MPC Custody vs Multisig vs Single-Key: Why Google Is Betting on Threshold Signatures

MPC Custody vs Multisig vs Single-Key: Why Google Is Betting on Threshold Signatures

MPC Custody vs Multisig vs Single-Key: Why Google Is Betting on Threshold Signatures

MPC Custody vs Multisig vs Single-Key: Why Google Is Betting on Threshold Signatures

Google is betting on threshold signatures because they let separate parties jointly sign transactions without any one employee, server or vendor holding the full private key. The chain still sees one standard signature, which is why Google is previewing an MPC custody service built on this design for banks and other institutions. The same family of cryptography already secures native cross-chain swaps on Chainflip, where threshold signatures split control of a private key across many parties so no one ever holds the whole thing. Here is how MPC compares to multisig and single-key setups, and where each one makes sense.

Single-key vs multisig vs MPC at a glance

All three models answer the same question. Who can move the funds, and what happens if one key is lost or stolen?

Criteria

Single-key

Multisig

MPC / TSS

Key structure

One private key

Several independent keys, each signs separately

Key shares held by separate parties; the full key is never assembled

Single point of failure

Yes, the one key

No, if the threshold is above one

No, if the threshold is above one

Onchain fees

Standard single-signature cost

Higher, due to extra signatures or contract calls

Standard, because one signature lands onchain

Chain support

Any chain

Depends on each chain's scripting or smart contract support

Any chain whose signature scheme the implementation supports

Changing signers

Move funds to a new key

Often requires a new address or a contract update

Handled offchain through a new key generation or share refresh

Transparency

Looks like a normal transaction

Signer set and threshold are visible onchain

Looks like a single-signer transaction; policy lives offchain

Single-key wallets

A single-key wallet is where most people start. One private key produces one signature and gives full control, which keeps it cheap, simple and compatible with every chain.

The problem is concentration. If the key leaks through phishing, malware or a careless backup, the funds are gone, and if it is lost there is no recovery path. For an individual using a hardware wallet that tradeoff can be acceptable, but for a treasury or protocol holding pooled funds it is not.

Pros: lowest fees, universal chain support, simple setup.

Cons: one point of failure, no shared control, no built-in policy layer.

Multisig wallets

Multisig spreads control across several independent keys and requires a set number of them, such as 3 of 5, to approve a transaction. Bitcoin supports this natively through script, while on Ethereum it usually lives in a smart contract wallet.

The upside is transparency. Anyone can inspect the signer set and threshold onchain, which makes audits straightforward. The downside is cost and fragmentation, since every extra signature or contract call adds fees and each chain implements multisig differently.

For a cross-chain protocol, that per-chain variation is a real constraint. A multisig on Bitcoin, Ethereum and Solana means three separate systems with three sets of failure modes. several high-profile bridge exploits also traced back to small signer sets where a handful of compromised keys was enough.

Pros: no single point of failure, onchain auditability, mature tooling.

Cons: higher fees, chain-specific setups, awkward signer rotation.

MPC custody and threshold signatures

Multi-party computation (MPC) is a broad field of cryptography where parties compute a result together without revealing their private inputs. A threshold signature scheme (TSS) applies MPC to signing. Each participant holds a key share, and any group above the threshold can produce a valid signature without the full private key existing in one place.

Onchain, the result looks like an ordinary single-signature transaction. That keeps fees at single-signer levels and works on any chain whose signature scheme the implementation supports. The tradeoff is visibility, because the signer set and policy are not recorded onchain, so you rely on the operator to be clear about how signing works.

Pros: no single key, standard fees, broad chain coverage, flexible signer changes.

Cons: complex cryptography to implement correctly, policy hidden from onchain view, still exposed to collusion above the threshold and to software bugs.

Why Google and other institutions are converging on MPC

According to Bitcoin Magazine, Google is previewing Google Multi-Party Compute, a custody system that splits private keys across separate environments, with several banks testing it ahead of a reported early-2027 launch. Google presented the product at its first digital assets symposium in New York, and KuCoin News reports that Alphabet already uses it to secure digital assets on Ethereum and Solana. Google has not named its design partners or disclosed pricing.

