3.4 24/7 Markets and DeFi Composability
What continuous settlement removes from the post-trade process and what it costs in liquidity, why a token that trades every hour on an asset that trades five days a week creates a new basis risk, and why composability is the one RWA capability with no traditional equivalent.
The previous two sections found the access and liquidity claims to be narrower than advertised. This one is different. Continuous settlement and composability are the two benefits where tokenization does something that traditional market infrastructure cannot replicate by working harder, and the honest treatment is not to deflate them but to be precise about what they cost.
Section 3.3 closed on the distinction between redemption and secondary trading, and on a market whose settlement rails work better than its order books. This section examines those rails directly, then turns to the property that makes an on-chain asset more than a faster version of an off-chain one.
The Settlement Baseline
The comparison has to start from where traditional settlement actually is, not from a caricature of it.
The United States moved equities and most corporate bonds to T plus one settlement on 28 May 2024, and the European Union, United Kingdom, and Switzerland have coordinated a joint transition to T plus one on 11 October 2027, with ESMA's high-level roadmap setting out the operational path 1. Conventional markets are not standing still, and the gap tokenization is closing is one business day rather than the multi-day delay that older arguments assume.
What one day of delay actually costs is counterparty exposure. Between trade and settlement, each side carries the risk that the other fails to deliver, and the machinery that manages that exposure is margin, central clearing, and capital. That machinery is the target.
What Continuous Settlement Removes
Settlement on a shared ledger can be atomic, meaning the asset leg and the cash leg execute simultaneously or neither executes. The BIS classifies this as delivery-versus-payment model 1, gross settlement of both legs, and its defining property is the elimination of principal risk 2. Neither party can deliver and fail to receive.
This is a genuine structural improvement rather than a speed improvement. Principal risk is not reduced by settling faster; it is removed by settling simultaneously. Everything built to manage it, including some margin requirements and some clearing intermediation, becomes proportionally less necessary.
Continuous operation compounds the effect. A settlement system that runs at all hours lets a participant move collateral to meet a margin call at three in the morning, or move cash across a weekend, which in conventional infrastructure would wait for the next business day. For collateral management in particular this is the most concrete operational benefit tokenization offers.
What It Costs: The Netting Problem
The cost is rarely stated alongside the benefit, and it is substantial.
Bech, Hancock, Rice and Wadsworth set it out directly in the BIS Quarterly Review. Tokenized systems are likely to result in more trades settling via DvP model 1, and settling trades individually generally requires more liquidity 2. Netting, which allows a day's obligations to be offset so that only the net difference moves, reduces liquidity needs but is in general more complex in a decentralised environment 2.
This is a real trade-off and not a transitional inconvenience. Conventional clearing houses net enormous gross volumes down to small net obligations, and that netting is precisely what lets participants support large trading volumes with modest cash balances. Gross atomic settlement discards it. Every trade must be fully funded at the moment it executes, so the same activity requires more cash sitting ready.
The authors identify a second-order version of the same problem that speaks directly to today's fragmented market. If each tokenised securities ledger had its own cash token, the need to hold cash tokens on several ledgers could increase aggregate liquidity requirements 2. A participant active across five chains must pre-position settlement assets on all five. This is the same fragmentation IOSCO identified when it named interoperability and the lack of high-quality settlement assets as the structural obstacles to scaling 3, arriving from the liquidity-management side.
So continuous atomic settlement removes principal risk and raises funding requirements. Whether that is a good trade depends on the participant, and it is not the unambiguous win the benefit framing implies.
What the Cash Leg Actually Is
Atomic settlement requires both legs on the same ledger, which raises a question the benefit framing tends to skip: what is the cash leg made of?
IOSCO named the lack of high-quality settlement assets as one of the two structural obstacles to scaling tokenisation 3. That phrasing is doing careful work. A settlement asset is high quality when it settles with finality and carries no meaningful credit or liquidity risk, which is why central bank money sits at the top of the hierarchy and why wholesale CBDC experiments exist at all.
Today most on-chain settlement happens in stablecoins, which are claims on a private issuer rather than central bank money. That is a workable settlement asset and a different one, and the difference is the subject of this paper's companion volume on stablecoins. For RWA settlement the practical consequence is that atomic DvP delivers finality against an issuer's promise, so the principal risk removed from the trade reappears, smaller and relocated, in the settlement asset itself.
This also sharpens the fragmentation problem. The BIS concern about holding cash tokens across several ledgers 2 is not only about liquidity efficiency. Each additional settlement asset a participant must hold is an additional issuer to assess. A market with one high-quality settlement asset on one ledger would be both cheaper to fund and simpler to analyse than the current arrangement, and neither condition holds.
The Asset Does Not Trade 24/7
There is a subtler problem that appears only once tokenization reaches assets with their own trading hours.
