Supply state

Where does the .eth supply sit today, and where is it heading?

Two views that bookend the supply state:

Backward - Active / Grace / Dead split per bucket. Active = expiry in the future. Grace = expired in the last 90 days (only the original owner can renew). Dead = expired ≥ 90 days ago, publicly registerable, but nobody has.
Forward - when does the active inventory's expiry cliff hit? Per-bucket mean / median / mode of months remaining, plus a stacked-bar chart of active names by future expiry month for the next 10 years.

What this tells you: A high dead % in a tier signals which supply nobody wanted; the forward chart shows when the next renewal-decision wave hits and how concentrated it is.

Snapshot taken 2026-10-04T03:03:06.456Z

Dead inventory by length

Names that were registered, expired past the 90-day grace period, and have not been re-registered by anyone. The 21-day premium dutch auction (90–111 days post-expiry) is part of "Dead" because the name is publicly registerable at that point.

Length	Total .eth SLDs	Active	Grace	Dead	Dead %
3 chars	50,164	2,773	1,226	46,165	92.03%
4 chars	113,922	9,305	4,047	100,570	88.28%
5 chars	482,761	97,216	16,173	369,372	76.51%
6–9 chars	1,701,208	385,202	58,590	1,257,416	73.91%
10+ chars	1,345,311	302,951	43,284	999,076	74.26%

Remaining expiry by bucket

Of names currently active in each bucket, how long do they have left before they'd hit grace if not renewed? Flips the page from looking backward (what's already dead) to looking forward (what's about to die).

Length	Active names	Mean (months)	Median (months)	Mode bin (months)
3 chars	2,773	13.9	6.7	<1
4 chars	9,305	16.2	6.8	<1
5 chars	97,216	36.2	12.1	7–8
6–9 chars	385,202	39.5	14.1	7–8
10+ chars	302,951	38.4	12.7	7–8

Mode bin "<1" means the most-common state is "expires within the next month" - i.e. names rolling over their initial 1-year term. A high mean with a low median says the bucket has a fat tail of multi-year up-front registrations dragging the average up.

Forward expiry cliff (next 10 years)

If nobody renewed anything from now on, when does each name expire? Bar height = total active names expiring that month, stacked by length bucket. The first ~12 months are usually a wall of 1-year-default renewers; deeper years carry the multi-year up-front registrations.

Toggle buckets below to isolate one tier — the Y-axis rescales automatically so smaller tiers (3/4/5 chars) become visible when the dominant 6–9 / 10+ buckets are turned off.

3 chars 4 chars 5 chars 6–9 chars 10+ chars

Active names by future expiry month (peak: 46,625 in 2027-05-01)

20262027202820292030203120322033203420352036

Methodology

All counts are unique names by namehash. Considers .eth SLDs (depth=2, parent='eth') with label resolved and expiry > 0.
Active = expiry >= now. Grace = expired in the last 90 days (only the original owner can renew). Dead = expired ≥ 90 days ago - anyone can register, but nobody has.
The 21-day premium dutch-auction window (90–111 days post-expiry) is part of Dead because the name is publicly registerable at that point - just at a higher price.
Doesn't distinguish 'never re-registered since first expiry' from 're-registered then re-expired again'. A name's current expiry determines its state today.
Forward expiry chart: capped at 10 years out. A handful of multi-decade registrations exist beyond that horizon and aren't shown.
Months-remaining stats use 30-day months (expiry_timestamp − now) ÷ 30 days. Mode bin is the most-common integer-month value; ties broken toward the smaller (closer-to-now) bin.