What are Custodial & Non-Custodial Systems?
Custodial and non-custodial systems are the two ways an organization can control digital assets. In a custodial system, a third party holds the private keys. In a non-custodial system, the organization holds them itself. The choice determines who can move funds, who carries liability, and what happens when something fails.
Who this is for
This work is for organizations that hold, issue, or move digital assets at a scale where a single compromised key would be a material event.
- Asset issuers. This includes commodity-backed and tokenized real-world asset issuers, particularly those operating across more than one jurisdiction. RCMG has built custody systems for gold issuers in exactly this position.
- Funds and corporate treasuries. These are organizations moving from an exchange account or a single hardware wallet to institutional-grade control.
- DAOs. This applies to treasuries where signer sets and approval rules have grown informally and no longer match the value they protect.
- Platforms. This covers exchanges, brokers, and fintechs deciding whether to self-custody, use a qualified custodian, or run a hybrid model.
Typical triggers include preparing for a regulatory license or audit, a board asking who can actually move the money, onboarding a first institutional client, or a signer leaving the organization.
Why this matters now
Institutional assets are moving on-chain faster than custody practices are maturing. On-chain real-world asset value grew from $5.42B at the start of 2025 to $19.32B by March 2026 (CoinGecko). In addition, 64% of asset managers now report interest in tokenizing, up from 40% (Coinbase/EY-Parthenon, 2026).
Wallet security is where institutions actually fail. Compromised wallet security was involved in 73% of institutional crypto failures from 2022 to 2025 (Chainalysis). Private-key compromises accounted for 43.8% of stolen crypto in 2024. The largest single theft, Bybit's $1.46B loss in February 2026, came from a compromise of signing infrastructure rather than of the blockchain itself (Chainalysis, 2026).
Regulators have also made clear that key control alone is not custody. A 2026 SEC/CFT report states that "private keys do not by themselves replace custodians." The report says legal control, segregation, key recovery, corporate actions, reconciliation, reporting, and insolvency scenarios must all be addressed.
What companies actually use it for
Cross-jurisdiction asset issuance
Issuers of tokenized commodities need custody that satisfies more than one regulator at once. This means segregated reserves, approval chains that reflect local signing authority, and records an auditor in each jurisdiction can rely on. Large-scale tokenized settlement is now operating at institutional volume; JP Morgan's Kinexys has processed over $1.5T through tokenized assets (SEC/CFT, August 2026).
Institutional treasury control
Multisig has become the default for serious holders. More than $180B is held in multisig wallets (Glassnode, early 2026), and 89% of funds over $100M use multisig (Fireblocks, 2025). The most common configurations are 3-of-5, at 47% of institutional deployments, and 2-of-3, at 31% (Safe, 2025). The real decisions lie elsewhere: who holds each key, where it is stored, and how the threshold relates to operational reality.
DAO treasury governance
DAO treasuries often hold substantial value behind signer sets chosen early and never revisited. Custody design here means aligning on-chain signing authority with the organization's actual governance, including how signers rotate and how emergency actions are constrained.
Platform custody and compliance
Exchanges and brokers holding client funds face enforcement pressure on top of security risk. AML/KYC penalties exceeded $5.1B between November 2023 and February 2025, which is more than the prior decade combined. Client asset segregation and auditable controls are now baseline expectations.
What happens without it
Without deliberate custody design, organizations typically discover their weaknesses during an incident. Only 34% of institutions test disaster recovery (BitGo, 2025). Most therefore do not know whether they can recover funds after losing a key or keyholder until they need to.
Platforms lost $3.4B to theft in 2025 (Chainalysis, 2026). Hardware security module integration reduces key compromise by 94%, yet many treasuries still rely on keys held on general-purpose devices. The less visible cost is commercial. Institutional counterparties, auditors, and regulators increasingly ask custody questions that informal setups cannot answer, which delays deals and licenses.
What an engagement looks like
- Discovery. We map what you hold, where, under whose control, and which regulatory regimes apply.
- Assessment. We evaluate custody model fit, key generation and storage, segregation, approval chains, recovery capability, and alignment with your legal and regulatory obligations.
- Design. We specify the target custody architecture: custodial, non-custodial, or hybrid. The design includes signing policies, signer roles, and recovery paths.
- Build and ceremony. We implement the system and conduct witnessed key ceremonies with full documentation.
- Handover. We train your team, test recovery, and transfer operational ownership.
What you receive. You receive an architecture specification and documented approval policies. You also receive ceremony records suitable for auditors, a tested recovery plan, and a team that can operate the system without us.
When this is not the right fit
- Your holdings are small and short-term. A reputable qualified custodian's standard offering will serve you better than a designed system.
- You want a product, not a design. RCMG does not sell custody software or act as a custodian. We design and build systems that you or your chosen custodian operate.
- You are seeking yield or investment strategy. RCMG does not offer financial products, advice, or investment programs.
Frequently asked questions
Should we self-custody or use a qualified custodian? The answer depends on your regulatory obligations, internal operational capacity, and counterparty expectations. Many institutions
