Back to DashboardIndustriesPhase 5 · Pilots & Migrationintermediate40 min

Custody PQC Migration

What a quantum attacker changes for digital-asset custody and how to migrate — HSM/MPC/enclave quantum deltas, PQ threshold MPC research, SLIP-39 seed backups, hardware wallet claims, PQ smart accounts, and phased migration.

Why this matters: None of today’s custody architectures is post-quantum on its own: classical threshold signatures output ordinary ECDSA/EdDSA signatures, share transport and backup envelopes are harvest-now-decrypt-later targets, and enclave attestation chains remain classical. Custodians can migrate platform-controlled channels, backup envelopes, and approval signatures now while preparing account- and chain-controlled paths separately.

Start here: Start with the Quantum Limits & HNDL Check to pinpoint what a CRQC breaks first in each custody combination and which party-to-party messages are exposed to harvest-now-decrypt-later today.

Practice in the Simulation

What a Quantum Attacker Changes for Custody — and How to Migrate

Prerequisites: Digital Assets (LM-045) and Module A1 (Custody Key Threats & Controls). This module links LM-045 for chain primitives, BIP-360/ERC-4337/Solana mechanics, and HD wallet fundamentals rather than repeating them, focusing strictly on what a cryptographically relevant quantum computer (CRQC) changes across HSMs, MPC, enclaves, seed backups, hardware wallets, and smart accounts.

Rule G1 (authority assumptions) & educational scope

Every claim that a custody control holds states which signer, guardian, recovery, and upgrade authorities an attacker must not control for it to hold. All demonstrations use fictional configurations for education and do not certify or verify deployed custody platforms.

7. Quantum Limits of HSM, MPC & Enclaves (Delta)

LM-045 PQC Defense Part 1 already establishes that Shor’s algorithm breaks exposed elliptic-curve public keys regardless of whether keys sit in an HSM, MPC cluster, or enclave, and that classical enclave ECDSA attestation is forgeable. Below are the custody-specific deltas beyond LM-045:

LM-045 LinkResearch

Shor’s algorithm recovers ECDSA and EdDSA private keys from exposed on-chain public keys regardless of whether the key is stored inside an HSM, MPC cluster, or hardware wallet.

Source: Project Eleven — Quantum Threat to Blockchains (2026-05-05), Executive Summary & Exposed Public Keys Analysis

LM-045 LinkResearch

Blockchain networks that rely on elliptic-curve public-key signatures face forgery once a CRQC can derive private keys from exposed public keys; storage isolation alone does not alter the public-key algorithm.

Source: Cambridge JBS / CCAF (2025-11-26), Cryptographic Exposure of Distributed Ledgers

HSM Delta — Classical Channels (HNDL)

Even when signing keys never leave the HSM, TLS/mTLS channels into HSM clusters carry key-share fragments, encrypted backups, and confidential policy payloads under classical key exchange — an immediate harvest-now-decrypt-later (HNDL) confidentiality risk.

HNDLResearch

Confidential data transmitted or stored under classical key exchange today can be recorded and decrypted later (HNDL) once a CRQC is available; in custody, share transport and encrypted backups are immediate HNDL targets.

Source: Federal Reserve FEDS Working Paper 2025-093 (2025-09-30), HNDL & Data Privacy Analysis

HNDL vs ForgeryStandard

NIST IR 8547 separates confidentiality risks requiring immediate HNDL mitigation from digital-signature authentication risks that must transition before a CRQC enables forgery.

Source: NIST IR 8547 (Final, October 2026), Transition Timeline & Risk Prioritization

MPC Delta — Classical Output, Factoring Internals & No NIST PQ Threshold Standard

Threshold Schnorr/EdDSA (RFC 9591 FROST) and threshold ECDSA (CMP, CGGMP21) emit ordinary classical signatures. Inside CMP and CGGMP21, Paillier encryption and Strong-RSA-style assumptions are factoring-based, and pairwise share transport over classical key exchange is an HNDL target.

MPC OutputStandard

RFC 9591 FROST produces a standard single Schnorr/EdDSA signature indistinguishable from a single-party signature, so its on-chain output carries the same elliptic-curve quantum exposure.

Source: IETF RFC 9591 (Informational, June 2024), §1 Introduction & Ciphersuites

CMP PaillierResearch

The CMP threshold-ECDSA protocol relies on Paillier homomorphic encryption and Strong-RSA-style assumptions (factoring-based), making its internal protocol messages quantum-vulnerable in addition to its ECDSA output.

Source: Canetti et al., IACR ePrint 2020/492 (2020-04-28), §1 Introduction & Cryptographic Assumptions

CGGMP21Research

CGGMP21 extends proactive threshold ECDSA with identifiable aborts while retaining Paillier/Strong-RSA algebraic assumptions and pairwise channels whose classical key exchange is an HNDL target.

Source: Canetti et al., IACR ePrint 2021/060 (updated 2024-10-21), §1 Overview & Protocol Setup

Negative · NIST ProcessStandardAs of 2026-11-20

As of 20 November 2026, no NIST-standardised threshold variant of ML-DSA, SLH-DSA, or FN-DSA exists; NIST IR 8214C opens a call for multi-party threshold schemes leading to public analysis and a characterization report rather than a competition selection.

Source: NIST IR 8214C (Final, January 2026), §1 Purpose & Submission Phases

Confidential Computing Delta — Classical Attestation Chains & Attested TLS

Today’s TEE attestation chains verify under classical elliptic-curve roots (SGX DCAP, Nitro COSE ES384, AMD SEV-SNP VCEK P-384, Intel TDX DCAP). Forging attestation lets a rogue enclave receive keys or signing approvals. Meanwhile, attested TLS requires hybrid key exchange to stop passive recording.

Intel SGXVendor shipped

Intel SGX DCAP remote attestation relies on ECDSA-signed quotes verified through the Intel Quote Verification Library API.

