Development Fund Proposal
Author: Amaury | Singularity Venture Hub
Status: Submitted
Created: 2026-04-17
Label: financial-workflows-composability
Champion: Canton Foundation (Bo Zhang)
Abstract
Canton has no reference implementation for on-chain data provenance, composable product lineage, or multi-party revenue distribution. As AI agents and data-driven DeFi protocols proliferate on Canton, builders need reusable contract patterns for verifying where data came from, tracking how derived products relate to their upstream sources, and splitting revenue across a composition chain — all while preserving Canton's sub-transaction privacy.
This proposal delivers these missing patterns as an open-source Canton Signal Exchange — a focused reference implementation that extends existing Canton tooling (cn-quickstart, Canton Utilities) with three contract modules that do not exist in the ecosystem today:
- Provenance tracking — An immutable on-chain record linking outputs to inputs, model versions, and timestamps, verifiable without exposing underlying data.
- Composable product lineage — A contract-level dependency graph where derived products formally reference their upstream sources, creating an auditable composition chain.
- Multi-party revenue splits — On-chain enforcement of fee distribution between original producers, curators, and the marketplace when derived products are purchased.
These patterns are wrapped in a working demo application with a lightweight UI and backend, enabling Canton builders to see, test, and reuse the contracts in their own projects.
Total grant request: $100,000 across 3 milestones — delivered in 6–8 weeks.
Specification
1. Objective
Problem: Canton's existing reference implementations (cn-quickstart, da-marketplace, Canton Utilities) cover atomic settlement, time-bounded subscriptions, and token lifecycle management. However, none address the contract patterns required for data commerce — specifically provenance verification, derived product composition with lineage, and revenue distribution across a multi-party supply chain. These are the patterns institutional data consumers and AI agents need before they can transact confidently on Canton.
Outcome: Three reusable Daml contract modules — provenance, lineage, and revenue splits — integrated into a working Signal Exchange demo application with a TypeScript backend and lightweight React frontend. Delivered open-source under a perpetual, non-exclusive license to Canton.
Success looks like:
- A producer can publish a data feed and push private updates visible only to entitled buyers (using Canton's observer model)
- A curator can create a derived product that formally references upstream feeds, with on-chain lineage
- When a buyer purchases a derived product, revenue splits automatically to upstream producers, the curator, and the marketplace — enforced by the contract
- A buyer or auditor can inspect full provenance: source lineage, model version, freshness, confidence, and input/output hashes — without seeing the underlying data
- All patterns are documented and tested for reuse by other Canton builders
2. Implementation Mechanics
The Canton Signal Exchange follows a strict separation between on-ledger state and off-ledger computation.
On-Ledger (Daml on Canton)
The implementation builds on existing Canton patterns where they exist and introduces new contract templates where they don't.
Leveraging existing patterns (not claiming novelty):
- SignalListing — Product metadata, pricing, visibility rules, and status lifecycle. Follows the listing pattern established in da-marketplace and cn-quickstart, adapted for data products with signal-specific fields (cadence, confidence thresholds, allowed buyer policies).
- PurchaseOffer + SettlementReceipt — Atomic purchase workflow. The
SettleAndIssueEntitlement choice creates both a settlement receipt and a subscription entitlement in a single transaction. Builds on the atomic settlement pattern from cn-quickstart's LicenseRenewalRequest / AllocationRequest integration, extended with an explicit on-chain receipt artifact for audit purposes.
- SubscriptionEntitlement — Buyer access rights with time bounds, usage rights, and status lifecycle. Extends the time-bounded access pattern from cn-quickstart's License contract.
Novel contract modules (the core contribution):
- SignalUpdateHeader — On-chain metadata for each signal update (payload hash, confidence, freshness expiry, model version). The
visibleTo observer array enables a recurring private content delivery pattern — producers push updates to entitled subscribers, with per-update visibility control. This continuous delivery workflow has no equivalent in existing Canton tooling.
- ProvenanceRecord — Immutable lineage record: source listing IDs, input hash set, output hash, compute job ID, model version, and timestamp. Creates a verifiable dependency DAG on-chain. No existing Canton reference implementation covers this pattern.
- RevenueSplit — Producer, curator, and marketplace fee shares with on-chain enforcement. When a derived product is purchased, revenue distributes automatically to all upstream contributors based on contract-defined splits. No existing Canton tooling handles multi-party, composition-aware revenue distribution.
Off-Ledger (TypeScript/Node.js + React)
- Backend (Express 5) — RESTful API layer integrating with Canton's Ledger API via gRPC. Handles listing management, purchase orchestration, update publication, provenance queries, and connector orchestration. The backend is intentionally lightweight — it demonstrates integration patterns, not production infrastructure.