Google Cloud's documentation describes the platform as combining threshold signing, transaction policy enforcement and enterprise governance controls. It uses TSS so mutually untrusting parties can jointly sign blockchain transactions without constructing or storing a full private key. That explains the institutional pull: no single employee, server or vendor becomes the key, and the chain still sees one standard signature.

How Chainflip uses threshold signatures for native swaps

Chainflip arrived at the same cryptography for a different reason. A cross-chain protocol needs to hold native BTC, ETH and SOL without wrapped tokens or a centralized custodian. Its validators use the FROST (Flexible Round Optimized Schnorr Threshold) scheme to jointly sign vault transactions, so no single party has the complete private key, as covered in Chainflip's decentralized infrastructure overview.

Per the Chainflip docs, signing requires 100 of 150 validators, and assets stay in native-chain vaults across Bitcoin, Ethereum, Solana, Polkadot, Arbitrum and other supported chains. This decentralized custody model means funds are secured by validators who must reach a supermajority to move anything, with no centralized custodian in the loop. The same vaults back Chainflip Lending, where native BTC collateral is held on the Bitcoin chain rather than wrapped.

Threshold signing removes single-key risk, not all risk. Validators colluding above the threshold, bugs in protocol logic and chain-specific integration issues remain possible failure modes. That is why staking and slashing add an economic layer on top of the cryptography. For more on the trust model, see why native settlement differs from solver networks.

You can watch these vaults at work on Chainflip Scan, which shows $10.16B in all-time swap volume.

Which one fits you

  • Individual long-term holder: a single-key hardware wallet works if you manage backups carefully. Add multisig if you want redundancy.

  • DAO or small team treasury: multisig, because onchain visibility of signers helps accountability.

  • Bank, fund or enterprise: MPC custody, either through a provider like the one Google is previewing or an in-house TSS setup, for policy controls and multi-chain coverage.

  • Cross-chain swapper: you don't need to run any of these yourself. Chainflip handles vault custody with threshold signatures, so you can buy native BTC with ETH or sell SOL for USDC without touching a wrapped token.

The takeaway

Google's entry signals that threshold signatures have moved from specialist cryptography into mainstream institutional infrastructure. Multisig still fits where onchain transparency matters most, and single-key wallets still work for individuals who manage their own risk. For pooled funds spread across many chains, removing the single key is becoming the baseline.

To see TSS-secured vaults in practice, start a native swap on Chainflip and follow it on Chainflip Scan.

Resources

  • Swap - Start swapping native assets

  • Lending - Borrow against native Bitcoin

  • Blog - Product updates and announcements

  • Chainflip Scan - Track swaps and network activity

  • Website - Explore Chainflip

Earn with Chainflip:

Find us:

FAQ

What is MPC custody?

MPC custody splits control of a private key into shares held by separate parties. Using a threshold signature scheme, a set number of those parties can sign a transaction together without the full key ever being assembled in one place.

What is the difference between MPC and multisig?

Multisig uses several independent keys and verifies each signature onchain, so the signer set is visible and fees rise with each signature. MPC produces one standard signature from key shares, which keeps fees low and works across chains, but the signing policy lives offchain.

Is Google's MPC wallet available yet?

Not commercially. Reports say Google Multi-Party Compute is in private preview with banks and other institutions testing it, and a commercial launch is planned for early 2027. Google has not disclosed pricing.

How does Chainflip use threshold signatures?

Chainflip's validators use the FROST threshold signature scheme to control native vaults on Bitcoin, Ethereum, Solana and other supported chains. Signing requires 100 of 150 validators, so assets are secured by validators with no centralized custodian.

Do threshold signatures make a protocol risk-free?

No. They remove the risk of a single compromised key, but collusion above the threshold, software bugs and integration issues remain possible. Economic measures like staking and slashing add another layer of protection.

Get Chainflip updates

Product news, protocol updates, and launches.