[LAST CALL] CHIPs signaling phase : Validator Vote Power Cap

Cosmos Hub Improvement Proposal (CHIP): Validator Vote Power Cap Summary This proposal introduces a vote power cap for validators in governance . The mechanism would ensure that no validator can unilaterally cast more than a fixed share of delegated voting power (e.g. 3%), while preserving both: • The autonomy of delegators , whose votes are never reduced or capped if cast directly. • The self-delegated stake of validators , which remains fully effective and is not subject to the cap. This aims to mitigate governance centralization risks while maintaining the security and economic integrity of the staking system. Motivation The Cosmos Hub governance process inherits validator voting power from the staking module. Today, large validators may control substantial fractions of the voting power, particularly if delegators do not vote directly. This concentration creates several issues: • Governance centralization : A handful of validators can disproportionately influence outcomes. • Reduced delegator sovereignty : Passive delegators effectively cede control to their validator, potentially against their intended interest. • Misalignment of roles : Validators…

Excerpt (1198 of 6578 characters). Read the whole post on the forum ↗

Thank you for this proposal aimed at greater decentralization of governance. The core issue, however, remains the same as the one you have already highlighted: validator duplication. Whatever method is used to improve decentralization or make governance more equitable, this risk inevitably persists. Personally, I’m not certain what the best path is to strengthen the Hub’s decentralization. It’s ultimately a political question: should we directly limit the voting power of validators who hold too much influence? Or should we introduce a Nakamoto bonus to let stakers themselves arbitrate decentralization, as @Guinch_Roze suggest here ? Either way, the underlying structural problem is the same validators may choose to duplicate themselves. Personally, I believe it’s time to end the completely open entry to the validator set by freezing the current list and preventing new competitors. Existing validators should gradually be required to meet stricter formal obligations to professionalize their operations and make them fully accountable, with clear duties mandated by governance. Failure to comply should trigger sanctions, including potential removal and replacement by a more…

Excerpt (1191 of 2089 characters). Read the whole post on the forum ↗

Quentin: The core issue, however, remains the same as the one you have already highlighted: validator duplication. We fully agree with this point and would add that custodial entities could potentially employ other bypass mechanisms as well. As we noted in our original post: Govmos: Additionally, it is worth exploring further mitigation measures to prevent custodial entities tempted to bypass this cap. It should be stated that custodians cannot self-delegate the assets of their clients, as well as casting direct votes on these accounts. The Core Challenge The real difficulty lies in identifying major threats and designing remedies that are both effective and resource-efficient. On-chain detection of fraud — such as validator duplication, bypassing caps through self-bonding, or manipulating direct votes — is technically infeasible. A Path Forward These issues are far easier to address on a human and legal level rather than purely through technical measures. We believe solutions should be anchored in legal settlement mechanisms , with governance acting as the coordinating layer. One possible approach would be to allocate the slashing penalty to…

Excerpt (1197 of 3242 characters). Read the whole post on the forum ↗

I fully understand and respect your point of view. That said, I advocate a different political approach that seems more viable and secure, both in the short and long term. The issue with your model is that it relies solely on limiting governance power without directly addressing voting power . As a result, an operator like Coinbase would still sign the same number of blocks its governance weight might decrease, but its influence on the network would remain unchanged. Another challenge, which you correctly mention, is that most of these entities provide custody services. Capping their delegated voting power at 2 % would not prevent them from accumulating substantial governance power . Finally, validator duplication would still be possible albeit less attractive because duplication would only affect governance power, not voting power or APR. In your model, the safeguard is essentially a constitution, a framework enforced by humans, with oversight committees and slashing penalties. While I agree on the importance of such a constitution, I don’t believe the Hub will adopt one until we have a clear PMF. This could take several years, during which we would have to rely on…

Excerpt (1195 of 3625 characters). Read the whole post on the forum ↗

We appreciate your thoughtful input and would like to clarify the scope and priorities of our proposal. Scope of the Proposal As you noted, our proposal focuses exclusively on governance power without altering voting power itself. Broader changes — such as block production or staking reward distribution — would have significant downstream consequences and must be addressed separately in a step-by-step sequence . We see governance decentralization as the first and most urgent step , ensuring that the decision-making process remains as decentralized as possible. The Governance Risk At present, large custodial validators and centralized exchanges rarely cast votes. However, the potential risk is clear: they could decide to exercise this power at any time, potentially altering governance outcomes or vetoing proposals outright. Our aim is to ensure that legitimate, self-custodied token holders remain the primary voice in governance, while diluting the disproportionate influence of custodial entities that do not allow clients to vote directly. This is about being proactive — addressing the problem before it materializes, not after. Why Not Quadratic Voting We also want to…

Excerpt (1196 of 3139 characters). Read the whole post on the forum ↗

I would emphasize the following key indicators:

voting participation (let's say at least 50%)

validator commission no higher than 20%

number of delegators (at least 500 delegators for every 100,000 Atoms)

If any of the parameters fail to meet the requirements, the validator is removed from the active set.