Source: Intel SGX DCAP ECDSA Quote Library API (2026-09), Overview & Quote Verification

AWS NitroVendor shipped

AWS Nitro Enclaves attestation documents are COSE-signed with ECDSA P-384 (ES384) chaining to the Nitro root of trust.

Source: AWS Nitro Enclaves Developer Guide, Verifying the Root of Trust

AMD SEV-SNPVendor shipped

AMD SEV-SNP attestation chains to the Versioned Chip Endorsement Key (VCEK) X.509 hierarchy using ECDSA P-384 signatures.

Source: AMD Publication #57230 Rev 1.05 (2026-09), VCEK Certificate Hierarchy

Intel TDXDraft

Intel TDX DCAP quoting libraries follow the ECDSA quote generation and verification API model.

Source: Intel TDX DCAP Quoting Library API Rev 0.91 (2026-09), Quote Generation & Verification Interfaces

RATSStandard

IETF RFC 9334 (RATS Architecture) separates Attester, Verifier, Relying Party, and Endorser roles in an algorithm-agile architecture.

Source: IETF RFC 9334 (Informational, January 2023), §3 Architectural Overview & Roles

EATStandard

IETF RFC 9711 defines the Entity Attestation Token (EAT) format over COSE/JOSE claims, allowing algorithm-agile attestation tokens.

Source: IETF RFC 9711 (Standards Track, April 2025), §1 Introduction & COSE/JOSE Encoding

Attested TLSResearch

Attested TLS binds remote enclave attestation to a TLS session channel; unless the TLS handshake uses hybrid PQ key exchange, recorded attested-TLS traffic remains vulnerable to HNDL.

Source: Knauth et al. (2018-01-17), Attested TLS Handshake Binding

8. How Each Custody Model Is Evolving (Standard / Draft / Vendor)

Control-plane signatures (backend approvals, policy rules, audit logs) and transport channels can move to post-quantum standards before Layer-1 blockchains do — while hardware FIDO2/WebAuthn authenticators and PQ threshold MPC remain on slower tracks.

HSM Firmware, PKCS#11 v3.2 & Hybrid TLS (RFC 10024)

CAVP ML-KEMVendor shipped

CAVP validation A6129 covers Securosys Primus HSM v3.1.0 ML-KEM firmware algorithm conformance; CAVP validates the algorithm implementation, not full CMVP module validation.

Source: NIST CAVP Record A6129 (2024-11-07), ML-KEM KeyGen & EncapDecap Validation

CAVP ML-DSAVendor shipped

CAVP validation A6130 covers Primus HSM v3.1.0 ML-DSA KeyGen, SigGen, and SigVer algorithm conformance.

Source: NIST CAVP Record A6130 (2024-11-07), ML-DSA Algorithm Validation

CAVP SLH-DSAVendor shipped

CAVP validation A6131 covers Primus HSM v3.1.0 SLH-DSA algorithm conformance.

Source: NIST CAVP Record A6131 (2024-11-07), SLH-DSA Algorithm Validation

HSM FirmwareVendor shipped

Thales Luna HSM Firmware v7.9 announces built-in ML-KEM and ML-DSA mechanisms and hybrid PQC protection for high-availability key synchronization.

Source: Thales Luna HSM Firmware v7.9 Release Note (2025-06-27), Firmware PQC Capabilities

PKCS#11 v3.2Standard

OASIS PKCS#11 v3.2 standardises ML-KEM encapsulate/decapsulate mechanisms alongside ML-DSA and SLH-DSA signature mechanisms.

Source: OASIS PKCS#11 v3.2 Standard (2026-06-03), PQC Mechanisms (ML-KEM, ML-DSA, SLH-DSA)

Hybrid TLSStandard

IETF RFC 10024 standardises post-quantum/traditional (PQ/T) hybrid key agreement mechanisms for TLS 1.3, enabling authenticated, transcript-bound hybrid channels today.

Source: IETF RFC 10024 (Proposed Standard, August 2026), §1 Introduction & Hybrid Key Shares

PQ Hardware Attestation & FIDO2 / WebAuthn Status

TPM 2.0 V185Standard

TCG TPM 2.0 Library V185 introduces post-quantum algorithm support across its architecture, structures, and commands.

Source: TCG TPM 2.0 Library V185 Part 0 (March 2026), Introduction & PQC Scope

TPM Part 2Standard

TPM 2.0 Library V185 Part 2 assigns algorithm identifiers for ML-DSA and ML-KEM.

Source: TCG TPM 2.0 Library V185 Part 2 (March 2026), Algorithm Identifiers (ML-DSA, ML-KEM)

TPM Part 3Standard

TPM 2.0 Library V185 Part 3 defines KEM-key commands such as TPM2_Encapsulate and TPM2_Decapsulate alongside ML-DSA signing commands.

Source: TCG TPM 2.0 Library V185 Part 3 (Published, March 2026), TPM2_Encapsulate & Signing Commands

SPDM 1.4.1Standard

DMTF SPDM 1.4.1 adds post-quantum signature and key-exchange support for device authentication and hardware measurement attestation.

Source: DMTF DSP0274 SPDM 1.4.1 (2026-06-26), PQC Algorithms & Measurements

Caliptra 2.1Standard

Caliptra 2.1 specifies dual ECDSA P-384 + ML-DSA-87 / LMS support for firmware manifest verification and command authorization while noting DICE certificate and quote extensions across revisions.

Source: Caliptra Root of Trust Revision 2.1, Firmware Signing, Authorization & DICE Certificates

TCG EK v2.7Standard

TCG EK Credential Profile v2.7 updates endorsement key credential structures for TPM 2.0 attestation roots.

Source: TCG EK Credential Profile v2.7 (2026-03-16), Endorsement Key Profiles

PQ RATS DraftDraftAs of 2026-11-20

Post-quantum guidance for RATS Evidence and Attestation Results is currently an individual IETF Internet-Draft, not a finalised standard.

Source: draft-elekula-rats-pqc-attestation-00 (2026-11-16), §1–§3 PQC Considerations for RATS

COSE ML-DSAStandard

IETF RFC 9964 registers ML-DSA algorithm identifiers for JOSE and COSE.