- AI/Data connectors — Mock sentiment, risk, and opportunity model services with deterministic SHA-256 provenance hashing. Designed as pluggable interfaces for future replacement with real data sources.
- React frontend (Vite) — Focused on demonstrating the novel contract patterns: provenance inspector (lineage graph, hashes, model versions), revenue split visualization, and the recurring update delivery flow. Marketplace catalog and purchase flow included for end-to-end demo completeness.
End-to-End Demo Flow
- Producer A lists a Sentiment Pulse Feed as a
SignalListing contract on Canton.
- Producer B lists a Vault Risk Feed as a separate on-ledger listing.
- Curator C creates a derived product — the Vault Opportunity Composite — referencing both upstream feeds in its
sourceListingIds. The lineage is on-chain via ProvenanceRecord.
- Buyer D purchases the Composite. The
SettleAndIssueEntitlement choice executes atomically: settlement receipt and subscription entitlement created in one transaction. RevenueSplit distributes revenue to Producers A, B, Curator C, and the marketplace.
- Curator C publishes a private update. The
SignalUpdateHeader is created with Buyer D in the visibleTo array. Non-entitled parties cannot see the payload.
- Buyer D inspects provenance: source feed lineage, model version, freshness timestamp, confidence level, input/output hashes — all verifiable without exposing the full data.
3. Architectural Alignment
This proposal targets three gaps in Canton's reference ecosystem:
- Provenance verification — Canton's sub-transaction privacy means provenance must be verifiable without revealing the underlying data. The
ProvenanceRecord pattern demonstrates this: buyers and auditors can inspect the full lineage chain (source IDs, hashes, model versions) while the actual data remains private to entitled parties. This is architecturally impossible on public chains where all contract state is visible.
- Composable contracts with lineage — Daml contracts can reference other contracts by ID, but no existing reference implementation demonstrates a multi-layer composition chain with formal lineage tracking. This proposal fills that gap.
- Multi-party atomic workflows — Canton's choice mechanism enables atomic multi-party settlement, but existing references are limited to bilateral transactions. The
RevenueSplit pattern extends this to N-party distribution within a single atomic transaction.
Per CIP-0082, the fund targets "core R&D, dev tools, security, audits, reference implementations, DeFi app(s), critical infra." This proposal delivers reusable reference patterns for the Canton ecosystem.
4. Backward Compatibility
No backward compatibility impact. The Canton Signal Exchange is a new, standalone reference implementation with no dependency on existing Canton deployments or contracts.
Milestones and Deliverables
Milestone 1: Novel Daml Contracts and Test Suite
- Estimated Delivery: Weeks 1–4 from project start
- Focus: The three novel contract modules (
ProvenanceRecord, RevenueSplit, SignalUpdateHeader) plus the adapted patterns (SignalListing, PurchaseOffer, SettlementReceipt, SubscriptionEntitlement), with full test suite
- Deliverables / Value Metrics:
- All Daml contract templates with full Daml Script test suite covering all choices, lifecycle transitions, and error conditions
- Integration test scenarios: multi-party atomic workflows, privacy enforcement, lineage tracking, revenue split distribution
- Non-entitled party visibility test: a party not in
visibleTo cannot observe SignalUpdateHeader contracts
- Contract documentation explicitly identifying which patterns are novel vs. adapted from existing tooling
Milestone 2: Backend, Frontend, and End-to-End Demo
- Estimated Delivery: Weeks 3–6 (parallel start, released after M1)
- Focus: Off-ledger integration layer, lightweight React UI focused on demonstrating novel patterns, seeded demo data, and scripted end-to-end scenario
- Deliverables / Value Metrics:
- Express 5 backend with Ledger API integration via gRPC
- Mock AI/data connectors with deterministic SHA-256 provenance hashing pipeline
- React web application: provenance inspector (lineage graph, hashes, model versions), revenue split visualization, marketplace catalog, purchase flow, buyer dashboard, producer console
- Seeded demo data and scripted end-to-end scenario runner
- Full demo completing list → purchase → revenue split → deliver → inspect provenance → derive flow
Milestone 3: Documentation, CI/CD, and Demo Hardening
- Estimated Delivery: Weeks 5–8 (final gate)
- Focus: Production hardening, documentation, live demo environment
- Deliverables / Value Metrics:
- GitHub Actions workflows: secret scanning, dependency audit, build, test, deploy
- Architecture guide explicitly documenting reusable patterns for other Canton builders
- API reference documentation
- Live demonstration environment with runbook, completing in under 5 minutes
- Pattern reuse guide: how to extract and adapt
ProvenanceRecord, RevenueSplit, and SignalUpdateHeader for other Canton applications
Acceptance Criteria
The Tech & Ops Committee will evaluate completion based on:
- Deliverables completed as specified for each milestone
- Demonstrated functionality or operational readiness
