ENSv2, the rebuilt Ethereum Name Service, went live on Ethereum mainnet on 27 October, according to ENS Labs, and the terms for anyone holding a .eth name have changed. A name of five or more characters costs $8 a year instead of $5. The grace period after expiry drops from 90 days to 28. Names that upgrade get a one-off 62-day extension to make up for it, and the upgrade itself is optional, with no deadline.
“Your existing ENS name will continue to work,” ENS Labs wrote, and no one has to re-register. Doing nothing does not keep the old price, though. The activation proposal says the new fees apply from execution “including to renewals of names that haven’t migrated yet”, and they are paid in USDC or DAI, not ETH.
The upgrade cannot be undone. Three audit reports on the code are public, but one of them covers a single pull request, and Blockful, whose review of the proposal sits on the DAO forum, says some live code is covered by none of them.
ENSv2 for .eth holders: staying put versus upgrading
The new price and the payment change apply on both paths. This table sets out what differs, from EP 6.43, the activation proposal and the ENS documentation. It describes the trade-offs; Coindaro makes no recommendation either way.
| What matters | Stay on v1 | Upgrade to v2 |
|---|---|---|
| Name keeps working | Yes, through a v1 fallback | Yes |
| Annual price, 5+ characters | $8 | $8 |
| Accepted payment | USDC or DAI | USDC or DAI |
| Grace period | 90 days to renew; resolves until about day 62, later days not documented | 28 days, after an expiry moved back 62 days; no resolving in grace |
| Open to new registrants | About 90 days after v1 expiry | About 90 days after v1 expiry |
| Deadline to act | None | None |
| Subname registry and role-based permissions | No | Yes |
| Commemorative NFT card | No | Yes, for eligible addresses |
| Can it be reversed? | Nothing to reverse | No |
What renewal costs now: $8 a year, less for multi-year terms
The fee change comes from EP 6.43, a social proposal that passed before launch. Three- and four-character names keep their prices of $640 and $160 a year. Only the five-plus tier moves, by 60%. The proposal also adds a discount for buying several years in one transaction, for new registrations and renewals alike.
| Term | Discount | 5+ characters | 4 characters | 3 characters |
|---|---|---|---|---|
| 1 year | 0% | $8 | $160 | $640 |
| 2 years | 12.5% | $14 | $280 | $1,120 |
| 3 years | about 31% | $16.50 | $330 | $1,320 |
| 4 years | about 31% | $22 | $440 | $1,760 |
| 5 years | about 31% | $27.50 | $550 | $2,200 |
| 6 years | about 44% | $27 | $540 | $2,160 |
The curve has one oddity. Six years costs less than five, $27 against $27.50 for a long name. EP 6.43 says registrations and renewals made before the v2 registrar contracts were deployed stay on the existing pricing.
ENS Labs’ 22 October post notes that base prices had not moved since May 2019, and says ENSv2 was a chance to ask whether the model still fits how the namespace is used. The post says the discounts aim to “optimize for long-term retention”, and that the DAO will review pricing every two years. The main purpose of fees, it adds, is to discourage speculative registration. They are paid directly to the ENS Endowment, which the ENS Foundation owns and Karpatkey manages, according to the activation proposal, and that includes renewals of names that never upgraded. Blockful says its test confirmed a renewal pays the Endowment.
No more paying in ETH, for now
The contracts accept USDC and DAI, both dollar-pegged stablecoins. The ENS App takes USDC on mainnet and USDC on layer 2 networks through crosschain intents, while the ENS Explorer takes mainnet USDC only, so DAI is a contract-level option. Blockful warned in its review of the proposal that owners holding only ETH must blockchain, where users swap tokens directly from their own wallets without handing custody of their funds to a company.">swap first, in a separate transaction, and that renewal tools paying in ETH stop working. The same review says ENS Labs plans ETH payments for the fourth quarter of 2026 or the first quarter of 2027. The ENS docs say only that more tokens will be added over time.
Any account can renew any name, the docs say, and doing so does not alter ownership or permissions, so a company that depends on a name can top it up. Renewals pay the base rate only, never a premium.
