[Temp Check / Pre-Proposal] Pocket Network’s New Governance Model

Proposals Pre-Proposals
1

The Pocket Network Foundation is proposing a new governance model and we’d like to gather feedback from the community before moving forward.

This post outlines the core principles, how the model works, and the safeguards in place.

Why a New Model?

Our goal is to design a governance system that is secure, fair, and representative. To guide this, we’ve built around four “North Stars”:

  • Use Existing Tools: Rely on proven systems, not custom complexity.
  • Captureless and Secure: Make hostile takeovers prohibitively expensive.
  • Anti-Plutocracy: Limit the ability of the wealthiest to dominate governance.
  • Represent All Stakeholders: Ensure builders, node runners, stakers, and community members all have a voice.

How the Model Works?

Validators as Delegates

  • Validators secure the network and vote on governance proposals.
  • Token holders delegate their stake to validators who best represent their views.
  • Delegation is flexible: you can change anytime.
  • Choosing validators as delegates significantly helps with sybil resistance and reduces which is curtail to our design due to considerations we have for anti-plutocracy

Quadratic Voting

Instead of “one token = one vote,” voting power grows with the square root of stake. This reduction of power will happen at the validator level.

  • 1M tokens = 1,000 votes
  • 16M tokens = 4,000 votes

This makes governance harder to capture by large holders and ensures broader participation.

Creating a proposal

To create a proposal a minimum amount of tokens must be staked, this ensures that proposal has a minimum support to begin with.

Community Caucuses

A caucus is a structured group within the community that represents a specific type of stakeholder. For example, builders, node runners, or stakers.

Each caucus is tied to a dedicated validator slot, operated by PNF, that casts votes on behalf of that group. Here’s how it works:

  • Membership
    • Once a year, PNF identifies required caucuses then community members can nominate themselves to join a caucus.
    • Validators vote on nominees. Those who reach quorum are added as caucus members.
  • Decision-Making
    • Within each caucus, decisions are made on a one-person-one-vote basis.
    • Members use a simple, familiar tool like Snapshot with wallet-based voting.
    • The outcome of these internal votes determines how the caucus validator votes on chain.
  • Support for Proposals
    • Proposers can approach a caucus to review and refine their ideas.
    • If the caucus supports it, they can back the proposal with their validator’s weight, signaling credibility to the wider community.
  • Incentives
    • Each caucus has a certain amount of incentive to compensate the members which will be distributed equally among the members of that caucus.
    • Less crowded caucuses mean higher rewards per member, balancing representation across groups.

In short, caucuses make sure that different types of contributors to Pocket have formal, recognized voices in governance without the need for running a validator or delegating tokens.

How secure is it?

  • Attacks require massive capital. A new validator must have enough stake to knock out an existing validator and get into a voting spot.
  • PNF at the moment operates nine validators and holds treasury tokens that can act as a safety brake.
  • An attack on this governance structure will be a very visible and noisy attack
  • Quadratic voting coupled with validators burden for breaching sybil resistance further reduces capture risk.
  • The system is designed to block malicious exploitation, though bad ideas could still pass if they have enough support (that’s democracy).
  • Here’s a simulator for further assessing how the model works in different situations: https://quadratic-voting-lab.lovable.app/

What is the 0 to 1 roadmap?

  1. Quadratic Voting: Modify Cosmos SDK governance → square root voting at validator level, integrate with proposal flow, build indexer for transparency, pilot test.
  2. Caucus System: CosmWasm multisig with supervisor key → manages validator + caucus membership; members vote via multisig (1 NFT = 1 vote).
  3. Governance UI/UX: Keplr + Mintscan dashboards, live governance feed on explorer, caucus voting support, quadratic voting simulator.
  4. Discussions Infra: New Discourse, migrate old threads, set up governance structure + caucuses.
  5. Caucus Deployment: Define caucuses, run nominations + validator vote, issue membership NFTs, integrate Snapshot for caucus voting.

More information is available here

FAQ

We put together a list of questions and answers that might further help you understand the model. Please check them here.

Call to Action

This post is a temperature check. We understand there might be still existing questions around what/how many caucuses we will have, what is the quorum threshold or minimum stake for creating a proposal.

