The first three parts established why TradFi cares and how the architectural fork breaks down. This part takes the capability layer. The four headline unlocks (fractionalisation, collateral mobility, programmability, cross-border settlement compression) are the lines on the slide deck that get an institutional programme funded. Each is grounded in production today, and each has visible limits the marketing material elides.
Fractionalisation
A private-credit fund interest is typically structured as a limited-partner commitment with a six- or seven-figure minimum. A commercial real-estate equity stake is bought in slabs that exclude most of the addressable allocator base. Tokenisation does not eliminate the regulatory wrappers that drive these minimums (qualified purchaser tests, accredited investor rules, professional-investor regimes), but it lowers the ticket size at the technical level, which lets issuers segment the same underlying across new wrappers more cheaply.
BUIDL is structured as a section 3(c)(7) private fund with a USD 5 million subscription minimum, which keeps it firmly institutional. Ondo's USDY sits adjacent as the same exposure repackaged for non-US accredited investors with smaller tickets. Franklin Templeton's FOBXX is a 1940 Act registered government MMF (money-market fund), which admits a wider US investor base than BUIDL with on-chain bookkeeping handling the transfer-agent function. The pattern: the same underlying exposure (T-bills plus repo plus cash) gets segmented across multiple tokenised wrappers, each pitched at a different allocator profile.
The same logic is starting to apply in on-chain credit, where a private-credit allocation that previously required a multi-million-dollar LP ticket is being repackaged via Centrifuge and similar platforms into smaller-denomination tokenised exposures. The legal wrappers still gate access. What changes is the cost of issuing and administering a smaller-denomination interest.
Collateral mobility
Collateral mobility is the operational property that the same instrument can be pledged, unpledged, substituted, and recalled across counterparties without leaving the chain. The pre-tokenised version of this lives at the major tri-party agents (BNY, J.P. Morgan, Euroclear, Clearstream) and runs on a daily batch cadence with manual recall workflows. The tokenised version runs on the same chain that holds the asset, with the smart contract enforcing the exclusivity of the pledge.
Kinexys's Tokenized Collateral Network (TCN) is the canonical production example. BUIDL has been posted as collateral against tokenised cash on TCN by institutional counterparties, notably for margin coverage on derivatives positions, with the substitution and recall mechanics enforced by the contract rather than a manual recall message. The April 2026 Nomura-Mizuho-JSCC tokenised JGB collateral trial on Canton is the APAC counterpart, designed to bring 24/7 collateral mobility to Japanese government bonds settled through JSCC (Ledger Insights).
The structural unlock is that a tokenised MMF interest can serve as cash collateral inside a venue without going through the redemption cycle that an off-chain MMF interest would require. This is the behavioural reason BUIDL has been treated by institutional counterparties as on-chain cash collateral despite being legally a fund interest, not a stablecoin. Tokenised collateral takes the topic in more depth.
Programmability
The chain holds the programme; the programme is what the bank's operations team would otherwise run by hand. The four canonical uses are intraday repo, conditional payments, automated waterfalls, and permissioned distributions.
Intraday repo is the most commercially active. Broadridge's Distributed Ledger Repo platform settled USD 339 billion in average daily repo volume in September 2025 on Canton, with the contract logic handling collateral substitution and recall inside the trading day. Fnality's Sterling product has been wired to Broadridge DLR for intraday DvP (delivery-versus-payment) repo against tokenised central-bank money. The point is not that intraday repo is impossible on legacy rails; it is that the funding cost was high enough that desks did not bother. Tokenisation makes it cheap enough to sit inside the trading day.
Conditional payments and automated waterfalls are more nascent but the same pattern. A coupon payment on a tokenised bond can route to the right tranche automatically. A trade-finance payment can release only when the bill of lading has been signed off. The MAS-led Project Guardian workstreams are the active testing grounds. Chapter VI on atomic DvP makes the boundary between "chain enforces atomicity" and "system inherits legal finality" precise.
Cross-border settlement compression
Cross-border payments still route, in 2026, through chains of correspondent banks for most of the world's currency pairs. The G20 cross-border roadmap has a 2027 target of 75% of cross-border payments settled in under one hour, which the well-served corridors will hit on the existing rail and the rest will not (FSB cross-border roadmap). The tokenised alternative is structurally different.
mBridge graduated from BIS Innovation Hub coordination in 2024 and is now operated by participating central banks (PBoC, HKMA, Bank of Thailand, Central Bank of the UAE, Saudi Central Bank) for cross-border wholesale settlement using a multi-CBDC platform. Project Agorá is the BIS-led successor, working with seven central banks plus major commercial banks to test cross-border tokenised wholesale settlement using both central-bank money and commercial-bank money on a unified ledger (BIS Project Agorá). Partior is the commercial-bank-led counterpart, settling tokenised deposits between member banks across multiple currencies in under 120 seconds versus the two-day correspondent baseline.
These are not vapourware. mBridge has settled real cross-border wholesale flows. Partior moves billions in annual notional. Agorá is design stage but has central-bank participation previous BIS work struggled to assemble. Each is a credible structural alternative to the SWIFT-plus-correspondent stack for the corridors it covers, with the limit that no single one yet covers all the corridors a global treasurer would need.
The capability layer above is what tokenisation unlocks. Part 5 sits on the other side of the equation: what it does not unlock, and what the marketing material conveniently leaves out.