A cross-chain swap moves value from one blockchain to another in a single workflow, often pairing a bridge transport with a DEX hop so the user lands on the destination chain holding the asset they actually wanted. The route a swap takes determines its price, settlement time, and counterparty risk. With about $312 billion in stablecoins circulating across at least 15 chains and Ethereum still anchoring roughly $48.8 billion in onchain TVL (DefiLlama, September 15, 2026), the routing layer has become the most important variable in cross-chain UX.
​

This guide ranks the routes worth using in 2026 for the three highest-volume cross-chain assets, USDC, USDT, and ETH, and walks through how each route handles fees, finality, and asset coverage. The comparisons assume an institutional or power-user lens: the goal is the lowest all-in cost path that does not introduce custodial or wrapped-asset risk.
​

The list below ranks routes by asset coverage, settlement speed, published fee model, and the security model of the underlying transport. Aggregators are ranked separately from native protocols because they compose routes rather than originate them.

The full route comparison sits in the table below. The fee column describes each route's published fee model rather than a quote, because every one of these routes prices per transfer and per chain pair.
​

Speeds above are the protocols' own published targets, not promised SLAs. Both fees and times shift with mempool conditions, solver inventory, and gas tokens, so quote the route at the moment you need it.
​

Eco Portal is the stablecoin-native swap UI on top of Eco's execution network. It treats cross-chain stablecoin movement as a single intent: the user expresses what they want to hold and where, and the network selects solvers, transport, and DEX hops behind a single signature. Portal currently routes USDC, USDT, USDS, FDUSD, PYUSD, RLUSD, and ETH across 15+ chains, including Ethereum, Base, Arbitrum, Optimism, Polygon, BNB Chain, Solana, and Avalanche.
​

Behind the scenes, Eco Routes (the developer-facing CLI and API) selects between Circle's CCTP and Hyperlane messaging based on origin and destination, CCTP for native USDC paths, Hyperlane for the long-tail of stablecoins and chains where CCTP is unavailable. The result is that a Portal user gets native USDC delivery on routes where CCTP exists, and high-throughput Hyperlane delivery elsewhere, without choosing the underlying transport.
​

Pricing is solver-competitive: solvers quote the user, the cheapest quote wins, and Portal collects a transparent service fee on top. Quotes are generated per request at portal.eco.com, and a Portal route bundles the destination DEX swap into the same intent rather than charging for a second transaction.
​

Best for: users who want a one-stop swap UX for stablecoins across 15 chains without managing separate bridges and DEX hops. Particularly strong for cross-stable swaps (USDC → USDT, USDS → DAI, etc.) where a pure CCTP route does not apply.
​

Circle's Cross-Chain Transfer Protocol V2 (CCTP V2) is the canonical native-USDC bridge. The mechanism is burn-and-mint: USDC is burned on the origin chain, Circle's attestation service signs the burn, and an equivalent amount is minted on the destination chain. There is no wrapped intermediary, no liquidity pool to drain, and no LP fee.
​

CCTP V2 supports Ethereum, Arbitrum, Avalanche, Base, OP Mainnet, Polygon PoS, Solana, Sui, Aptos, Unichain, Linea, World Chain, and Codex. Standard transfer finality is around 13 minutes on Ethereum origin (driven by Ethereum's hard finality assumptions); Fast Transfers using Circle's allowance API can complete in under a minute on most destination chains. Circle's fee documentation states that Standard Transfers are free and Fast Transfers cost 0 to 13 basis points depending on the source chain.
​

Best for: USDC-only flows where the user controls both ends and does not need a destination-chain swap. Treasuries that move USDC from a corporate account on Ethereum to a Base operating wallet should default to CCTP. The trade-off is that CCTP only handles USDC. Any other token requires an additional swap leg, which is where aggregators and intent-based routes pull ahead.
​

Across is an optimistic intent-based bridge that has become the default fast path between Ethereum and major L2s for USDC, USDT, and ETH. Relayers fill user transfers on the destination chain in 2–10 seconds, then claim repayment from the origin chain after a UMA dispute window. The user experiences an instant transfer; the relayer takes the time-value risk.
​

Across supports Ethereum, Arbitrum, Optimism, Base, Polygon, ZkSync, Linea, Blast, and Mode. Its fee documentation describes a utilization-based LP fee plus a relayer fee covering destination gas and capital risk, quoted per transfer through the suggested-fees API rather than fixed at a headline rate. Across does not support cross-stable swaps directly, the protocol moves the same asset between chains; cross-stable conversion requires composing Across with a destination-chain DEX.
​

Best for: moving USDC, USDT, or ETH between EVM L2s when speed is the priority and the user does not need to change asset.
​

