Bridges have accounted for a disproportionate share of total value lost, and the post-mortems show a small number of recurring causes.
Why bridges concentrate risk
They hold locked assets backing representations issued elsewhere.
Which accumulates very large balances in a single contract or custody arrangement.
The value at risk is visible on chain, which makes them identifiable targets.
Validator key compromise
Several incidents involved attackers obtaining enough signing keys to authorise withdrawals.
Which is not a cryptographic failure but an operational security one.
Social engineering of individuals holding keys has featured in more than one case.
Small validator sets
Where a threshold of a handful of signers authorises transfers.
Which means compromising a few parties is sufficient regardless of the value secured.
The mismatch between the cost of compromise and the value at stake was visible in advance.
Signature verification bugs
Implementation errors accepting invalid proofs.
Which allowed forged withdrawal authorisations.
These are code defects and were present in audited contracts in some cases.
Upgrade and initialisation errors
Contracts left in a state allowing takeover.
Which includes uninitialised proxy contracts and misconfigured permissions after upgrades.
Deployment procedure rather than contract logic was the failure point.
Monitoring and response
Time between exploitation and detection varied from minutes to days.
Which correlated directly with how much was lost.
Protocols with automated anomaly detection and pause capability limited damage substantially.
Recovery outcomes
Some funds were recovered through negotiation, chain analysis and exchange cooperation.
Which was the exception rather than the rule.
Reimbursement from treasuries or new token issuance was the more common resolution.
What the record indicates
The attestation mechanism, not the contracts on either side, is where these systems fail.
Published trackers document these arrangements per bridge, and the information is public.
The scale relative to other categories
Bridge losses have accounted for a substantial share of all funds stolen in the field.
Which is disproportionate to the amount of capital they hold.
Incident databases track this and the concentration is clear.
State-linked activity
Some incidents have been attributed publicly to state-affiliated groups.
Which changes the resources available to an attacker considerably.
Attribution is made by governments rather than by technical analysis alone.
Recovery negotiations
Public bounty offers in exchange for returning funds.
Which succeeded in at least one large case.
The legal status of such arrangements is not straightforward in most jurisdictions.
Design responses
Larger validator sets, native verification, rate limits and value caps.
Which have been adopted unevenly across the sector.
Application-level caps are cheap and have limited losses where implemented.
Assessing a bridge today
Public risk trackers document validator arrangements, upgrade keys and audit history for major bridges.
Insurance and reimbursement
Some protocols reimbursed users from treasury or through new token issuance.
Which spreads the loss to token holders.
Others could not, and users bore the loss entirely.
Audit coverage
Several exploited bridges had been audited.
Which indicates that review of complex cross-chain logic is genuinely difficult.
Scope exclusions and post-audit changes featured in more than one case.
Wrapped asset risk
Holders of representations backed by a compromised bridge held claims against nothing.
Which is the practical consequence for users who never interacted with the bridge directly.
Different wrapped versions of the same asset are not interchangeable for this reason.
Alternatives that reduce exposure
Natively issued assets on multiple chains, and protocol-level interoperability frameworks.
Which move the trust to the issuer or to the consensus layer.
The practical advice
Minimise bridged balances, check the risk trackers, and prefer native assets where available.
Why it still matters
The pattern is consistent enough to be predictive: a small signing set securing a large balance is an arithmetic problem waiting to be solved by someone.
That arithmetic is publicly checkable for every bridge in operation, which is why the risk trackers exist.
Reading the primary material
Post-mortems, exploit transactions and public risk trackers document each incident in detail.
Which allows the recurring causes to be identified directly rather than taken on trust.
Before using one
Check who authorises releases, how many independent parties are required, and whether administrative keys exist.
Where the category stands now
Losses have continued while the largest individual incidents have become less frequent.
Which reflects better practice at the largest bridges rather than a solved problem.
Newer and smaller bridges reproduce the original weaknesses regularly.
The one sentence version
A handful of signers securing an enormous balance is an arithmetic problem, and it has been solved repeatedly by people who were not supposed to solve it.
A note on the underlying difficulty
Verifying one chain's state on another is genuinely hard, and the shortcuts that make it practical are where the risk concentrates.
Native verification solves it properly and costs more to build and to run, which is why simpler designs persist.
The trackers documenting these arrangements exist precisely because the information is public and nobody was looking at it.