The MPC vs multisig question is about where a company's signing rules are enforced. Multisig writes the rule, such as two of three keys, into the Bitcoin address itself, so the network enforces it. Multi-party computation (MPC) splits one key into shares held by different parties and enforces the rule inside software, so the network only ever sees a single ordinary signature.
Both can deliver dual control. They differ in transparency, portability and how much the company depends on a particular vendor's design.
Two ways to avoid a single point of failure
The governance aim is the same in both models: no single person, device or organisation should be able to move the company's Bitcoin alone.
Multisig in one paragraph
Several independent private keys exist. The address records how many must sign. When enough keys sign, the network accepts the transaction. Our guide to multisig custody for companies covers quorum design and key placement in detail.
MPC in one paragraph
A single private key is never assembled in one place. Instead, separate key shares are generated and held by different parties or machines. To sign, the share holders run a joint calculation that produces a valid signature without any of them revealing their share. A policy engine decides which combination of shares, and which approvers, must take part.
MPC vs multisig: the comparison
| Factor | Multisig | MPC |
|---|---|---|
| Where the quorum is enforced | Bitcoin network rules | Provider's software and policy engine |
| What appears on the blockchain | Multi-key spending conditions (visibility depends on address type) | A single standard signature |
| Evidence of who signed | Visible in transaction data | In the provider's logs |
| Changing the quorum or signers | Usually requires moving funds to a new address | Shares can be refreshed without moving funds |
| Vendor dependency | Low, if built on open standards | Higher: shares typically only work with the vendor's software |
| Maturity of the method | Long-established in Bitcoin | Newer, with designs that vary by vendor |
Neither column is better in every row. The table is a set of trade-offs.
What the board should weigh
Transparency and audit evidence
With multisig, an auditor can see from the transaction itself that several keys signed. That is strong, independent evidence that dual control operated. With MPC, the signature looks like any single-key transaction, so the evidence comes from the provider's logs and its SOC 1 Type II or SOC 2 Type II reports. That is acceptable, but it means relying more on the provider's control environment. See our audit readiness checklist.
Vendor dependency and exit
Multisig built on open standards can, in principle, be signed by any compatible wallet software. If a provider disappears, the company can rebuild access with its own keys and the published wallet configuration.
MPC key shares are usually tied to the vendor's implementation. If the vendor fails or the relationship ends, the company needs a documented way to move funds out. Ask how that works before signing, and write it into your custody exit plan.
Flexibility in operations
MPC is often easier to operate day to day. Signers can be added or removed by refreshing shares, without moving Bitcoin to a new address. Approval policies can be detailed: limits by amount, time of day, destination or approver group.
Multisig changes are more deliberate. Replacing a key generally means creating a new address and moving funds, which is slower but leaves a clear on-chain record.
Security assumptions
Multisig security rests on the Bitcoin protocol and on how well each key is protected. MPC security also rests on the correctness of the vendor's protocol and code. Flaws in specific MPC implementations have been reported publicly over the years and patched. That is not a reason to reject MPC, but it is a reason to ask for independent security reviews and to understand how the vendor handles vulnerabilities.
Practical tip: ask any provider to walk you through, in writing, exactly how you would move your Bitcoin if their company ceased trading tomorrow. The clarity of that answer is a good test of either model.
How the models are used in practice
Many arrangements mix the two:
- Custodians may use MPC internally to protect operational wallets, and multisig or hardware-based schemes for deep reserves.
- Shared-control setups often use multisig, so that a company key, an independent key agent and a custodian each hold one key.
- Some providers combine MPC shares with hardware security modules, so each share is itself protected in certified hardware.
For a company, the useful question is not "which technology?" but "which risks does each design leave with us, and can we evidence our controls?"
Questions to ask a provider
- Which signing model protects our assets, and does it differ between online and offline wallets?
- How many parties hold key material, and can any of them sign without the others?
- Can we hold one key or share ourselves?
- How are approval policies configured, and who can change them?
- What independent security reviews have been performed on the signing scheme?
- What happens to our access if your company is acquired, loses a licence or fails?
- How does the model interact with segregation and bankruptcy remoteness in your legal structure?
The answers should feed directly into the company's key management and signing policy.
Choosing between MPC and multisig
A rough guide:
- Multisig tends to suit long-term reserves, boards that want on-chain evidence of dual control, and companies that want to reduce reliance on any single vendor.
- MPC tends to suit balances that move more often, teams that need detailed approval workflows, and arrangements where a regulated custodian runs the signing infrastructure.
Many companies use both, matched to the purpose of each wallet. Whatever the choice, take legal advice on ownership and insolvency treatment.
How The Bitcoin World helps
We help finance teams compare signing models against their governance, audit and continuity requirements, and document the decision for the board as part of our custody architecture work. We do not hold keys or client assets. Custody and execution are carried out by licensed third-party institutions.