Google is assembling a vertically integrated crypto infrastructure stack that spans custody, settlement, and blockchain services. On October 28, 2026, the company previewed Google Multi-Party Compute (MPC), a key-splitting custody service for banks, at its inaugural digital assets symposium in Ne...
"GCUL could deliver significant efficiencies for collateral, margin, settlement and fee payments as the world moves toward 24/7 trading." — Terry Duffy, Chairman and CEO, CME Group
Google is assembling a vertically integrated crypto infrastructure stack that spans custody, settlement, and blockchain services. On October 28, 2026, the company previewed Google Multi-Party Compute (MPC), a key-splitting custody service for banks, at its inaugural digital assets symposium in New York. The product joins Google Cloud Universal Ledger (GCUL), a permissioned Layer 1 blockchain already piloting with CME Group and BMO, and an expanding Web3 services division hiring blockchain architects across Asia-Pacific and North America.
The combined offering positions Google against native crypto custody providers — Coinbase ($300 billion in assets under custody), BitGo ($90 billion), and Fireblocks ($10 trillion in cumulative transaction volume) — as well as Amazon Web Services and Microsoft Azure, neither of which has shipped a comparable end-to-end stack. The timing is deliberate: private-key compromises caused 40% of crypto's cumulative $16.7 billion in hack losses, according to CoinDesk, and DeFi protocols alone lost $1.3 billion to exploits in the first eight months of 2026. Google is betting that institutional capital will not flow to digital assets at scale until custody infrastructure meets enterprise-grade security standards — and that a hyperscaler is better positioned to set those standards than a crypto-native startup.
Google Multi-Party Compute distributes transaction-signing authority across separate environments instead of concentrating control in a single private key. A bank using the service can store key fragments in Google Cloud, on its own hardware, and with a third-party cloud provider. No single party — including Google — holds enough fragments to unilaterally sign a transaction. Institutional rules are applied before any transaction is authorized.
The product is currently in private preview. Several design partners described by Google as "major financial institutions" are testing the technology, though no names have been disclosed. Alphabet itself has begun using MPC to secure its own digital-asset holdings on Ethereum and Solana. General availability is expected around early 2027, according to reporting by CryptoNews and CoinDesk.
Google says the service can eventually support post-quantum cryptography, but current protections focus on compromised credentials and unauthorized transaction signing — the attack vectors responsible for the majority of dollar losses in crypto security breaches.
The pricing model has not been publicly disclosed. Google's existing cloud security services typically follow consumption-based billing, but custody infrastructure may adopt a different structure given the regulatory requirements around segregated assets.
Google Cloud Universal Ledger is a permissioned Layer 1 blockchain designed for cross-border payments and asset settlement. First announced in March 2025, the network supports smart contracts written in Python — a deliberate choice aimed at reducing the learning curve for bank developers accustomed to data science and risk-modeling toolchains rather than Solidity or Rust.
Key architectural features include:
GCUL's most advanced pilot involves CME Group and Bank of Montreal (BMO). On March 24, 2026, BMO announced tokenized cash capabilities that will allow institutional clients to move value using CME Group's permissioned network on GCUL. The bank plans to offer an institutional settlement instrument to regulated financial services firms in the second half of 2026, pending regulatory approval.
On October 16, 2026, Jack Henry & Associates became the first core banking technology provider to integrate with GCUL, opening a potential path for more than 7,200 community banks and credit unions to offer tokenized deposits. No client bank or launch date was named.
Google charges a single API integration with predictable monthly fees — a SaaS model that contrasts with the variable gas-fee economics of public blockchains.
Google's entry creates a three-way contest for institutional digital-asset infrastructure: crypto-native firms, traditional custodians, and hyperscale cloud providers.
Crypto-Native Custody
| Provider | Assets Under Custody/Platform | Key Metric | |---|---|---| | Coinbase Custody | ~$300 billion | Custodian for 9 of 11 spot BTC ETFs | | BitGo | ~$90 billion | IPO in early 2026 at $2B valuation | | Fireblocks | $10 trillion cumulative tx volume | 2,400+ institutional clients |
These firms built infrastructure from the crypto side up. Coinbase Custody Trust Company is the largest regulated qualified custodian in crypto and serves as the primary custodian for the majority of U.S. spot Bitcoin and Ethereum ETFs. BitGo completed its IPO in early 2026, pricing at $18 per share for a $2 billion valuation, and reported 19% second-quarter revenue growth. Fireblocks operates as an infrastructure-layer platform securing over 300 million wallets.
Their advantage is maturity: years of production operation, regulatory licenses, and established relationships with asset managers. Their disadvantage is that they are single-purpose. None offers a blockchain, a general-purpose cloud, and custody in a single contract.
Hyperscale Cloud Providers
Amazon Web Services offers Managed Blockchain supporting Ethereum, Polygon, Bitcoin, and Hyperledger Fabric, plus API-based blockchain data access through AMB Query. Microsoft Azure provides blockchain services through ConsenSys Quorum. Neither has launched a proprietary blockchain or a dedicated MPC custody product.
Google's stack is wider: a proprietary blockchain (GCUL), a custody layer (MPC), cloud infrastructure (Google Cloud), and AI services (Vertex AI, Gemini) that can be layered on top. The bundle creates lock-in potential that neither AWS nor Azure currently matches in digital assets.
The Digital Asset Custody Market
The global institutional crypto custody market was valued at $1.41 billion in 2025 and is projected to reach $1.83 billion in 2026, growing to $14.4 billion by 2034 at a CAGR of 29.4%, according to Intel Market Research. The broader digital asset custody market — including tokenized real-world assets — is estimated at $0.7 trillion in 2026 by Mordor Intelligence, growing to $2.12 trillion by 2031.
These projections suggest an infrastructure market large enough to attract hyperscaler investment but still small relative to Google Cloud's $44.1 billion in annualized revenue as of Q2 2026.
Google's timing corresponds to a measurable escalation in private-key-related losses. CoinDesk reported in June 2026 that private-key compromises — not smart-contract bugs — caused 40% of crypto's cumulative $16.7 billion in hack losses. TRM Labs found that infrastructure compromises accounted for approximately 15% of incidents but 76% of dollar losses.
Specific 2026 incidents illustrate the scale:
The shift in attack methodology — from code exploits to social engineering, session hijacking, and validator key theft — plays directly to Google's strengths in identity management, hardware security modules, and enterprise access controls. MPC addresses the specific failure mode (single-key compromise) responsible for the largest losses.
What it gets:
What it doesn't get:
Google's crypto infrastructure play is not a single product launch. It is a stack: custody (MPC), settlement (GCUL), and compute (Google Cloud), integrated under a single vendor relationship and priced as a service. The approach mirrors Google's enterprise playbook in other verticals — offer infrastructure that creates switching costs, then expand the footprint.
The question is whether regulated institutions want their digital-asset infrastructure from a hyperscaler or from firms that built it specifically for crypto. Coinbase, BitGo, and Fireblocks have operational track records measured in years and hundreds of billions in assets. Google has distribution, AI, and the engineering resources of a $2.3 trillion company, but zero production custody history.
The market is large enough for both models to coexist. But Google's entry changes the competitive dynamics: crypto-native custody firms must now defend their positions against a competitor that can bundle blockchain, custody, cloud, and AI into a single enterprise contract. The firms that survive will be those whose specialization in digital assets creates more value than Google's breadth can replicate.