Source: IETF RFC 9964 (Published, May 2026), COSE & JOSE Algorithm Registrations

Negative · WebAuthn HWStandardAs of 2026-11-20

W3C WebAuthn Level 3 accepts registered COSE algorithm identifiers, but hardware FIDO2/CTAP authenticator PQ deployment is not yet established as standard across commercial devices as of 20 November 2026.

Source: W3C WebAuthn Level 3 (May 2026), COSEAlgorithmIdentifier Registry Usage

FIDO Server DraftDraft

The FIDO Server Requirements v2.3 review draft recommends ML-DSA COSE identifiers at the server layer while hardware authenticator / CTAP rollout remains pending.

Source: FIDO Server Requirements v2.3 Review Draft (2026-02-26), Supported COSE Algorithms

Chain-Side Proposals (Linked to LM-045)

Bitcoin DraftDraft

Bitcoin BIP-360 (Draft) proposes Pay-to-Merkle-Root (P2MR), a soft-fork output type removing the vulnerable Taproot key-path spend; mechanics are covered in LM-045 Part 2.

Source: Bitcoin BIP-360 (Draft, v0.12.1), Abstract & Specification

Bitcoin DraftDraft

Bitcoin BIP-361 (Draft) outlines a phased sunset of legacy ECDSA/Schnorr spends once a post-quantum output type is active.

Source: Bitcoin BIP-361 (Draft, 2026-02-11), Migration Phases & Legacy Sunset

EthereumResearch

The Ethereum post-quantum roadmap tracks consensus and execution-layer migration paths including account abstraction and signature-scheme upgrades.

Source: ethereum.org Post-Quantum Cryptography (Living Document), Roadmap & Account Abstraction

8b. Post-Quantum Threshold Signing & MPC (Research — Not Standardised)

All schemes below are research as of 20 November 2026. Splitting ML-DSA’s rejection sampling and secret nonce vectors across parties without leaking information or inflating signatures is much harder than FROST; Falcon’s trapdoor Gaussian sampling is harder still. Meanwhile, upgrading MPC pairwise channels to RFC 10024 hybrid TLS is the near-term quantum fix that does not wait for PQ threshold signatures.

FamilyScheme / PaperVerifier CompatibilityRounds & PartiesSignature Size
Classical Threshold Schnorr / EdDSA (today)RFC 9591 FROST (secp256k1 / Ed25519)Standard Schnorr / EdDSA verifier (quantum-vulnerable)2 rounds (1 with preprocessing) · Arbitrary t-of-n threshold64 B
Classical Threshold ECDSA (today)CMP (ePrint 2020/492) / CGGMP21 (ePrint 2021/060)Standard ECDSA verifier (quantum-vulnerable; Paillier internal)3–4 rounds (presigning + 1 online) · Dishonest majority (t-of-n)64–65 B
FIPS 204-Compatible Threshold ML-DSAQuorus (ePrint 2025/1163, USENIX Security 2026)Unmodified FIPS 204 ML-DSA verifierMulti-round MPC with rejection-sampling retries · Honest majority (scales to larger groups)2,420 B (ML-DSA-44) / 3,309 B (ML-DSA-65)
FIPS 204-Compatible Threshold ML-DSAMithril / Efficient Threshold ML-DSA (ePrint 2026/013)Unmodified FIPS 204 ML-DSA verifier3–4 rounds (small-group optimized) · Small signing quorums (e.g., 2-of-3, 3-of-5)2,420 B (ML-DSA-44) / 3,309 B (ML-DSA-65)
Threshold-Friendly Lattice Variant (non-FIPS-204)Threshold Raccoon / TRaccoon / Hermine / RingtailRequires new chain / verifier algorithm (not FIPS 204)2–3 rounds (no rejection-sampling restarts) · Varies by scheme (some with identifiable aborts)Scheme-specific (larger than classical, non-FIPS)
Early Hash-Based Threshold ResearchAker (ePrint 2026/2899) / Discrete-CKKS FHE (ePrint 2026/2360)FIPS 205 SLH-DSA / XMSS (research trade-offs)2-party stateful cap (Aker) or 1-round FHE (2026/2360) · 2-party or threshold FHE committee7,856 B (SLH-DSA-128s) / 17,088 B (128f)
Why It Is HardResearch

Thresholdizing lattice and hash-based signatures is significantly harder than Schnorr/FROST because ML-DSA relies on rejection sampling over secret nonce vectors and Falcon requires discrete Gaussian sampling over a secret trapdoor.

Source: Comprehensive Survey of Threshold Signatures (arXiv:2311.05514), Post-Quantum Threshold Signatures Section

FIPS 204-CompatibleResearch

Quorus (ePrint 2025/1163, USENIX Security 2026) constructs FIPS 204-compatible threshold ML-DSA under an honest-majority MPC model, outputting signatures verifiable by unmodified ML-DSA verifiers.

Source: IACR ePrint 2025/1163 (2025-06-19; USENIX Security 2026), Abstract & Evaluation Tables

FIPS 204-CompatibleResearch

Efficient Threshold ML-DSA / Mithril (ePrint 2026/013, USENIX Security 2026) targets small signing groups with FIPS 204-compatible output.

Source: Celi et al., IACR ePrint 2026/013 (2026-01-05), §1 Introduction & Benchmarks

FIPS 204-CompatibleResearch

Brave’s published research summary highlights Efficient Threshold ML-DSA as a FIPS 204-compatible threshold signing construction suitable for small quorums, still at research maturity.

Source: Brave Research Summary (2026), Overview & Parameter Trade-offs

Preprint · FIPS 204Research

TALUS (arXiv:2603.22109) is a preprint proposing FIPS 204-exact threshold ML-DSA using boundary clearance techniques.