However, we’d like the community’s feedback on the core of the governance. We want to hear from you:

  • Do you agree with the four North Stars?
  • Does the validator + quadratic voting system seem fair and secure?
  • Do caucuses make sense as a representation mechanism?
  • What vectors of attack do you see that is not addressed?
  • What risks or challenges do you see in the scope of work?

Please share your thoughts in the comments. Your feedback will guide the next iteration of this governance proposal.

References

Governance Committee

This work has been produced by the governance committee. Special thanks to everyone involved.

4 Likes
2

Thanks for the detailed proposal, it does show throughout research, and also sorry for the late response…

I’ll firs answer the questions directly:

Yes, they are indeed what I wish for the DAO governance.

Yes, but…
The devil is in the details, more below…

They do, and they are an interesting use of the DAO validators.

DAO / PNF “break” or stopping mechanism sees a little overconfident, but the base security is high anyways.
Shifting of voting power without economic means could be achieved (attacking the validators hardware, DDoS before voting events)

A crosstalk between voting power and economic returns might appear as validators set the rewards they give to their delegated stakers.
The definition of the caucuses allow the PNF to take control of their voting power arbitrary, this undermines their effective representation.

The Long Answer

(oh I know you were not expecting a short one…)

As I understand this, the validators will be fighting for governance power based on tokens. There will be only a set amount of validator slots, that will be low according to the FAQ, and based on pure stake validators will fight for them.
From these validator slots the DAO will hold (or try to hold) a number of them to be used as the caucuses validators/vote slots. The economic pressure this will have in the DAO treasury is unknown, but currently is manageable.

The free-validators

The “free” validators, those not attached to a caucus, will have a double play:

  • Generate revenue for their stake and any delegated tokens.
  • Vote according to their convictions and represent all delegated token owners.

This creates a cross-talk, where one user might not be politically active but the act of delegating their tokens to a validator results in a political action. This is OK, as every action is political some say…
Some validators could be apolitical themselves, abstaining from voting (they are here for the moniz), and they can hold lots of voting power, specially with a small validator pool, which leads to my first question:

  1. How does the abstaining of a validator affects the quorum and/or general voting?

The other case could also be true, a validator that shares 100% of its rewards to lure voting power from token delegators that are apolitical (or greedy depending on how you interpret their actions), inflating their voting power based on unaccounted economic power (the money they spent for keeping the hardware running). So, and this is more a context question than a problem:

  1. Will there be rules on validators bad behavior (populism)?

The DAO Validators / Caucuses

The DAO, through PNF, will hold some validator slots, which will be used to create caucuses. These will have several uses, (I hope on of them will take into account Servicers stake), that we are not discussing here and that’s fine. I just need clarification on this:

  1. Caucuses, in the eyes of the voting system, are validators just like the “free” validators described before?

If the voting system will treat these validators just as “free” validators, meaning that the DAO will need to be a strong economical player in the validation landscape (and I will assume this for the rest of this text). While it is arguable that “free” validators also represent the community, there will be a clashing of economical firepower between the DAO, politically active members and purely capitalist members (just like everywhere on earth I think… were you would like to live at least).
Achieving a balance, with a honest DAO and the current economy should be easy, now, how this will delve into the future is very complex, more if the POKT token reaches success. If we drain the DAO treasury (again), we will not only lose economic sustainability but also lose grip of governance. The stakes are high so we, the DAO, need to be smart (something difficult in politics).

Moving on, caucuses’ members will be of an specific “kind”. It seems that is expected that caucuses will be specialized in a matter and that their member will act as “knowledgeable” people in that matter. If this is the case (just seeking clarification):

  1. How will the competence of the caucus members will be assessed? A popularity contest, a.k.a. voting?

Also, they will be paid for their work, so:

  1. Where does the tokens for the caucus member payment comes from? DAO Treasury? caucus validator earnings?

Finally, the caucus vote is decided using a egalitarian vote among their members, BUT the final vote of the caucus is controlled by PNF, this leads to another question:

  1. wtf?

