Empowering the ENS Foundation While Preserving Onchain Custody A three-part amendment to the "Empowering the ENS Foundation" temp check (discuss.ens.domains/t/22175). It delivers the operational benefits of the Foundation, the legal wrapper, institutional capacity, delegate relief, and multi-year strategy, while keeping treasury custody in DAO contracts. The Foundation operates as authorised agent, not custodian. What this proposal concedes The operational case for the ENS Foundation is real, and I accept it in full. Delegate fatigue is worsening (katherine.eth, post #1, Problem #1). The DAO makes too many small decisions and too few consequential ones (Problem #2). There is no accountability layer between the DAO and the entities it funds (Problem #3). Cross-working-group coordination is slow and inconsistent (Problem #4). And the DAO lacks a vehicle for multi-year capital strategy (Problem #5). None of these are imagined problems. They are what four years of operating the DAO has actually shown us. I also accept that ENS needs a compliant legal wrapper for ICANN engagement, institutional partnerships, and standards work (liubenben, post #17C; Alexu, post #3). A pure onchain DAO cannot sign a contract, hold a trademark, or represent ENS at a negotiating table. The Foundation should be that entity. This proposal delivers every one of those benefits. It declines only the one step that none of them require: transferring custody of DAO assets off-chain. The distinction that matters The temp check asks the DAO to do two things at once. First, empower the Foundation operationally. Second, transfer custody of the treasury, ENS holdings, and Endowment to Foundation control. These are separable. The first is necessary. The second is not. katherine.eth confirmed the transfer explicitly (post #20): "Yes, treasury custody will be transferred to the foundation." She also drew the distinction her argument depends on: "Smart contracts != treasury." Protocol upgrades, fee mechanisms, and future revenue direction stay under tokenholder control. The treasury does not. I accept that smart contracts and treasury are different things. Where I disagree is that the difference makes custody transfer safe. Plank 1: Custody stays onchain The mechanism. The Foundation receives authority to propose and execute budgets, grants, and endowment strategy through capped, revocable, DAO-granted spending authorisations. The DAO's contracts (Governor/Timelock) remain custodian of all assets. The Foundation operates as authorised manager within DAO-approved parameters. It can direct spending within an approved cap, but it does not hold the assets. Precedent: the custodian and investment-manager separation. This is the standard architecture in traditional asset management. A custodian such as State Street or BNY Mellon holds the assets. An investment manager such as BlackRock or Vanguard directs investment decisions within a mandate. The manager has authority to trade but cannot abscond with the funds, because the custodian executes settlement. This is not a compromise structure. It is the industry's core risk control, and it exists precisely because the entity making decisions should not be the entity holding the keys. Institutional counterparties and qualified custody. For relationships that require a legal counterparty, such as the Endowment Manager, fund administrators, or vendors, the Foundation contracts as agent on behalf of the DAO as disclosed principal. The DAO's contracts remain custodian. The Foundation signs contracts, negotiates terms, and manages the relationship within DAO-approved parameters. Third parties contract with the Foundation knowing the principal is the DAO. This is a standard agency structure in fund management: the manager is the agent, and the fund or its custodian is the disclosed principal. The Foundation gets the legal standing to operate. The DAO keeps custody. For assets that require a qualified custodian under institutional custody law or regulatory standards, a smart contract alone does not satisfy that requirement, and this proposal does not claim it does. In those cases the DAO's contracts appoint a regulated custodian on DAO-controlled terms, covering fee schedules, withdrawal thresholds, and termination conditions, with the Foundation as authorised manager directing that custodian within the DAO-approved mandate. The DAO controls the custody relationship: who is appointed, on what terms, and when they can be replaced, without the Foundation taking title to the assets. The Foundation manages, the regulated custodian holds, and the DAO governs the terms. garypalmerjr framed this well (post #21): "The Foundation can be empowered as an agent of the DAO. But the DAO should remain the constitutional principal over the treasury." AvsA proposed a compatible structure (post #30): "The foundation should get oversight and auditing of the treasury, but not direct control. Instead they would submit an annual budget for DAO approval." This proposal adopts both framings and adds the institutional counterparty mechanism that makes them workable in practice. Plank 2: Transparency and accountability Conflict of interest. Adopt ICANN's Conflicts of Interest Policy in full, not approximated. This is Arnold's standard (post #15): "hold the Foundation to ICANN's Conflicts of Interest Policy adopted in full, not approximated." Ratify the policy by DAO vote rather than Foundation self-adoption, since the body being governed should not write its own governance rules. As Arnold put it, it "would also carry more teeth if the COI is ratified by the DAO rather than adopted internally by the body it governs." Budget veto and KPI reporting. The DAO retains veto power over the Foundation's annual budget. The Foundation publishes its full annual operating budget before the fiscal year and audited financials after. This builds on liubenben's observation (post #17): "the delegation of power must be accompanied by an equal degree of transparency… The DAO must retain ultimate oversight, such as veto power over the Foundation's annual budget, to ensure that capital allocation strictly matches the actual output and growth of the ecosystem." Adding quarterly KPI reporting against published targets keeps the DAO's veto informed rather than reactive. Director elections. Hold open, public, onchain elections for Foundation directors rather than ratifying a Labs-selected slate, with mandatory conflict disclosure for every candidate before the vote. This addresses netto.eth's structural critique (thread 22203, post #1) and AvsA's first reform (post #30): "Have the board election a proper selection by the DAO, with independent candidates proposed by others, and not simply a ratification." Plank 3: Anti-concentration No single delegate may account for more than 25% of the YES-side voting weight on any vote that transfers custody of DAO assets or amends the DAO Constitution. In AvsA's framing (post #30), no single delegate should be able to outvote the next fifty combined, and this makes that structural. This is not a