For a US asset manager that holds tokenised treasuries on Ethereum and trades perpetual swaps on a Solana DEX, moving value between the two chains is not a feature. It is the entire job. Cross-chain bridges are the rails that make that movement possible, and they have also been the single most expensive failure point in crypto. According to Chainalysis research, more than $2 billion was stolen from bridge exploits in 2022 alone. The category has cleaned itself up since, but the trust still has to be earned.
What a bridge actually does
A cross-chain bridge takes an asset on one chain and produces a representation of that asset on another. The original is locked in a contract on the source chain, and a wrapped or canonical version is minted on the destination. When the user wants to come back, the wrapped token is burned and the original is released. The mechanism is straightforward. The hard part is who, or what, decides that the lock and the mint are consistent.
Three architectures dominate. Lock-and-mint bridges rely on a multisig or validator set to attest that funds were locked. Light client bridges run a small on-chain replica of the source chain so the destination can verify the lock natively. Optimistic bridges use a fraud-proof window where attestations can be challenged. Each model trades latency, capital efficiency and trust assumptions differently. The first model is the easiest to ship and the easiest to attack. The second is the slowest and the most secure. The third is the current pragmatic middle ground.
A fourth architecture, intent-based settlement, is now scaling fast. Instead of locking the source asset and waiting for a confirmation, a market of solvers fills the user’s request on the destination chain from inventory, then settles with the source chain in the background. The user sees a near-instant fill. The protocol settles atomically once the source confirmation arrives. The model is a re-architecture of bridging into something closer to wholesale FX market-making, and it changes which firms own the rail.
The risk record, in numbers
Bridge security has been the most painful subplot of the last four years in crypto. The Ronin bridge lost $625 million in March 2022 when the Sky Mavis multisig was compromised. The Wormhole bridge lost $325 million the previous month after a smart contract bug let an attacker mint wrapped ETH without locking real ETH. Nomad lost $190 million in August 2022 in a chaotic free-for-all triggered by a misconfigured proxy upgrade. The pattern is not random: it is the cost of bridges sitting at the most valuable concentration point in a multi-chain stack.
The 2023-2025 trend is more encouraging. Exploit losses dropped sharply once protocols started moving away from custodial multisigs and toward light-client and zero-knowledge architectures. Total bridge volume kept growing through that period, which means the surviving bridges processed more value with fewer incidents. That is the kind of operational maturity that matters to a US risk committee approving a digital-asset desk.
The remaining attack surface is concentrated in a smaller set of components: validator-set rotations, governance contracts that can pause the bridge, and the off-chain relayer infrastructure that moves messages between chains. Every bridge that has shipped through 2025 with a clean track record has invested heavily in monitoring and rapid pause capability for exactly these components. The lesson from the 2022 wave is that detection speed mattered as much as cryptographic robustness, because the largest losses were always the ones that ran for hours before anyone noticed.
Where US institutions sit today
US institutional bridge use is still concentrated among crypto-native firms, but the buyer profile is widening. Asset managers running on-chain treasury strategies need to move stablecoins between Ethereum and the Layer 2 their preferred yield protocol sits on. Custodians supporting the tokenized US Treasuries market that reached roughly $7 billion in late 2025 need a settlement path between Ethereum mainnet and the rollup the fund actually trades on. Payment processors experimenting with how merchants are starting to accept stablecoin payments need bridges to move USDC between issuance chains and consumer-wallet chains without retail users noticing.
The volume picture is dominated by two flows. Stablecoin redistribution between mainnet and Layer 2s accounts for the largest share, followed by liquid-staking-token migration as users chase yield differences across networks. Native token bridging for users buying into a new ecosystem is a distant third by value, even though it gets the most press. For the institutional buyer, the relevant question is which bridge a regulated counterparty actually uses for its largest flows, not which one a retail user sees in a wallet UI.
A small handful of named bridges now dominate institutional flows. Across the Layer 2 stack, the Across, LayerZero and Hyperlane stacks have built the largest non-custodial volumes. The Circle CCTP bridge for native USDC has become the default for stablecoin movement once a treasury team can take on the issuer relationship. The big custodians have largely declined to operate bridges themselves and instead route institutional flow through a small slate of audited third-party rails.
How US regulators are framing the question
US regulators have not produced bridge-specific guidance. They do not need to. Cross-chain bridges are evaluated under the existing framework for money transmission, securities settlement and operational risk. The Office of the Comptroller of the Currency has signalled, through interpretive letters since 2020, that national banks can participate in public blockchain infrastructure under standard risk-management controls. The Securities and Exchange Commission has focused on the asset side, asking whether a wrapped token retains the security status of its underlying. The Financial Crimes Enforcement Network has looked at bridges through the AML lens, particularly when sanctioned protocols are involved.
The Tornado Cash sanctions in 2022 created a long-running compliance question for bridge operators about how to filter prohibited counterparties without breaking the permissionless nature of the rail. The answer the market has settled on is to push compliance to the edge: enforce screening at the on-ramps and off-ramps used by US persons, rather than at the bridge contract itself. That is the same pattern emerging in the digital euro and what it means for crypto and banks debates in Europe, and the convergence is not coincidental.
| Date | Bridge | Reported loss | Primary source |
|---|---|---|---|
| February 2022 | Wormhole | ~$325M | Chainalysis |
| March 2022 | Ronin Network | ~$625M | Chainalysis |
| August 2022 | Nomad | ~$190M | Chainalysis |
| March 2024 | BlackRock BUIDL fund launch (Ethereum) | , | Securitize / PR Newswire |
Chainalysis estimated cross-chain bridges accounted for ~69% of all stolen crypto in 2022.
Where bridges go from here
The next phase of bridge design is being shaped by three forces. Zero-knowledge proofs are starting to underwrite light-client verification at production scale, which removes the validator-set trust assumption that drove the worst exploits. Shared sequencer and shared settlement layers, such as those developing around the Ethereum rollup ecosystem, are reducing the number of places a bridge actually needs to sit. And intent-based architectures are letting users specify a desired outcome rather than a specific route, which abstracts the bridge entirely from the user experience.
For US institutions, the implication is that the bridge stops being a product they choose and starts being infrastructure they audit. The right diligence questions are the same ones any treasurer would ask of a correspondent bank: who is on the hook for a failed transfer, how fast does the rail recover from a stuck transaction, and what is the operational record over a multi-year window. The crypto-specific overlay is which cryptographic guarantees the bridge actually provides, and how those guarantees survive the worst day in its history.
By the end of 2026, the bridge category will likely look more like wholesale payments infrastructure than the wild experiment it began as. The technology is converging on a smaller number of safer designs, the institutional buyer is becoming more discerning, and the regulatory frame is settling. The remaining question is which firms own the bridge-as-infrastructure layer, and whether they will be crypto-native protocols or the same custodians and clearers that already sit between US institutions and the rest of the financial system.
