Aurigraph Research Team · Institutional Finance · Frontier Technology · status as of 2026-10-20
Cross-border payment and trade finance still run on infrastructure designed for a world of correspondent-banking hops, batch settlement, and paper documents. A wire crosses several intermediaries and settles on a T+2 (or slower) cycle; a letter of credit is a stack of physical documents inspected by hand, and its lifecycle is a fertile ground for fraud, duplicate financing, and dispute. Aurigraph applies a permissioned distributed ledger to both problems at once: central-bank money as a tokenized, natively-settling instrument, and trade finance instruments as programmable Active Contracts that release funds against verifiable conditions rather than manual sign-off. Both are issued as Composite Token Framework (CTF) credentials — carrying identity, compliance, and provenance in the token itself — settled with deterministic finality on HyperRAFT++ consensus, recorded under triple-entry accounting, and designed to interoperate with existing bank rails through ISO 20022 messaging. This paper describes the architecture and is explicit about which figures are engineering/capability targets versus shipped behavior. It describes no specific central-bank pilot, bank partnership, or regulatory approval — none are claimed.
Two structural inefficiencies sit at the centre of institutional finance.
Cross-border settlement is slow and opaque. A cross-border payment typically traverses a chain of correspondent banks, each holding funds, applying its own cut-off windows, and reconciling on its own books. Securities settle on a T+2 cycle in most major markets — meaning value and delivery are separated by days during which counterparty and settlement risk accrue. The parties cannot see, in real time, where a payment is or whether the offsetting leg has moved. Liquidity is trapped in transit; reconciliation is a manual, after-the-fact exercise across ledgers that were never designed to agree.
Trade finance is paper-bound and fraud-prone. A letter of credit (LC) is a promise by a bank to pay a seller once documentary conditions are met — bills of lading, inspection certificates, invoices. Today that promise is administered on paper (or scanned paper), inspected by hand, and reconciled across the issuing bank, advising bank, buyer, and seller — none of whom share a single source of truth. The gaps are exploited: the same invoice financed twice at two banks, forged bills of lading, and disputes over whether conditions were actually met. Because no party holds an authoritative, tamper-evident record of the instrument's state, each must trust its own copy and chase the others.
Both problems share a root cause: value and its governing conditions live in separate, unsynchronized systems. Money moves on one rail; the document that says when it should move lives on another; identity and compliance live in a third. Aurigraph's thesis is that when the instrument, its money, its legal conditions, and its compliance state are the same on-ledger object, settlement collapses from days to sub-second and the fraud surface collapses with it.
A central-bank digital currency on Aurigraph is central-bank money issued as a tokenized instrument on a permissioned ledger, not a claim on a commercial intermediary. Issuance, circulation, and redemption are ledger operations that a central bank controls directly, with commercial banks and regulated participants holding permissioned roles beneath it.
Instant settlement. Because the CBDC token is the money and the ledger is the settlement layer, a transfer settles when consensus commits it — there is no correspondent chain and no separate clearing step. Settlement finality is deterministic (see §6): once a transfer is committed to the HyperRAFT++ Write-Ahead Log and reaches quorum, it is final and irreversible, not probabilistically confirmed over many blocks.
Retail and wholesale tiers. Aurigraph's finance vertical models both variants a central bank actually needs:
Privacy and compliance in balance. A CBDC design lives or dies on the tension between citizen privacy and regulatory oversight. Aurigraph does not resolve this by choosing one — it resolves it structurally. Because each token is a CTF credential (§5) whose compliance and identity data are a discrete, selectively-disclosable layer, a commercial-bank role can verify that a counterparty is KYC-cleared and a transfer is within policy without exposing the full transaction graph, while the central bank retains the permissioned visibility its mandate requires. Selective disclosure (via BBS+ signatures in the CTF envelope) is designed to make "prove compliant, reveal nothing else" a native operation rather than a policy promise; today the BBS+ engine is a development simulation (AV11-3116), so the privacy guarantee in production is the layer separation, not a zero-knowledge proof.
Trade finance is where programmability pays off most directly, because the instruments are already conditional promises — they simply lack a machine that can evaluate the conditions and act. Aurigraph models them as Active Contracts: library templates whose declarative automation rules mint and transition tokens on state entry, so the legal instrument and the token lifecycle are the same object.
Letters of credit. An LC becomes an Active Contract whose states mirror its real lifecycle — issued → documents presented → conditions verified → paid (or rejected). Documentary conditions are encoded as machine-evaluable rules; when the required evidence (a bill-of-lading hash, an inspection attestation, a customs event) is anchored on-ledger, the contract transitions automatically and releases payment. No manual document inspection stands between condition-met and funds-moved.
Escrow and milestone release. Trade payments are frequently milestone-based — a percentage on shipment, a percentage on delivery, the balance on acceptance. Aurigraph expresses this as escrow held by an Active Contract that releases tranches as each milestone's evidence is anchored. Funds are locked in the token itself, not in a third-party account, so release is atomic with the condition rather than dependent on an operator.
Anti-fraud by construction. Because each instrument is a single authoritative on-ledger record with an append-only history, the classic trade-finance frauds lose their footing: an invoice tokenized once cannot be financed twice without the second attempt colliding with the existing record; a bill of lading is bound cryptographically to the instrument (CTF Binding layer, §5); and every state transition is a signed, timestamped, tamper-evident event visible to all authorized parties on the same ledger.