Source: arXiv:2603.22109 Preprint (March 2026), Abstract & Design Overview

Preprint · FIPS 204Research

Shamir-nonce DKG (arXiv:2601.20917) is a preprint exploring FIPS 204-compatible threshold ML-DSA via distributed Shamir nonce generation.

Source: arXiv:2601.20917 Preprint (January 2026), Abstract & Nonce DKG Protocol

Non-FIPS VariantResearch

Threshold Raccoon (ePrint 2024/184) modifies the signature scheme to eliminate rejection sampling via additive masking, improving round efficiency at the cost of FIPS 204 verifier incompatibility.

Source: IACR ePrint 2024/184 (2024-02-07), Abstract & Masking Design

Non-FIPS VariantResearch

TRaccoon (ePrint 2025/849, CRYPTO 2025) adds efficient identifiable aborts to lattice threshold signing; chains or verifiers would need to adopt a new non-FIPS-204 verification algorithm.

Source: IACR ePrint 2025/849 (2025-05-13), Abstract & Identifiable Aborts

Non-FIPS VariantResearch

Hermine (ePrint 2026/419) proposes a FROST-like lattice threshold signature outside the FIPS 204 wire format.

Source: IACR ePrint 2026/419 (2026-03-02), Abstract & Protocol Summary

Non-FIPS VariantResearch

Ringtail (ePrint 2024/1113, IEEE S&P 2025) achieves practical two-round LWE-based threshold signatures, but its outputs are not verifiable by FIPS 204 ML-DSA verifiers.

Source: IACR ePrint 2024/1113 (2024-07-09; IEEE S&P 2025), Abstract & Comparison Table

SLH-DSA ResearchResearch

Aker (ePrint 2026/2899, 15 November 2026 preprint) explores a two-party SLH-DSA construction verifying under FIPS 205 with signature-count and statefulness trade-offs.

Source: IACR ePrint 2026/2899 Preprint (2026-11-15), Abstract & Construction Trade-offs

Threshold FHE/CKKSResearch

ePrint 2026/2360 investigates one-round threshold XMSS and SPHINCS+ built over threshold fully homomorphic encryption instantiated with discrete CKKS rather than TFHE.

Source: IACR ePrint 2026/2360 Preprint (2026-10-05), Abstract & Discrete-CKKS Instantiation

NIST MPTS 2026Standard

NIST’s MPTS 2026 workshop (January 2026) convened threshold ML-DSA researchers as part of the IR 8214C multi-party threshold characterization effort.

Source: NIST MPTS 2026 Workshop Proceedings (2026-01-26), Program & Threshold PQC Sessions

Custody AlternativeResearch

Signature-agnostic MPC custody research (arXiv:2607.08226) separates PQ member authentication (ML-DSA / SLH-DSA) from secret-shared quorum authorization while threshold PQ signing matures.

Source: arXiv:2607.08226 Preprint (July 2026), Abstract & Custody Architecture

5c. Shamir Secret Sharing for Seeds (SLIP-39) & VSS Deltas

LM-045 (Cold Wallet & HD Wallet steps) covers Grover on seeds, Shamir-vs-MPC, and PQ-capable wrapping; Module A2 owns recovery operations. Here we examine the format internals, quantum reach into BIP-32/VSS, and envelope deltas.

SLIP-39 InternalsStandard

SLIP-0039 replaces BIP-39 mnemonics with byte-wise GF(256) Shamir secret sharing, a four-round PBKDF2-HMAC-SHA256 Feistel passphrase layer, an interpolated 4-byte digest, and an RS1024 checksum.

Source: SatoshiLabs SLIP-0039 (Final, 2017-12-18), Specification: Secret Sharing, Feistel & RS1024

Ideal ShamirStandard

In ideal Shamir secret sharing over a finite field, any subset of fewer than t shares leaves the probability distribution of the secret completely unchanged (information-theoretic secrecy).

Source: Adi Shamir, Comm. ACM (May 1979), §2 Polynomial Interpolation Scheme

BIP-32 xpub + ShorStandard

Under BIP-32 non-hardened derivation, knowledge of one child private key plus the parent extended public key (xpub) recovers the parent private key; Shor supplies the child private key from its on-chain public key, whereas hardened derivation blocks parent recovery.

Source: Bitcoin BIP-32 (Final, 2012-02-11), Private Parent Key -> Private Child Key & Security Note

Hybrid KEM-DEMDraft

Hybrid HPKE with post-quantum KEMs (draft-ietf-hpke-pq) provides a KEM-DEM envelope (encapsulation + KDF + AEAD) to replace RSA/ECDH share wrapping for stored backups.

Source: IETF draft-ietf-hpke-pq (2026-07-06), Hybrid KEM-DEM Suite Definitions

Feldman VSSResearch

Feldman Verifiable Secret Sharing publishes discrete-log commitments g^s; a CRQC running Shor’s algorithm inverts g^s and recovers the secret directly from public verification data.

Source: Paul Feldman, FOCS 1987, Non-Interactive VSS Protocol

Pedersen VSSResearch

Pedersen VSS provides information-theoretic hiding (resisting secret recovery even against a CRQC), but its computational binding relies on discrete logarithms and breaks under Shor.

Source: Torben Pryds Pedersen, CRYPTO 1991, Commitment Hiding vs Binding Properties

Lattice PVSSResearch

ePrint 2025/901 proposes a generic framework for lattice-based non-interactive publicly verifiable secret sharing (PVSS).

Source: IACR ePrint 2025/901 (2025-05-20), Abstract & Framework

Lattice PVSSResearch

arXiv:2504.14381 studies generic PVSS constructions and lattice-based instantiations for post-quantum verifiability.

Source: arXiv:2504.14381 Preprint (April 2025), Abstract & Lattice Instantiations

PVSS Binding DeltaResearch

ePrint 2026/813 distinguishes a lattice-IBE + hash-commitment PVSS construction from a second construction using Pedersen commitments whose binding falls to Shor (yielding long-lasting privacy, not full post-quantum binding).

