Suit Up · Asia's institutional desk on tokenisation

TL;DR

When a bank holds, settles against, or extends credit using a tokenised asset, the capital cost is gated by which Basel SCO60 bucket the asset lands in. Group 1a means broadly the same capital treatment as the underlying: a tokenised bond is treated as a bond, a tokenised MMF (money-market fund) as a fund interest. Group 1b inherits the underlying treatment plus an add-on risk weight reflecting gaps in the redemption mechanism, issuer supervision, or technology infrastructure. Group 2 is sharply higher, with an exposure cap (the headline is one percent of a bank's Tier 1 capital, with conditions under which it can rise to two percent) that effectively forecloses meaningful balance-sheet inventory in unbacked crypto. The classification is gating: it determines whether an activity is economically viable on a bank balance sheet at all. The Basel Committee endorsed a targeted review of SCO60 at the March 2026 GHOS meeting (see the regulation page); the review is the principal Basel-side process to track over the next 12 to 18 months.

Decision frame

SCO60 does not classify crypto-assets by their technology (public versus permissioned, EVM versus Move, settled on Ethereum versus settled on Canton). It classifies them by a structured set of conditions about the underlying asset, the redemption mechanism, the supervisory framework over the issuer, and the technology infrastructure. The same on-chain instrument can fall into different buckets depending on how those four pillars are met.

The decision frame for a bank evaluating a tokenised exposure runs in four cuts. What is the underlying: a claim on a regulated security, on regulated bank money, on a basket or commodity, or on an unbacked digital asset. What is the redemption mechanism: delivery of the underlying at par on a defined cadence, or something weaker. Who is the issuer and under what supervision: a federally chartered bank, an SEC-registered transfer agent paired with a broker-dealer, a fund manager licensed under FIEA or SFA, an unsupervised offshore vehicle. What is the technology infrastructure: a permissioned ledger with named participants and a documented operator, a public ledger with a credible operator and explicit governance, or a public ledger with no operator at all.

Most institutional readers know SCO60 exists. The playbook below helps decide which bucket a specific structure lands in and what the structural levers are if it lands in the wrong one.

The three Group classifications

Group 1a: tokenised traditional assets meeting all conditions

Group 1a covers tokenised representations of traditional assets that meet the standard's full conditions: the on-chain instrument represents a claim on the underlying asset itself; redemption delivers the underlying at full face value on a defined cadence; the issuer is supervised under a recognised framework that imposes prudential, conduct, or fund-management discipline; the technology infrastructure is robust enough to satisfy a supervisor; and the legal framework gives the on-chain entry operative force, per the legal-control discussion in Chapter I, Part 3.

Capital treatment is broadly the same as the underlying. A tokenised US Treasury bill is treated as a Treasury bill, a tokenised investment-grade bond as that bond, a tokenised MMF interest as a fund interest. There is no tokenisation tax: the chain is the recordkeeping rail, not a separate asset class.

Two worked examples sit cleanly here. BlackRock's BUIDL (BlackRock USD Institutional Digital Liquidity Fund) is a 3(c)(7) Delaware LLC operating as a regulated MMF, with Securitize as the SEC-registered transfer agent, BNY as fund administrator and custodian of the underlying T-bill, repo, and cash sleeve, and contract-level whitelist gating administered by the transfer agent. Underlying regulated, issuer and platform supervised, on-chain entry operative for SEC transfer-agency purposes, redemption at par. A bank holding BUIDL holds it at the capital cost of the underlying MMF interest, subject to its national supervisor confirming the SCO60 classification. HSBC's Orion tokenised-bond programme is the second example: a debt security on a permissioned ledger run by HSBC with named participants, the on-chain entry placed as the operative register for transfer. A bank holding an Orion-issued bond holds it at the same capital cost as the equivalent off-chain bond.

Group 1b: tokenised arrangements meeting weaker conditions

Group 1b covers tokenised arrangements that satisfy the broad shape of Group 1 (a credible underlying, an identifiable issuer, an enforceable claim) but where redemption is weaker, issuer supervision is partial, or the technology infrastructure has identified gaps. Capital treatment is the underlying-asset treatment plus an additional risk weight reflecting the structural weaknesses. The calibration runs through the standard's text and through national-supervisor implementation; the operating point is that there is a real capital-cost premium relative to Group 1a, even where the asset is structurally close.

