What is MPC (Multi-Party Computation)? MPC 101

Multi-party computation (MPC) is a cryptographic method that splits a private key into separate shares distributed across multiple independent parties. The complete key is never assembled in one place, at any point in time. The result: no single compromised device, insider, or attacker can access your funds.

Three things every digital asset operator should know about MPC:

1. No single point of compromise. Key shares are generated, stored, and used independently. The private key never exists as a whole.
2. Fireblocks’ MPC-CMP signs transactions 8x faster than standard MPC and requires just 1 signing round versus 9 in the GG18 industry standard. All without sacrificing security.
3. MPC is blockchain-agnostic. Because it operates at the cryptographic layer (ECDSA and EdDSA), a single MPC implementation secures wallets across all major chains simultaneously.

MPC has become the security standard for institutional digital asset operations, protecting trillions of dollars across exchanges, payment processors, banks, and fintechs. Here’s how it works, and why it matters.

What Is Cryptography, and Where Does MPC Come From?

The field of cryptography gives its users a method for sending messages that only the intended recipient can read, preventing unauthorized third parties from reading them if intercepted, and verifying the identity of the sender.

Digital assets depend on cryptography in a specific way: every wallet is controlled by a private key, i.e. a secret number that authorizes transactions. Whoever controls the private key controls the assets. Keeping that key secure is the central problem of digital asset custody.

MPC’s theoretical foundations were laid in the early 1980s:

• 1982 — Secure two-party computation is formally introduced as a solution to the “Millionaire’s Problem”
• 1986 — Andrew Yao extends the framework to any feasible computation
• 2016–present — MPC enters the digital asset space and becomes the dominant security technology for institutional wallet infrastructure

Before MPC, most cryptography focused on concealing content. MPC introduced something different: the ability to run computations using data from multiple sources without any party revealing their private input, and without requiring a trusted third party.

How Does MPC Work?

In a general sense, MPC enables a group of participants, each holding their own private data, to jointly compute a function on that combined data without any party revealing what they individually hold.

Applied to digital asset wallets, this works as follows:

Key generation: Individual key shares are created and randomized across multiple endpoints, e.g. cloud servers, on-premises hardware, or mobile devices. The shares are never combined or shared with one another. Together, the endpoints compute the wallet’s public key (its blockchain address) without any single endpoint knowing the full private key.

Transaction signing: When a transaction needs to be signed, a quorum of at least 3 endpoints each independently validate the transaction against the organization’s policy rules and contribute their share of the signature. The complete signature is produced through a distributed computation. At no point is the private key assembled.

The result: Even if one endpoint is compromised by an external attacker, a malicious insider, hardware failure or otherwise, the key shares held elsewhere are useless in isolation. There is no single point of compromise across the entire key lifecycle.

What Makes Fireblocks’ MPC-CMP Different?

Fireblocks developed MPC-CMP, an advanced MPC protocol that improves on the previous industry standard (GG18) across speed, flexibility, and operational security.

Speed: Standard MPC protocols require multiple communication rounds between key share holders before a transaction signature is complete. MPC-CMP reduces this to a single round, making it up to 8x faster and enabling institutional-grade operations at high throughput.

Cold storage signing: MPC-CMP supports hot and cold wallet signing configurations, allowing at least one key share to be stored offline in an air-gapped device. This addresses regulatory requirements in jurisdictions that mandate offline key storage, without the operational friction of traditional cold wallets.

Open source and audited: Fireblocks has open-sourced the MPC-CMP protocol under a limited license, enabling cryptographers, auditors, and security researchers to inspect and validate the implementation. The code is periodically reviewed by independent auditors and has been implemented for Intel SGX and AWS Nitro secure enclaves.

How Does MPC Protect Against Different Threats?

MPC’s distributed architecture eliminates the three primary attack vectors targeting digital asset wallets:

External attackers: Because key shares are distributed across multiple independent environments, an attacker would need to simultaneously compromise every share location. This is far more difficult than stealing a single private key or seed phrase.

Internal threats and collusion: No single employee or administrator has access to all key shares. A rogue insider cannot unilaterally access funds. Policy rules are enforced inside hardware-isolated secure enclaves (Intel SGX, AWS Nitro), so even privileged infrastructure access cannot override transaction approvals.

Human error: Unlike seed phrases or hardware wallets, MPC removes the risk of a single catastrophic mistake, for example a lost device, a mistyped address, or an accidental export of a private key. Key shares can be refreshed and rotated without changing the wallet’s blockchain address.

MPC vs. Multi-Sig: What’s the Difference?

Both MPC and multi-signature (multi-sig) require multiple parties to authorize a transaction. The critical difference is where signing happens.

Multi-sig is enforced onchain: the blockchain itself validates that multiple separate private keys have signed. This creates protocol dependencies since not all blockchains support multi-sig, and those that do implement it differently. Moving to a new quorum configuration requires changing the wallet’s on-chain address, which introduces operational risk (funds sent to the old address may be lost).

MPC operates off-chain: signing is a cryptographic computation performed between key share holders before anything is broadcast to the blockchain. This means:

• Chain agnostic: The same MPC implementation secures wallets across every blockchain that uses ECDSA or EdDSA (which is virtually all of them)
• Lower fees: Signing happens off-chain, so there’s no on-chain multi-sig overhead
• Flexible quorums: Changing from a 3-of-5 to a 4-of-6 setup doesn’t require changing wallet addresses
• Reduced attack surface: A single MPC implementation covers all supported chains; multi-sig bugs must be patched separately per protocol

How Does Fireblocks Layer MPC into a Broader Security Architecture?

MPC is Fireblocks’ foundation but not the whole picture. Fireblocks uses a multi-layer security approach, because no single technology is unbreakable.

MPC-CMP

Key shares distributed across multiple isolated cloud environments (or hybrid on-prem/cloud), with cryptographic guarantees that the complete key is never assembled.

Secure enclaves (TEEs)

MPC key shares, policy logic, and API credentials are stored and executed inside hardware-isolated trusted execution environments (Intel SGX, AWS Nitro, GCP Confidential Spaces). Even if an attacker gains OS-level access to a server, encrypted enclave memory cannot be extracted.

Policy Engine

Granular, cryptographically enforced transaction authorization rules — including source, destination, asset type, and amount thresholds — with separation of duties to prevent internal collusion. Policy rules are signed by a quorum of admins and enforced inside SGX.

Fireblocks Network

Automated deposit address authentication that eliminates copy-paste errors and address-substitution attacks (one of the most common sources of fund loss in institutional operations).

For a deeper dive, read our Guide to Digital Asset Wallets and Service Providers.