Exploring trust-minimized fast synchronization for relayer infrastructure in Cosmos

Hi everyone,

We are Teragone Factory, a distributed systems and blockchain infrastructure engineering team working on synchronization, verification and operational scalability challenges in decentralized environments. Over the past few years, we have been working on production-grade distributed state certification, aggregated signatures and fast synchronization systems in another large PoS ecosystem.

As we’ve spent more time exploring Cosmos, one operational question increasingly stands out: could interchain infrastructure eventually require more trust-minimized synchronization mechanisms as ecosystem complexity grows?

This is not a proposal, product announcement or funding request, we are mostly trying to understand whether this direction resonates with existing operational pain points, and whether similar work already exists somewhere in the stack.

The problem

One concrete area that feels increasingly important is relayer infrastructure recovery and bootstrapping.

Today, relayers interacting with multiple chains must continuously maintain synchronized and verified views of many distributed states. When incidents, upgrades or migrations occur, recovery still relies on heavy assumptions: replaying significant chain history, or depending on trusted checkpoint and snapshot providers.

At current ecosystem scale, this is still manageable, but as the number of sovereign IBC-connected chains grows, the operational cost of synchronization and recovery may eventually become a real bottleneck.

One possible direction

CometBFT already includes a state sync mechanism, but it currently depends on externally trusted checkpoint providers to bootstrap trust. What feels worth exploring is whether validators could periodically co-sign compact cryptographic certificates attesting to specific chain snapshots or state roots, effectively replacing that external trust dependency with validator-backed attestations, without modifying the underlying consensus or IBC assumptions.

Work such as Mithril on Cardano has shown this class of mechanism can operate at production scale in PoS environments. Mithril relies on stake-based threshold signatures, whether a similar model is compatible with CometBFT’s validator key infrastructure and operational constraints is one of the open questions we’d most value feedback on, and where we suspect the Cosmos context requires a meaningfully distinct approach.

In practice, this could allow faster relayer recovery, lighter verification flows, and more operationally resilient interchain infrastructure, while keeping trust anchored to validator consensus rather than centralized providers.

Three open questions

State certification and IBC

For protocol engineers: could periodic validator-certified state snapshots complement existing IBC light client models for synchronization workflows without weakening current trust assumptions?

Signature aggregation

On signature aggregation: would aggregated signature schemes realistically provide meaningful operational gains within existing Cosmos validator infrastructure, and is BLS or a threshold scheme compatible with current CometBFT key management constraints?

Relayer operations

For relayer operators: is synchronization and recovery latency becoming a meaningful operational constraint as chain count increases?

Curious to hear from validators, relayer operators and protocol engineers working close to these constraints, and to learn whether related work already exists somewhere within the Cosmos ecosystem.

Would be very interested in hearing contradictory opinions as well here.

Part of the reason we opened this discussion is precisely to understand whether this problem is actually perceived as meaningful by people operating close to the infrastructure layer, or whether we may be overestimating its operational importance from outside the ecosystem.

Particularly interested in feedback from relayer operators, validators or protocol engineers on:

  • whether synchronization/recovery workflows are becoming painful in practice,

  • whether the trusted checkpoint assumption is considered acceptable operationally,

  • and whether this direction feels useful at all from a Cosmos perspective.

Also very open to being pointed toward existing work if similar approaches are already being explored elsewhere in the ecosystem.

Thank you !