Atomic delivery-versus-payment. The delivery leg (the tokenized instrument or asset) and the payment leg (CBDC or tokenized cash) can be committed as a single ledger mutation, so either both move or neither does — eliminating the principal risk that a T+2 gap creates. Where both legs live in the same channel's registries, the consensus layer commits both as one atomic update rather than requiring a third-party router.
A CBDC or trade-finance rail that cannot speak to the existing banking system is a walled garden. The world's payment infrastructure is converging on ISO 20022 — the structured, data-rich XML message standard (pacs.* for interbank clearing and settlement, pain.* for payment initiation, camt.* for cash management and reporting) now adopted by SWIFT, major RTGS systems, and the Bank for International Settlements' CBDC work.
Aurigraph's finance vertical is designed for ISO 20022 message interoperability: on-ledger CBDC and trade-finance operations map to and from ISO 20022 messages, so a participating bank's existing systems can originate and reconcile against the ledger without a bespoke integration. A cross-border transfer initiated as an ISO 20022 payment instruction can settle on-ledger and emit the corresponding settlement and reporting messages back to the counterparty's rails.
Interoperability extends to cross-border settlement between distinct CBDC systems: bridge tokens act as an atomic-exchange instrument between two participating central-bank ledgers, so a payment can settle across two currency zones without either side surrendering control of its own issuance.
Status. ISO 20022 interoperability is a design aligned to the platform's finance roadmap, not a certified conformance claim: the V12 tree contains no ISO 20022 message mapping yet, and emitted messages carry no ML-DSA co-signature (AV11-3855). This paper describes the designed message-mapping and bridge model; production ISO 20022 certification and any specific network integration are forward-looking.
V12
Every CBDC and trade-finance instrument on Aurigraph is issued as a Composite Token Framework (CTF) credential — a single W3C Verifiable Credential that binds decentralized identity, regulatory compliance, provenance, a cryptographic binding to the underlying artefact, and post-quantum key material into one signed envelope. Compliance is therefore inside the instrument, not a spreadsheet a verifier must chase.
Concretely, for these instruments the CTF layers carry:
Travel Rule, natively. Cross-border value transfer triggers the FATF Travel Rule (Recommendation 16), which requires originator and beneficiary information to travel with the payment. Aurigraph's compliance layer carries Travel Rule applicability policy (compliance/tfr in V12) and is designed to emit IVMS 101 messages, with the live Travel Rule exchange and sanctions screening (OFAC SDN, UN consolidated list) performed by the Battua service today. Because the required identity data already lives in the token's compliance and identity layers, Travel Rule emission is designed as a read off the instrument rather than a separately-assembled report. KYC records are currently signed with Ed25519 — a named classical exception on the road to ML-DSA.
compliance/tfr
For the full five-layer construction, the post-quantum signature envelope, and the selective-disclosure model, see the Composite Token Framework whitepaper.
Deterministic finality. Aurigraph settlement is committed by HyperRAFT++, a RAFT-derived, crash-fault-tolerant consensus engine with a majority quorum and a memory-mapped Write-Ahead Log. A transaction is written to the WAL and reaches quorum before it is acknowledged; once committed it is final — there is no probabilistic reorg window, which is precisely the property institutional settlement requires. Two caveats, stated plainly: production consensus currently runs in SIMULATE mode on a single node (real multi-node consensus, q = ⌈2N/3⌉ with N ≥ 7, is planned), and the platform's throughput and finality figures are engineering/capability targets — the public stats endpoint reports them as "not measured" — to be confirmed by published, reproducible benchmarks rather than read as measured production numbers.
SIMULATE
Triple-entry accounting. Beneath the instruments, Aurigraph records value movements under a Triple-Entry Accounting (TEA) model, implemented in the platform's finance service (/api/v11/finance). Conventional double-entry bookkeeping records a DEBIT on the sender's account and an offsetting CREDIT on the receiver's, applied atomically. Aurigraph adds the third entry: each ledger entry embeds an independent cryptographic hash anchor, so the entry is not only balanced against its counterpart but also chained into a tamper-evident record that any party can verify. The result is a shared, cryptographically-notarized set of books — the "third entry" is the ledger's own signed proof — which is exactly what removes the reconciliation gap that correspondent banking leaves open. Accounts, balances, ledger histories, and reconciliation are first-class operations on the finance API.
/api/v11/finance
Together, deterministic finality and triple-entry accounting mean a settled transfer is simultaneously final, balanced, and provable — the three things a T+2 correspondent flow cannot offer at once.
The inefficiencies of cross-border settlement and trade finance are not failures of effort — they are the natural consequence of keeping money, the conditions that govern its movement, identity, and compliance in four separate systems that were never designed to agree. Aurigraph collapses them into one on-ledger object: central-bank money as a tokenized instrument that settles with deterministic finality, trade-finance instruments as programmable Active Contracts that release value against verifiable conditions, compliance embedded in the token via the Composite Token Framework, and a shared, cryptographically-notarized record under triple-entry accounting — all designed to interoperate with the existing banking system through ISO 20022 rather than to replace it. Throughput, finality, and ISO 20022 conformance figures in this paper are engineering and capability targets, and no specific central-bank pilot, bank partnership, or regulatory approval is claimed. What is claimed is an architecture in which settlement is fast, finality is real, and fraud loses the seams it depends on.
For the platform this rail runs on — HyperRAFT++ consensus, NIST Category 5 post-quantum cryptography, RocksDB ledger stores, and real-world-asset settlement — see the Aurigraph DLT whitepaper.
Explore: About · Products · Features · Whitepapers · Sustainability · Tokenized carbon credits · Real-world asset tokenization · SDK · Blog · Plain-text site index · Sitemap · Contact: contact@aurigraph.io