Source: IACR ePrint 2026/813 (2026-04-24), Constructions 1 & 2 Security Comparison

Analysis · Seed -> ML-DSAStandardAs of 2026-11-20

FIPS 204 defines deterministic ML-DSA key generation from a single 32-byte random seed ξ; deriving ξ from a wallet master seed via hardened HMAC is a proposed custody construction, not FIPS behaviour.

Source: NIST FIPS 204 (Final Standard, August 2024), Algorithm 6 (ML-DSA.KeyGen_internal) & Parameter Sizes

FIPS 203 SeedsStandard

FIPS 203 defines deterministic ML-KEM key generation from two 32-byte random seeds (d, z); compact seed backup works only when the HSM or wallet retains and restores that seed form.

Source: NIST FIPS 203 (Final Standard, August 2024), Algorithm 16 (ML-KEM.KeyGen_internal)

Negative status & proposed construction (as of 20 November 2026)

No standard PQ HD-wallet derivation path or xpub-style public derivation equivalent exists as of 20 November 2026. Deriving ML-DSA (ξ) or ML-KEM (d, z) seeds from a wallet master seed via hardened HMAC requires domain separation and explicit backup compatibility — it is a proposed construction, not FIPS 203/204 behaviour.

5d. What “Quantum-Ready” Means on a Hardware Wallet

Always split any hardware-wallet post-quantum claim into four parts: (1) device integrity & attestation, (2) chain-accepted transaction signing, (3) seed & derivation, and (4) device-to-host channel & clear-signing display. In institutional custody (Module A1 §5), hardware wallets also act as approver devices with trusted displays: PQ device attestation does not establish PQ approval signing without separate vendor evidence. For physical side channels during PQ signing, see PQC Implementation Security (LM-081).

Part 1 · Device IntegrityVendor shipped

Trezor Safe 7 documents SLH-DSA-128 (hybrid with Ed25519) for boardloader firmware verification and ML-DSA-44 for device attestation — covering Part 1 (device integrity), not chain transaction signing or PQ approval signing.

Source: Trezor Safe 7 Technical Article, Quantum-Ready Bootloader & Device Attestation

Part 2 · AnnouncedVendor announced

Freedom Factory PQ1 (announced) describes a vendor-defined compressed SPHINCS+ variant ("SPHINCS+ C10", not a FIPS 205 parameter set) for EVM smart-account transaction signing; shipping status is not established by the whitepaper alone.

Source: Freedom Factory PQ1 Whitepaper Rev 1.0 (2026-05-01), Architecture & Signature Scheme

Part 4 · Sig SizesStandard

FIPS 205 SLH-DSA-128s signatures are 7,856 bytes and SLH-DSA-128f signatures are 17,088 bytes (versus 2,420 bytes for FIPS 204 ML-DSA-44), stressing QR/air-gap transport, USB/BLE framing, and on-device clear signing.

Source: NIST FIPS 205 (Final Standard, August 2024), Table 2 (SLH-DSA Parameter Sets)

MCU TimingsResearch

Blockstream Research benchmarks hash-based signature schemes (SLH-DSA, reduced SPHINCS+, stateful variants) on wallet-class microcontrollers (Jade, Trezor, Ledger, BitBox02) to measure signing latency and memory pressure.

Source: Blockstream Research (2026-08-19), Hardware Wallet Hash-Based Benchmarks

SE Side ChannelsResearch

Ledger Donjon’s analysis examines secure-element RAM/flash limits and electromagnetic/power side-channel key recovery against unprotected PQC implementations during signing (cross-reference LM-081).

Source: Ledger Donjon (2026-04-29), Side-Channel Attacks on PQC Implementations

SE AnnouncementVendor announced

A Thales press release announces support for Samsung’s S3SSE2A embedded secure element as a post-quantum-capable chip without naming specific PQC algorithms in the release text.

Source: Thales Press Release (2026-01-06), S3SSE2A Embedded Secure Element Announcement

ANSSI Scope CheckStandard

ANSSI certification report ANSSI-CC-2024/26-R01 for Samsung S3SSE2A lists an ML-DSA library in documentation outside the evaluated certified cryptographic services, illustrating why cert scope must be read closely.

Source: ANSSI-CC-2024/26-R01 (2025-10-01), Scope & Documentation List

Five questions to test any hardware wallet PQC claim

  • 1. Which of the four parts does the claim actually cover (device integrity, transaction signing, seed/derivation, or host channel/display)?
  • 2. Which exact algorithms and parameter sets are used (e.g., FIPS 204 ML-DSA-44 / FIPS 205 SLH-DSA-128s vs a vendor-defined variant)?
  • 3. Is the post-quantum path enforced unconditionally, or is it optional / hybrid with a classical fallback?
  • 4. Which target blockchains or smart accounts actually accept the PQ signatures the device produces?
  • 5. Is the implementation validated (CAVP algorithm conformance vs full CMVP / CC module certification) or only vendor-stated?

8d. Post-Quantum Smart Accounts from the Custody Point of View

Mechanics of ERC-4337, EIP-7702, EIP-8141, and Solana are covered in LM-045 PQC Defense Parts 3–4. From a custody perspective, a smart account enforces on-chain quorum, timelock, allowlist, and recovery rules that hold against a compromised custodian stack only if (Rule G1) that stack does not hold enough signer, guardian, recovery, or upgrade authority to alter them — combining with Module A1 §5 off-chain policy boundaries.

ERC-4337 PremiseStandard

ERC-4337 lets smart accounts supply arbitrary validation logic inside validateUserOp (premise only; the ERC itself names no post-quantum algorithm). Bundler ECDSA transactions are a liveness/fee dependency, not an authorization dependency.

Source: Ethereum ERC-4337 (Final), EntryPoint & validateUserOp Specification

EIP-7702 Residual KeyStandard

Under EIP-7702, an EOA’s secp256k1 key writes a persistent delegation indicator (0xef0100 || address) that remains until replaced or cleared — so that legacy secp256k1 key retains re-delegation control unless explicitly deactivated.

