Custody providers and Arc: qualified custody, MPC, and validator relationships compared

"Custody provider is connected to Arc" is not one claim — it can mean a named integration partnership, an operating validator seat, or nothing more than being a well-known custodian in the same industry conversation. This guide separates those, name by name, because the institutional stakes in this category make the distinction matter more than in almost any other part of the Arc ecosystem.

Three different things this category conflates

This directory covers Wallets & Custody → Custodians specifically — institutional custodians, not consumer wallets or wallet-as-a-service infrastructure; see Built on Arc's wallet support status guide for that half of the Wallets & Custody category.

Qualified custody is a regulated status — a custodian meeting specific legal requirements (in the US, typically under securities or trust-company frameworks) to hold assets on behalf of institutional clients with defined fiduciary and regulatory obligations. It is a legal and compliance designation, not a technology choice.

MPC (multi-party computation) custody is a technology approach — private key material is split and never fully reconstructed in one place, reducing single-point-of-failure risk. A provider can offer MPC custody without being a qualified custodian in the regulatory sense, and vice versa; the two are frequently bundled in marketing but are answering different questions (how is the key protected, versus who is legally accountable for the asset).

Self-custody infrastructure — wallet and key-management tooling that lets an institution hold its own keys rather than delegating to a third party, often still using MPC or hardware security modules under the hood, but without handing custody to an external qualified custodian at all.

A provider can sit in more than one of these categories simultaneously, and a headline like "X brings institutional custody to Arc" does not by itself tell you which one you're getting. Read past the headline to what's actually being offered.

A separate, and separately important, question: validator seat vs. custody relationship

Arc's founding validator cohort — 11 named institutions per Circle's 5 August 2026 press release (BlackRock, DTCC, Galaxy, Global Payments, ICE, Mastercard, MoneyGram, SBI Group, Standard Chartered, Sumitomo Corporation, and Visa) — operate validator nodes that participate in Arc's permissioned consensus. That is a network-security and governance role, not a custody relationship with end users or builders: "Founding validator" is a mark, not a role. All eleven hold Role: Integrates plus the Founding validator mark — Visa and Mastercard under Banks & Enterprises → Card networks, MoneyGram under Payments & Payouts → Remittance, BlackRock under Banks & Enterprises → Asset managers, Galaxy under Exchanges & Trading → Market makers, and DTCC, ICE, Global Payments, SBI Group, and Sumitomo Corporation under Banks & Enterprises. None of the named custody-specific providers in this guide (Fireblocks, Copper, BitGo, Cactus Custody, Ledger Enterprise) appear on that 11-name validator list. Being a validator and being a custodian for assets on the chain are two different jobs, and this guide will not treat validator status as evidence of a custody product, or vice versa.

What's actually confirmed, provider by provider

Fireblocks is named directly in Circle's founding-validator-and-integrations press release as an Arc integration partner — the strongest tier of confirmation in this category. Fireblocks separately operates as MPC-based institutional wallet and custody infrastructure broadly in the industry; the confirmed fact here is the named Arc relationship, not a specific description of what the integration covers technically.

Builds on Arc: Provides institutional MPC-based custody and wallet infrastructure; named as an Arc integration partner. · Role: Serves · Category: Wallets & Custody → Custodians

See Built on Arc's wallet support status guide for how institutional custody differs from consumer wallet integration.

BNY is named in the same release as a company "exploring integrations" with Arc — explicitly a weaker, more provisional status than a confirmed integration partner. Circle's own language distinguishes BNY's status from the harder integration-partner list, and this guide preserves that distinction rather than flattening it into "BNY supports Arc." (Integrates — Banks & Enterprises → Banks; no Founding validator mark — Announced/exploring only, per the canonical role map.)

Copper has a confirmed, dated relationship with Circle generally — a 28 April 2025 partnership enabling institutional clients to convert off-exchange balances into USDC for trading and settlement while assets remain in custody through Copper's ClearLoop system. That partnership, as published, does not mention Arc specifically. This guide will not extend a general Circle-USDC partnership into an Arc-specific claim without direct evidence; as of 2 Sep 2026, an Arc-specific Copper relationship is not yet confirmed. (Serves — Wallets & Custody → Custodians, per the canonical role map, though the Arc-specific tie itself is unconfirmed.)

BitGo, Cactus Custody, and Ledger Enterprise (Ledger Enterprise as the distinct institutional-custody arm, separate from Ledger's consumer hardware-wallet integration named elsewhere in Circle's release) — no independently verifiable Arc-specific relationship, announcement, or documentation was found for any of the three as of 2 Sep 2026. All three are established, credible custody providers in the broader industry; none of that establishes an Arc-specific claim on its own, and this guide will not infer one. BitGo and Cactus Custody are Serves (Wallets & Custody → Custodians) per the canonical role map; Ledger Enterprise isn't on that list, but the same role applies by analogy pending confirmation.

The directory

Cross-reference the status column against the categories above: is this a qualified custodian, an MPC technology provider, self-custody infrastructure, or some combination — and is the Arc relationship a named integration, a validator seat, an "exploring" status, or unconfirmed. Most of the value in this category, as of 2 Sep 2026, is in keeping those distinctions separate rather than merging every "institutional-grade" name into one undifferentiated list.

Why the permissioned validator set changes the custody conversation

Arc's validator set is permissioned, not open — roughly 100 validators running Malachite, a Tendermint-based BFT consensus engine targeting sub-second finality (around 780ms per published figures). That design choice matters for how an institutional custody decision should actually be evaluated on Arc, compared to a permissionless chain. On a permissionless network, a custodian's technical job includes navigating validator diversity and slashing-risk considerations that don't really apply the same way here. On Arc, the relevant question shifts toward who is actually running the permissioned validator infrastructure — the 11 founding validators named above — and what governance or upgrade path exists around that set, which is a governance question more than a custody-technology one. A custody provider's marketing that leans on "battle-tested on decentralized networks" credentials is answering a slightly different question than what a permissioned, institutional-grade chain like Arc actually requires from its custody partners.

What counts as verification here, concretely

For an institutional buyer actually evaluating this category rather than just reading about it, the checklist is narrower than marketing copy usually implies: a named, dated relationship with Circle or Arc specifically (not just general industry credibility); a stated scope (custody of USDC specifically, broader multi-asset custody, or wallet infrastructure only); and ideally, once mainnet exists, a demonstrable integration — API documentation referencing Arc's transaction format, or a case study naming a specific institutional client using the service on Arc. None of the six providers named in this guide currently clear all three as of 2 Sep 2026; Fireblocks clears the first two.

What a custody relationship to Arc will actually need to demonstrate at mainnet

Once Arc mainnet exists, a meaningful custody claim should be checkable against something concrete: a documented integration (API support, signing infrastructure specific to Arc's transaction format), a stated regulatory scope (which jurisdictions, which client types), and ideally independent confirmation beyond the custodian's own marketing. Pre-mainnet, none of that is fully testable — which is exactly why this guide draws a hard line between "named as a partner" and "has shipped a working, verifiable integration."

Built on Arc is an independent directory. Arc is a Circle product; we are not affiliated with, endorsed by, or operated by Circle.

Sources: - Circle Announces Founding Validator Cohort and Major Integrations for Arc - Copper partners with Circle — Asset Servicing Times - Arc | The Economic OS — official site