Two examples illustrate. Pre-GENIUS-Act stablecoins held by US banks (before the GENIUS Act established the federal payment-stablecoin issuer perimeter) sat in this bucket: issuer supervision was state-level and patchy, redemption depended on issuer discretion in stress, and disclosure cadence was below what a federally supervised regime would require. They were not Group 2 (identifiable reserves, identifiable issuers) but they did not clear the Group 1a test. Bilateral tokenisation pilots in jurisdictions where the supervisory framework over the issuer was less explicit than the major regimes (a sandbox-only pilot in a jurisdiction without finalised rules, a tokenised arrangement issued by an unregulated SPV with a credible underlying) often sit similarly. The structural lesson: issuer supervision and the technology-infrastructure controls are the two most common reasons a structure falls from 1a into 1b.

Group 2: unbacked crypto and structures failing both Group 1 tests

Group 2 covers unbacked crypto and tokenised arrangements that fail the Group 1 conditions outright. Capital treatment is sharply higher, with an exposure cap typically set at one percent of a bank's Tier 1 capital and conditions under which it can rise to two percent. The cap is a hard ceiling: a bank approaching the limit cannot add inventory regardless of risk-weight optimisation.

This is where bitcoin, ether, and most pre-stablecoin-regime crypto-assets land. Almost no institutional tokenisation programme intends to fall here. The structural significance is that Group 2 effectively forecloses meaningful balance-sheet exposure to unbacked crypto. Banks running client-facing crypto businesses do so through agency arrangements, custody that does not constitute principal exposure, or specific approval pathways outside the inventory model. The cap defines the institutional treatment of unbacked crypto.

Implications for product design

If you are designing a tokenisation product that will be held by a bank counterparty, the goal is Group 1a treatment. Five structural levers, each pointed at one of the SCO60 conditions.

Pick a regulated underlying. Treasuries, investment-grade bonds, MMF interests, regulated bank deposits. The underlying-asset choice is the largest single determinant of which bucket the structure can land in. An unregulated underlying (an unbacked digital asset, a synthetic exposure, a non-fund collective vehicle) forecloses Group 1a regardless of how well the other conditions are met. The cross-cutting reference is asset-class regulatory treatment.

Ensure issuer supervision is explicit and recognisable. A federally chartered bank under OCC supervision, an SEC-registered transfer agent paired with a registered broker-dealer, a fund manager licensed under FIEA in Japan or SFA in Singapore, an MAS Major Payment Institution, or an HKMA-licensed stablecoin issuer (such as HSBC under the Stablecoins Ordinance). The supervisory framework needs to be one a Basel-implementing supervisor will recognise without argument. Sandbox participation is not a substitute.

Build redemption-at-par mechanics on a defined cadence. T+0 or T+1 redemption against the underlying at par is the design pattern Group 1a is built around. Redemption-at-market, redemption with a discretionary fee, or redemption at issuer convenience are weaker structures that pull the arrangement toward Group 1b. The Hong Kong Ordinance's one-business-day statutory floor is the kind of constraint a Group 1a-aspiring structure should meet without strain.

Pick a permissioned ledger or a public ledger with explicit operator and governance. The technology-infrastructure condition turns on documenting the controls a supervisor expects: who operates the ledger, how upgrades and incident response are governed, the access-control model, key management, dispute resolution. A permissioned ledger (Canton, Project Ensemble participants, Kinexys) makes this tractable. A public ledger with an explicit operator (a supervised transfer agent administering the allowlist, as with BUIDL) can satisfy it too, though the supervisory conversation is longer.

Have a clear legal opinion that the chain entry is the operative record. The legal-control test in Chapter I, Part 3 is the load-bearing legal piece. The convergence pattern across major jurisdictions is toward a control or possession-equivalent test (UCC Article 12 in the US, the third-category-of-personal-property doctrine in England, the equivalent practice under Singapore SFA and trust law), but codification timing and certainty are not aligned, and an issuer running the same structure across four jurisdictions will need four opinions.

The targeted review and what to track

The Basel Committee endorsed a targeted review of SCO60 at the March 2026 GHOS meeting and is expediting the work per the November 2025 Mexico City meeting. The substantive direction is not yet public. Four areas are credibly on the table.

Calibration around the Group 1a and 1b boundary. The current text places it at a structured set of conditions, so a marginal failure on one condition pushes the structure from 1a to 1b. Whether the review tightens, loosens, or restructures that boundary is the single most consequential open question for bank-side product design.

The Group 2 exposure cap. The one-percent-of-Tier-1 headline (with conditions for two percent) is the ceiling that defines bank-side treatment of unbacked crypto. Whether the cap or its tier conditions are recalibrated is a second-order question for institutional tokenisation programmes but a first-order one for any bank running a meaningful unbacked-crypto franchise.

Treatment of public-permissionless infrastructure. The standard does not formally exclude public-permissionless ledgers from Group 1a, but the technology-infrastructure conditions are easier to satisfy on a permissioned ledger. Whether the review issues guidance clarifying how a public-permissionless ledger can satisfy the conditions, or whether the practical effect is to keep institutional issuance on permissioned rails, is the third area to track.