I like the system of representatives, it gives power to a different kind of community member, but:

  1. how is it justified that PNF (probably by only the director’s rule) can take hold of all the caucuses votes, when the caucuses votes emerge from a very decentralized voting scheme? in which scenario the caucuses are the rouge element that need to be overruled?

On the Economic Game

While currently the DAO is very strong in POKT tokens, this could be an spurious state. The actual hold of the DAO on the supply, with liquid tokens (immediate action) is actually around ~14 M tokens, the rest is staked, I’m not sure where the 500M number comes from (not doubting this exists) but it probably is not liquid. Doing math with the gross amount of DAO tokens seems wrong, as the actual strength will be lower.
I do not dare to make any projections here, since now validators will not only be an economical game but also a political one, so any modelling is futile. I cannot overstate how delicate the managing of the validators will become in case the POKT token succeeds, and we must use the “peace” (or poverty) time we have now to setup the tools to be on top of this at every block…

A Note on DDoS

I’m not versed in this, but:

  1. what would happen if before a voting event some validators are brought down or jailed, either maliciously or due to negligence of the operator?
    Can they loose their spot, shifting the voting power to other potential validators with lower stake?

If this is possible, this is an attack vector that could be much cheaper and silent than taking over the network.

2 Likes
3

I appreciate the thought that went into your reply!

I know different people will have different thoughts around the caucus system, but given that I was the one who thought that up, figured I’d add my thoughts. Those validators will already have the f-chains stake to start with, so they’ll be “preloaded” in favor of these community groups, and can also be delegated to. The Foundation as custodian would execute that validator vote in favor of their collective decision.

But, that doesn’t mean the Foundation has to physically hold the voting power. I’m open to structures that allow for a multisig or other form of decentralized voting for that caucus.

4

That’s a good question to make sure we’ve got a good answer for. The default voting window is 14 days with Cosmos, so I would think that’s enough time to deal with an issue like that, but it’s the kind of scenario that we should be able to account for.

5

Thank you @RawthiL, I truly appreciate your response and attention.
I’m going to just dive right into answering your questions.

Since we plan to leverage the default (x/gov) as much as possible. Abstaining is still considered as part of the vote counted for the quorum.
Counting abstentions in quorum has upsides and downsides. It makes participation more inclusive, discourages manipulation, and signals legitimacy through higher turnout. But it can also lead to deadlock if many abstain and encourage “lazy voting,” where members avoid taking a stance.
We can implement a different logic if we decide to change the default setting, and I also do prefer having different types of voting but for now the committee opted for an MVP rather than building out features like this. However, if the community requires we can certainly extend the scope.

While you are right, I believe that the supply side is always more profitable than the validator side, so truly greedy delegators should opt for the supplier side rather than the validator side. However, we did consider that there might be uninformed greedy delegators who simply delegate to this side, but their population should not break the system.
Regarding bad behavior (populism), we currently do not have such rules and are not planning to, as this would also create an incentive for validators to compete. Higher returns will also encourage more stake and less circulating supply, which are all desired outcomes from my perspective.

Yes, in the eyes of the voting system a caucus is just a validator same as others. The only difference is that its vote is defined by a group in a transparent and public manner rather than an individual or a private group.

Your remark on free validators is true, but at the end of the day the power is there to be given and if a truly distributed economy shapes the system should work efficiently. However, as you mentioned it is extremely important how we treat the funds and circulating tokens.

True. The assumption here is that validators will pick the right people for caucuses however, much like any governance, you can be incompetent and popular at the same time.

Since the caucuses are operated by PNF and they are managing PNF validators, the compensation will also come from the DAO Treasury. Free validators can also create the same scheme but they are in charge of creating their caucus and incentivize participation.
Additionally, to ensure funds are being put to good use. Instead of promising an upfront payment we can retroactively compensate caucus members based on their performance and engagement.

Apologies for the confusion. I was trying to convay that, the final vote reflects the caucus decision and is not controlled or influenced by PNF. However, since the validator is run and managed by PNF, they ultimately have the technical ability to stop the vote if necessary.

On the Economic Game

I share the same sentiment. Would you suggest that we expand the scope of work a little bit right now and include the setup/building the monitoring tools now?
Initially, I was considering to improve through small iterations but due the sensitivity of the matter it might not be a bad idea to expand, and include the monitoring in this phase.

