Proposal: Canton NFT Standard

Development Fund Proposal

Author: Fairmint Org: Fairmint, Inc. Status: Submitted Created: 2026-08-12 Label: token-asset-standards Champion: Jonathan Mayeur, IntellectEU (@jonathan-ieu)

---

Abstract

This proposal funds a DAML NFT standard for Canton, submitted as a new CIP and built on CIP-0056 / CIP-0112 (Token Standard V2). The profile adds the item-level semantics the token standards leave open: stable per-item identity, collection uniqueness, per-item metadata, provenance, privacy rules, and whole-asset (including soulbound) lifecycle invariants. Transferable NFTs move over the standard V2 transfer rail. This proposal defines no second transfer protocol.

Canton has no ratified NFT-specific standard. CIP-0056 and CIP-0112 specify holdings, transfer instructions, allocations, and settlement, and are agnostic to what the asset is. They do not specify how a unique item keeps a durable identity across ContractId churn, how duplicate token ids are prevented within a collection, how wallets render per-item metadata, or how non-divisible transfer works. Teams that need those semantics today (marketplaces, credential issuers, registries, RWA with per-unit identity) each invent their own, whether they use a bespoke template or a CIP-0056 Holding with amount = 1.

Fairmint's SEC-registered Transfer Agent begins issuing production certificates on Canton under the new standard, demonstrating a new end-to-end reference design (in Milestone 3).

---

Specification

1. Objective

Problem: Canton lacks an NFT-specific standard that wallets, issuers, registries, and marketplaces can implement consistently. This proposal will define and seek approval for a CIP covering the NFT behavior described under Implementation Mechanics. The work is limited to the NFT profile, reference implementation, conformance tests, integration documentation, and initial adoption. Marketplace mechanics, custody models, public discovery, and regulated issuance policies remain outside scope.

2. Implementation Mechanics

The standard is a new CIP defining an NFT profile on Token Standard V2: pinned interface packages that a template implements alongside the CIP-0112 holding and transfer interfaces (the existing versioned splice-api-* packaging convention). One template is both a V2 holding and a conforming NFT: V2 wallets, venues, and indexers already understand its ownership and settlement, and NFT-aware applications read item semantics from the same asset.

The profile adds only what the token standards leave open (specified with RFC 2119 language in the CIP):

1. Stable identity. One NFT, one identity (registry_id/collection_id/token_id), anchored to the V2 InstrumentId and preserved across lifecycle changes; ContractId is never the durable identifier.
2. Whole-asset invariants. Exactly one live holding, amount = 1.0, supply 0 or 1. Split, merge, fractional amounts, and partial transfer are rejected on every exposed path.
3. Registry uniqueness. A dual-authority NftRegistry with collection-scoped ledgers rejects duplicate token ids per collection lineage.
4. Standard transfer. Transferable NFTs use the CIP-0112 transfer flow; no second protocol. Soulbound NFTs advertise no transfer support and reject every ownership-changing path.
5. Consent-based issuance. Third-party mint requires designated-holder acceptance; issuance provenance appears in the standard view. (Funded work; Milestone 2.)
6. Standard read model. A thin NFT interface on the same template exposes registry, collection, item id, holder, metadata, confidentiality reference, and provenance.
7. Per-item metadata. Namespaced standard keys (nft.canton.network/), compatible with Splice token-metadata-v1 at the asset level, plus an optional ConfidentialityRef so confidential payloads stay off-ledger.
8. Private by default. Visibility bounded to stakeholders; no enumeration, no public holder disclosure, no leakage of NFT existence through standard discovery surfaces.

Standards compliance matrix. Interface package versions are pinned, so conformance is verifiable against packages:

Standard	How the NFT profile relates
CIP-0056 (Token Standard)	Conforming NFTs are holdings; V1 wallets interoperate per CIP-0112's cross-version compatibility rules
CIP-0112 (Token Standard V2, ratified 2026-06-12)	Primary rail: holdings, transfer instructions, events; profile invariants constrain V2 choices for amount = 1.0 assets
Splice token-metadata-v1	Asset-level metadata surface reused; profile adds the per-item metadata layer
CIP-0103 (dApp Standard)	Wallet guide specifies signing-flow composition for mint acceptance and transfers
CIP-0104 (Traffic-Based App Rewards)	Produces direct traffic attribution
ERC-721 / EIP-4906	Normative semantic mapping (Appendix A) for teams and assets migrating from EVM, including metadata-update semantics
DAML smart-contract upgrade (SCU)	Interface packages versioned and pinned for upgrade compatibility, per Splice packaging conventions

Delivery approach: All code develops in a public open-source repository; the specification is submitted as a CIP to canton-foundation/cips, so the network governs the standard. Reference implementation, conformance tests, and integration docs ship with the spec.