A tokenized Treasury fund or a tokenized equity can change hands at any hour. The underlying instrument cannot. Equity markets close on weeknights and close entirely at weekends, and a fund's net asset value is struck on a schedule. The token's market is continuous, the asset's market is not, and the difference between them is a basis that did not previously exist.
This is now a live category rather than a hypothetical. Tokenized stocks held $486.69 million in market capitalisation at the end of Q1 2026 with $15.12 billion of quarterly spot volume 4, which means substantial trading occurs against an underlying that is frequently closed. When the reference market is shut, on-chain price discovery is the only price discovery, and it runs on whatever liquidity is present at that hour.
The dependency this creates is the oracle dependency from Section 2.4. A contract that liquidates collateral, or a protocol that values a position, acts on the price an oracle reports. If that price is stale because the source market is closed, the contract acts on facts that no longer hold, and it does so automatically. The FSB warns that reliance on unregulated or non-compliant oracles can hinder accurate assessment of token prices and asset quality 5, and a closed reference market is the condition under which that warning has the most force.
Continuous markets are therefore straightforwardly good for assets that genuinely trade continuously, and they introduce a new risk surface for assets that do not. Section 5.1 treats the oracle version of this problem in full.
Composability Is the Genuinely New Capability
Composability is the property with no traditional analogue, and it is where the strongest case for tokenization sits.
Because a tokenized asset is a contract on a shared ledger, other contracts can use it without asking permission or building an integration. A tokenized money market fund can be posted as collateral, wrapped, lent against, or embedded in another product, and the holder keeps earning the underlying yield while the position does other work. In conventional infrastructure each of those uses requires a bilateral arrangement, a custodial agreement, and an operational build.
This has moved from argument to practice. In June 2025 Securitize announced that BlackRock's BUIDL would be accepted as collateral on the Crypto.com Exchange for qualified institutional clients and on Deribit for institutional futures and options trading 6. Securitize's chief executive described BUIDL as evolving from a yield-bearing token into a core component of crypto market infrastructure 6.
The economic content of that is worth stating plainly. An institution posting Treasuries as derivatives margin has historically forfeited the convenience of cash or accepted operational friction to avoid it. A tokenized Treasury fund used as margin earns its yield and satisfies the margin requirement simultaneously, and can be moved at any hour. That is capital efficiency of a kind the conventional stack cannot easily reproduce, and it explains why tokenized Treasuries grew into the largest RWA category rather than tokenized real estate.
It also explains the concentration Section 3.1 documented. The tokenized assets that succeeded are the ones whose value increases when they become composable, which means high-quality, fungible, yield-bearing collateral. That is not a coincidence of early adoption. It is the category where composability pays.
Composability Transmits Stress
The same property that makes collateral efficient makes exposures travel.
When a tokenized fund is collateral inside several lending markets and exchanges at once, a problem in the fund becomes a problem in all of them simultaneously and automatically. There is no intermediary to pause, no operational delay that functions as a circuit breaker, and no requirement that anyone notice. The FSB's assessment of tokenisation identifies exactly this concern, and the BIS Financial Stability Institute attributes the currently limited financial stability risk in part to low interconnectedness 5 7, which is a description of a market that has not yet composed itself very much.
The reasoning runs in an uncomfortable direction. Composability is the main benefit, interconnectedness is the same fact viewed from the supervisor's chair, and the reason the risk is tolerable today is that the benefit has not yet been realised at scale. If composability delivers what its advocates expect, the low-interconnectedness condition that makes regulators comfortable stops holding.
That is not an argument against it. It is the reason Chapter 5 treats smart contract and composability risk as first-order rather than incidental, and the reason the regulatory question in Chapter 6 is about how these markets are supervised rather than whether they are permitted.
- The settlement baseline is one business day, not several. The US moved to T plus one on 28 May 2024 and the EU, UK, and Switzerland follow on 11 October 2027, so tokenization competes against improving conventional infrastructure.
- Atomic settlement removes principal risk structurally rather than reducing it, because both legs execute simultaneously. This is a real improvement that speed alone cannot deliver.
- The cost is netting. BIS research finds gross settlement generally requires more liquidity than netted settlement, and holding separate cash tokens across multiple ledgers could raise aggregate liquidity requirements further, which makes fragmentation a funding problem as well as a liquidity one.
- A token that trades continuously on an asset that does not creates a new basis and a stale-oracle risk. Tokenized stocks traded $15.12 billion in Q1 2026 against a $486.69 million market cap, much of it while the reference market was closed.
- Composability is the genuinely novel capability, demonstrated by BUIDL being accepted as collateral on Crypto.com and Deribit in June 2025, letting a position earn yield and serve as margin at once. It is also interconnectedness, which is what supervisors cite as currently limiting systemic risk.