programme on Centrifuge has cleared USD 1bn TVL on a tokenised CLO product, and Apollo's ACRED has brought a top-five global private-credit franchise on-chain through Securitize. Both prove the structure works for specific configurations. "We will just put the existing fund on-chain" is wrong: the SPV-plus-tokenisation pattern is intentional architecture, not an accident of how the first products happened to launch. The off-chain SPV is doing real work (holding the credit, contracting with the asset manager, providing the legal claim that the on-chain token references), and skipping that layer collapses the regulatory and operational discipline that makes the wrapper viable. "Composability is free upside" is wrong: composability with permissionless DeFi protocols introduces tail risk (oracle failure, lender-pool insolvency, bridge compromise, smart-contract exploit on the borrowing venue) that you may not want to absorb on behalf of your underlying credit position. "Tokenisation will solve our distribution problem" is wrong: it changes your distribution problem. Whether that change is an improvement depends on your starting position, your existing institutional channels, and whether DeFi-native demand is incremental or substitutional for your existing buyer base.

The right starting point is the distribution intent. If the goal is incremental DeFi-native demand (Sky / Grove allocator deployments, prime-broker collateral demand against composable wrappers, stablecoin-issuer reserve allocations, allocator-side deployments from on-chain treasuries), the architecture leans toward public-chain native (Centrifuge V3 plus Aave Horizon plus a multi-chain footprint that reaches the destination chains where institutional DeFi liquidity concentrates). If the goal is institutional-only distribution with the on-chain wrapper as an operational efficiency rather than a distribution channel (faster settlement, programmable cap-table, T+0 secondary in a permissioned venue), the architecture leans toward US-perimeter regulated (Securitize as transfer agent, broker-dealer, ATS bundle; whitelisted secondary inside Securitize Markets or an equivalent venue). Both are live; the choice is structural, not aesthetic.

The eight design decisions

The legal wrapper

Cayman SPC (segregated portfolio company), Delaware LLC, Luxembourg SCS (société en commandite spéciale), Irish ICAV (Irish collective asset-management vehicle), Singapore VCC (variable capital company). Each has different SPV-and-token coupling implications and different downstream consequences for distribution perimeter and tax treatment. Cayman SPC dominates for offshore credit funds where the buyer base is global qualified-purchaser and the credit underlying is dollar-denominated; the SPC structure lets a single legal entity host multiple segregated cells, each backing a different on-chain wrapper. Delaware LLC fits where US-perimeter SEC-registered transfer-agent integration is the priority and the buyer base sits inside the US §3(c)(7) qualified-purchaser perimeter. Luxembourg SCS and Irish ICAV fit when European institutional distribution is in scope, with the AIFMD passport as the operative regulatory perimeter. Singapore VCC fits when MAS-perimeter institutional distribution is the lead and the credit underlying has APAC concentration. The wrapper hosts the credit underlying; the on-chain token exposes the SPV's interest. The wrapper choice precedes the platform choice, not the other way around.

SPV-plus-tokenisation architecture

The structural primitive across Centrifuge, Securitize, Provenance, and adjacent platforms. The off-chain SPV holds the loans, notes, or credit positions, contracts with the asset manager that runs the strategy, and provides the legal claim against which the on-chain token is exposed. The on-chain wrapper is a token representing a beneficial interest in the SPV; the legal documentation in the SPV's offering memorandum maps the on-chain holder's rights to the off-chain underlying. The clean separation of credit administration (off-chain, conventional fund-admin and trustee operations) from distribution and secondary trading (on-chain, programmable, composable where the wrapper allows it) is the design that makes the architecture work at institutional scale. Skip this layer at your peril; the alternatives (direct on-chain credit issuance without an SPV, or representations of an off-chain register where the off-chain register controls in any disputed case) collapse either the regulatory perimeter or the structural reason to tokenise in the first place.

Issuance platform

Three serious counterparties, each with a different centre of gravity. Centrifuge is public-chain-native, with the SPV-plus-tokenisation pattern as core architecture and the JAAA / Janus Henderson / Aave Horizon stack as the canonical reference. V3 completed its EVM (Ethereum Virtual Machine) migration in April 2026, putting the protocol natively across nine ledgers including Ethereum, Base, Arbitrum, Optimism, Polygon, Avalanche, BNB Chain, Solana via wrapped representations, and the original Centrifuge Chain. Securitize is US-perimeter, with the SEC-registered transfer-agent plus broker-dealer plus ATS (alternative trading system) bundle in one operating company; Apollo's ACRED is the canonical credit-side reference, with the Provenance and Solana deployments as the chain footprint. Maple Finance is DeFi-native, with institutional lending pools (syrupUSDC and syrupUSDT at meaningful scale on the productive-collateral side) but without the traditional asset-manager partnership pattern that Centrifuge and Securitize have built. See Evaluating issuance platforms for the full evaluation frame; for credit specifically, the platform shortlist falls out of the distribution intent rather than from generic platform features.