Current implementation status: Experimentation by the Fairmint dev team confirmed feasibility to build a private-by-default NFT standard for the Canton ecosystem using existing Canton rails.

3. Architectural Alignment

Relationship to CIP-0056 and CIP-0112 (Token Standard V2). CIP-0056 describes itself as a generic standard for tokenized assets, and CIP-0112 extends it with privacy-enhanced batch settlement, account structures, and TradFi-aligned timing. A unique asset can be represented as a Holding with amount = 1 and settled through the standard's transfer and allocation rails. What CIP-0056/CIP-0112 do not specify are the NFT-specific semantics this proposal standardizes:

• Stable per-item identity: no standard field or rule ties a holding to a durable token identity across ContractId churn; nothing prevents two amount = 1 holdings from claiming the same logical item.
• Collection-scoped uniqueness: no standard mint lineage or duplicate-rejection authority per collection.
• Per-item metadata: CIP-0056 metadata is asset-class-level; there is no standard way for a wallet to render item-level name/description/image or a confidentiality reference.
• Whole-asset lifecycle: non-divisibility, consent-based whole-item transfer, and issuance provenance are conventions each implementer re-invents today.

A conforming NFT is a CIP-0112 holding. Transferable NFTs use the standard V2 transfer flow; the new CIP adds only the item-semantics interfaces and invariants. Wallets keep one transfer rail and indexers keep consuming standard holding-change events.

The standard mirrors the interface-package pattern the ecosystem already trusts (versioned splice-api-* packages, Kaiko Data Standard) so adopting teams face a familiar shape: implement the interface, read through the views.

Wallet connectivity (CIP-0103). The integration documentation includes a wallet guide showing how conforming wallets surface the NFT view, render nft.canton.network/ standard metadata, execute standard CIP-0112 transfers for transferable NFTs, and present mint-acceptance requests for consent, including how these flows compose with CIP-0103 dApp connectivity for signing requests. A wallet with an existing V2 integration needs only the thin NFT read layer.

Ecosystem priority fit. The work falls under the published "App Building and Developer Experience" priority area: token standards, interoperability across wallets, assets, and dApps, and documentation/examples.

Relationship to other NFT efforts in the ecosystem (non-duplication):

Several ecosystem teams are building NFT applications and libraries (marketplaces, audited component libraries). Those efforts implement and consume NFT semantics; none of them publish a specification. This proposal contributes to the network-governed specification those implementations can share, with a reference implementation and a production consumer: Fairmint's equity certificates.

Fairmint's equity certificates (soulbound credentials over unique assets) are the first production consumer application. Production certificates today are ERC-721-based on an EVM chain; a DAML EquityCertificate prototype already implements the draft Nft interface. Milestone 3 migrates production issuance onto the Canton standard, exercising the soulbound path.

Featured App relevance. Featured Asset Issuer qualification currently references CIP-0056 compliance; unique-asset issuers have no equivalent on-ramp today. A ratified NFT standard gives Tokenomics and Accountability a conformance target for future unique-asset issuer categories.

4. Backward Compatibility

Fully Backwards Compatible. The standard ships as new interface and template packages that implement the published CIP-0056/CIP-0112 interfaces without modifying Splice, the token standard packages, or any deployed asset. Existing V2 wallets, venues, and indexers interact with conforming NFTs through the interfaces they already support; existing V2 wallets can process standard holdings and transfers; NFT rendering requires support for the new read interface.

---

Milestones and Deliverables

All demos run on the V2-profile implementation produced by this grant. Each milestone ends with an acceptance check. Fairmint builds milestones 2–4 in parallel with CIP review; if governance amends the spec, Fairmint reworks the implementation within the stated milestones without supplemental funds.