We would like to invite all community members to share their feedback on this proposal in the coming week. Following the standard 14-day governance discussion period on this forum, we intend to move the proposal on-chain.

If no major revisions are suggested, our current timeline would place the on-chain vote around mid-October.

With all respect, but there’s more traffic in Tg chats or X. If you want more feedback I’d invite you to tg chats..

This is an interesting idea. I have some questions though, from the practical standpoint.

• All validators would see their governance VP capped to 2% of their actual VP?
  If so, doesn’t it simply displace the issue? 2% of 2M is still wildly superior to 2% of 500K and if all the biggest validators vote with 2% of their VP it will remain more than 2% VP of the rest of the set.
• We can maybe expect that the votes of the delegators will nudge the outcome a bit, but their participation is traditionally low and it is unclear that this new rule would really drive a significant fraction of them to vote individually.
• And therefore: what about the quorum? If the number of individual votes does not massively increase, don’t we risk seeing proposals fail because we can’t get 40% of the overall VP to vote?
  Among the last proposals, most just barely reached the quorum, and a few actually did not.

(I read the PR diff in the linked gdoc but it doesn’t look like there is a planned workaround to this, although admittedly i’m not skilled at go and went through it quickly so I might have missed something.)

highstakes: All validators would see their governance VP capped to 2% of their actual VP? That cap is global , not relative to each validator’s own stake. So if the Hub has 100 million ATOM bonded, and `VotePowerCap = 0.03` (3%), then: • `capTokensInt = 3,000,000 ATOM`. • That value is the maximum amount of defaulted voting power that any single validator can wield . • It applies equally to every validator, regardless of how large or small their delegation set is. How it works in practice For each validator in the tally loop: • Calculate their `delegatorDefault` (tokens from delegators who didn’t vote). • Compare against the global cap (`capTokensInt`). • If `delegatorDefault <= capTokensInt` → all of it is counted. • If `delegatorDefault > capTokensInt` → only `capTokensInt` is counted. Self-delegations are added on top of this capped value, since they are explicitly excluded from the cap. highstakes: We can maybe expect that the votes of the delegators will nudge the outcome a bit, but their participation is traditionally low We agree with this point. However, we believe the risk may very well materialize at some point,…

Excerpt (1197 of 2378 characters). Read the whole post on the forum ↗

Got it, thanks for clarifying! I had indeed overlooked the “global cap” part.

Sounds good.

Imo, this proposal needs a tweak to ensure it does not significantly impact the ability to pass proposals due to a failure of quorum. Looking at the top 12 validators (the ones who will be impacted by this proposal), 5 are very active voters. Collectively, this proposal would cap their total quorum contribution at 10%, removing 7.34% of total vote power contributing to quorum. Looking at the last 10 gov proposals on the Hub, only 1 passed by more than this margin, and 4 of them were rejected due to a failure to reach quorum. This proposal could make it such that no future gov proposals are able to pass. Instead, I’d propose applying the VP cap to proposal results only, and not to quorum requirements. More broadly, however, I think this proposal is premature. As Mag has mentioned, Cosmos Labs is on the verge of making a number of significant announcements that could impact the course of the Hub’s / ATOM’s future. Without clarity on those announcements, it’s hard to know what might break with this proposal. Given there’s no real urgency in getting this passed, I’d propose waiting. There are also a number of other ways to solve the problem of governance centralization. One…

Excerpt (1198 of 1302 characters). Read the whole post on the forum ↗

Thank you for raising these valuable points. On quorum risks You are absolutely correct to note that this proposal could potentially affect quorum under certain circumstances. However, it is important to emphasize that the proposal was intentionally designed with adjustable parameters . A simple way to mitigate this concern would be to increase the initial cap parameter from 2% to 3% . This would ensure that the introduction of the cap does not disturb quorum while allowing us to later reduce it, once we have confirmed that quorum remains stable. On applying the cap only to results While this idea may seem appealing at first glance, it would introduce additional complexity in the codebase and create dependencies tied to variable quorum thresholds. Our goal is to keep the mechanism as simple, transparent, and self-contained as possible within the governance module. On the timing of this proposal At Govmos, we firmly believe that innovation should not rely solely on foundation-led initiatives . The Cosmos Hub’s strength lies precisely in its decentralized governance — a principle we must continue to embody. This proposal establishes the foundation for a…

