Crypto Gloss

Across Protocol: Optimistic Cross-Chain Bridge

Authors Risk Labs / UMA Protocol
Year 2021
Project Across Protocol
License GPL-3.0
Official Source https://docs.across.to/

This page is an educational summary and analysis of an official whitepaper or technical paper, written for reference purposes. It is not a verbatim reproduction. CryptoGloss does not claim authorship of the original work. All intellectual property rights remain with the original author(s). The official document is linked above.

Across Protocol is a cross-chain bridge described in its 2021 documentation by Risk Labs (the company behind the UMA Protocol). Across’s primary differentiator from other cross-chain bridges is its optimistic oracle + bonder architecture:

  1. Bonders (relayers) front capital on the destination chain immediately, so users receive funds without waiting for L1 finality
  2. The bonder’s payment on the source chain is not released until UMA’s optimistic oracle confirms the transfer was valid
  3. A single Hub Pool on Ethereum L1 manages all liquidity; Spoke Pools on L2s collect deposits

> Documentation: Available at docs.across.to.

Publication and Context

In 2021–2022, cross-chain bridges were either:

  • Validator-based (Wormhole, Axelar): Secure but require waiting for validator confirmation
  • Canonical (Optimism/Arbitrum official bridges): Secure but 7-day withdrawal wait (fraud proof window)
  • AMM bridges (Hop): Fast but expensive for large amounts and require deep per-chain liquidity pools

Across proposed an optimistic design: let specialized bonders take the short-term liquidity risk for users (giving them instant transfers), then settle with cryptographic finality via UMA’s dispute resolution system.

Architecture: Hub and Spoke

Hub Pool (Ethereum L1):

  • Centralized liquidity pool on Ethereum mainnet
  • Holds USDC, WETH, and other bridgeable assets
  • Manages liquidity allocation across spoke pools
  • Settles claims from relayers/bonders

Spoke Pools (L2s and other chains):

  • Light contracts on Arbitrum, Optimism, Polygon, etc.
  • Receive user deposits on the source chain
  • Verify relayer claims and process destination-chain payments

Transfer Flow (One-Click Cross-Chain)

Step 1: User deposit on source chain

User sends USDC to the Arbitrum Spoke Pool, specifying destination chain (Optimism) and recipient.

Step 2: Bonder (Relayer) fronts funds

A bonder monitoring the deposit queue immediately sends USDC from their own wallet to the user on Optimism. The user has their funds in seconds — no waiting for L1 finality.

Step 3: Relayer submits fill claim

The bonder submits their fill claim to the Hub Pool on Ethereum L1, including proof of the fill.

Step 4: Optimistic verification

UMA’s optimistic oracle accepts the claim. If no one disputes it within the dispute window (2 hours), the claim is considered valid.

Step 5: Bonder repayment

The Hub Pool releases the bonder’s fronted funds plus a fee from the source chain deposit.

Dispute path: If an incorrect fill claim is submitted, any challenger can post a bond and dispute it. UMA’s dispute resolution process verifies via its DVM (Data Verification Mechanism); the losing party’s bond is slashed; legitimate disputes are compensated.

UMA Optimistic Oracle Integration

Across uses UMA’s optimistic oracle — UMA’s mechanism for verifying real-world assertions on-chain:

  • A claim is made: “Bonder X filled deposit Y with Z USDC on chain A”
  • Default assumption: the claim is correct (optimistic assumption)
  • Dispute window: anyone can challenge within 2 hours
  • If challenged: UMA’s DVM (a token-holder vote on the assertion) resolves the dispute
  • The UMA DVM is the ultimate source of truth; DVM votes are incentivized to be honest via UMA tokenomics

This is an optimistic rather than cryptographic bridge — security rests on the assumption that at least one honest disputer exists and has sufficient incentive to challenge fraudulent claims.

LP Capital Efficiency

Unlike Hop Protocol (which requires separate AMM pools per chain), Across’s single Hub Pool design means liquidity is pooled in one place — the Hub Pool on Ethereum L1. A dollar deposited is available to relay transfers on any spoke chain, not just one.

LP returns: LPs earn fees from all bridge transfers regardless of which chains are active, making capital more efficient than per-chain liquidity pools.

Utilization-based fee model: Transfer fees in Across are set algorithmically based on pool utilization — fees increase when demand is high relative to available capital, incentivizing more LP deposits.

Across v2 / Intent Architecture

Across v2 (2024) pivoted toward an intent-based bridge model:

  • Users express an intent (transfer X from chain A to chain B) rather than executing a specific path
  • Solvers (competitive relayers) compete to fill intents optimally
  • Best-priced fill wins; losing solver bids are not executed

This is the same “intents” paradigm emerging across DeFi (UniswapX, 1inch Fusion, CoW Protocol) — outsourcing route optimization to competitive solvers rather than requiring on-chain route specification.

Reality Check

Across is widely considered one of the most capital-efficient and secure bridges for transferring assets between Ethereum and its L2s:

  • No major hacks as of 2024
  • Best-in-class fees for moderate transfer sizes
  • UMA optimistic oracle provides dispute resolution without requiring 24/7 validator uptime

Caveats:

  • Optimistic security assumptions: Security depends on honest disputers existing and being incentivized. A scenario where all potential disputers are compromised is theoretically possible.
  • Bonder capital requirements: Bonders must front significant capital to fill multiple transfers; they earn fees but have liquidity tied up between deposit and repayment.
  • Limited to Ethereum ecosystem: Across originally supported only Ethereum L2s; cross-chain support for other ecosystems (Solana, Cosmos) is limited compared to Axelar or Wormhole.

Legacy

Across pioneered the optimistic oracle bridge model and the intent-based bridge architecture that has become a major design trend in 2024. Its single Hub Pool design demonstrated that capital efficiency doesn’t require per-chain AMM pools. ACX token governance gives bridge users a path to fee revenue.

Related Terms

Research

  • Risk Labs / UMA. (2021). Across Protocol Documentation. docs.across.to.

— Primary documentation. Covers Hub/Spoke architecture, bonder flow, UMA oracle integration, and fee model.

  • Buterin, V. (2021). Why remote validation across domains is impractical. Ethereum Research Forum.

— Blog post outlining fundamental limitations of cross-chain bridges; provides theoretical context for why Across’s L2-to-L2 focus (within Ethereum ecosystem) is safer than arbitrary chain-to-chain bridges.

  • Hart, J., et al. (2023). Intents-Based Architecture for Cross-Chain Protocols. Risk Labs.

— Technical paper on intent-based bridge design; describes the solver competition model underlying Across v2.