Milestone 1: Feasibility Spike + Specification + CIP Submission

• Estimated Delivery: 1-2 Months After Approval
• Focus: Prove the profile is thin on the real V2 packages, lock the normative spec, open network governance, and make everything public.
• Deliverables / Value Metrics:
  • Feasibility spike on current CIP-0112 packages: one transferable and one soulbound asset, each a single template implementing both the V2 holding and the NFT profile, demonstrating a standard V2 transfer with stable identity at amount = 1.0; rejection of fractions, split, merge, aggregation, and partial transfer; standard holding-change events; and the metadata / provenance view. Spike results (pass, or the exact V2 constraint found) published; any hard incompatibility is raised with the Token Standards SIG before the spec locks.
  • Public open-source repository containing the specification, the existing prototype packages (re-published from Fairmint's private repository), the spike code, tests, and built .dar artifacts.
  • CIP draft PR opened against canton-foundation/cips with a CIP number assigned, announced on cip-discuss, with pinned interface-package versions per the compliance matrix.
  • Spec review session held with the Token Standards / Asset Standards SIG (calendar evidence + published minutes).
  • Explicit "enforceability boundary" section in the spec documenting what is enforced on-ledger versus trust-based.

Milestone 2: Reference Implementation on the V2 Rails

• Estimated Delivery: 1-2 Months
• Focus: Build the new reference design: the profile's interface packages and reference templates on CIP-0112, closing the known authorization gaps from the prototype.
• Deliverables / Value Metrics:
  • Profile interface packages (identity, registry/uniqueness, per-item metadata, provenance, soulbound marker) plus reference templates implementing them alongside the V2 holding and transfer interfaces.
  • Whole-asset invariants enforced across every exposed V2 choice (supply 0/1, no split/merge/partial paths), with negative tests proving rejection.
  • Holder-authorized third-party issuance (mint acceptance path) and issuance provenance in the standard view. A user-held NFT cannot appear without designated-holder consent.
  • Documented mitigation or explicit trust-boundary statement for the dual-authority off-path-create gap, plus registry-scaling benchmarks for large collections.
  • End-to-end demo on DevNet on the new reference design: mint acceptance → standard V2 transfer of the transferable NFT → rejected ownership change on the soulbound NFT, executed from a generic V2 wallet flow, with on-chain transaction references published so a second, independent reviewer can verify without deep context.

Milestone 3: CIP Acceptance + Production Migration + Ecosystem Adoption

• Estimated Delivery: 1-2 Months
• Focus: CIP approval, production migration, external integrations, and the integration guides.
• Deliverables / Value Metrics:
  • Production migration on the new reference design: Fairmint's equity certificates, ERC-721-based on an EVM chain today, issued as conforming soulbound NFTs on Canton, with on-chain evidence (DevNet, MainNet where applicable) and the migration documented as a reusable EVM-to-Canton case study.
  • At least 2 independent integrators (wallet, marketplace, registry, or issuer) with verifiable DevNet or MainNet integrations against the standard, confirmed via written attestations or on-chain/repository evidence from entities independent of Fairmint. Target profile: one Featured Wallet rendering the NFT view via its existing V2 integration, plus one marketplace, registry, or issuer.
  • CIP Accepted: the CIP reaches "Approved" status in canton-foundation/cips. Fairmint carries the governance-timing risk on this gate (see Acceptance Criteria).
  • Public adoption report: integrations, feedback received, changes made, and a maintenance roadmap.
  • Wallet integration guide: reading the NFT view, rendering standard metadata, executing standard V2 transfers for NFTs, handling mint acceptance, CIP-0103 signing-flow composition.
  • Application builder guide with the equity-certificate worked example and the ERC-721 / EIP-4906 semantic mapping for teams and assets arriving from EVM.
  • Docs usability check: a named external developer completes a working integration from the published docs alone and attests to it in writing.

Milestone 4: Production Equity Volume at Scale

• Estimated Delivery: 12/31/2026
• Focus: Issue certificates against Fairmint's production cap tables at scale. Fairmint operates cap tables for hundreds of companies.
• Deliverables / Value Metrics:
  • Equity certificates representing at least 18M shares (aggregate across Fairmint-managed cap tables) issued as conforming soulbound NFTs on Canton.
  • Privacy-preserving verification methodology published with the milestone claim: aggregate share counts, certificate counts, and issuer counts reported with on-chain evidence sufficient for an independent reviewer to confirm scale without seeing per-holder positions.
  • Registry and issuance throughput validated at production scale: collection-scoped uniqueness ledgers and mint-acceptance flows benchmarked under the full certificate load, with results published (extending the Milestone 2 registry-scaling benchmarks from synthetic to production volumes).
  • Ongoing issuance wired to the standard: new equity positions on Fairmint-managed cap tables receive conforming certificates by default, so the volume figure grows after the grant rather than being a one-time migration snapshot (pending holder's KYC/KYB completion).

