The short answerIBC is what a bridge should be: each chain runs a light client of the other and verifies proofs cryptographically, so there is no committee to bribe or key to steal. It costs nothing beyond gas and has never suffered a protocol-level theft. Its reach is confined to Cosmos-SDK chains and a few adapters.
Best forTrust-minimised transfers within Cosmos
Trust assumptionsweight 30%
9.5avg 7.2
Exploit historyweight 25%
9.5avg 8.8
Settlement speedweight 15%
8.5avg 8.3
Chain coverageweight 15%
6.5avg 8.1
Costweight 15%
9.5avg 8.2
this service category average across 15Score = Σ mark × weight. Recommendable to most readers, with stated caveats. Methodology.
| Criterion | Mark | What we found | Source | Checked |
|---|
| Trust assumptions | 9.5 | not sourced yet | not sourced yet | — |
| Exploit history | 9.5 | not sourced yet | not sourced yet | — |
| Settlement speed | 8.5 | not sourced yet | not sourced yet | — |
| Chain coverage | 6.5 | not sourced yet | not sourced yet | — |
| Cost | 9.5 | not sourced yet | not sourced yet | — |
What we liked, and what fell short
Liked- Light-client verification with no external validator set to compromise
- No fee beyond gas
- Four years of heavy use with no protocol-level theft
Fell short- Confined to Cosmos-SDK chains and a few adapters
- Security of a route is only as good as the weaker chain's consensus
Why the marks landed where they did
Almost every bridge asks you to trust a group of entities to attest that something happened on another chain. IBC does not. Each chain runs a light client of the counterparty chain and verifies consensus proofs directly, which means the security of a transfer reduces to the security of the two chains involved.
Why that is categorically different
The multi-billion-dollar bridge losses in this sector — Ronin, Wormhole, Harmony, Nomad — all involved compromising a validator set, a signature check or a key. IBC has no such component. There is no set of signers, no multisig, no attestation service. The failure mode is a chain's own consensus failing, which is a different and much higher bar.
The limits
IBC works between chains that can run light clients of each other, which in practice means Cosmos-SDK chains and a growing set of adapted implementations. Connecting to Ethereum or Solana requires additional machinery that reintroduces some trust. And a connection is only as strong as the weaker chain's consensus — IBC to a chain with three validators inherits that weakness.
Cost and practice
Gas only. Transfers typically complete in seconds to a minute depending on block times, and wallets such as Keplr handle routing so users rarely see the mechanics.
Who should use it
Everyone operating within Cosmos, which is where it works. Its wider significance is as the reference standard other cross-chain designs are measured against.
FAQ
What makes IBC more secure than other bridges?+
Each chain verifies the other's consensus directly through a light client, so there is no external validator set, multisig or attestation service to compromise.
Has IBC ever been exploited?+
No protocol-level theft in four years of heavy use, though implementation bugs in individual chains have been found and patched.
Can I use IBC to reach Ethereum?+
Only through additional adapters and bridges that reintroduce trust assumptions. Native IBC works between chains capable of running light clients of each other.
What does IBC cost?+
Gas only. There is no protocol fee.
