Dissolution is a proposal type, not an event

Most writing about DAOs covers how bodies are created: a charter is drafted, seats are elected, a budget is approved. Far less covers how they end. Yet a mature DAO constitution usually names the ending explicitly. Optimism's Operating Manual carries a proposal type called Council Dissolution, described as applying when a council “is no longer fulfilling its Charter”: permissionless to propose, active approval, 51% threshold, no veto, a two-week review period and a one-week vote.

Two structural facts follow from that entry, and they shape everything else on this page.

Continuation is the default. The dissolution proposals themselves state the rule: “a persistent Council is expected to continue into the next Season unless a Dissolution proposal is approved.” Nothing expires on its own. A body that has outlived its purpose keeps drawing its budget until somebody spends governance capital to close it, which is why dormant working groups accumulate in DAOs that never write the closing clause.

Dissolution is cheaper than removal. In the same table, dissolving an elected council needs 51%. Removing a director of the Optimism Foundation needs 76%. The community body is the easier thing to take apart, by constitutional design, and that asymmetry is not hidden anywhere: it is two rows of the same published table.

The 2026 Optimism unwind

On 25 June 2026 the Optimism Foundation posted three dissolution proposals in a single voting cycle, targeting the Grants Council, the Milestones and Metrics Council and the Developer Advisory Board. All three share an identical opening and an identical diagnosis: after multiple seasons, “the current structure creates coordination friction without proportional benefit to grantees or the broader ecosystem.”

The Grants Council proposal is the one that states the general principle. It lists the overhead by name – “running elections, onboarding members, operating budgets” – and then draws a line that other DAOs can borrow directly:

“We believe those costs only make sense when there is a strong reason that a function being fulfilled by an independent group of people meaningfully increases decentralization where it matters most (ie. the Security Council.)”

That is a usable test. An elected body is worth its overhead when independence from the core team is the point of the body, and not when the body merely distributes work the core team could do faster. The same proposal discloses that Retro Funding, the Foundation's own programme, was paused on the same cost/benefit reasoning – so the retrenchment is not only aimed at community structures.

The other two proposals show the two ways a body dies. The Milestones and Metrics Council died by dependency: it existed to verify milestones on grants made by the Grants Council, so “without any community-led grant programs, the Charter of the Milestone and Metrics Council no longer serves any purpose.” Kill the upstream body and the downstream one has nothing to do. Delegates noticed the gap this leaves and asked, in the Grants Council thread, who would handle pending milestone verification after dissolution; the answer given was that the second dissolution was a separate proposal in the same cycle, which does not answer the question.

The Developer Advisory Board died by obsolescence argument, and its proposal contains the most striking sentence of the three. One of the Board's charter purposes was to help non-technical delegates understand protocol upgrades. The Foundation's case for closing it: “With rapid advances in artificial intelligence, we believe non-technical delegates are now sufficiently empowered to understand the implications of technical proposals without a dedicated Board required to fulfill this role.” Whatever one makes of it, it is a dated, checkable claim about what a DAO now expects its delegates to do unaided, and it belongs in any honest record of how governance structures were justified away in 2026.

Read the constitution's diff, not the proposal

The proposals describe three closures. The document underneath them describes something larger. On the same day the proposals were posted, the Operating Manual was rewritten as v2.0.0, a commit of 32 additions and 96 deletions, merged three weeks later. Reading that diff is the fastest way to see what a governance unwind actually removes, because the deletions are the parts nobody wrote a proposal about.

The v1.0.1 manual opened: “The Optimism Collective is governed by two houses, the Token House and the Citizens' House.” The v2.0.0 manual opens: “The Optimism Collective is governed by the Token House.” Everything downstream of that sentence went with it:

• The Citizens' House in full – the citizen attestation, the End Users / Apps / Chains stakeholder split, weighted quorum, weighted approval, and every Joint House calculation.
• The stakeholder veto with dynamic thresholds, where two groups needed 17% each to veto, three needed 14%, four needed 11%, escalating to 23/20/17% on a re-submission. It is replaced by a flat 20% veto threshold and a single paragraph of appeals process.
• The Retro Funding section, including the sentence recording that Retro Funding was “currently paused through the end of 2026.”
• Two whole proposal types: Token Allocations (the type the Budget Board proposed under) and Reflection Period.
• The Developer Advisory Board's role in protocol upgrades. Under v1.0.1 an upgrade needed active approval from the Board plus a stakeholder veto window. Under v2.0.0 the Protocol and Governor Upgrades row has no approving party at all: voting type N/A, approval Veto, one week.
• The sentence “Over time, key parameters of the Operating Manual will be maintained by the community (metagovernance.)”

Two renames in that diff are worth more than they look. The proposal type once called Persistent Structure Dissolution is now called Council Dissolution: a general power to wind down any persistent structure, narrowed in name to councils. And under v1.0.1 that power could be routed to either house or to a joint vote, at 51% single-house or 60% joint. With one chamber gone, only the 51% single-house path remains. Separately, on 8 June, the Anticapture Commission's charter had already been deleted from the repository with the commit message “No longer needed”, with no dissolution proposal attached to it at all.