Acceptance for M4: aggregate certificate coverage ≥ 18M shares confirmed by the published verification methodology against on-chain evidence; throughput benchmark report published; default-issuance path live in production.

---

Acceptance Criteria

The Tech & Ops Committee will evaluate completion based on:

• Deliverables completed as specified for each milestone.
• Demonstrated functionality: working demos backed by on-chain transaction references that an independent reviewer can verify without deep context.
• Documentation and knowledge transfer: an external team can implement or integrate the standard from published materials alone (explicitly tested in M3).
• Each milestone lists an acceptance check the committee can answer yes/no from published evidence (repository state, CIP status, attestations, on-chain transaction references) without interpretation.
• Alignment with stated value metrics: M3 is accepted on CIP Acceptance plus external integrations, and M4 on production equity volume (at least 18M shares represented on the standard).

Project-specific conditions:

• CIP Acceptance is a payment gate. Milestone 3 payment releases only when the CIP reaches "Approved" status. Fairmint accepts that governance timing is not fully within its control and carries that risk; build work (M2) proceeds in parallel, and if governance requests spec changes, Fairmint implements them within the stated milestones without supplemental funds.
• Independent verification. For M2–M4, Fairmint will provide demo access and on-chain references sufficient for a second reviewer designated by the committee to confirm the functionality, consistent with established accountability practice. For M4 specifically, the reviewer receives the aggregate-verification methodology and on-chain evidence to confirm the equity-volume claim without access to per-holder data.

---

Sustainability

Fairmint operates production infrastructure on Canton today (validator, Featured App "Onchain Transfer Agent", the Open Cap Table Protocol DAML packages) and is the standard's first production consumer, so maintenance is tied to Fairmint's own operations:

• Fairmint commits to maintaining the reference implementation, conformance suite, and documentation for a minimum of 12 months after final milestone acceptance (security fixes, compatibility with Canton/Splice releases, integration support for adopters).
• Once the CIP is ratified, evolution of the standard itself is governed by the CIP process, not by Fairmint unilaterally.
• Extension proposals (DVP above the core, discovery, soulbound/credential semantics, regulated registry policy) will be brought as separate, independently fundable proposals.

---

Funding

Total Funding Request: 4,500,000 CC

Payment Breakdown by Milestone

• Milestone 1 (Feasibility Spike + Specification + CIP Submission): 700,000 CC upon committee acceptance
• Milestone 2 (Reference Implementation on the V2 Rails): 300,000 CC upon committee acceptance
• Milestone 3 (CIP Acceptance + Production Migration + Ecosystem Adoption): 1,500,000 CC upon committee acceptance
• Milestone 4 (Production Equity Volume at Scale): 2,000,000 CC upon final release and acceptance

Adoption-based share: 3,500,000 of 4,500,000 CC (78%) is gated on adoption outcomes (CIP Acceptance, independent third-party integrations, and production equity volume), exceeding the fund's ≥50% adoption-based requirement. Build milestones (M1–M2, 1,000,000 CC) are priced at development cost; the adoption milestones carry the margin because Fairmint bears the risk of missing them.

Cost Basis

Build milestone pricing, by engineering effort and skillset:

• M1 (700,000 CC): senior DAML architecture, 2 engineers × 1.5 months for the spike, spec finalization, and CIP drafting, plus governance/SIG coordination through CIP review.
• M2 (300,000 CC): mixed DAML and test engineering, 1 engineer × 1.5 months for the profile packages, reference templates, and conformance/negative test suites.
• M3 / M4: integration and production engineering (≈ 2 engineers × 1.5 months) for the migration and adopter support, plus the adoption-risk margin described above.

CC amounts convert from USD estimates at the 30-day average CC/USD price (CoinGecko) on the submission date, to be recalculated upon committee approval.

Volatility Stipulation

The project duration is under 6 months. Should the timeline extend beyond 6 months, remaining milestone payments will be revalued for USD/CC price volatility per Development Fund policy; scope remains as stated in the Milestones section, and Fairmint absorbs rework from governance-requested spec changes without supplemental funds.

---

Demand Evidence

• Fairmint (production issuer): equity certificates issued under its SEC-registered transfer agent operations, per Milestones 3–4.
• Several data providers are in direct conversations with Fairmint to expose relevant NFT data into their data pipelines for public consumption.

