An Arbitrum bridge moves assets between Ethereum (or another chain) and Arbitrum One, an Ethereum Layer 2 with multi-billion-dollar onchain TVL per DeFiLlama as of Q1 2026. Bridging Arbitrum is unusual: the chain ships with its own rollup bridge whose withdrawal path takes seven days by design, so most users pick a faster third-party route in at least one direction. The deposit path is fast on every option; the withdrawal path is where bridges diverge sharply.
​

An Arbitrum bridge is any protocol that lets a user move tokens between Arbitrum One (or Arbitrum Nova) and another chain. Routes split into two families: the native rollup bridge run by the Arbitrum Foundation, which inherits Ethereum security but takes about seven days to withdraw, and faster third-party bridges that pre-fund destination liquidity and settle later.
​

Arbitrum One processed over two billion lifetime transactions by 2026, so bridge volume there sits among the highest of any L2. The native bridge minimizes trust assumptions; third-party bridges minimize wait time. See Eco's fastest bridge overview if speed is the only criterion.
​

Arbitrum's native bridge handles deposits in roughly 10 to 15 minutes and withdrawals in about seven days. The asymmetry is structural. Deposits inherit Ethereum's finality almost immediately, while withdrawals must wait through the optimistic rollup challenge period so verifiers can submit fraud proofs against any invalid state root that was posted to L1.
​

That seven-day window is documented by the Arbitrum Foundation's bridge quickstart and is the same security model used by Optimism, Base, and other optimistic rollups. Fast bridges sidestep the wait by holding inventory on both chains: a relayer fronts the user's destination asset within seconds, then claims the source-chain funds when the message verifies. The trust model shifts from rollup proofs to the relayer's economic security. For deposits, the native bridge is hard to beat on cost. For withdrawals, third-party routes are usually faster and often cheaper than the L1 gas needed to finalize a native exit.
​

The table below summarizes the six most-used route families. Fees and speeds are typical ranges for small to mid-size transfers; gas varies with mainnet conditions. Each fits a different combination of asset, direction, and tolerance for waiting.
​

Eco publishes a separate bridge fees breakdown with size-tier analysis if cost is the deciding factor.
​

Circle's Cross-Chain Transfer Protocol burns USDC on the source chain and mints fresh native USDC on Arbitrum, so no wrapped representation is created. Circle deployed native USDC on Arbitrum in June 2023, which makes Arbitrum a first-class CCTP destination rather than a bridged-USDC chain.
​

CCTP V2 launched on Ethereum and Avalanche on March 11, 2025 and has since extended across the EVM mainstream, with Arbitrum among the 13+ supported chains in Circle's CCTP overview. V2 adds a Fast Transfer mode that finalizes in 8 to 20 seconds on many corridors, plus programmable hooks that let a destination contract react atomically on mint. Burn-and-mint avoids the pool-drain risk of older lock-and-mint bridges. CCTP is the route most teams pick for native USDC in either direction. See the Eco CCTP V2 guide for the walkthrough.
​

Across is an intent-based bridge with deep Arbitrum roots, built by the Risk Labs team behind UMA. A user signs a single intent specifying source chain, destination chain, token, and amount. Relayers compete to fill it on the destination, then claim repayment after the message is verified by UMA's Optimistic Oracle. The UMA case study walks through how the oracle handles disputes.
​

Across typically finalizes Arbitrum withdrawals in seconds to a few minutes, which is the largest delta from the native bridge's seven-day exit. Supported assets include ETH, WETH, USDC, USDT, DAI, and WBTC across Arbitrum, Optimism, Base, Polygon, and Ethereum. It is one example of the intent-and-solver mechanism described in Eco's intents and solvers guide; deBridge and Squid Coral use related mechanisms. The trade-off is the trust model: native exits inherit Ethereum security, while intent fills depend on the solver's bond plus the oracle's challenge process.
​

Hyperlane warp routes are permissionless token bridges built on Hyperlane's interchain messaging. Each warp route is deployed for a specific token across a specific set of chains; the source chain locks or burns supply, the destination mints a synthetic representation or releases locked collateral. Hyperlane's warp routes documentation describes the four asset types: native, collateral, synthetic, and rebase.
​