Source: Ethereum EIP-7702 (Final), Specification & Persistent Delegation Semantics

Permit / ecrecoverStandard

ERC-20 tokens implementing EIP-2612 permit verify off-chain EIP-712 secp256k1 signatures via ecrecover, meaning an EIP-7702 account’s legacy secp256k1 key can still authorize token transfers on permit-enabled contracts.

Source: Ethereum EIP-2612 (Final, 2020-04-13), Specification (permit & ecrecover)

Draft · Key RevocationDraft

EIP-7851 (Draft) proposes allowing EIP-7702 delegated accounts to update delegation via contract code and permanently revoke the residual secp256k1 key.

Source: Ethereum EIP-7851 (Draft, 2024-12-27), Abstract & Key Revocation

Negative · No Mainnet PrecompileDraftAs of 2026-11-20

EIP-8051 (Draft) proposes Ethereum precompiles separating a FIPS 204-compatible ML-DSA verifier from an EVM-friendly ML-DSA-ETH variant; as of 20 November 2026, no PQ precompile is live on Ethereum mainnet.

Source: Ethereum EIP-8051 (Draft, 2025-10-15), Specification (FIPS 204 vs ML-DSA-ETH)

Draft · Falcon PrecompileDraft

EIP-8052 (Draft) proposes an Ethereum precompiled contract for Falcon-512 verification.

Source: Ethereum EIP-8052 (Draft, 2025-10-17), Abstract & Precompile Interface

Draft · ML-DSA PrecompilesDraft

EIP-8355 (Draft) proposes three Ethereum precompiles for ML-DSA-44, ML-DSA-65, and ML-DSA-87 verification.

Source: Ethereum EIP-8355 (Draft, 2026-07-30), Specification (ML-DSA-44/65/87)

Draft · EIP-8141Draft

EIP-8141 (Draft) proposes frame transactions splitting validation, execution, and gas payment at the protocol level.

Source: Ethereum EIP-8141 (Draft, 2026), Abstract & Frame Structure

Draft · EIP-7932Draft

EIP-7932 (Draft) proposes a secondary signature algorithm registry and sigrecover precompile for Ethereum accounts.

Source: Ethereum EIP-7932 (Draft, 2025), Algorithm Registry & sigrecover

Prototype · ML-DSAPrototype

ZKNOX ETHDILITHIUM provides experimental Solidity on-chain verifiers for FIPS 204 ML-DSA and an EVM-optimized variant as unaudited prototypes with multi-million-gas execution costs.

Source: ZKNoxHQ/ETHDILITHIUM Prototype Repository, Solidity Verifier & Gas Notes

Prototype · FalconPrototype

ZKNOX ETHFALCON provides experimental Solidity Falcon signature verifiers as an unaudited prototype.

Source: ZKNoxHQ/ETHFALCON Prototype Repository, Solidity Verifier Contracts

Contrast · AlgorandVendor announced

Algorand’s post-quantum roadmap outlines native Falcon accounts, cryptographic agility, and hybrid/multisig account evolution as a chain-level contrast to EVM smart-account verification.

Source: Algorand Foundation PQC Roadmap (2026-06-18), Native Falcon Accounts & Roadmap

Contrast · AlgorandVendor shipped

Algorand’s technical brief documents a mainnet transaction verified with Falcon post-quantum signatures.

Source: Algorand Technical Brief (2025-11-03), Falcon Mainnet Transaction Verification

Hybrid accounts & key-storage recursion (Rule G1)

A hybrid smart account requiring both ECDSA and ML-DSA protects against a break of either signature only if no classical recovery guardian, upgrade admin, or EIP-7702 legacy key can bypass the check. Remember also that the smart account’s PQ signing key must itself live inside a protected boundary (hardware wallet §5d, ML-DSA HSM, or enclave) — Module A1’s signing-authority rules still apply.

8c. Phased PQC Migration Path for a Custody Platform

Builds on LM-045 component readiness notes and links Migration Program, Crypto Agility, and CBOM. Separate every custody dependency into three clocks: platform-controlled, account-controlled, and chain-controlled.

  1. Phase 0 — Inventory (CBOM): Catalog all HSM/MPC/enclave channels, backups, approval keys, and exposed on-chain addresses; separate confidentiality (HNDL) from authentication (future forgery).
  2. Phase 1 — No-Regret Now (Platform-Controlled, HNDL First): Enable RFC 10024 hybrid TLS on HSM/MPC/enclave channels; re-wrap stored backups under hybrid KEM-DEM; upgrade to CAVP/CMVP-scoped ML-KEM/ML-DSA HSM firmware; move backend approvals and audit archives to ML-DSA + RFC 4998 renewal; enforce address hygiene.
  3. Phase 2 — Chain-Dependent (Track & Prepare): Maintain per-chain watch lists (BIP-360/361, Ethereum, Solana via LM-045 Parts 2–4), prepare sweep tooling, and budget for downtime/fee congestion.
  4. Phase 3 — MPC & Threshold: Keep classical threshold for L1 output wrapped in Phase 1 PQ channels/backups; evaluate separated ML-DSA member auth or smart-account multisig until NIST IR 8214C / threshold standards mature.
  5. Phase 4 — Crypto Agility & Evidence: Keep algorithm choice in configuration (NIST CSWP 39), verify ACVP test vectors, and map vendor evidence to regulatory frameworks.
Phase 0 · DiscoveryDraft

NIST NCCoE SP 1800-38B outlines cryptographic discovery and inventory workflows for building a complete dependency baseline before migration.

Source: NIST NCCoE SP 1800-38B (Preliminary Draft), Cryptographic Discovery Architecture

Phase 0 · CBOMStandard

CycloneDX Cryptography Registry (v1.7) provides structured CBOM fields to inventory keys, algorithms, protocols, and certificates across custody components.

Source: CycloneDX Cryptography Registry v1.7 (2025-10-21), Crypto-Asset Schema Fields

Phase 4 · AgilityStandard

NIST CSWP 39 defines crypto-agility practices: treating algorithm choice as configuration rather than hardcoded logic so custody stacks can swap primitives cleanly.

