.ghost
Is .ghost a real TLD? Who’s applying to run it?
.ghost is not contested. One application for it is on ICANN’s Reveal Day list, as of 7 October 2026. It is Oinkadot. Starlight (Namecheap) named .ghost as a backup string for its .droid application. ICANN deactivated it, because another applicant applied for the same string or named it as a backup.
Who is applying, and what they run today
None of the 2 is tied to a registry that runs TLDs today. Starlight (Namecheap) and Oinkadot filed 65 applications between them across the round.
Will names sell?
A playful, spooky word-domain that will sell on price and novelty; few serious projects will make it their main address.
Last round the AI scored the 2012 strings blind to their fate. Strings it scored 60 or more hold a median 27,168 names today; those under 30, a median of 30. .ghost’s closest 2012 strings hold .ninja 25K, .rip 14K and .boo 4K.
The nearest 2012 strings, and how many names each holds today
The closest 2012 string, .ninja, held 25,488 names, June 2026. .rip held 13,578; .boo held 3,620. Chosen by AI from every string in the 2012 round, without being told what happened to any of them.
Same bet: a vivid, non-industry English noun that no trade or company owns, sold as attitude and persona to anyone who wants a memorable identity rather than a category address, so its value rests entirely on brandable resonance and volume, not on serving a defined vertical.
Same bet in structure: a culturally loaded short word carrying one obvious literal use (memorial/absence) plus a wide meme-and-slang halo, betting that mood and connotation, not a buyer segment, drive registrations.
Same bet: a tiny evocative word with spooky and affectionate double meaning, monetized on novelty, wordplay and short-string appeal, with no professional constituency whose primary identity would ever sit on it.
hobbyists, indie game and horror creators, ghost-kitchen and ghostwriting services, and privacy-minded developers who want an evocative one-word address; secondarily, brand-protection buyers covering the word.
What a .ghost domain is
.ghost is not a TLD yet. It is a string on our record of the ICANN 2026 new gTLD round, in our Culture category. One application for it is on ICANN’s Reveal Day list.
Is the .ghost domain extension real yet?
.ghost is not yet a domain extension you can register: nothing resolves under it, and no registrar sells names in it. It is a string applicants have put forward in the ICANN 2026 new gTLD round, the process that decides whether it becomes one.
When can you buy a .ghost domain?
You can’t yet. A .ghost name can only be registered once the string is delegated (added to the DNS root) and its registry opens. Last round's closest parallel to .ghost was .ninja, delegated on 28 December 2013.
Is .ghost a real TLD?
Not yet. .ghost is not in the DNS root, so nothing resolves under it. It’s here because .ghost is on ICANN’s Reveal Day list of applied-for strings. Every entry on this page links to its source.
- Starlight (Namecheap) · first recordedICANN
- Oinkadot · now confirmedICANN
- Oinkadot · first recordedDomain Name Wire (Andrew Allemann)
“CEI2647T-T96247 · Castor Exploration, Inc. · replacement string of .droid · Deactivated OI2692T-T80659 · Oinkadot Inc · primary”ICANN
“A new partnership has applied for 25 top level domain names in this year's ICANN TLD expansion round.”Domain Name Wire (Andrew Allemann)
- Band chosen: 30-49, usable but limited, at its low edge. The string is a single evocative dictionary word with several qualitative demand pockets: seasonal and horror culture, gaming, ghost kitchens, ghostwriting, and a privacy/anonymity connotation that could appeal to security and pseudonymity tools; there is also existing software-brand affinity for the word, which suggests some defensive and identity interest (a qualitative inference, not a verified fact). By analogy with other playful dictionary-word TLDs (e.g., novelty and hobbyist strings), such spaces have historically produced price-driven registration volume, weak renewal quality, and thin aftermarkets, so registrations at volume are plausible but a strong brand signal is not; that analogy is flagged as such, and specific volumes are unknown. It is not in the 50-69 band because there is no identifiable category where .ghost would plausibly become a sustained primary-identity choice needing only light explanation; the privacy angle is speculative and contested by established alternatives. It is not in the 10-29 band because the word's multiple commercial uses (ghost kitchens, ghostwriting, gaming, privacy tooling) plus the applicant context, stated neutrally, that both Oinkadot principals run consumer registrars, which favors retail distribution and promotion, make volume registrations and occasional genuine end-use more likely than for a pure joke string. Main market uncertainty: whether the privacy/anonymity connotation converts into real primary-identity adoption by security and pseudonymity products, or whether demand stays seasonal and novelty-driven with weak renewals.