- Documentation and knowledge transfer provided
- Alignment with stated value metrics
Milestone 1 — specific conditions:
- All Daml Script tests pass in CI
- A committee member or delegate can review contract source and verify:
SettleAndIssueEntitlement atomic choice creates both SettlementReceipt and SubscriptionEntitlement in a single transaction; ProvenanceRecord correctly links derived products to upstream source listing IDs with hash verification; RevenueSplit distributes funds to all upstream contributors in a single atomic transaction
- Non-entitled party visibility test passes: a party not in
visibleTo cannot observe SignalUpdateHeader contracts
Milestone 2 — specific conditions:
- Full end-to-end flow (list → purchase → revenue split → deliver → inspect provenance → derive) completes in a UI walkthrough
- Provenance inspector displays source lineage, model version, freshness, confidence, and hash chain
- Revenue split visualization shows correct distribution across producers, curator, and marketplace
Milestone 3 — specific conditions:
- CI pipeline passes on clean checkout (all Daml, backend, and E2E tests green)
- Live demo completes full scenario in under 5 minutes before the committee
- Pattern reuse guide reviewed and confirmed usable by a committee member or delegate
Funding
Total Funding Request: $100,000
Payment Breakdown by Milestone
| Milestone | Deliverable | Grant Release |
|---|
| Milestone 1 | Daml contracts and test suite | $40,000 upon acceptance |
| Milestone 2 | Backend, frontend, end-to-end demo | $35,000 upon acceptance |
| Milestone 3 | Docs, CI/CD, demo hardened | $25,000 upon acceptance |
| TOTAL | 6–8 weeks | $100,000 |
Funding is denominated in USD, payable in Canton Coin per CIP-0100.
Volatility Stipulation
The project duration is under 6 months. Should the project timeline extend beyond 6 months due to Committee-requested scope changes, any remaining milestones must be renegotiated to account for significant USD/CC price volatility. Milestones 2 and 3 are additionally subject to renegotiation if CC/USD moves ±40% during delivery.
Co-Marketing
Upon release, SVH will collaborate with the Foundation on:
- Announcement coordination — Coordinated social media campaign from SingularityNET, ASI Alliance, Fetch.ai, Cudos, and Intellistake, positioning Canton as infrastructure for confidential, provenance-verified data commerce.
- Case study or technical blog — Technical deep-dive on the novel contract patterns (provenance, lineage, revenue splits) and how Canton builders can reuse them.
- Developer workshops — Hands-on sessions with Canton-selected projects walking through pattern reuse, signal publication, provenance setup, and data pipeline integration.
Motivation
Why These Patterns Matter for Canton
Canton's existing reference implementations cover atomic settlement, time-bounded subscriptions, and token lifecycle. But as AI agents and data-driven DeFi protocols arrive on Canton, builders need patterns the ecosystem doesn't have yet: verifiable data provenance, composable product lineage, and multi-party revenue distribution.
These are not niche requirements. Any application where one party produces data, another curates or enriches it, and a third consumes it — risk analytics, compliance scoring, market intelligence, oracle feeds — needs exactly these patterns. Building them once as a reusable reference saves every subsequent Canton builder from reinventing them.
The Signal Exchange wraps these patterns in a working demo application, so builders can see them in context, test them end-to-end, and extract the contracts for their own projects.
Why a Data Marketplace Specifically
A data marketplace is the simplest application that exercises all three novel patterns simultaneously: provenance (where did the data come from?), lineage (what was it derived from?), and revenue splits (who gets paid when derived products sell?). It serves as a natural integration test for the patterns, not just an isolated contract library.
Rationale
Why SVH?
Singularity Venture Hub is the venture and integration hub of the ASI ecosystem, working directly with SingularityNET, ASI Alliance, Fetch.ai, and Cudos. This provides direct access to AI/data engineering talent and a distribution network that can bring Canton visibility in the Web3 AI community.
SVH's engineering team has deep expertise in meTTa, a functional programming language with similar type-theoretic foundations to Daml. To strengthen Daml-specific delivery confidence, SVH will engage Daml-experienced engineering support for contract architecture review and validation.
Beyond delivery, SVH will serve as a go-to-market facilitator for the Signal Exchange on Canton — driving developer education, integration support, and ecosystem activation through its ASI network.
Why This Scope
This proposal intentionally builds on existing Canton patterns (cn-quickstart for subscriptions, Canton Utilities for settlement) rather than rebuilding them from scratch. The budget and scope are focused on the novel contract modules — provenance, lineage, and revenue splits — that don't exist in the ecosystem today. The demo application provides just enough context to make the patterns usable, without duplicating infrastructure Canton already has.