Unfortunately, the default x/gov module uses the current state of the network instead of an snapshot of the network at the time the proposal was published hence what you mentioned is actually a risk but this is a risk that other ecosystems who are using default x/gov are barring.
To fix it the best solution is to implement the snapshot but for short term mitigation we can have a higher quorum requirement and harsher penalties for down time.
Again, would you prefer expanding the scope and implement this?

Hope I had been able to answer the questions, please feel free to ask more.

1 Like
6

This is an important and sweeping change, and I’d rather take the time to get it right, so I’m open to extending the scope of implementation on stage 1 if needed.

1 Like
7

Thanks for the answers!

Yes, this is true, but it should be just that. The existence of the mechanism is unavoidable but it should be only used on extreme situations. The text seems to take this too lightly.

I think that we need to advance and adapt here, no need to overbuild at this point.

Currently PNF is having more ordered finances and it can protect spending and treasury lockups for staking. The part that I don’t know is how is the validators staking monitoring, we should have visibility on staking movements and maybe some alarms set for drastic movements on the validator set. Nothing too fancy.
However none of this will provide any security, only the ability to react, the real firepower is the liquid tokens that PNF controls.

How does this work? allowing a snapshot of validators to vote only, the ones present at voting start?

Also, how much work are we talking about here? seems like big change.

8

Kudos to the Governance Committee for this well-thought-out initiative.

I have some questions. Here are a few :

Meaning of Quorum

What is meant by “quorom”?

Is quorum the % of total eligible voting power that must participate in order for the vote to be valid, with voting power calculated at the end of the voting window based on the sum of all possible quadratic votes?

Regarding quorum, does abstention consume the square-root-weight of that validator’s staked tokens?

Would quorum be the same across all vote types? What percentage is being considered for the quorum?

What Validators?

The FAQ says: “Which validators can vote? Only validators in the verified slots can vote. Currently, there are 15 slots, and for a new validator to gain voting access, it must replace an existing one by staking more power.”

PNF’s 9 validators take up 9 verified slots, so only 6 are up for grabs?

The snapshot to determine the six most highly staked “free validators” should be taken at the outset of voting so that community members know which validators are in play and with whom they should delegate their stakes. This snapshot would be separate from the one that determines voting results.

Validator Voting

Is the idea that each validator will apply all his votes to his/her position on a proposal (for, against or abstain)? Or is split voting allowed (e.g. 60% for, 40% against).

And what’s to stop a malicious “free validator” from voting differently on a proposal from what it publicly announced to attract delegated stakes to boost its voting power?

Delegating to Validators for Voting

Assuming verified validator set is known when voting begins, community members will delegate to the validator whose stated position on any vote coincides with that of the community member. However, won’t a longer voting window than 14 days be needed to enable community members to un-stake from suppliers and re-stake with validators for voting?

Caucus Questions

Don’t we know already that there are only 4 caucuses: node runners, builders, stakers, other community members? Doesn’t it therefore make sense to create caucuses as soon as the new governance system is implemented and if there’s a need for a new caucus, PNF can add at it any time. (As to caucus membership, this should be subject to periodic voting.)

Don’t you mean vote “against” the proposal, not “stop” the vote?

Proposal Formulation

From the FAQ:

“How do caucuses serve as a replacement for the old pre-proposal system? A person seeking a grant or proposing an idea can approach the relevant caucus first. If the caucus, which consists of people deeply involved in that specific domain, agrees the proposal is valid, they will work with the proposer to refine it. The caucus can then put its weight behind the proposal, acting as a qualified reviewer and increasing its chances of success.

“How do members of caucuses determine the vote of their validator?

Each caucus will have a dedicated forum for discussion where members can post topics and debate proposals. This will lead to a vote to determine the caucus’s official position.“

Don’t we want to enable the community at large to give feedback on gelling proposals? While input from those “deeply involved in that specific domain” is perhaps most valuable, other community members can offer important insights too. So non-caucus members, including those who self-nominated but were not elected, should have access to the dedicated forum.

NOTE: I have read the preproposal and the FAQ but not the other documents listed under References.