The manual's own change process is the reason all of this can travel in a commit: releases “go into effect as releases are published on GitHub”, and “only the removal of a proposal type requires a governance vote.” Anyone auditing a DAO's structures should therefore diff the constitution on every release rather than track the proposals, and Optimism deserves credit for keeping the constitution in public version control where the diff can be read at all. Most DAOs keep theirs in a document that silently overwrites itself.

The arithmetic of an abstention

The dissolutions were not uncontested, and the shape of the opposition is instructive. L2BEAT's governance team published its reasoning in the Grants Council thread on 15 July: it voted abstain. The rationale is worth quoting because it names the structural trap precisely:

“A community led council cannot function without the backing and funding of the Foundation, and forcing its continuation under these circumstances would only lead to friction.”

An elected body funded by the entity proposing its dissolution has no real veto. Voting the proposal down keeps the body alive on paper while the counterparty that pays for it has publicly said it does not want it. That is the general lesson: a council's independence is bounded by who controls its budget line, which is the same dependency that shapes service provider renewals and delegate incentive programmes.

The abstention also has mechanical consequences that the manual spells out and that are easy to get backwards. Optimism requires 30% quorum, measured against active votable supply, and the manual states plainly that quorum “includes abstain votes.” The approval threshold is computed differently: it is the share of yes votes “relative to the total number of yes/no votes cast”, and it “does not include abstain votes.” So an abstention helps a proposal clear the participation bar it must clear, and is absent from the ratio it must win. Under this rule set, abstaining is not a neutral act with respect to a proposal's passage – it is a contribution to quorum with the opposition removed. Anyone designing a ballot should decide deliberately whether that is what they want abstention to mean; see quorum and threshold design for the wider family of choices.

Overhead was the argument. The surviving body is the expensive one

The dissolutions were argued on cost of overhead. It is worth checking the overhead against the DAO's own budget lines, because the figures point the other way.

As budgeted for Season 7 and reported by SEEDGov, the Collective's governance operating cost was 1,055,000 OP, split across four bodies: Grants Council 400,000, Milestones and Metrics 170,000, Developer Advisory Board 190,000, and the Security Council 295,000. The three bodies dissolved in 2026 are the first three, and together they were 760,000 OP, the cheaper 72% of that structure. The one carved out as worth its overhead is the fourth.

What that carve-out now costs is a separate proposal. The Security Council operating budget for Seasons 10 and 11 requests 5,040,000 OP, covering twelve months from 1 July 2026, for 14 members (13 signers and a lead) at 30,000 OP per member per month. The periods are not directly comparable – one season against twelve months – so this is not a like-for-like multiple, and the token's price moved a long way over the interval. The per-member rate is the cleaner number, and a delegate raised exactly that point in the thread: paying a body in a depreciating governance token makes the token line escalate even when the real compensation does not, which is an argument for denominating in stablecoins rather than for shrinking the council.

None of this makes the dissolutions wrong. It does mean that “reducing governance overhead” is a description of a structural preference, not of a saving: what was removed was the community-elected layer, and what grew was the signer set that executes upgrades. A DAO that wants the cost argument to carry weight should publish the before-and-after in one table, in one currency, over one period. Almost none do, for the reasons set out in DAO governance operating costs.

Where the mandate goes: the operating foundation

Dissolution answers what closes, not who does the work afterwards. In 2026 the common destination is a foundation with staff, and ENS is running the clearest public version of that argument.

A temp check posted on 19 June 2026 proposed establishing the ENS Foundation as an operating foundation with a full-time Executive Director, staff, and a five-seat board ratified by the DAO. Its diagnosis is the sharpest statement of the case for retrenchment anywhere in the 2026 record: the dysfunction comes from “the gap between what the DAO was meant to do (steward credibly neutral infrastructure) and what it actually does day to day (act as a budget committee for an organization).” Token votes are “slow, they carry little context, and they force a yes or no choice”, while running an organisation “needs repeated adjustment, judgment, and a specific person who can be held accountable.”

The executable proposal that followed on 30 July shows what community pressure moved and what it did not. Three scope reductions are itemised in its own abstract: DAO-held ENS tokens stay under tokenholder control, with a single carve-out of 1,000,000 ENS restricted to Foundation employee compensation; the operational wallet (wallet.ensdao.eth, roughly $16M in ETH and stablecoins as of July 2026) stays with the DAO rather than being delegated; and the Endowment Safe (endowment.ensdao.eth, roughly $65M) passes to Foundation Board control but with every transaction routed through a timelock that the Security Council can cancel. Tokenholders keep protocol control and the power to appoint and remove directors.

