Remove ICS from the Cosmos Hub (Software Upgrades to Follow)

ATOM Economic Zone
1

This forum post is meant to notify the community of two upcoming software upgrade proposals that will impact the operation of Interchain Security on the Cosmos Hub. The first proposal will prevent the creation of new consumer chains, and the second proposal will remove the ICS module from the Gaia codebase.

Background

Interchain Security launched in May 2023 to significant fanfare within Cosmos. With two high-value chains (Neutron and Stride) publicly precommitted to use ICS at launch, the feature promised high revenues for ATOM stakers and a path to sustainability for the Cosmos Hub and ATOM, as well as a way to reduce security costs for consumer chains.

Despite the initial excitement, ICS struggled to find product market fit. Few additional chains expressed an interest in launching consumer chains on the Hub. The consumer chains that did launch failed to provide sufficient revenue to the Hub to offset the cost of validating them. Even after the launch of Partial Set Security (PSS) allowed for permissionless deployment and opt-in validation models, both validators and prospective consumer chains viewed the product as more of a burden than a benefit. Eventually, consumer chains began to leave, beginning with Neutron in March of last year.

Today, there is no meaningful demand for new consumer chain creation. Most companies building on the Cosmos SDK do so because they do not want the third-party security assumptions incumbent in shared security solutions like ICS. Those that can tolerate those security assumptions prefer alternative shared security models like Ethereum or Solana. In addition, most potential consumer chains don’t have the scale, budget, or impact to drive meaningful value accrual to ATOM via a shared security model.

Meanwhile, ICS imposes significant technical and operational debt on the Cosmos Hub for all parties involved:

  • It is expensive to operate, since validators must run additional infrastructure for each chain they support, for only a fraction of the full staking rewards they would receive from a sovereign chain

  • Validators take on additional risk for equivocation on consumer chains with minimal associated benefit.

  • ICS makes it extremely difficult to make needed performance improvements to the Cosmos Hub because of the deep dependency it has on CometBFT internals. These performance improvements will be needed to bring new features and functionality to the Cosmos Hub in 2026.

  • Incident responses on consumer chains use up valuable developer time and resources that could be spent on further optimizing the Cosmos SDK, adding new features to the Hub, or growing enterprise adoption of the Cosmos Ecosystem

The drawbacks of keeping ICS on the Cosmos Hub are clear. The benefits are not. In light of this, ICS should be deprecated.

ICS Deprecation Plan

ICS deprecation will proceed in two phases, with each phase occurring over its own independent software upgrade proposal that will be voted on by the community:

  • Phase 1: Disable new consumer chain creation

  • Phase 2: Disable ICS Module

Phase 1 requires the integration of an antehandler to block new consumer chain creation. This will be implemented as part of the v27.0.0 software upgrade proposal, expected to go on-chain in mid-March.

Phase 2 will remove the ICS module itself from the Cosmos Hub via the following actions:

  • Remove provider code or replace with staking module

  • Update IBC router wiring

  • Add module removal and store deletions in upgrade handler

  • Remove interchain-security from go.mod

  • Test app execution

Phase 2 will proceed after all existing consumer chains have offboarded, provided that they do so in a reasonable timeframe. Hypha and Cosmos Labs are currently assisting Stride and Elys in the offboarding process. All other consumer chains have already been offboarded.

Final Thoughts

Interchain security was a worthwhile experiment for the Hub, and launched at a time where shared security was a strong value proposition for highly decentralized, well capitalized chains like the Cosmos Hub. However, the future direction of the space is changing, and the Hub should modernize with it. These software upgrades will help improve Hub performance, lower cost of validation for opted-in validators, and help prepare the Hub for a role as a high-performance interchain service provider for the next generation of Cosmos chains.

12 Likes