ENS DAO

The ENS DAO governs the Ethereum Name Service, the dominant naming and identity layer for Ethereum, which maps human-readable .eth names to wallet addresses and other records. Since a November 2021 airdrop of the ENS token, the protocol's treasury, working groups, and parameters have been governed on-chain by token holders and their delegates. ENS governance is frequently cited as a benchmark – and, in 2026, as a cautionary tale about voting-power concentration.

Structure

The DAO funds work through working groups, and since Term 7 began on 1 July 2026 it has run one. The EP 6.44 Snapshot vote ranked “Metagov only” first, and the Term 7 election seated three Meta-Governance stewards: netto.eth, sovereignsignal.eth and abdullahumar.eth. The Ecosystem and Public Goods groups dissolved, and the stewards are still settling how much of their funding and grants work Meta-Governance takes on (Meta-Gov minutes, 21 August 2026). The DAO also holds an endowment managed by the asset manager Karpatkey to give it a self-sustaining runway. Its legal interface is the ENS Foundation, a Cayman Islands nonprofit that provides limited liability and lets the DAO contract with the real world. (ENS Foundation – ENS docs)

The 2026 governance battle

In June 2026 ENS became the year's most-discussed governance dispute, and the record of it is unusually complete: every step ran through the DAO's own forum, Snapshot space, and governor contract.

On 19 June a proposal introduced by the delegate katherine.eth sought to transfer operational control, grants, and treasury management – the operational wallet, the DAO's ENS holdings, and the Karpatkey-managed endowment – to a five-seat ENS Foundation board. In parallel, the DAO was voting on whether to renew its Security Council, the 4-of-8 multisig whose single power is to cancel a proposal already sitting in the timelock. The two questions were inseparable: a council empowered to cancel a passed proposal could, in principle, cancel that one.

The advisory Snapshot vote on renewal (23–28 June, 73 voters) read as a clear pass: 1,274,195 ENS for to 85,204 against. It also carried 3,386,139 ENS abstaining – 71% of everything cast – of which 3,261,042 came from a single voter, ENS lead developer nick.eth, whose self-delegation that month was itself the subject of the dispute (The Defiant).

The binding vote went the other way. Read from the ENS Governor, the executable that would have granted the renewed council its cancel role opened at block 25,424,114 (29 June 2026, 15:07 UTC) and closed at block 25,469,932 (6 July 2026, 00:34 UTC), and it was defeated 3,859,981 ENS against to 1,137,702 for. The DAO's advisory instrument and its binding one had pointed in opposite directions.

The replacement was tabled while that vote was still running, not after it. On 30 June katherine.eth posted a framework for a new council: eight members, the threshold for any cancellation raised from 4-of-8 to 5-of-8, a tighter written mandate, a public pledge each nominee must affirm, a removal mechanism for members who act outside it, and an Appointment Agreement with the ENS Foundation drafted by counsel – with the scope of authority deliberately unchanged, “cancellation of timelocked proposals, nothing more”. Nominations were open to anyone and closed on 6 July. Eight of fifteen candidates were elected in a Copeland-method vote (7–12 July, 50 voters), and the executable seating them passed 4,228,716 to 94,526.

As of 8 August 2026 the outcome is verifiable on-chain: the new council contract holds PROPOSER_ROLE on the DAO timelock, its owner Safe reports a threshold of 5 over 8 owners, and the term runs to 16 July 2028, when the same expiry logic fires again. ENS did not lose its veto and did not simply renew it; it made the body re-argue its mandate in public and wrote the answer down.

The Foundation question reached the same governor in August, revised after its temp check. The executable made the ENS Foundation the DAO's operating arm, with a full-time executive director, staff and a five-member board that token holders appoint and can remove. The endowment keeps its assets in place, but its transactions now pass through a nine-day timelock in which the Security Council can cancel anything outside the Foundation's mandate. The DAO keeps its ENS, 54.6% of total supply, apart from a one-time transfer of 1m ENS to the Foundation for employee compensation (ENS blog).

The delegate Curia voted against because, after execution, the DAO holds no role on that new endowment timelock (forum thread). The proposal was executed, and Alexander Urbelis became the Foundation's first executive director (ENS DAO Newsletter #118). By 15 September custody of the endowment Safe had moved to the Foundation, and the same newsletter put the endowment’s NAV at $93M, with August strategy earnings of $223K at a 3.08% blended APY and premium sales of $113K (Newsletter #120).