For Arbitrum, warp routes are most relevant for ERC-20s that lack a dedicated bridge: governance tokens, restaking derivatives, long-tail stablecoins. Warp Routes 2.0 introduced native rebalancing via Everclear, letting routes pull liquidity from multiple source chains so a single destination route never runs dry. Warp routes are not the cheapest path for ETH or native USDC; CCTP and Across usually win there. The niche is the ERC-20 long tail.
​

LayerZero is a messaging protocol; Stargate is the canonical liquidity protocol built on top of it. Stargate pools liquidity across chains so a transfer is paid out from the destination pool and reconciled via LayerZero messaging. For Arbitrum, Stargate runs unified pools for USDC, USDT, ETH, and other assets, with transfers settling in minutes. USDT0, the Tether-issued omnichain USDT using LayerZero's OFT standard, is live on 15+ chains including Arbitrum and is the closest USDT analog to native CCTP USDC. The LayerZero stack is one of the cleanest USDT routes onto Arbitrum, especially from BNB Chain or Tron. LayerZeroScan publishes live message volumes.
​

Aggregators do not move funds themselves. They quote across many of the bridges above, then route through the chosen one in a single transaction. For Arbitrum that often means CCTP for native USDC, Across for ETH or WBTC withdrawals, Stargate or LayerZero for USDT0, and a warp route for an obscure ERC-20.
​

The three most active aggregators on Arbitrum are LI.FI, Squid (built on Axelar), and Jumper (LI.FI's consumer front end). All three quote multiple rails and surface price plus speed per route. Their fees are small markups on the underlying route. See Squid and Jumper for differences and the LI.FI guide for its routing engine.
​

The Arbitrum sequencer orders L2 transactions before they are batched to L1. Deposits enter via L1-to-L2 messages the sequencer processes once confirmed on Ethereum; withdrawals must clear the rollup's challenge window before they can be claimed on L1. Sequencer backlog or downtime can stretch the usual 10 to 15 minute deposit window. In April 2025, Offchain Labs deployed Timeboost on Arbitrum One: a sealed-bid second-price auction grants the winner a 200-millisecond ordering advantage for one minute at a time, per the Arbitrum docs. For ordinary bridge users Timeboost is invisible; for combined bridge plus DEX flows it changes block-top competition. More in Eco's sequencer explainer.
​

The right Arbitrum route depends on direction and asset, not on a universal ranking. Deposits from Ethereum settle fast on every option, so the native bridge or CCTP usually wins on cost. Withdrawals split sharply: the native exit takes seven days but minimizes trust, while CCTP, Across, and Hyperlane finalize in minutes by relying on intermediary security. A practical decomposition:
​

For builders routing payments across L2s programmatically, Eco Routes exposes a single intent API that aggregates CCTP, Hyperlane, and ERC-7683-compatible rails, so the choice above is made per transaction rather than per integration.
​

Arbitrum is an optimistic rollup. State roots posted to Ethereum are assumed valid unless challenged within a seven-day window during which verifiers can submit fraud proofs. Third-party bridges skip the wait by fronting destination liquidity and accepting different trust assumptions.
​

For USDC, CCTP is far faster in the withdrawal direction. CCTP V2 Fast Transfer finalizes in 8 to 20 seconds on many corridors, including Arbitrum to Ethereum. The native bridge withdrawal is about seven days. Deposits are roughly comparable on both, with CCTP in seconds to minutes and the native bridge in 10 to 15 minutes.
​

USDT0, the LayerZero-based omnichain USDT, lives natively on Arbitrum. Standard USDT is also widely supported via Stargate, Across, and aggregator-routed paths. Hyperlane warp routes add a permissionless option. USDT does not have a CCTP-equivalent burn-and-mint protocol, so route choice matters more than for USDC.
​

Both are optimistic rollups with seven-day withdrawal windows on their native bridges. Both support CCTP for native USDC. Optimism sits inside the OP Superchain with intra-Superchain bridging paths to Base, Mode, and others; Arbitrum has its own Orbit chain ecosystem. See the Eco Arbitrum vs Optimism guide.
​

Sources and methodology. Bridge mechanics verified against the Arbitrum Foundation docs, Circle CCTP documentation, Hyperlane docs, and Across Protocol research at UMA. TVL pulled from DeFiLlama in Q1 2026. Timeboost launch and fee figures from Arbitrum docs and reporting by The Block. CCTP V2 launch date from Circle's March 2025 announcement. Figures refresh quarterly.
​