A 28-day grace period, and a name that goes dark during it
What happens inside the window changes too. Under v1 a lapsed name kept resolving during the grace period and its records could still be edited. Under v2, according to ENS Labs, a name in grace appears expired. It does not resolve, it cannot be transferred, and its records cannot be edited. The owner can still recover it at the normal annual price.
After day 28 comes a 21-day cooldown, during which a decaying premium sits on top of the base price for anyone registering the name. ENS documentation says this exists to prevent sniping, starts near $100 million, halves every day and reaches zero at the end of the window. If nobody takes the name, it becomes available at the base price.
The 62-day extension exists so that upgrading does not bring anyone’s deadline forward. It is applied automatically and only once, and EP 6.43 says owners keep the same total amount of time to renew. As an illustration, using our own arithmetic rather than an ENS example: a name due to expire on 1 January 2027 would, once upgraded, have an expiry of about 4 March. Grace would run to about 1 April, which is also 90 days after 1 January, and the premium window would last until about 22 April.
Names left on v1 can be renewed for the full 90 days. They keep resolving through a fallback until their v2 reservation expires, about 62 days after the v1 expiry. The migration documentation does not say what lookups return in the last 28 days of the old window, and neither do the other ENS pages we read. ENS Labs says it will flag expiry in the App and through notification channels, and calls renewing well in advance the safest approach.
Upgrading is optional, and it is one-way
CryptoSlate put this catch in its headline on 13 October, while the plan was still a draft. It holds up against ENS’s own texts. In the activation proposal, ENS co-founder Nick Johnson (nick.eth) wrote: “The v1 freeze is reversible; migrating a name is not.” The ENS migration documentation describes the mechanism. An upgraded name’s v1 token goes to a contract called the Graveyard, of which the docs say: “It has no mechanism to transfer or upgrade the tokens it holds.”
The old token stays in that contract, and the name is managed on v2 from then on. The owner does not lose the name: migration turns its reserved v2 slot into a full registration with an owner. There is no way back to v1.
In exchange an owner gets control over subnames. With v2 a name can run its own registry, and permissions become fine-grained roles, so one person can be allowed to set a resolver but not transfer a subname. The docs say names left on v1 go without per-record permissions, record linking and the new resolver. Those tools are aimed at projects issuing names to their users. A name with one owner and no subnames has little use for them, and the upgrade still cannot be reversed. Wrapped names with locked permissions keep them, the proposal says.
A few cases need more care:
- The commemorative NFT card is revealed only once an address has upgraded all its eligible names. Eligibility rests on holding an active .eth name at the 1 June 2026 snapshot, with one NFT per address. Minting is optional and costs only Ethereum network fees, and ENS Labs states no deadline.
- Blockful says about 306 names were skipped in the copy to v2 because ENS Labs could not read their text, along with names longer than 255 bytes. An owner who knows the text can register such a name again on v2, it adds.
- Names already past their grace period were not carried across and became open to anyone on v2, the proposal says.
- ENS has set no deadline, so a message claiming you must migrate by a certain date does not match its published plan. ENS Labs says upgrades start in the ENS App, at app.ens.domains.
ENSv2 audits: what three reports cover, and the gap Blockful flags
Early coverage named Certora and Nethermind audits without showing them. Three reports were added to ENS’s contracts-v2 repository in a pull request merged on 14 October, on its post-audit-2 branch, and we read all three. An audit records what its authors reviewed at a given commit. It does not show that the live contracts are free of bugs.
| Report | Date | Scope | Findings | Status |
|---|---|---|---|---|
| Certora, assessment and formal verification | May 2026 | contracts-v2 | 8: 3 medium, 1 low, 4 informational | 7 fixed, 1 acknowledged |
| Nethermind NM-0591 | 1 June 2026 | contracts-v2, verifiable-factory | 10: 1 medium, 4 low, 4 informational, 1 best practice | All 10 fixed |
| Nethermind NM-0965 | 13 October 2026, updated 19 October | One pull request (#417), not the whole code | 7: 1 high, 1 medium, 4 low, 1 informational | 3 fixed, 4 acknowledged |
The one high-severity finding, in NM-0965, concerned a malicious gateway able to forge responses and spoof name lookups; the report marks it fixed. The four items left acknowledged are three low-severity issues and one informational note. Nethermind says this review looked at the pull-request changes “rather than re-auditing the complete ENS V2 codebase”. The repository holds two versions of it. The first, dated 13 October, gives c676a9c as its final commit. An update dated 19 October gives 81869c3, the commit Blockful cites. We compared the text of the two, and only the date and the final commit differ.
