Part 3 closed on the unit economics. This part answers two questions a TradFi or Web3 reader is likely to bring into the next conversation. First, why this cycle. The answer is not that tokenisation suddenly works; it is that three slow-moving pieces of plumbing have lined up at once. Second, what mistakes still derail atomic delivery-versus-payment (DvP) discussions when readers from different backgrounds meet. Most of the confusions are corrections to slogans that were never quite right.
Why it matters now
Three convergences make this the cycle in which atomic DvP graduates from pilot to production.
First, the prudential floor is set. The basel sco60 cryptoasset standard gives banks a defined capital regime for holding tokenised assets and their custody risk (Basel SCO60). Until 2023 the question "what does it cost to hold this on the balance sheet" was unanswerable; now it is, and capital cost can be priced into the DvP product. Internal credit committees that previously declined tokenised exposures on prudential grounds can now reference a known capital figure, which moves the decision from "no, in principle" to "yes, at this price".
Second, the settlement-layer experiments are converging. Project Agorá, Project Ensemble in hong kong, the Bank of Japan-led Payment Innovation Project sandbox in japan, and the Korean and Australian wholesale CBDC tracks all assume the same architectural pattern: tokenised asset ledger plus tokenised cash ledger plus coordinator. The pattern is no longer up for debate; the live questions are interoperability, governance, and which jurisdictions ship first. The BIS Annual Economic Report 2024 chapter on tokenisation is the cleanest articulation of why the tiered design is dominant, and the speech by Hyun Song Shin alongside the AER chapter is the operator-targeted version of the same argument (BIS speech).
Third, the production examples have multiplied. The April 2026 announcement that Nomura, Mizuho, and JSCC will trial tokenised JGB collateral on Canton with Digital Asset, with two of the three previously confirmed as Payment Innovation Project sandbox participants, is the latest in a sequence that started with Kinexys intraday repo and now spans most of the major Asian bond markets. The interesting analytical question for a 2026 desk is no longer "is atomic DvP real" but "which jurisdiction owns the standard for the cash leg". That is a much more useful question, and it has different answers in Hong Kong (the regulator-coordinated tiered ledger), Japan (mixed trust-issued and bank-issued routes), and Singapore (consortium-led with regulator oversight via Project Guardian).
Common confusions
"Atomic settlement on Ethereum is risk-free DvP." It is not. Atomicity in the cryptographic sense (the chain commits both legs or neither) does not give you settlement finality (the legal property that the trade cannot be unwound in insolvency or by court order). For finality you need a designated system, settled cash that the law treats as settled, and a recognised legal framework around the chain. A Uniswap-style atomic swap between two arbitrary tokens is atomic; whether it is "DvP" in the regulated sense depends on what the tokens legally represent and on whether the venue is a designated settlement system. Mostly it is not. The legal property bundle described in Tokenisation, defined is the prerequisite, not the cryptographic property.
"DvP requires a central counterparty." It does not. Bilateral DvP existed before central counterparties (CCPs) and continues to exist alongside them. CCPs are valuable for multilateral netting and default management, both concerns at the Model 3 end of the taxonomy. A pure Model 1 atomic DvP has no need of a CCP because there is no inter-trade netting and no surviving exposure. The CCP question is independent of the DvP question, and conflating the two is one of the easier ways to muddle a rulebook.
"Smart contracts make DvP free." They do not. They redistribute the cost. The legal infrastructure that makes the chain entry operative still has to exist (rulebook, designation, finality regime). Reconciliation between the chain and the participants' systems of record still has to be run. KYC, anti-money-laundering (AML), and sanctions screening at the boundary still cost money. What smart contracts remove is the manual sequencing and confirmation work in the back office. That is a real saving, but only one operational line out of many, and the savings on that line have to outweigh the new costs of running the chain.
"Atomic DvP solves the cross-border settlement problem." It solves the within-system version. Cross-border atomic DvP requires a shared settlement substrate spanning the jurisdictions involved, which is exactly what Project Agorá and Project Ensemble are working towards but which does not yet exist in any general-purpose form. The cross-border case in 2026 is still a portfolio of bilateral and small-multilateral arrangements, with the regulator-coordinated work pointing toward a more general substrate later in the decade.
"You cannot have atomic DvP without a wholesale CBDC." You can. Tokenised deposits on a shared ledger achieve it (the Project Ensemble tokenised deposit leg works this way, even before the wholesale CBDC layer is fully in place). Bank-issued or trust-issued stablecoins on the same ledger achieve it. Fnality's settlement coin achieves it under a hybrid characterisation. The wholesale CBDC option is one path, not the only one, and treating it as the only option ignores the live Asian programmes that are running today on tokenised commercial bank money. Chapter VIII compares the wholesale CBDC and tokenised deposit paths directly.
Part 5 closes the chapter with the production examples a 2026 operator will actually meet, from the Kinexys deposit-and-asset DvP through Project Mariana's wholesale CBDC AMM, the Eurex repo work, and the most recent Asian named pilots.