---

Co-Marketing

Upon release, Fairmint will collaborate with the Foundation on:

• Announcement coordination for the CIP submission and for the standard's ratification.
• A technical blog / case study: designing an NFT standard for a privacy-first network, including the equity-certificate worked example.
• Developer promotion: a walkthrough session or workshop for wallet and application teams (offered through the Token Standards SIG), and joint announcements with the first external adopters.

---

Motivation

Why this is valuable to the Canton ecosystem:

• Ecosystem teams already define their own item conventions for lack of a standard. Each additional bespoke NFT model is another integration wallets and indexers must build separately.
• Beneficiaries: wallets (one integration renders any conforming unique asset) and unique-asset applications: marketplaces, credential issuers, RWA registries, equity and fund tokenization.
• The item-semantics layer runs today as a prototype with tests and built artifacts. The grant funds the V2-rails implementation, CIP governance, and adoption.
• The privacy properties (no global enumeration, no public holder disclosure, consent-based transfer) are not available in public-chain NFT standards and matter to institutions holding regulated unique assets.

---

Rationale

Need Beyond CIP-0056. An amount = 1 holding is a valid settlement representation of a unique item, but CIP-0056/CIP-0112 say nothing about how that holding's identity survives ContractId churn, how a collection rejects duplicate token ids, how a wallet finds item-level metadata, or what non-divisible lifecycle looks like. Two independent amount = 1 implementations do not interoperate on any of those dimensions. This profile specifies them on the same rails.

Use Token Standard V2. Fairmint's original prototype had its own transfer lifecycle (offer / accept / expire) with zero Splice dependencies. We abandoned that shape after CIP-0112 ratified: a standalone standard would force wallets, custodians, venues, and indexers to support a second ownership and transfer model. The profile reuses the V2 account, holding, transfer, and event surfaces and adds only identity, uniqueness, metadata, provenance, privacy, and whole-asset rules. The Milestone 1 spike tests this on the real V2 packages; if a hard incompatibility surfaces, we document the constraint and raise it with the Token Standards SIG.

Doesn't ratified CIP-0112 already make this unnecessary? No. CIP-0112 deepens settlement (batch privacy, accounts, timing). It neither addresses nor intends to address per-item identity, uniqueness, per-item metadata, or whole-asset lifecycle.

Why not wait for a library to define the pattern? Library components in funded ecosystem grants define no identity model, uniqueness authority, consent-based mint path, or privacy visibility rules, and a library import does not guarantee interoperability between wallets, registries, and marketplaces. A ratified specification does, and gives those libraries a target to implement.

Why is the registry first-class? The profile needs an on-ledger mint lineage and uniqueness authority. Keeping NftRegistry / CollectionRegistry separate from live asset state means transfer lifecycle changes never weaken uniqueness tracking, and collection-scoped ledgers keep write costs bounded per collection rather than per registry.

Why move beyond authority-only mint? A user-held asset should not appear in a holder's private asset set solely because the registry authority decided to place it there. Requiring designated-holder acceptance makes issuance match the transfer security model and makes provenance meaningful to wallets and auditors. This is the main funded implementation change (M2).

Fairmint Implementation Role. Fairmint operates a Canton validator and a Featured App in production, authored the Open Cap Table Protocol DAML packages, and built the NFT prototype because its equity-certificate roadmap requires it. This grant funds the work for the broader Canton ecosystem.

---

Appendix A — ERC-721 / EIP-4906 semantic mapping

Normative mapping for teams migrating EVM NFT assets to Canton. Fairmint uses it in Milestones 3–4 for its own production ERC-721 equity certificates.

ERC-721 / EVM concept	Canton NFT profile equivalent
(contractAddress, tokenId)	Stable item identity anchored to the V2 InstrumentId, independent of ledger ContractId
ownerOf	Owner of the single live amount = 1.0 V2 holding
tokenURI	Structured per-item metadata plus optional confidentiality_ref
ERC-721 Metadata JSON (name/description/image)	Reserved nft.canton.network/ standard metadata keys
safeTransferFrom	Standard CIP-0112 transfer flow (recipient consent native to Canton)
EIP-4906 MetadataUpdate events	Standard holding-change events plus versioned metadata in the profile view
Soulbound (e.g. EIP-5192-style)	Profile soulbound mode: no advertised transfer support; every ownership-changing path rejected
Minting to a third party	Mint acceptance path (designated-holder consent required)
Per-token approval / operator-for-all / enumeration	Not standardized in the profile (enumeration conflicts with privacy-by-default)
Atomic swap / DVP	Native via CIP-0112 allocations; no NFT-specific extension needed