Chain choice

Public chain for DeFi composability (Ethereum mainnet plus the major L2s plus Solana for the deepest institutional DeFi liquidity) versus permissioned chain for closed-loop institutional distribution (Provenance, Canton, the Kinexys EVM rail, Singapore's GL1-aligned permissioned ledgers). The composability decision is binary in practice: a token issued on a permissioned chain cannot be borrowed against on Aave Horizon, used as collateral by Sky / Grove, or reach the broader DeFi-native allocator base; a token issued on a public chain without whitelist enforcement cannot meet the regulatory perimeter that most credit-fund wrappers require. The hybrid (issue on public chain with whitelist gating at the contract level) is the dominant production pattern for credit-fund tokenisation today, and it is what both Centrifuge V3 and Securitize's Vault Registrar architecture are designed to support.

Distribution intent

Two patterns dominate. DeFi-native: the token trades on a public lending venue (Aave Horizon, Morpho's curated institutional markets, Euler v2), gets used as collateral against stablecoin borrowing, and gets bought by allocators like Sky / Grove under governance-approved deployment envelopes. Institutional-only: the token transfers are restricted to whitelisted addresses, primary issuance and redemption clear through the platform's regulated venue, and secondary is whitelisted-only on a venue like Securitize Markets. The hybrid (whitelisted institutional perimeter with optional borrowing-against on a venue like Aave Horizon's institutional-collateral lane) is structurally available but operationally non-trivial; it requires the whitelist administration to extend across the lender venue, which is what Aave Horizon's institutional-collateral lane was built to support.

Liquidity and redemption design

Private credit does not have natural secondary liquidity. Tokenisation does not change that; it changes the form of illiquidity. Pre-defined redemption windows (quarterly is conventional for direct-lending and asset-backed credit funds, annually for distressed and mezzanine, daily is structurally inappropriate for any private-credit underlying) work in the tokenised context the same as the off-chain context. The temptation to promise daily redemption against a quarterly-liquidity underlying is the most common design error and the one that surfaces in stress. The on-chain wrapper can provide intraday secondary (a holder selling the token to another whitelisted holder on Securitize Markets, or borrowing-against on Aave Horizon as a substitute for outright redemption) but cannot manufacture liquidity in the underlying credit position. Document the redemption mechanics in the offering memorandum at the cadence the underlying actually supports; price the on-chain liquidity primitives accordingly.

Custody

Two layers, two counterparty sets. Bank-grade custodian for the underlying credit position (BNY, State Street, Northern Trust, the conventional fund-administrator-and-custodian counterparties); the asset class is no different from a non-tokenised private-credit fund at this layer. On-chain custody for the token: Anchorage Digital Bank as the OCC-chartered specialist, BNY Digital Assets as the GSIB option, Securitize-as-platform-with-custody for mandates inside the Securitize perimeter, or self-custody for institutional buyers with their own qualified-custody arrangements. The on-chain custodian needs whitelist consistency with the issuer's transfer-agent recordkeeping; the integration handshake is where most pitches fall apart in production. See Evaluating custody providers for the role-by-role evaluation frame.

Reporting and disclosure

Quarterly NAV (net asset value) is more typical for private credit than the daily NAV that tokenised MMF programmes have normalised; the underlying valuation does not refresh at higher cadence than the credit administration supports. On-chain attestation of underlying composition (Merkle-tree-based attestation of the SPV's holdings, refreshed on the same cadence as the NAV) is the institutional discipline that distinguishes credible tokenised credit programmes from marketing-led ones. The transparency question is whether the disclosure is restricted to whitelisted institutional holders (the conventional offering-memorandum-confidential pattern) or extended to permissionless DeFi users (which becomes structurally relevant when the token is composable into a DeFi venue where non-whitelist participants can read the on-chain attestation). The hybrid (institutional disclosure to whitelisted holders, summary-level on-chain attestation visible to any reader) is the dominant pattern emerging through 2025-2026.