Stargate is the canonical stablecoin bridge built on LayerZero messaging. It uses a unified-liquidity model: each supported asset has a single global pool that tracks a delta across chains, so users withdraw the same canonical token on the destination side rather than a wrapped variant. Stargate handles USDC, USDT, ETH, FRAX, and a handful of other assets across 70+ chains via LayerZero's DVN-secured messaging.
​

LayerZero is not a live Eco Routes partner, it is a parallel infrastructure layer with its own messaging stack and security set. Users comparing Stargate against CCTP or Eco Portal should weigh the security tradeoff: Stargate inherits LayerZero's DVN model, in which configurable verifier sets (often Google Cloud, Polyhedra, and Nethermind) sign the cross-chain message before it executes. CCTP relies on Circle's centralized attestation. Eco Portal selects between CCTP and Hyperlane per route.
​

Best for: non-EVM destinations and long-tail chains where Stargate has unified liquidity but CCTP and Across do not. Less competitive on Ethereum-to-major-L2 paths where Across and CCTP are faster and cheaper.
​

Meta-aggregators do not originate routes. They compose underlying bridges and DEXs into a best-quote path. The four worth knowing in 2026:
​

LI.FI (docs) indexes 30+ bridges (Across, Stargate, CCTP, Hop, Connext, deBridge, Symbiosis, and others) and 20+ DEXs across 30+ chains. The SDK and widget are embedded in MetaMask, Phantom, Robinhood, and other consumer wallets. Pricing is best-quote at request time. Trade-off: route quality depends on which underlying bridges are integrated and which DEX hops are quoted at the moment of request.
​

Squid is built on Axelar's General Message Passing layer. Strongest on EVM-to-Cosmos paths and Axelar-native assets like axlUSDC. Less competitive on EVM-only routes where CCTP, Across, or Stargate dominate.
​

Jumper is LI.FI's flagship retail UI. Same routing engine, consumer-grade UX, gas-token abstraction, and a single-signature swap-and-bridge flow. It is the highest-traffic consumer front end built on LI.FI's routing engine.
​

Mayan is a Wormhole-anchored aggregator with strong Solana coverage. Auction-based fillers compete to settle the user's swap, similar to an intent design. Best for SOL-paired routes and any path that touches Solana, Sui, or Base from a non-EVM origin.
​

The right route depends on what the user is moving.
​

For pure USDC origin to USDC destination on a chain pair where both sides natively support USDC, Circle CCTP V2 is the cheapest correct answer: Circle charges no protocol fee on a Standard Transfer, finality is about 13 minutes, and there is no wrapped intermediary. For USDC where the user wants a destination-chain swap (USDC on Ethereum to USDT on Base, for example) bundled into one signature, Eco Portal wins because it uses CCTP for the bridge leg and a destination-chain DEX hop for the swap, atomically. Across is the fast alternative on EVM L2 pairs when the user wants USDC to USDC and 2–10 second settlement.
​

USDT does not have a CCTP analogue. USDT0, built by Everdawn Labs on LayerZero's OFT standard, is the canonical native-USDT cross-chain rail on supported chains, but coverage is narrower than CCTP. For most user-facing transfers in 2026, the practical route is Eco Portal or Across for EVM L2 pairs, and Stargate for non-EVM destinations. Total USDT supply sat at $183.3 billion on September 15, 2026 per DefiLlama, so liquidity is rarely the constraint, the constraint is which chain pair the user needs.
​

ETH bridges sit in two camps: native-ETH bridges (Across, Stargate, the canonical L2 bridges) and wrapped-ETH bridges that mint a wrapped representation. Native is preferable: no wrapped-asset risk, no peg-recovery scenario. Across is the practical default for ETH between Ethereum and major L2s, sub-10-second fills, 0.04–0.10% fees. The canonical L2 bridges (Optimism, Arbitrum, Base) are slower (7-day withdrawal window unless using a fast-bridge layer) but have the strongest security model for large transfers because they inherit Ethereum's data availability. Eco Portal handles ETH cross-chain when the user also needs a destination-chain swap or a non-EVM endpoint.
​

Four criteria separate the routes worth using from the ones to avoid.
​

Speed. Optimistic and intent-based routes (Across, Eco Portal, Mayan) settle in seconds because a relayer or solver fronts the destination-chain liquidity. CCTP V2 standard finality requires Ethereum hard-finality (~13 minutes); CCTP V2 Fast Transfers reduce this to under a minute at the cost of a small fast-fee. Lock-and-mint bridges (Stargate, Squid) settle in 1–3 minutes after destination-chain message verification.
​