Excerpt (1193 of 2689 characters). Read the whole post on the forum ↗

Govmos: This proposal establishes the foundation for a resilient governance framework , one that protects against the growing risks of delegation-based centralization . This is not only a matter of decentralization ethics but also of regulatory prudence. With the upcoming Clarity Act in the U.S. — which stipulates that no group of entities should control more than 20% of a network to qualify as a commodity — it becomes clear that we need such safeguards in place. Thank you for this response but, unfortunately, this proposed solution won’t have meaningful bearing on whether or not the Hub is sufficiently decentralized to qualify for commodity status under the clarity act. Regardless of what happens with on-chain governance, a supermajority of consensus power can alter chain state by forking the chain at any time. This is what will matter for the purposes of the clarity act. Adjusting on-chain vote power doesn’t impact consensus power in a dpos system. The proposal really changes governance in name only. This is why I’m urging a bit of patience here. At the end of the day, this proposal won’t have immediate positive impact but may have an outsized negative impact.…

Excerpt (1193 of 2496 characters). Read the whole post on the forum ↗

On the relevance of the Clarity Act to this proposal, we must respectfully disagree with your statement. Here is a direct excerpt from the Clarity Act : “No person or group of persons under common control— (i) has the unilateral authority, directly or indirectly, through any contract, arrangement, understanding, relationship, or otherwise, to control or materially alter the functionality, operation, or rules of consensus or agreement of the blockchain system or its related digital commodity; or (ii) has the unilateral authority to direct the voting, in the aggregate, of 20 percent or more of the outstanding voting power of such blockchain system by means of a related digital commodity, nodes or validators, a decentralized governance system, or otherwise, in a blockchain system which can be altered by a voting system.” Source: Clarity Act — H.R.3633 (119th Congress) This section clearly refers to outstanding voting power , not to asset ownership. Many in the ecosystem have mistakenly interpreted this as a 20% ownership threshold, whereas the law explicitly concerns the aggregate votepower —regardless of how it is acquired. In practice, this means that any…

Excerpt (1191 of 2973 characters). Read the whole post on the forum ↗

Respectfully the text you’ve quoted proves my point. Particularly the following:

Govmos:

has the unilateral authority to direct the voting, in the aggregate, of 20 percent or more of the outstanding voting power of such blockchain system by means of a related digital commodity, nodes or validators, a decentralized governance system, or otherwise,

Onchain governance is interpreted quite broadly here, and most people familiar with governance outside of Cosmos (e.g., how Ethereum governance works) know that this is inclusive of node upgradeability to a designated binary. We’re not as used to that in Cosmos, but it’s absolutely applicable here.

Again, I respect your right to put this on chain, but if you intend to do so I would strongly recommend setting a higher initial cap of at least 5%, which has the same effect of preventing a single party’s on-chain vote power from exceeding 20% while being far less likely to have unintended negative impacts on the overall governance of the chain.

We hope you don’t mind us respectfully challenging your points — after all, that’s precisely the purpose of these discussions: to exchange perspectives and foster constructive debate on complex or contentious topics. RoboMcGobo: Onchain governance is interpreted quite broadly here, and most people familiar with governance outside of Cosmos (e.g., how Ethereum governance works) know that this is inclusive of node upgradeability to a designated binary. We’re not as used to that in Cosmos, but it’s absolutely applicable here. We understand the comparison, but we believe this interpretation cannot be directly applied to Cosmos chains for structural reasons. In the Cosmos framework, a validator failing to upgrade the chain — even with a 20% of VP — would not constitute a governance or consensus failure. With 33% VP entities would only prevent block production. For any meaningful fork or divergence from the designated binary, the validator (or group of validators) would need to control two-thirds of the consensus . This is a much higher threshold, inherent to the BFT consensus model that underpins Cosmos chains. Furthermore, each software upgrade in Cosmos is…

Excerpt (1193 of 1914 characters). Read the whole post on the forum ↗

Thank you for this write up, we learnt alot while reading it.

Any update on this topic ?