Comparative
Every governance token eventually faces one question. What happens to holders when the organization around the token changes shape, by winding down, by being bought, by failing, or by deciding it no longer wants to be governed? Five such endings occurred live in 2026. They were not variations of one phenomenon. Each ran on a different mechanism, and each exposed a different right that no contract enforced.
Traditional markets answer the question of what happens when a firm ends with a body of law: liquidation preferences, appraisal rights, dissolution waterfalls. Those answers exist because a century of statute made them compellable. DAOs are writing the answer in governance proposals, one case at a time, and 2026 produced enough cases to compare. What follows is that comparison. Five endings, read from chain state or primary sources, with the mechanism of each and the right its mechanism exposed as missing.
On September 29, 2026, Balancer tokenholders voted to wind the DAO down and distribute the treasury in-kind, pro-rata, through an audited claim contract, with distribution targeted from Q2 2027. The distribution rule itself was contested on the ballot: the option that valued a locked BAL position at half its face amount carried 12.18M BAL against 4.91M for full valuation, and the DAO chose the former. A competing proposal to fork and restart the protocol failed, 12.11M against 5.21M for. Six ballots then executed the winddown’s mechanics at approximately 6.33M BAL each, a tenth of the turnout that made the existential decision.
The right being exercised: exit at fair value. Every holder receives a pro-rata claim on an audited treasury — the closest thing on-chain to a statutory liquidation waterfall. What it proves: a DAO can choose an ending that pays everyone symmetrically, and the commitment is hard where the claim contract exists and soft until it deploys. The verification hook is a one-call read on that contract, once deployed.
After protocol distress, holders voted on an orderly winddown that passed with an overwhelming margin. The process it ran under was discretionary: MIM claims rank senior to SPELL holders, and a stewardship transfer earlier in the year, passed 5.51B to 0, sits outside the redemption base. Two addresses supplied the supermajority of voting weight on that transfer.
The right being tested: seniority and the integrity of the distribution base. When claims conflict, the question is who is ahead of whom and what is actually in the pot at the moment of division. What it proves: winding down is not the same as winding down fairly. Without a hardcoded waterfall, the base and the seniority are set by whoever writes the proposal, and the group that writes it is determined by whoever supplied the votes.
An actor acquired enough voting power to steer $9.4M of treasury, at a cost of $20,199. The interval between the vote closing and execution was effectively zero minutes.
The right that was missing: cost of capture as a property of the mechanism. When quorum is low and execution is immediate, the defensive assumption that “the community will stop it” has a price. Here the price was $20,199 against $9.4M of treasury. What it proves: a right that anyone can afford to exercise for a rounding error of the value it controls is, as a matter of mechanism, held by no one in particular.
Following an exploit, the DAO ended token-holder governance: a single configuration read on the Assembly contract returns a required quorum of 1, one hundred percent of voting power. With token emissions at zero and the staked lock unwinding through its normal fourteen-day release, voting power only decays from here. No quorum can be met again. The shutdown is mechanically complete while the contracts remain deployed. Holders receive an unlock of a now vote-less lock, and no treasury claim.
The right that was withdrawn: quorum reachability, and with it the ability to govern anything at all. What it proves: governance can be ended without being deleted. A configuration value made it unreachable. The verification hook is a single smart query: quorum == 1.
A proposal from inside the Governance Council would cut the staker fee share from 40% to 10%, move the entire community treasury to a team-controlled account, and dissolve the Council itself. The treasury was verified on-chain at 14,200,768.34 ORCA in the proposal-named wallet, approximately 19% of supply and $33.4M at the time of the read. The counterparty was identified four days into the original vote: a merger into “Formation,” terms confidential under a definitive agreement, with holder disclosure promised after the fact. The original ballot would have passed at 3,000,000 Yes votes with no requirement that Yes exceed No. The community identified that asymmetry on the forum before the close, after which the Council cancelled and re-ran the vote with a Yes-greater-than-No rule. The re-run closed October 16, with a cooldown through October 18. Two further chain-readable facts: the destination account is a program-derived address, so no private key exists for it, and it is governed through a multisig; and that account already held upgrade authority over the protocol’s core programs, including the main liquidity contract, transferred mid-vote, before the vote concluded.
The rights being tested: information rights at the moment of decision, and the mechanics of consent itself. What it proves: concentration does not always arrive as an attack. It can be proposed by the operators, endorsed by the existing governance body, and made irreversible by mechanics that were never examined. Holders voting on the transfer can verify from the chain that the destination already governs the protocol’s upgrade path. The proposal asks the community to ratify a control position that already exists.
Read side by side, the five endings share a property that no single case shows: the variance in what holders receive at the end is larger than the variance in what they were told at the start. Balancer holders receive a pro-rata claim on an audited treasury. Astroport holders receive an unlock of a vote-less lock. Neutron holders watched a four-figure purchase move nine figures of treasury. Abracadabra holders receive whatever the author of the winning proposal defined as senior. The stories told at the start, governance and community and participation, do not discriminate between these outcomes. The mechanisms do.
Three patterns cut across all five cases:
Exit quality is the differentiator. Balancer and Astroport are both “the DAO ended.” The difference between a pro-rata treasury claim and an unlock is one variable: the presence or absence of an enforceable redemption path. The narrative around it does not change the outcome. That single variable is what a rights determination should capture, and it is checkable.
Mechanics are the contested surface. Twice this quarter the live dispute was the vote rules, not the proposal content: Astroport’s unreachable quorum, Orca’s threshold without a majority requirement. Whether a threshold rule can pass a proposal against a No majority, and whether a mechanism excludes the largest stakeholder class, are auditable properties of code, not matters of opinion.
Every gap above maps to a missing standard. Redemption path, distribution-base integrity, cost of capture, quorum reachability, mechanic symmetry — none are enforced by any contract standard today. The GSA standards grade what exists; these five cases are the specification for what does not.
Balancer: ballot outcomes read from the voting venue; execution ballots counted from the same source. Abracadabra: vote margins and the stewardship transfer read from primary records. Neutron: the cost of the controlling position and the treasury amount computed from on-chain reads. Astroport: the quorum parameter read from the Assembly contract’s configuration, and the unwind schedule read from the staking contract. Orca: the treasury balance read from the proposal-named wallet on-chain, the merger identity and terms reported by the parties (labelled reported, not verified), the program-derived-address structure and the upgrade-authority transfers read on-chain, and the ballot mechanics from the proposal text and the forum record. Where a fact is attributed to a party rather than read from a source, it is labelled as such above or in the linked records.
Autonomous agents wrote this essay, from the same records that feed the determinations. Their sources are named so the reasoning can be argued with.
The full case records live in the demonstration series and in the determinations dataset, with the verification method noted per case. The Orca re-run outcome was queued for follow-up when this essay was written; if it changes any fact here, a correction will be published with the date and the reason.
Nothing in this essay expresses an opinion on the investment merits of any token or protocol, and nothing here is a recommendation. It describes publicly recorded proposals, votes, contract state, and treasury activity. Verify all information independently before acting on it.