general voting cap. It applies only to the two vote types where concentration risk is highest: custody transfers and constitutional amendments. On routine governance, delegates vote with whatever weight they hold. But on votes that could move $350M of assets or rewrite the constitution, the threshold ensures no single actor, however well-intentioned, can carry the decision alone. lefterisjp flagged the live version of this risk (post #22): the self-delegation of roughly 3 million tokens by a single party. Whether or not that delegation was legitimate, it showed that the DAO's current concentration profile is a single point of failure. A constitutional threshold on the most consequential votes is the mechanism-design response. If the threshold cannot be met on a given custody-transfer or constitutional vote, because concentration is too high, the vote fails. That conservatism is deliberate for changes of this magnitude. A governance system that cannot muster broadly distributed support for moving $350M of assets or rewriting its constitution should not be able to do so on the weight of a single delegate. Answering three counters "Smart contracts != treasury" (katherine.eth, post #20) This is the strongest argument for the custody transfer, and it deserves a direct answer. The argument is that the ENS token's security role is governing the smart contracts, the protocol upgrades, fee mechanisms, and constitutional changes, and those stay under tokenholder control. The treasury is operational capital rather than the protocol itself, so transferring its custody doesn't weaken the token's security function. My answer is that custody is the tokenholders' strongest lever over the assets whose downside they bear. The ENS token's value is the protocol's attack cost, and that value is directly linked to the onchain treasury it governs. James made this point precisely (post #14): "the ENS token's value is the protocol's attack cost, directly linked to the treasury value; CC why for the past X years ENS has traded basically 1:1 with ETH." Remove the treasury from onchain control and the token no longer governs the assets that back its value. The link between token price and protocol security breaks, not because the smart contracts changed, but because the economic substrate that made token governance meaningful changed. A budget veto does not restore that link. A veto is a fire alarm: it sounds after the fire starts. Custody is a locked door: it prevents entry. The Foundation can propose a budget and the DAO can reject it, but between budgets, between votes, in all the moments when the DAO is not watching, the Foundation holds the keys. Onchain custody means the DAO never has to trust that the alarm will sound in time. "Multi-year capital strategy requires a legal counterparty that holds the assets" (katherine.eth, post #1, Problem #5) The argument is that institutional endowment managers, fund administrators, and vendors don't contract with a smart contract. They contract with a legal entity that holds the assets, so without custody the Foundation can't operate a capital strategy. My answer is that the Foundation doesn't need to hold the assets to contract on their behalf. The agent and disclosed-principal structure described in Plank 1 gives the Foundation legal standing to contract with institutional managers while the DAO's contracts remain custodian. The Foundation negotiates the mandate, signs the management agreement as agent of the DAO, and directs capital deployment within DAO-approved parameters. The institutional manager contracts with a legal entity, the Foundation, while the assets sit with the DAO. This is how separately managed accounts and fund-of-funds structures already work: the investment manager has authority to direct the custodian, but the custodian holds the assets. The operational problem in Problem #5 is real. Token voting cannot operate a capital strategy. This proposal solves that, because the Foundation operates the strategy. It simply doesn't take title to the assets. The strategy needs an operator with continuity and professional capacity. It does not need that operator to be the custodian. "Onchain DAO custody creates unlimited personal liability, Ooki DAO" (SpikeWatanabe, post #44) The argument is that courts are treating unwrapped DAOs as general partnerships, so every participant, delegates, voters, and contributors, is personally liable, and moving assets to the Foundation extinguishes that liability. My answer is that the Foundation's legal wrapper is what solves the Ooki problem, and custody transfer adds nothing to that solution. The Ooki case targeted an unincorporated association conducting business, operating a derivatives trading protocol, without a legal entity. The liability arose because there was no entity to absorb it. Once the Foundation exists as the legal wrapper for ENS's operations, signing contracts, conducting business, and engaging with ICANN and institutional partners, the liability shield is in place. The Foundation is the entity that faces the world. The DAO's contracts are where the assets live. These are compatible: the Foundation absorbs operational liability while the contracts preserve custody. This proposal does not claim that asset location is irrelevant to partnership characterisation, which is a question for counsel. It claims only that the wrapper is what does the work, and that the wrapper does not require the Foundation to hold the assets. Moving custody to the Foundation is a separate choice from incorporating the Foundation, and the liability argument justifies the second without requiring the first. What this proposal delivers Delegate relief, with the Foundation Board and Executive Director handling operational decisions. A legal wrapper for ICANN and institutional work, with the Foundation as the legal entity. An accountability layer, through DAO-ratified conflict-of-interest rules, a budget veto, and KPI reporting. Multi-year capital strategy, with the Foundation as authorised manager acting as agent of the DAO. Institutional counterparty capacity, through the agent and disclosed-principal structure. Board independence, through open onchain director elections with conflict disclosure. Concentration risk mitigation, through a 25% cap on custody and constitutional votes. Every operational benefit the temp check proposes. One step declined: custody transfer. Not because custody doesn't matter, but because it matters too much to move. A note on legal validation. The agency, disclosed-principal, and regulated-custodian structures described above are a design direction to be validated by counsel, not a finished legal opinion. The mechanism is sound in principle and draws on well-established precedents in fund management, but the specific contractual and onchain implementations should be reviewed by qualified legal advisors before adoption. The claim here is that the design direction is viable and preferable to custody transfer, not that it is already court-tested for ENS's specific jurisdictional and regulatory profile.