The objections in that thread are as much a part of the record as the proposal. Delegates argued that board composition and accountability structures were not addressed; that ENS Labs is simultaneously the DAO's largest funding recipient and the holder of enough ENS to represent a majority of active voting power; and that a foundation staffed and funded this way is accountable to the entity that populates it rather than to tokenholders. A competing temp check argued for an independent foundation on accountability grounds, which is a reminder that the choice is rarely DAO-versus-foundation but which foundation, appointed how. The DAO had already priced the interim: pausing its working groups for a four-month retrospective needed a $234,000 continuity budget, 83% of it compensation for the roster.

The pattern across both DAOs is the same three-step: elected bodies are closed on overhead grounds, execution consolidates into a foundation or core team, and tokenholders retain protocol-level control plus the power to remove directors. Whether that is maturation or capture depends entirely on whether the retained powers are usable in practice – which is a question about thresholds, quorum and the distribution of voting power, not about intentions. Compare progressive decentralization, which describes the same institutional arc pointed the other way, and governance minimization, which removes discretion from the protocol rather than relocating it into a company.

A checklist for a live dissolution

If a dissolution proposal appears in a DAO you participate in, these are the questions the published documents can answer:

• Who proposes, and who funds? If the proposer also controls the body's budget line, the vote is a formality and should be read as one.
• Is the mandate transferred or extinguished? A closure that names the successor is an org change. A closure that names nobody is a capability being dropped, and the work it was doing (milestone verification, technical review of upgrades) becomes an unowned gap.
• What happens to unspent budget and to signing keys? Multisig membership, streams, and any permissions the body held outlive the charter unless the proposal revokes them explicitly.
• What obligations are mid-flight? Grants pending verification, appeals in progress, elections already scheduled.
• Does the constitution change in the same window? Diff it. The proposals are the visible part; the manual revision is where a chamber, a veto class or a threshold quietly leaves.
• What does the threshold table say about the reverse move? If closing a community body is 51% and removing a director is 76%, the structure has a direction of travel independent of anyone's intent.
• Is there an exit for members who disagree? In most token DAOs, no – only selling. Where a rage-quit right exists, a restructuring vote is precisely the moment it is for.

How Caper approaches this

Caper has nothing of this shape to dissolve, which is a design choice with its own trade-offs rather than a solution. Searching the contract sources for a delegate, council or committee concept returns nothing: there is no elected body, no seat, no charter and no roster whose overhead has to be justified season by season. The governance surface is a proposal and a ballot, and both are paid for at the point of use – the configured proposal fee and vote fee are asserted in XRD and deposited into the caper's own treasury, so the running cost of governance falls on the participants who use it rather than on a standing budget line.

The consequence that matters here is the remedy available to a member who loses a restructuring argument. Where a token DAO's dissenting delegate can vote no, abstain, or sell, a caper holder can exit: an individual action, requiring no vote and no counterparty's approval, that redeems a member's claim on the treasury and burns their position. That claim is the same canonical vote weight the protocol uses everywhere else, and it consumes both buckets – the governance tokens and the soulbound vote tokens earned by participating – so it is not a pro-rata share of the treasury by balance, and a large position alone does not enlarge it.

What Caper does not offer is the thing the councils were for. There is no independent group standing between a proposal and execution, no technical board reviewing upgrades on a non-expert's behalf, and no elected reviewer of grants. A structure with no bodies never faces a dissolution vote, and it also never had the layer whose loss the Optimism delegates were objecting to. See what a caper is and how DAOs fail.

References

• Council Dissolution Proposal: Dissolve the Grants Council – Optimism Collective forum, 25 June 2026, with the L2BEAT abstention rationale of 15 July.
• Council Dissolution Proposal: Dissolve the Milestones and Metrics Council – Optimism Collective forum, 25 June 2026.
• Council Dissolution Proposal: Dissolve the Developer Advisory Board – Optimism Collective forum, 25 June 2026.
• Optimism Operating Manual v2.0.0 – proposal types, thresholds, quorum definition and change process.
• Operating Manual v1.0.1 → v2.0.0 diff – the commit removing the Citizens' House, stakeholder vetoes, Retro Funding and two proposal types.
• Anticapture Commission charter deletion – repository commit, 8 June 2026.
• Optimism Security Council Operating Budget for Seasons 10 and 11 – 5,040,000 OP for 12 months, 14 members at 30,000 OP per month.
• Token House participation and incentives: Season 7 – SEEDGov, the per-council operating budgets used above.
• Guide to Season 9 – “From Experiment to Organization”, the season framing the dissolutions cite.
• [Temp Check] Next Era of ENS DAO: Empowering the ENS Foundation – ENS forum, 19 June 2026.
• [Draft][Executable] Next Era of ENS DAO: Empowering the ENS Foundation – ENS forum, 30 July 2026, with delegate objections.
• [Temp Check] Empowering an Independent ENS Foundation for Accountability – the competing proposal, 24 June 2026.