Cost. Native protocols charge little or nothing at the protocol layer: Circle's CCTP is free on Standard Transfers and 0 to 13 bps on Fast Transfers, and the user pays origin and destination gas either way. Liquidity-pool bridges add an LP fee that scales with pool utilization. Aggregators add an integrator fee on top of whichever underlying route they select. Eco Portal sits in the same range as CCTP for USDC paths because it uses CCTP underneath; for non-USDC paths it competes via solver pricing.
​

Coverage. CCTP supports 13 chains for native USDC. LayerZero / Stargate supports 70+ chains but with a unified-pool security model. Eco Portal supports 15+ chains across 7+ stablecoins and ETH. Across supports the major EVM L2s. The right choice depends on which chain pair the user needs.
​

Security model. Bridges have been the largest source of crypto theft historically: Ronin ($624M, March 2022), Wormhole ($326M, February 2022), Nomad ($190M, August 2022), and Poly Network ($611M, August 2021), per the Rekt leaderboard. The lesson is that bridge security is a function of the underlying message-passing layer. Burn-and-mint (CCTP) is the simplest model: no pooled liquidity to drain. Optimistic models (Across) require a dispute window. DVN-secured messaging (LayerZero) is configurable per app but inherits the chosen verifier set. Hyperlane uses an interchain security module that apps configure per route.
​

Several routes that were viable in 2022–2023 have been retired or quarantined and should be avoided in 2026:

The pattern across these incidents is that bridges with custodial multisig or under-audited verification logic carry concentrated tail risk. The 2026 routes worth using all rely on either burn-and-mint (no pooled funds to steal), optimistic verification with dispute windows, or DVN sets with configurable verifier diversity.
​

Three patterns that experienced users apply when picking a route:
​

Use CCTP directly when both ends are USDC. If the user controls origin and destination wallets and wants USDC on both sides, calling CCTP directly through Circle's contracts skips aggregator fees entirely. Eco Portal and other aggregators add value when the user wants a destination-chain swap, gas abstraction, or a single signature for a multi-hop flow.
​

Match the route to the asset's native rail. USDC has CCTP. USDT has USDT0 on supported chains. ETH has the canonical L2 bridges. When a native rail exists for the asset, it is almost always cheaper and safer than a generic liquidity-pool bridge for the same pair.
​

Aggregate when uncertain, originate when certain. If the user does not know which underlying route is cheapest at this moment, a meta-aggregator (LI.FI via Jumper, Eco Portal) finds it for them. If the user runs the same path repeatedly and knows the cheapest underlying route, they save the aggregator fee by calling the underlying protocol directly.
​

Eco operates at the orchestration layer above transports like CCTP and Hyperlane. Rather than choosing between a bridge, a swap aggregator, and a settlement layer, teams integrate Eco once and get unified stablecoin routing across 15 chains. Eco Portal is the retail surface; Eco Routes (CLI and API) is the developer surface. Both consume the same execution network and the same solver set, which means a Portal user and a Routes API caller see consistent pricing and chain coverage. For background on the underlying primitives, see cross-chain stablecoin swap infra and guaranteed 1:1 stablecoin swaps across chains.
​

For native USDC between two CCTP-supported chains, Circle CCTP V2 is the cheapest: no protocol fee on a Standard Transfer, so the user pays only gas. For pairs that include a destination-chain swap, Eco Portal is competitive because it uses CCTP underneath where applicable and adds intent-based solver pricing. Across is fastest for L2-to-L2 same-asset transfers.
​

CCTP is a native USDC transport, not a generic bridge, it burns USDC on origin and mints on destination, so there is no wrapped intermediary or pooled liquidity. For USDC-to-USDC flows it is structurally safer than lock-and-mint bridges. For non-USDC assets or cross-stable swaps, an orchestrator or aggregator is still required.
​

Range is wide. Optimistic and intent-based routes (Across, Eco Portal, Mayan) settle in 2–60 seconds. CCTP V2 standard finality is about 13 minutes on Ethereum origin; CCTP V2 Fast Transfers settle in under a minute. Canonical L2 bridges have 7-day withdrawal windows for L2-to-L1, unless using a third-party fast bridge.
​

Bridges have historically been the largest target for crypto theft. The 2026 routes worth using rely on burn-and-mint (CCTP), optimistic verification with dispute windows (Across), or DVN-secured messaging with configurable verifier sets (LayerZero, Hyperlane). Avoid multisig-only bridges and any wrapped asset issued by a discontinued protocol.
​

Yes. Cross-stable cross-chain swaps require either an aggregator that bundles a bridge plus a destination-chain DEX hop, or an intent-based router like Eco Portal that handles both legs under one user signature. Pure CCTP cannot do this on its own because CCTP only moves USDC.
​