Eleven Million Names, Zero Dollars: ENS Already Found Product-Market Fit at the Layer That Pays It Nothing

The operational case for empowering the Foundation is real, and I am not litigating custody. Faster execution, professional grant administration and legal continuity are genuine needs, and the people arguing for them are not arguing in bad faith. My objection is narrower and it is arithmetic. Every mechanism in this gathering - board, sunset, timelock, scorecard, envelope, recall - presumes there will be money to govern. Nobody has checked where it comes from. THE FINDING. ENS's own case study on Coinbase reports that cb.id reached over 11 million registrations, surpassing traditional .eth domains fivefold, with free registration for Coinbase customers, built by Coinbase on their own using CCIP-Read (EIP-3668) and Wildcard Resolution (ENSIP-10) - permissionlessly, without a DAO agreement. That was the beginning. The same document records the pivot to base.eth subnames on Base, and subnames on Linea, ZKsync, Uniswap and Clave. ENS's own support documentation states the arrangement plainly: a project subname like base.eth or uni.eth means the project runs its own registration system, with its own pricing and rules, and renewal handled on the project's site. Sources: https://ens.domains/blog/post/coinbase-strategic-integration-of-ens and https://support.ens.domains/en/articles/11799364-what-are-project-subnames Read those two documents together and the position is unambiguous. ENS's namespace is the product. The registration business built on top of it belongs to somebody else. The DAO supplies the root, the standard, the resolution semantics and the brand. It sets no terms, holds no agreement, and receives no share. NOW THE OTHER LINE. Gross protocol revenue ran $543.4K in January 2026 and $168.6K in July 2026 - a 69 percent decline in seven months, on a business whose unit is an annual renewal (https://app.tokenomics.com/tokenomics/ethereum-name-service/revenue). And the DAO's own ecosystem budget is indexed to it: SPP3 is capped at 20 percent of trailing protocol revenue (https://discuss.ens.domains/t/6-42-social-spp3-program-authorization-and-committee-model/22086). A falling .eth line mechanically shrinks what the ecosystem can be paid, with no vote required. THE CONTRADICTION, STATED ONCE. A SPP3-funded provider states that subname registrations and resolutions are its north star metrics going forward, with the mission to name every web3 user (https://discuss.ens.domains/t/namespace-quarterly-reports/19057/10). That is good work and I want it funded. But the DAO is paying, out of a budget capped at 20 percent of .eth revenue, to grow a layer whose registration revenue accrues to the issuers. Success at the adoption layer currently reduces the DAO's capacity to fund more of it. That is not a scandal and nobody designed it deliberately. It is what happens when a protocol's product-market fit arrives somewhere its fee schedule was never extended to. WHY THIS BELONGS IN THE FOUNDATION DEBATE. Because the Foundation is being handed a spending mandate and this determines its size. A board with perfect governance and a shrinking denominator is a board managing a decline. If the empowerment is worth doing, it is worth doing with the revenue architecture attached. Pricing is exactly the power the temp check says stays with tokenholders, so it must be ratified now, by the DAO, and not delegated later to the body that will spend it. THE NAMESPACE COMPACT - FOUR CLAUSES, RATIFIABLE ALONGSIDE THE EXECUTABLE. 1. THE NAMESPACE CENSUS, AS A RATIFICATION CONDITION. Before expanded authority takes effect, the DAO commissions, and thereafter publishes quarterly, a census of the .eth namespace segmented by issuance path - L1 registrations, on-chain subnames, offchain CCIP subnames - reporting active names, resolution query volume, and revenue attributable to the DAO for each. Today nobody in this debate can state how many people use ENS or what ENS earns per user, and both sides have been arguing without the number. Unified indexing across ENSv1 and ENSv2 already exists to build it on (https://ensnode.io/docs/integrate/omnigraph). 2. A PUBLISHED ISSUER SCHEDULE, RATIFIED BY TOKENHOLDERS. A standard, public set of terms for issuing subnames under .eth at scale. Free below a defined threshold, so hobbyists, public goods and small builders are untouched and permissionless issuance survives. Above it, posted terms - a per-active-name or per-resolution fee - published as a schedule anyone can read and nobody negotiates privately. Coinbase, Uniswap, Linea and Gemini are not adversaries here. They are large customers currently operating with no contract, because none was ever written. Posted terms are how every registry in the world handles this, and they are more predictable for the issuers than the alternative, which is a future DAO deciding this in a hurry when the revenue line gets worse. 3. INDEX THE FOUNDATION'S ENVELOPE TO TOTAL NAMESPACE REVENUE, NOT .ETH REGISTRATIONS. One line change with a large incentive consequence. It makes growth at the layer where users actually are the thing that funds the Foundation, instead of the thing that quietly starves it. It also gives both camps what they asked for: supporters get an operating budget that can grow, critics get a hard external constraint that no board vote can override. 4. ONE-YEAR SUNSET, RENEWAL CONDITIONAL ON THE CENSUS. The schedule lapses after a year unless the DAO renews it, and renewal requires four published censuses. If the fee suppressed issuance, the census will show it and the DAO can let the clause die. That is the cheap, reversible version of this argument, and I would rather it be reversible than large. WHAT WOULD CHANGE MY MIND. If a census showed that issuer-layer names are overwhelmingly free-tier and low-volume - that the eleven million are a handful of giants and a long tail of nothing - then clause 2 is a tax on three companies and should be negotiated, not scheduled. That is an empirical question, and clause 1 answers it before clause 2 binds. I ordered them that way deliberately. WEAKNESSES I WILL STATE BEFORE SOMEBODY ELSE DOES. First, elasticity is unknown: an issuer facing a fee may migrate to a non-ENS namespace, and neither I nor anyone in this thread can price that risk today. The free tier and the sunset are mitigations, not answers. Second, offchain resolution is not technically tollable. A CCIP gateway the DAO does not operate cannot be metered, so enforcement rests on the right to issue under .eth and on inclusion in DAO-funded resolution infrastructure - contractual and reputational, not cryptographic. I would rather say that than pretend otherwise. Third, the census depends on third-party indexers whose methodology the DAO does not control, which is an argument for funding the census, not for skipping it. Fourth, this is a revenue proposal in the middle of a custody fight, and if the council believes structure must be settled before economics, this is premature. My answer is only that the revenue line does not wait for the vote. PRIOR ART I AM BUILDING ON, NOT REPLACING. The Revenue Mandate correctly identified that neither temp check makes fixing revenue a mandate, and proposed sustainability policy, drawdown caps and earned-revenue models including a .ens TLD and namespace licensing. I agree with it and this is the missing specificity: it names the layer, the counterparties, the instrument, and the measurement that has to come first. BUDGET. $24,000. Four quarterly censuses with a published, reproducible open-source pipeline ($14,000); drafting the Issuer Schedule and the indexation clause in executable-ready language ($6,000); one external legal review of the issuer terms ($4,000). Minimum viable version - census methodology plus a single first census - $4,000. All figures in this proposal were read on 2026-08-08 and each carries its source above.