Blockful rebuilt all 15 calls of the proposal and found they matched what was on chain. Of its nine findings, six are resolved, two acknowledged and one open, in a review last edited on 29 October. The most serious, rated high and resolved, was a plan to hand the .eth token to the Endowment Safe, where the DAO has no role; Blockful says the token reached a DAO Safe on 27 October. A finding about post-audit fixes was closed by the updated Nethermind report.
The open item, which Blockful labels L04 and rates low, reads: “Some contracts now live include code that no published audit report covers.” Blockful names WrapperRegistry, LockedMigrationController and UserRegistry, deployed on 26 and 28 October and already in use, with changes merged after the audited code. It adds two ens-contracts changes in the 18 October contracts that fall outside the report’s scope. ENS Labs reviewed and approved these changes itself, Blockful says, and its proposed fix is for the auditors to review them and update the report. The forum thread shows no later reply on this.
Blockful’s acknowledged medium item concerns upgraders who used the migration helper. New accounts called HCA let users register a name with one signature, relying on a service run by Rhinestone, and a single key can change that service’s code. If the key were stolen or misused, Blockful says, it could move the remaining v1 names of users who approved the helper and set the primary name of addresses that have none. Blockful suggested the ENS app remove that approval right after each migration, and does not say whether it does. It reports that Rhinestone will make the service unchangeable in the fourth quarter of 2026 or the first quarter of 2027. “Who holds the key until then is still open,” the review says. The ENS Labs site links a bug bounty on Immunefi. We found no report of an exploit since launch, in a search that was not exhaustive.
How ENSv2 reached mainnet, and what the ENS token did
ENS is a naming system that maps a readable name to a wallet address and other records, through smart contracts. ENS Labs counts more than 36 million names across .eth and ecosystem namespaces and over 600 applications that support them.
The project first planned to run v2 on its own network, Namechain. On 6 February 2026 ENS Labs dropped that plan and moved everything to Ethereum itself, The Block reported, citing a base layer that had scaled faster than expected. Readers new to that trade-off can start with our guide to layer 1 versus layer 2.
The activation draft reached the DAO forum on 11 October. Voting opened on 18 October and closed on 25 October at about 06:01 UTC, with about 1.69 million votes for and 24.94 against, above the quorum of 1 million, The Merkle reported. The three largest supporters gave about 33.8% of the total in favour. After a two-day timelock the proposal could run from about 06:00 UTC on 27 October, and TokenPost reports it was carried out that day. Blockful says it checked on 27 October that the ENS Labs deployer kept no role.
The ENS token was used to vote, which is its role here as for any governance token. Nothing in the launch post or the proposal changes that, and names are not paid for in ENS.
What apps have to do
Owners do not need to touch anything for lookups to work, but apps must resolve against the Universal Resolver to be compatible with v2. The proposal said on 11 October that most integrations were ready. Wu Blockchain reported on 31 October that more than 40 teams had finished integration, including 1inch, Arbitrum, Safe and MetaMask. The count is Wu Blockchain’s alone. ENS Labs’ launch post lists those companies as featured partners but gives no such number and does not say each has finished. We have not tested any of these apps.
What is still unknown, and the next dates
ENS Labs gave no time of day for the launch. The proposal expected registrations to pause for a few hours between execution and the v2 opening; how long that lasted has not been reported. No date is attached to the removal of the broad .eth roles the DAO kept to fix errors after launch, which waits for a follow-up vote once v2 has run safely.
Blockful puts ETH payments and the Rhinestone change in the fourth quarter of 2026 or the first quarter of 2027, and lists the follow-up vote on those .eth roles as the next thing it will check.