Google is preparing to launch a crypto custody system that helps banks protect and transfer digital assets without storing complete private keys on a single computer. The service uses multi-party computation, or MPC — a branch of cryptography developed in the 1980s — to reduce exposure to stolen credentials, one of the most common failure points in digital asset storage.

Google Tests MPC Custody Service With Banks

Google will present the system, branded Google Multi-Party Compute, at a digital asset security symposium in New York on October 28. The company expects commercial availability in early 2027, following a private testing period.

Grosse news : Google va proposer de la conservation de cryptos

Le géant américain a dévoilé aujourd'hui le Google Multi-Party Compute. Derrière ce nom (un peu technique) se cache un service pensé pour les banques, qui doit protéger leurs cryptos et leurs actifs tokenisés.

Le… pic.twitter.com/gKcvYOhBjQ

— Grégory Raymond (@gregory_raymond) October 28, 2026

News of the announcement circulated quickly on X, where Grégory Raymond posted that Google had unveiled Google Multi-Party Compute, a service designed for banks to protect their cryptocurrencies and tokenized assets. Tokenized assets are conventional financial instruments represented as digital tokens on a blockchain.

Major financial institutions are already testing Google Multi-Party Compute, although the company has not named them. Pricing and future bank customers remain undisclosed. Those commercial terms are among the details to watch between now and the planned early 2027 launch.

Alphabet, Google's parent company, already uses the technology to secure some of its digital assets on Ethereum and Solana. It has not identified the assets or their value.

How the MPC Security System Works

MPC divides control of a private key among separate computers. Under a typical setup, a bank could keep one share on its own equipment, another with Google Cloud, and a third with a different provider. The shares work together to approve transactions. Because no single machine ever holds the full key, the system reduces the risk of unauthorized transfers even if one environment is compromised. The structure resembles the dual-control, or maker-checker, arrangements banks already use for high-value payments.

The design addresses risks illustrated by a recent CoinDCX wallet security breach. The exchange reported losing about $15.99 million in USDT from an old wallet setup, while stating that customer balances remained safe.

Google says customers control their key shares, including those stored on its cloud infrastructure. The service also checks transfer requests against rules set by the institution before approving them.

Recent Attacks Raise Crypto Security Concerns

Stolen credentials remain a common problem in digital asset storage. Attackers who obtain a working private key can authorize transfers, which are difficult to reverse once confirmed on a blockchain.

Separately, hardware wallet maker Foundation reported a Cloudflare API key breach on October 26. Attackers sent fake emails from company domains, but Foundation said its Passport devices and stored Bitcoin were not compromised. The incidents underscore the operational risks that custody providers are working to engineer away.

Google's current service targets unauthorized transactions and compromised credentials. It does not claim to stop future attacks that could break existing cryptographic methods through advances in artificial intelligence or quantum computing. The company says the system could eventually support post-quantum cryptography, but it has not provided a date for that feature.

The announcement follows fresh institutional payment projects. Circle and ACI Worldwide announced a USDC settlement test in Colombia, examining how banks could use stablecoins alongside existing payment systems.