Source: NIST CSWP 39 (Final, 2025-12-19), Crypto-Agility Design Principles

Phase 2 · DowntimeResearch

Pont et al. (preprint) estimate the block-space and downtime required to sweep Bitcoin UTXOs to post-quantum outputs under congestion.

Source: Downtime Required for Bitcoin Quantum-Safety (2024-10-22), Migration Throughput & Downtime Model

Phase 2 · FeesResearch

Fukuda (2025) analyzes signature-size, throughput, and fee trade-offs when migrating major blockchains to post-quantum signatures.

Source: Fukuda (2025), Blockchain PQC Migration Cost Analysis

Regulator · MASStandard

MAS’s 2024 quantum advisory urges financial institutions to inventory cryptographic assets, assess HNDL exposure, and engage vendors on PQC roadmaps.

Source: MAS Advisory (2024-02-20), Recommended Actions for Financial Institutions

Regulator · HKMAStandard

HKMA’s Fintech 2030 Blueprint (February 2026) sets out quantum-preparedness expectations for Hong Kong financial institutions.

Source: HKMA Fintech Promotion Blueprint (2026-02-03), Quantum Preparedness Pillar

Regulator · BISStandard

BIS Paper 158 outlines a phased quantum-readiness roadmap for financial systems emphasizing inventory, governance, and testing.

Source: BIS Paper 158 (July 2025), Financial System Migration Roadmap

Financial · ASC X9Standard

ASC X9’s 2025 readiness assessment identifies financial-sector PQC migration gaps across HSMs, payments, and key management.

Source: ASC X9 Readiness Assessment (2025-08-05), Financial Sector Readiness Findings

CISA / NSA / NISTStandard

CISA/NSA/NIST Quantum-Readiness guidance recommends starting with cryptographic discovery, HNDL prioritization, and structured vendor questionnaires.

Source: CISA/NSA/NIST Quantum-Readiness (2023-08-21), Discovery & Vendor Engagement Steps

Negative · MiCAStandardAs of 2026-11-20

As of 20 November 2026, EU MiCA (Regulation EU 2023/1114) contains no explicit post-quantum cryptography mandate or cutover date.

Source: EU Regulation 2023/1114 (MiCA), Custody & ICT Security Provisions (negative check)

Negative · DORAStandardAs of 2026-11-20

As of 20 November 2026, EU DORA (Regulation EU 2022/2554) requires sound cryptographic lifecycle governance but contains no explicit PQC algorithm mandate.

Source: EU Regulation 2022/2554 (DORA), ICT Risk Management Provisions (negative check)

Worked Example · Vendor RoadmapVendor announced

Circle’s published 2026 PQC roadmap illustrates one digital-asset issuer/custodian’s public plan across infrastructure, smart contracts, and chain tracking (vendor roadmap, not an industry standard).

Source: Circle PQC Security Roadmap (2026-05-30), Phased Infrastructure & Chain Plan

7b. Quantum Views of Custody Operations (Cross-Module Deltas)

MPC + HSM Hybrids (Module A1 §5b)

Output remains classical ECDSA/EdDSA; replace RSA/ECDH share wrapping with a KEM-DEM envelope (ML-KEM encapsulation + KDF + AEAD, not ML-KEM wrapping alone; PKCS#11 v3.2 defines encapsulate/decapsulate), secure peer channels with RFC 10024, and use ML-DSA HSM firmware for protocol auth. Future threshold ML-DSA would still need in-HSM party computation to preserve Variant (b) (Section 8b).

Rotation & Revocation (Module A2 §5e)

After a CRQC arrives, an exposed on-chain key cannot be revoked — only outrun. MPC refresh and classical HSM rotation do not remove on-chain Shor exposure; planned migration (Section 8c) and control-plane ML-DSA rotation must precede CRQC availability.

Recovery & Inheritance (Module A2 §5f)

Recovery envelopes are long-lived HNDL targets — re-wrap them under hybrid KEM-DEM now (§5c). A classical heir key behind a multi-year timelock may outlive CRQC arrival; a hidden Taproot script leaf does not help if the tweaked key-path public key is broken by Shor, whereas P2WSH/P2SH or draft BIP-360 P2MR hide keys only until first spend (mempool race). PQ smart accounts can rotate guardians without moving funds (Section 8d).

Key Ceremonies & Backup DR (Module A2 §5g, §5j)

Dual-algorithm ceremonies generate ML-DSA (ξ) and ML-KEM (d, z) inputs per FIPS 204/203. Stateful hash-based firmware keys (LMS/XMSS; see Stateful Signatures) must never be restored twice (RFC 10033).

Audit Evidence, Monitoring & Insurance (Module A2 §5h; Module C §6c, optional further reading)

Ceremony minutes, signed logs, and Proof-of-Reserves records that must remain verifiable for years must enter an RFC 4998 / Wathiqa PQ timestamp renewal chain before classical signatures fail (see Trust Services PQC). As of 20 November 2026, no standard insurer/auditor PQC questionnaire mandate is established across the custody market.

Access Control & Destruction (Module A2 §5i, §5k)

FIDO2/WebAuthn hardware PQ deployment is not yet established despite RFC 9964 COSE IDs (Section 8). Destroying a classical private key does not protect funds behind an already exposed on-chain public key, and retired classical encrypted backups stay HNDL targets until destroyed.

Evidence RenewalStandard

RFC 4998 (Evidence Record Syntax) preserves long-term integrity of signed records through hash trees and periodic timestamp renewal before the underlying signature algorithm weakens.

Source: IETF RFC 4998 (Standards Track, August 2007), §1–§5 Hash-Tree Renewal & Timestamps

Wathiqa ProfileDraft

The Wathiqa profile (individual Internet-Draft) extends RFC 4998 Evidence Records with algorithm agility for post-quantum long-term verification; re-signing after classical algorithms become forgeable does not restore history.