Why it matters

The episode is a textbook live case study in plutocracy and capture. One delegate, Lefteris Karapetsas, wrote that Johnson had "delegated ~50% of the voting supply to himself, essentially becoming the DAO," and Security Council member Brantly Millegan called the Foundation proposal "the equivalent of treasury capture by ENS Labs." It crystallizes the core problem of token-weighted voting: when one party can marshal a majority of active votes, the decentralized form offers little protection for everyone else – the same dynamic documented across large DAOs in 2026. (EtherWorld)

ENSv2 and the scrapped Namechain L2

Beyond governance, 2026 also reset ENS's technical roadmap. ENSv2 – a ground-up rewrite of the ENS contracts and architecture for better usability, flexibility, and developer tooling – was originally meant to live on Namechain, a dedicated ENS Layer-2 that ENS Labs and the Linea Association agreed to build in December 2024. On 5 February 2026 ENS Labs scrapped the L2 entirely and moved ENSv2 to Ethereum mainnet.

Nick Johnson's framing: "the landscape has changed between when we first decided to pursue an L2" – the base layer had scaled enough (ENS reports ~99% lower registration gas costs over the prior year alongside Ethereum's 2025 gas-limit rise) to make a separate rollup's cost savings no longer worth its complexity, echoing Vitalik Buterin's renewed emphasis on scaling L1 itself. ENS Labs says roughly 80% of development effort can now go to ENSv2's core features rather than maintaining custom chain infrastructure. It is a notable case of a DAO-governed protocol reversing a flagship architectural bet as the surrounding ecosystem shifted under it.

Pricing names

The DAO's constitution says what registration fees are for. Article II of the ENS DAO constitution makes their primary purpose preventing the namespace from being overwhelmed by speculatively registered names, with funding the DAO second, and its examples allow governance to raise prices for either reason. The failure that article guards against had been measured before ENS launched.

Harry Kalodner, Miles Carlsten, Paul Ellenbogen, Joseph Bonneau and Arvind Narayanan studied Namecoin, the first fork of Bitcoin with its own blockchain, which runs a decentralised registry for .bit domain names (An empirical study of Namecoin and lessons for decentralized namespace design, 2015). Of roughly 120,000 registered names they found 28 that were not squatted (held unused, usually for resale) and had nontrivial content. They found evidence of at most about 250 sales of names in Namecoin's history, and concluded that allocating the scarce supply of memorable names is a mechanism-design problem far harder than had been realised.

The .eth registrar addresses the same allocation problem with four mechanisms (ETH registrar documentation). Names are leased by the year and priced by length: $5 a year for five characters or more, $160 for four and $640 for three, paid in ETH to the DAO's treasury. A name that expires and passes its 90-day grace period returns through a 21-day Dutch auction, in which a premium starting near $100m falls towards zero until someone registers the name. Registration is split into a commit and a reveal at least 60 seconds apart, so nobody watching pending transactions can register the name first.

How Caper approaches this

ENS shows what happens when a dominant voter can override a DAO and minority members have no clean way out. A caper responds with two design choices described on its governance pages: voting weight that grows with demonstrated participation, not stake alone, so influence is harder to simply buy or delegate to oneself; and a built-in exit right that lets any member who has voted redeem a treasury share sized by that same participation-weighted vote weight rather than be locked in under a controlling party. Those mechanics are stated neutrally on the linked Caper pages.

References

• ENS DAO Governance Forum.
• ENS docs: The ENS Foundation.
• The Defiant – ENS delegates call the Foundation proposal a governance attack as Johnson self-delegates.
• Cryptopolitan – ENS DAO governance reform proposal drama.
• EP 6.45 – Renewal of the Security Council, the renewal thread carrying the mandate argument in full.
• Establishing a new Security Council – the executable, with the deployed contract and Safe addresses.
• Harry Kalodner, Miles Carlsten, Paul Ellenbogen, Joseph Bonneau and Arvind Narayanan, An empirical study of Namecoin and lessons for decentralized namespace design (Workshop on the Economics of Information Security, 2015).
• ENS DAO constitution (Article II, fees) and ENS docs: ETH registrar (pricing, premium auction, commit-reveal).