What is an MPC wallet? A plain-English guide for businesses
An MPC wallet signs blockchain transactions without ever creating a whole private key. Learn how multi-party computation and threshold signatures work, and why businesses use them.
Key takeaways
- An MPC wallet splits signing power across several parties, each holding a key share. The whole private key never exists.
- Shares jointly compute an ordinary signature, so the blockchain sees a normal transaction and fees stay the same.
- A 2-of-3 setup survives the loss or theft of any one share.
- MPC is a technology, not a custody model. Who holds the shares decides whether a wallet is custodial.
Every blockchain account is controlled by a private key. Whoever can produce a valid signature with that key can move the funds, and there is no bank to call if it is stolen. For a business holding meaningful balances, the question is not whether to protect that key but how.
For years the choices were awkward. You could guard a seed phrase yourself and accept that one person, one laptop or one safe was a single point of failure. You could hand your assets to a custodian and accept their counterparty risk and their pace. Or you could build an on-chain multisig and live with its per-chain quirks. Multi-party computation (MPC) offers a different answer: a wallet whose private key never exists in the first place.
The one-sentence definition
An MPC wallet is a wallet where several independent parties each hold a share of the signing key, and they run a cryptographic protocol that produces a valid signature together, without any of them ever learning, or reassembling, the full key.
How it works, step by step
1. Distributed key generation
Instead of one machine generating a key and then splitting it, the parties run a distributed key generation (DKG) ceremony. Each party contributes randomness and ends up with its own secret share. Together the shares define one public key, and therefore one blockchain address, but the matching private key is never computed anywhere. There is no “dealer” who saw it, because nobody did.
2. Threshold signing
When a transaction needs signing, a sufficient number of parties (the threshold) run a signing protocol. Each uses its share to compute a partial result, and the results combine into a single, ordinary ECDSA or EdDSA signature. Modern protocols such as CGGMP21 (for ECDSA chains like Bitcoin and Ethereum) and FROST (for EdDSA chains like Solana) are designed so that even a malicious participant cannot extract another party’s share or forge a signature.
3. Share refresh
Because shares are just numbers, the parties can periodically re-randomise them while keeping the same public key. After a refresh, old shares are useless. An attacker who quietly steals one share has to steal enough others before the next refresh, which shrinks their window dramatically.
4. Recovery
If a share is lost, the remaining parties can run a reshare that issues a fresh share to the party that lost theirs, again without reconstructing the key. Good designs gate this behind time locks and approvals so it can’t be abused.
What the blockchain sees
Nothing unusual. The output of an MPC signing ceremony is a standard signature from a standard address. That has three practical benefits:
- Chain coverage. One MPC system works on every chain that uses the same signature scheme, with no smart contract to deploy or audit per chain.
- Normal fees. A transaction costs the same as any single-signer transaction, unlike on-chain multisig, which can cost more.
- Privacy. Your approval structure, signer list and policy changes are not published on-chain.
Why threshold matters: 2-of-3
The most common business setup is 2-of-3: three shares, any two can sign. It balances two risks that pull in opposite directions:
| Scenario | 1-of-1 key | 2-of-2 MPC | 2-of-3 MPC |
|---|---|---|---|
| One share stolen | Funds lost | Safe | Safe |
| One share lost | Funds lost | Funds locked | Safe, reshare |
| One party offline | Can’t sign | Can’t sign | Other two can sign |
A 2-of-2 design protects against theft but turns any lost share into a lockout. A 2-of-3 design survives either a theft or a loss of any single share.
MPC is a technology, not a custody model
This is the point most guides skip. MPC describes how signatures are produced, not who controls them. If a provider holds two of three shares and can sign without you, that wallet is effectively custodial, however advanced the cryptography.
What matters is the signing policy: which combinations of shares are allowed to sign, and under what conditions. At Merkle Vault the three shares sit with you, with us and with an independent insurer, and the platform is configured so that your share or your approval is required for every transfer. Merkle Vault and the insurer cannot combine to move your funds. The insurer’s share takes part only in a recovery that your own admins start.
When an MPC wallet makes sense
MPC is a strong fit when you:
- hold balances large enough that a single compromised device would be a material loss;
- need several people to approve transfers, with limits that differ by amount, asset or destination;
- operate across several chains and don’t want a different multisig setup for each one;
- want to automate payouts or sweeps without giving a server a whole key;
- need to show auditors, investors or regulators exactly who can move funds.
It is probably overkill for a single individual’s small hot wallet, where a good hardware wallet is simpler.
Questions to ask any MPC provider
- Who holds each share, and which combinations can sign? Get it in writing.
- Where do policies run? Rules enforced only in a web dashboard can be bypassed; rules enforced next to a key share cannot.
- Which protocols, and who audited the implementation? Ask for the protocol names and the audit report.
- How are shares refreshed, and how often?
- What happens if the provider disappears? There should be a documented exit procedure you can verify before you deposit.
The bottom line
An MPC wallet replaces “protect the key” with “make sure no key exists.” Done well, with a sensible threshold and a client-first signing policy, it gives a business the controls of a bank portal while keeping ownership where it belongs: with the business.
Ready to see it working? Read how Merkle Vault’s security model works or book a live 2-of-3 signing demo.
Frequently asked questions
Is an MPC wallet the same as a multisig wallet?
No. A multisig uses several whole private keys and a smart contract or script that counts signatures on-chain. An MPC wallet has no whole keys at all: key shares compute a single, standard signature off-chain, so it works on any chain that uses that signature type.
Can an MPC wallet be hacked?
Any system can be attacked, but an MPC wallet removes the single most common target: a whole private key or seed phrase. An attacker must compromise enough separate parties, before the shares are refreshed, to reach the signing threshold.
What does 2-of-3 mean?
There are three key shares, and any two of them can produce a signature together. One share alone can do nothing, and losing one share does not lock you out.