Source: Wathiqa Profile Draft (2026-09-06), Algorithm-Agile Evidence Renewal

Stateful Backup LimitStandard

RFC 10033 details why stateful hash-based signature keys (LMS/XMSS) must never be restored from a stale backup or reused across state forks — a critical constraint for ceremony and DR planning.

Source: IETF RFC 10033 (Informational, September 2026), State & Backup Management Guidance

LMS StatefulnessStandard

RFC 8554 specifies the stateful Leighton-Micali Hash-Based Signature (LMS) scheme used in hardware/firmware roots of trust.

Source: IETF RFC 8554 (Informational, April 2019), LMS One-Time State Requirement

9B. Takeaways Matrix

ComponentQuantum LimitHNDL or ForgeryWhat Can Migrate NowMaturity
HSM / MPC / Enclave channels (TLS / mTLS / attested TLS)Classical ECDHE key exchange exposes recorded shares & approvalsHNDL (Confidentiality)Yes — deploy RFC 10024 PQ/T hybrid TLS 1.3 nowstandard
Key & share backups (RSA / ECDH / ECIES envelopes)Recorded ciphertext envelopes can be decrypted later to rebuild seeds/sharesHNDL (Confidentiality)Yes — re-wrap under hybrid KEM-DEM (draft-ietf-hpke-pq) with ML-KEM HSMsdraft
Control-plane approvals, policy rules & audit logsClassical ECDSA/RSA signatures become forgeable once a CRQC existsFuture Forgery (Auth)Yes — backend/HSM approvals to ML-DSA; archive evidence in RFC 4998 chainsstandard
Enclave & hardware attestation chains (SGX/Nitro/SEV-SNP/TDX)Platform endorsement roots use classical ECDSA P-256 / P-384 todayFuture Forgery (Auth)Track vendor migration (TPM 2.0 V185, SPDM 1.4.1, Caliptra 2.1); hybrid TLS nowvendor-shipped
Hardware FIDO2 / WebAuthn approversCOSE ML-DSA IDs registered (RFC 9964), hardware authenticator rollout pendingFuture Forgery (Auth)Pair with HSM/backend ML-DSA approvals; track CTAP/authenticator updatesdraft
MPC threshold transaction signing (ECDSA / FROST)Classical output + Paillier/Strong-RSA internals; no NIST PQ threshold standardBoth (HNDL + Forgery)Protect channels & backups now; evaluate separate auth or smart-account multisigresearch
SLIP-39 / Shamir seed backups & BIP-32 HD walletsNon-hardened xpub + Shor child key recovers parent; no standard PQ HD pathBoth (HNDL + Forgery)Enforce hardened derivation, stop address reuse, hybrid KEM-DEM share envelopesstandard
PQ smart accounts (ERC-4337 / EIP-7702)EIP-7702 retains secp256k1 control & permit paths; no Ethereum PQ precompile liveFuture Forgery (Auth)Prototype hybrid validation on testnets; track EIP-7851 / 8051 / 8052 / 8141prototype
Layer-1 chain outputs (Bitcoin / Ethereum / Solana)Exposed on-chain public keys vulnerable to Shor forgeryFuture Forgery (Auth)Address hygiene & sweep tooling prep; track BIP-360/361 & chain roadmapsdraft

Check your understanding

55 questions on Custody PQC Migration, each with its answer and the reason.

Take the quiz

Next step

Produce the artifact: Deployment Playbook

This module belongs to phase 5 (Pilots & Migration); Deployment Playbook produces a deliverable of that phase in the Command Center.

Production Tool Capabilities & Verification Checklist

Capabilities to require in your production tools:

  • Authenticated, transcript-bound PQ/T hybrid TLS (RFC 10024) for HSM clusters, MPC peer channels, and enclave attested TLS, paired with hybrid KEM-DEM (draft-ietf-hpke-pq) share and recovery backup envelopes.
  • CAVP/CMVP-scoped HSM firmware (PKCS#11 v3.2 ML-KEM, ML-DSA, SLH-DSA) moving control-plane approvals, policy rules, and RFC 4998 evidence renewal chains to post-quantum signatures before CRQC arrival.
  • Explicit separation of platform-, account-, and chain-controlled migration clocks, including Rule G1 verification of residual classical authorization dependencies on EIP-7702 and ERC-4337 smart accounts.

How to evaluate in your environment

Audit custody CBOM classifications in staging: verify RFC 10024 hybrid handshakes on HSM/MPC links, confirm hybrid KEM-DEM backup re-wrapping with ceremony metadata binding, and test negative smart-account legacy-key bypass cases.

Production evidence to collect

Custody CBOM with exposed-address register, HSM firmware CAVP vs CMVP module-boundary records, hybrid TLS/KEM-DEM test transcripts, and account-type residual-dependency sign-off.

Completing this Hub module and its browser-based sample exercises teaches the technical and governance decision criteria for your role; it does not test, validate, or certify your deployed production systems.

Record evidence for your environment

Identify the tested build and configuration. Use reference IDs for observations. For a proposed exclusion, record the system boundary or rule and why it excludes this build.

Authenticated, transcript-bound PQ/T hybrid TLS (RFC 10024) for HSM clusters, MPC peer channels, and enclave attested TLS, paired with hybrid KEM-DEM (draft-ietf-hpke-pq) share and recovery backup envelopes.
CAVP/CMVP-scoped HSM firmware (PKCS#11 v3.2 ML-KEM, ML-DSA, SLH-DSA) moving control-plane approvals, policy rules, and RFC 4998 evidence renewal chains to post-quantum signatures before CRQC arrival.
Explicit separation of platform-, account-, and chain-controlled migration clocks, including Rule G1 verification of residual classical authorization dependencies on EIP-7702 and ERC-4337 smart accounts.

Evidence missing. 0 of 3 requirements have a reference or scope rationale.

Learning module content can be inaccurate. Please double-check its information. Report inaccuracies in PQC Today GitHub Discussions.