Go-To-Market
A marketplace is only as valuable as the supply, demand, and liquidity flowing through it. SVH commits to a comprehensive, long-term GTM plan that goes far beyond a launch announcement — we own the end-to-end work of making the Signal Exchange a living, high-utility venue on Canton.
1. Business Development: Recurring Supply–Demand Matching
SVH will operate a dedicated BD function whose sole purpose is to keep the marketplace alive with real transaction volume:
- Active outreach to Canton participants — SVH will systematically engage both live Canton projects and incoming applicants to understand (a) what data they need to operate better and (b) what data or signals they can publish. This is a recurring, hands-on effort, not a one-time survey.
- Ongoing matching work — SVH's BD team acts as a marketplace operator — matching supply with demand across Canton participants, seeding initial transactions, and ensuring continuity of listings and subscriptions. This is the discipline that turns a technical POC into a living exchange.
- Anchor tenants from day one — SVH will recruit and onboard initial anchor producers and anchor buyers at launch, ensuring the marketplace is populated with real listings and real transactions from week one — avoiding the "empty marketplace" problem that kills most new venues.
2. Leveraging the ASI Ecosystem
SVH's position as the venture hub of the ASI ecosystem gives the Signal Exchange a distribution advantage no other grantee can offer:
- ASI Alliance AI project network — Through SingularityNET, ASI Alliance, Fetch.ai, and Cudos, SVH has direct relationships with dozens of AI and data projects. SVH will bring these projects onto the Exchange as both producers (publishing their models, analytics, and signals) and buyers (consuming data to power their own AI products).
- 300+ Web3 AI builder distribution channel — SVH owns a distribution channel of 300+ Web3 AI builders — a direct line to the developer community most likely to become early producers, curators, and consumers on the Signal Exchange. This channel will be activated through targeted technical content, office hours, and one-to-one onboarding.
- Cross-ecosystem social media — Coordinated content from SingularityNET, ASI Alliance, Fetch.ai, Cudos, and Intellistake across their combined audiences, positioning Canton as the first institutional DeFi network with a confidential, provenance-verified data marketplace.
3. PR & Media
- Intellistake-led PR — Intellistake (publicly listed company in Canada) leads press outreach, leveraging investor relations and media channels to generate coverage of the Canton Signal Exchange in Web3 media outlets.
- Milestone-based PR waves — Additional PR pushes triggered upon reaching key milestones — data volume thresholds, number of active listings, and transaction volume. This creates a sustained narrative of growth rather than a single launch announcement.
- Conference presence — Co-marketing at Token2049, Consensus, and ASI-ecosystem events. SVH will present the Signal Exchange as a flagship Canton use case, bringing new eyes to the network.
4. DeFi Data Flywheel
- Data-driven DeFi on Canton — Risk scores, opportunity signals, and sentiment analytics on the Exchange provide the decision-grade intelligence that DeFi protocols and institutional participants need. More data products drive better decision-making, which drives more transactions on Canton.
- DeFi interoperability outreach — Active outreach to DeFi actors on Canton to expand both the range of data sources published and the range of signal consumers, directly increasing marketplace transaction volume and network utility.
5. SVH as First Buyer & Future Product Roadmap
A critical commitment that distinguishes this grant: SVH is not just the builder of the Signal Exchange — we are also committing to be a major buyer on it, building our own product line on top of the marketplace.
SVH as a Buyer: Autonomous DeFi AI Agents
SVH will purchase signals (risk scores, opportunity composites, sentiment feeds) on the Signal Exchange to power the next generation of autonomous AI agents that optimize DeFi strategies on Canton. These agents consume on-chain signals through verified entitlements, make decisions, and execute on Canton's DeFi primitives — delivering optimized yield to end users while generating transaction volume on both the Signal Exchange (signal purchases) and Canton's DeFi layer (agent-driven transactions).
The Full-Stack Value Chain
This creates a complete, self-reinforcing value chain on Canton:
- DeFi builders earn — By publishing risk analytics, opportunity scores, and market signals as paid listings on the Exchange.
- SVH (and other AI agent builders) buy — Subscribing to those signals to power autonomous DeFi agents — generating recurring revenue for signal producers.
- Retail end users benefit — From AI-optimized yield strategies on Canton, with the underlying signals and decisions fully auditable via on-chain provenance.
- Canton captures volume — From signal subscriptions, agent-driven DeFi transactions, and settlement activity — across multiple layers of the stack.
The Signal Exchange is therefore not just a standalone marketplace — it is the foundational data layer that unlocks a whole class of AI-native DeFi products on Canton, built by SVH and others. This ensures persistent transaction volume long after the POC period ends, and aligns SVH's long-term business interests with the success of Canton.
