|
||
|
||
On October 7, ICANN published the full list of applications for its 2026 round of new domain extensions. There are 1,615 in total, including 333 from brands and 16 from communities. Everyone is busy counting names and picking winners. I keep thinking about something else: who is actually going to use all of these?
The 2012 round had 1,930 applications for 1,409 different extensions. This time the total is lower, but the fights over each name look harder. ICANN’s list shows 263 extensions where more than one applicant wants the same name.
One independent analysis counts 891 applications, about 55% of the round, caught in these fights. In 2012 it was roughly 39%. That’s one analyst’s own count and not an ICANN number, so treat it as a rough guide.
The field is also lopsided. Of the 1,615 applications, 864 come from North America and 506 from Europe. Africa has just 16.
It wasn’t cheap to get in, either. Domain Incite reports the application fee was $227,000, and 47 of the 1,663 applications submitted never paid it.
In 2012, a crowded extension often ended in a private auction, and the losers shared the winner’s money. That’s not allowed this time. If applicants can’t settle the fight by switching to a backup name or dropping out, it goes to an ICANN auction or a community priority review.
Most applicants saw this coming. Nearly three-quarters filed a backup name. They can switch between October 8 and 21, and the final list is locked in on November 17. Applicants who want out can still withdraw after that and get 65% of the evaluation fee back, but only until November 27. The objection period starts on November 17 and runs until March 16, 2027.
The most fought-over extension is .agent, with 13 applications, according to ICANN’s own list. Google’s registry arm (listed as Charleston Road Registry), Meta Registry Holdings and OpenAI are all in the fight, along with Radix, XYZ.COM and eight others. .agi has 7.
The pattern is hard to miss. These companies don’t just want to protect a brand. They want to own a word that people might one day say to an AI.
Now look at where the money goes. In DN Journal’s latest chart, covering the two weeks ending October 4, 13 of the top 20 sales were .ai names. The biggest was Act.ai at $375,000. But the single biggest sale overall was a .com: Mogul.com for $1.3 million.
Buyers are curious about new extensions. When it’s time to write a big check, they still go for .com.
The first new extensions have been around for years now, and the DNIB Quarterly Report for Q2 2026 shows how it’s going. New extensions ended June with 52.9 million registrations, up 34% from a year earlier. That’s solid growth.
The catch is renewals. The latest combined estimate for new extensions is 31.0%. For .com it’s 74.9%. Two of the biggest new ones, .xyz and .top, are at 17.5% and 18.9%. So plenty of people register these names, and most don’t keep them.
Out of 401.6 million registrations worldwide, new extensions hold about 13%. .com alone holds more than 40%.
A registration only shows that someone bought a name once. A renewal shows they still wanted it when the bill came. On that test, the new extensions have work to do.
In my last CircleID piece, I wrote that AI answer engines seem to favor distinctive names over generic ones. I think the same idea applies here.
When an AI summarizes a whole category, a generic ending like .agent or .labs gives it very little to hold on to. A distinctive brand name on a well-known extension gives it something specific to repeat. The name in front of the dot may matter more than what comes after it.
If you’re buying a name, don’t assume a new extension is an upgrade. Ask yourself whether the name would still stand out without it.
If you’re investing, remember that a crowded fight isn’t the same as value. Thirteen bidders for one extension only proves thirteen companies want to run it. It doesn’t prove anyone will type it or renew it.
If you work in policy or run a registry, watch renewal numbers more than launch numbers. The objection period starting November 17 is a good chance for the community to ask which extensions will really serve users, and which mainly serve the people who applied.
The next few months will decide which extensions make it to the starting line. The real fights, the auctions, come later, and the winners will still have to build real businesses and convince customers to stay.
Sponsored byWhoisXML API
Sponsored byVerisign
Sponsored byDNIB.com
Sponsored byRadix
Sponsored byCSC
Sponsored byVerisign
Sponsored byIPv4.Global