CCIP and native bridges cover the institutional default and the chain-pair specialist. Two further mechanisms sit at the edges and matter for a complete picture: atomic swaps, the cryptographically pure answer that almost no institutional flow uses, and intent-based routing, the newer architecture that retail and crypto-native flows have migrated to. This part covers how each works, where each fits, and where the other generalised messaging protocols (LayerZero, Wormhole, Axelar) sit relative to CCIP.
Atomic swaps via HTLCs
An atomic swap is a bilateral cross-chain trade with no intermediary. The mechanism is a Hash Time-Locked Contract (HTLC) on each chain, set up so that revealing a single secret unlocks both. Party A locks their asset on chain 1 in a contract that pays out to B if B can produce a pre-image whose hash matches a known value, otherwise refunds A after a timeout. B locks the mirror on chain 2 under the same hash condition. When B reveals the pre-image to claim A's asset on chain 1, A learns the pre-image and uses it to claim B's asset on chain 2. Either both legs settle or both refund.
The structural property is genuine cross-chain atomicity. There is no certifying network, no escrow operator, no trust assumption beyond the cryptographic primitives and chain-level finality on each side. Where CCIP has a window between source-chain lock and destination-chain mint, an atomic swap collapses that window to zero.
Institutional finance does not use the pattern at scale because of operational cost. Both counterparties have to be online during the swap window. Both have to manage HTLC timeouts so a failed counterparty does not strand assets. Multi-party flows do not extend: adapting a fundamentally two-party operation to flows involving issuer, transfer agent, custodian, and counterparty introduces enough complexity to wipe out the atomicity benefit. The pattern lives in some DEX implementations and bilateral DeFi arrangements. It does not live in tokenised MMF or tokenised deposit settlement, and is unlikely to.
Hold atomic swaps as the theoretical reference point. They prove cryptographically pure cross-chain settlement is possible. Every other mechanism is a trade-off against the operational cost of that purity.
Intent-based routing
Intent-based routing inverts the question: the user specifies an outcome, and a competitive solver network finds the path. A user submits an intent ("I want X tokens on chain B in exchange for Y tokens on chain A"). A network of solvers competes to fulfil it. The winning solver fronts the destination-chain balance immediately, then unwinds the source-chain side through whatever mechanism is most efficient: a canonical bridge, an inventory rebalance, a separate cross-chain operation.
Across, deBridge, and CoW Protocol are the prominent examples. The user experience is fast and cheap: solvers compete on price and speed, the winning solver delivers the destination tokens within seconds to minutes, and the fee is bounded by the spread the solver can extract. For retail and crypto-native flows, this has become the dominant cross-chain UX, displacing direct bridge interactions on most volume.
The institutional reservations are three. First, the solver layer introduces counterparty risk: the user trusts the solver to deliver, the solver trusts the source-chain side to eventually unwind. Most solver networks have collateral or stake requirements, but the risk shape is different from a CCIP transfer where the protocol itself stands behind the message. Second, the centralisation profile is non-trivial: a small number of well-capitalised solvers tends to dominate any given chain pair, and "who are the solvers, what is their balance sheet, what is their incident history" does not have a stable answer. Third, the regulatory framing is unsettled. How regulators treat an intent transaction (series of trades, single cross-chain transfer, swap) is jurisdiction-specific and evolving.
For institutional DvP at MMF-settlement scale, intent-based routing is worth tracking but is not yet the default. The category is moving fast, the security record is still being built, and institutional patience for a short track record is limited. Likely path: it matures, picks up institutional partnerships, and becomes a credible third option alongside CCIP and native bridges over the next few years. Not there today.
LayerZero, Wormhole, Axelar
The other generalised messaging protocols compete with CCIP on the same surface with weaker institutional positioning. LayerZero uses a two-of-two model where the application chooses an oracle and a relayer, both of which must agree before a message executes. The model is flexible and the protocol has strong adoption in DeFi, but the application-chosen verifier model is harder to underwrite for institutional diligence than CCIP's single specified Risk Management Network.
Axelar runs its own proof-of-stake validator network as a sovereign chain, with messages relayed through that network to and from connected chains. The architecture is clean and the validator set is non-trivial, but the institutional footprint is smaller than Chainlink's and the production deployments at the names that anchor the institutional case (Swift, DTCC, JPM) are not there.
Wormhole uses a Guardian network of nineteen signers attesting to cross-chain events. The protocol moves significant volume and has matured since its early incidents. The episode that anchors institutional reluctance is the February 2022 exploit: a signature-validation bug in the Solana-side contract let an attacker mint 120,000 wrapped ETH on Solana without the corresponding Ethereum deposit, a loss of roughly USD 326 million. Jump Crypto, the project's main backer at the time, replaced the funds within hours to keep the protocol solvent. The exploit was a contract bug rather than a Guardian-network compromise, and Wormhole's subsequent security investments have been substantial, but the incident is in the file every institutional counsel reads on the cross-chain category.
The strategic picture: CCIP is the institutional default because of the RMN architecture, the regulatory engagement, and the production deployments at the right names. LayerZero, Axelar, and Wormhole are credible protocols but are not yet displacing CCIP on the institutional surface. The cross-chain category has been the most heavily exploited DeFi surface in dollar terms, which is why the institutional bar for any new mechanism is high.