Worked example: APAC asset manager launching a tokenised AAA CLO equity-tranche product

An APAC-headquartered asset manager (or the APAC arm of a global manager) is launching a tokenised AAA CLO (collateralised loan obligation) equity-tranche product, targeting both institutional buyers (allocator demand from Sky / Grove and adjacent DeFi-native stablecoin issuers, prime-broker demand for the wrapper as collateral against repo or margin facilities) and DeFi-native distribution (Aave Horizon listing as institutional collateral). The eight design decisions resolve as follows.

Cayman SPC for the legal wrapper, with each segregated cell backing a different on-chain wrapper if the manager runs multiple credit strategies through the same legal entity. The SPC choice keeps the offshore qualified-purchaser distribution perimeter clean and lets the platform layer remain agnostic to the specific cell. Centrifuge V3 as the issuance platform, following the JAAA pattern: the platform handles the protocol layer (issuance, redemption, transfer mechanics, cross-chain bridging), the SPV handles the legal wrapper, and the asset manager runs the underlying CLO strategy. Ethereum mainnet plus Solana for the chain footprint, with extension to Avalanche if the Sky / Grove allocator deployment thread becomes operational on the same template as JAAA. Janus Henderson is not a literal partner here, but the JAAA template is the structural reference for how an asset manager partners into a public-chain protocol stack rather than building proprietary issuance rails.

Aave Horizon as the named DeFi-native distribution venue, with the institutional-collateral lane handling the whitelist gating at the lane level rather than the wrapper level. BNY as the conventional custodian for the underlying CLO position, leveraging the existing custody relationship the asset manager almost certainly already has on the off-chain fund side. Anchorage Digital Bank as the on-chain custodian for the tokenised wrapper, with the Securitize-style transfer-agent handshake replaced by Centrifuge's institutional-protocol equivalent. Quarterly NAV from the fund administrator, with monthly portfolio attestation on-chain (Merkle-tree-based, refreshed on the same cadence as the underlying CLO loan-tape).

A workable phased rollout runs 6-12 months. Months 1-3: legal-wrapper standup (Cayman SPC, offering memorandum, SPV documentation, qualified-purchaser distribution architecture). Months 3-6: platform integration with Centrifuge V3, on-chain contract deployment, whitelist administration setup, custodian onboarding (BNY for the underlying, Anchorage for the on-chain token). Months 6-9: primary issuance to a small named allocator base (anchor allocator commitments from the asset manager's existing institutional channel; initial Sky / Grove conversation if the structural fit is right). Months 9-12: Aave Horizon listing application, prime-broker bilateral acceptance for collateral use, secondary expansion of the allocator base. The full rollout is faster than a conventional fund launch because the regulatory wrapper is offshore and the platform integration is non-trivial but not greenfield.

Red flags

Going DeFi-native distribution without an established allocator demand source. You ship the wrapper, the token sits at zero TVL, and the whole programme reads as a marketing exercise. Sky / Grove demand has to be either committed in advance or structurally probable; absent that, an institutional-only first-phase rollout is the cleaner path.
Promising daily redemption against a quarterly-liquidity underlying. The most common design error in private-credit tokenisation, and the one that surfaces in stress when the on-chain holder expects the wrapper to provide liquidity the underlying cannot. Document the redemption cadence at what the underlying actually supports.
Using a permissionless chain without whitelist enforcement on a private-credit token. The regulatory perimeter for credit-fund interests requires gating to qualified-purchaser or accredited-investor categories in most jurisdictions; a permissionless token without contract-level whitelist enforcement is a regulatory perimeter risk that surfaces in the first regulator inquiry, not in steady-state operations.
Letting the platform partner own the credit administration. The asset manager's franchise is the credit underwriting and the credit servicing; if the platform partner ends up doing the loan-tape attestation, the cashflow administration, or the credit-event handling, the manager has become a vendor for its own credit franchise. The right division of labour keeps the credit administration with the asset manager and the on-chain plumbing with the platform.
Treating tokenisation as a marketing exercise without restructuring the operational handoffs. The existing fund admin (the off-chain accounting, NAV calculation, investor reporting) and the tokenisation platform (the on-chain registry, the primary issuance and redemption rails, the whitelist administration) must integrate at the data-flow level. Pitches that describe tokenisation as a parallel rail without addressing the operational handoff are missing the load-bearing piece.

[Suit Up]

Tokenising private credit in practice