Credit-risk weights for novel tokenisation structures. Tokenised private credit, tokenised structured products, and tokenised composability arrangements (bank-deposit-backed tokens used as collateral in institutional-DeFi venues; see the institutional composability paradox) sit at the edge of the existing classification. Whether the review adds calibration for novel structures rather than leaving them to be back-solved is the fourth.

Worked example

A G-SIB Asia-Pacific treasury team is evaluating three tokenised-asset additions to its inventory: BUIDL for the trading book, JPYC for FX settlement on its Tokyo-Singapore corridor, and direct bitcoin for a client-accommodation flow. The team needs SCO60 classification of each before any clears the new-product committee.

BUIDL. The underlying is a regulated MMF (a 3(c)(7) Delaware LLC with the T-bill, repo, and cash sleeve held by BNY in conventional custody). The issuer is supervised: BlackRock as fund manager, Securitize as SEC-registered transfer agent. Technology infrastructure is a public chain (Ethereum mainnet plus multi-chain extension) with contract-level whitelist gating administered by the transfer agent. Legal control is operative: under SEC transfer-agency rules, the on-chain transfer of a BUIDL token is the change of legal ownership. Redemption is at par on the institutional cadence the fund offers. The classification analysis points cleanly toward Group 1a, subject to home-supervisor confirmation. The strategic conclusion: the addition is economically viable, with no tokenisation premium over the off-chain MMF position the bank could otherwise hold.

JPYC. The underlying is a yen-denominated stablecoin with 1:1 reserves of cash deposits and Japanese government bonds. The issuer is JPYC Inc., registered as a fund-transfer service provider (FTSP) under the Japan PSA, supervised by the FSA (see Japan: JPYC's FTSP route). Technology infrastructure is a public chain with issuer-administered allowlist controls. Redemption is at par, mediated by FTSP statutory mechanics. The classification analysis is more nuanced. The supervisory framework is real and recognisable (the FSA-supervised FTSP route is the cleanest non-bank, non-trust path under Japanese law) but structurally lighter than the bank-direct or trust routes, and the historical FTSP per-transaction cap has not been tested at institutional payment scale. The team would likely land on Group 1a if the home supervisor accepts FTSP supervision as sufficient, and on Group 1b if the supervisor takes a stricter view. The strategic conclusion: the addition is workable but carries supervisor-discretion risk; validate with the national supervisor before sizing inventory.

Direct bitcoin. The underlying is unbacked. There is no issuer. The classification is Group 2, with the one-percent-of-Tier-1 exposure cap. The strategic conclusion: direct bitcoin inventory is not an economically viable balance-sheet position for a G-SIB at any meaningful scale. The client-accommodation flow has to run through agency or custody structures that do not constitute principal exposure, or through a separately approved business line sized within the cap. The Group 2 treatment is the operative reason institutional bitcoin businesses run as agency or custody franchises rather than principal-inventory desks.

Synthesis: BUIDL clears at Group 1a and is added at the same capital cost as an off-chain MMF position. JPYC clears at Group 1a or 1b depending on supervisor view; size contingent on confirmation. Direct bitcoin does not clear as principal inventory; restructure to run through the custody franchise. The Basel classification is the gating prudential filter, and the product-design conversation downstream is conditional on it.

Red flags

Assuming "tokenised" automatically grants Group 1a. The conditions are detailed and a marginal failure on any one of them pulls the structure into Group 1b. The on-chain wrapper does not, by itself, do the Basel work.
Treating sandbox participation as evidence of supervisory framework. A sandbox is a permissioned environment for testing under regulatory observation; it is not authorisation. An issuer whose only credentials are sandbox participation does not satisfy the Group 1a issuer-supervision condition.
Ignoring jurisdictional implementation variation. SCO60 is the global standard; the implementation is national, and the timing and texture vary across supervisors. A structure that satisfies the standard in concept may be classified differently by different national supervisors. The bank's own supervisor's view is the operative one for the bank's capital calculation.
Building product structures that pass SCO60 conditions in concept but fail on the technology-infrastructure controls in practice. The condition turns on documented operator identity, governance over upgrades and incidents, key management, and access-control models. A structure that ticks the conceptual box but cannot pass a supervisor's operational examination of the controls will not stay in Group 1a.
Pricing tokenised products without modelling the capital cost on the bank counterparty. A product that lands in Group 1b on the bank counterparty's books is materially more expensive for the bank to hold than a Group 1a equivalent, and the bid the bank can support reflects that. Asset managers presenting tokenised products to bank counterparties without a Basel-classification view are leaving the pricing conversation incomplete.

[Suit Up]

Capital treatment of tokenised assets in practice