2 Likes
9

Thanks @RawthiL, I’ve shared some answered previously in Discord but for consistency I will reiterate here.

For sure. We definitely need to have a lawyer or someone expert in writing to rewrite the actual constitution for the governance. The previously shared version was an early draft for community temp check.

As shared in Discord, we’ve considered it in the proposal for developing the governance system. This is an expansion on the initially shared scope which we can go for now or postpone to the next development phase.

The detailed technical specification is above my pay grade but tldr is that we will take a snapshot of the state of the network when proposal is posted and the vote will be based on the snapshot not current state of network. So for example if a node is jailed after the proposal is posted, it is still able to vote. For details check the proposal.

10

Thank you for the feedback and attention.

Yes, to reiterate, quorum is the minimum percentage of total voting power (sum of square root of each validators staked amount) that must participate in a vote for it to be valid.

I’m not sure what do you mean by “consume” but abstain will be considered towards quorum. Regardless of the vote (for, against or abstain), the square root of the staked amount will be considered as validators voting power.
If by “consume”, you mean “spend” or “burn” the stake, no the staked amount will be intact.

Yes it will be the same for all proposals. The cosmos default standard is 40% but it is up to the community. What is your suggestion?

Not necessarily. At the moment PNF has 9 validators and they are in within the top 15 slots. However, anyone can create a new validator and with enough stake they can knock out other validators and get a spot in the top 15.

The snapshot is not actually something separate, we will just consider the specific block number in which the proposal was published instead of using the latest block number.
Delegetors don’t need to move their delegation, this mechanism was considered to prevent post proposal manipulation of the network.

There’s no split voting, validators can choose only one of the available options.

By delegating to a validator you are entrusting them with your vote, and they are entitled to make the decision. Now if they decide to act maliciously we can’t stop them. While we’re prepared to prevent them and keep the treasury safe, we can’t directly influence their vote.

Voting validators are always known as they are the top 15 in terms of staked amount. The mechanism at the moment doesn’t consider this. It assumes that governance participants will keep a certain amount locked in validators to participate. Our assumption is that most of the ecosystem members will keep the bigger portion of their tokens in the supplier side and stake a small amount in validators to have exposure to governance.
Additionally, with snapshot voting even if there’s enough time to move from supplier to validator, the system will not recognize the newly added voting power and will consider the state of the chain at the time proposal has been posted.

If other community members are happy with the four pillars you mentioned we can go ahead with them. We left it open so the community can decide on it.

Yes that is a correct interpretation. However, depending on situation it can be for or against.

The discussions of the caucus members are public and anyone from the community is welcome to participate. However, they are the final decision makers for the validator in their control.

Thank you for the detailed feedback, I hope I was able to address all of the questions and concerns.

11

Mehrdad, thanks for your answers.

Quorum

That seems reasonable. (But this does not account for the issue raised by Ramiro on validators being jailed or down at the moment of snapshot in regard to which you suggest a higher quorum for short term mitigation.)

Caucus Validators

The caucuses created by the community presumably including for stakers, builders and node runners, will each match to a PNF validator. Since validators with enough stake can knock other validators out of the top 15, doesn’t this mean that PNF will need to be able to monitor and maintain sufficient stake to ensure its caucus validators are among the top 15 upon proposal publication?

Your FAQ (above) says “The foundation’s nine validators act as a safety brake. If a malicious proposal were to pass, the foundation could dump its treasury into these validators, increasing their weight and providing a buffer to stop a malicious vote.”

Since PNF validators can be knocked out of the “Group of 15” voting pool, doesn’t “these validators” necessarily refer only to one or all of the caucus validators (assuming PNF ensures these stay in the pool)?

Proposal Refinement

A proposer can approach all the caucuses at the same time and have each refine his/her ideas, correct?

Support for Proposals

The caucus would have to show support by announcing that it will back the proposal once it’s published. Correct? While it could show backing for the proposal with its actual vote - that may be too late to have any influence depending on when its validator votes and whether the other validators have already voted.

Staking with Validators

Since voting power is fixed at the moment of proposal publication, community members/stakers would need to know “free” validators’ views far enough in advance to have time to delegate or unstake their stakes accordingly. Correct?