Thirteen separate applicants, more than for any other string in ICANN’s current new generic top-level domain expansion round, are competing for the rights to run .agent. Domain Name Wire’s rundown of the applicants reads like a who’s-who of companies betting on AI agents becoming a defining part of how the internet works, and the list alone is a reasonable barometer of how seriously the biggest names in tech are taking that bet.
For a small business owner who registers domains for a living, so to speak, without ever thinking about ICANN’s application process, the honest question this raises isn’t “should I get a .agent domain,” it’s “does any of this actually affect me, and when.”
Who’s Actually in the Running
The applicant list is heavy with names that need no introduction. Google, through its Charleston Road Registry subsidiary, is applying, and the company has history here: it reportedly paid $25 million to win the .app string in an earlier expansion round, which gives a sense of what a contested, strategically important TLD can cost a company willing to outbid everyone else. OpenAI is also in, applying for .agent alongside .gpt, .agi, .asi, and several others, which tells you this isn’t a one-off; it’s a deliberate land grab across every string connected to the company’s core brand territory. Meta is pursuing five generic strings including .agent, .aiglasses, and .metaverse. Alibaba, EchoStar’s DishDBS, Intercap, Radix Technologies, XYZ.COM, Meitu-linked FireLabs Capital, Nano Labs, Internet Land & Cattle, and Identity Digital round out the rest of the field.
One applicant stands apart from the rest: Open Agent Registry, Inc., a community-focused bid pursuing something called Community Priority Evaluation, a mechanism that can let a qualifying community-based application win over a straight commercial auction if it meets a fairly demanding set of criteria. Domain Name Wire’s own assessment is that success looks unlikely for that path here, which matters because it means the realistic outcome is a commercial auction among a field of well-capitalized tech companies, not a community-governed registry winning on the strength of its mission statement.
What an Auction This Contested Actually Costs
When this many well-funded applicants want the same string and there’s no way to split it, ICANN’s process resolves it through an auction, and the estimates here aren’t modest: the author estimates the auction could easily exceed $50 million, and one reader commenting on the piece guessed a bidder could go past $100 million, which is speculation rather than reporting. Only two of the thirteen applicants, XYZ.COM (which has .village as a fallback) and Identity Digital (which has .kernel as a fallback), have named a replacement string they could switch to, according to the article.
That’s a number worth sitting with, because it immediately tells you who this fight is for. It is not a contest a small business, a regional reseller, or even a mid-sized SaaS company could realistically enter. This is three or four of the largest technology companies on earth deciding which of them gets to own a piece of naming infrastructure for a category of software they all believe is about to matter enormously.
The Timeline Is Longer Than the Hype Suggests
It’s easy to read coverage like this and feel a flicker of urgency, the same instinct that drove early land grabs for .com names decades ago. The actual ICANN process doesn’t move that fast. A proposed new TLD has to pass applicable reviews and reach a formal delegation decision before a registry can even begin operating it and selling registrations, a process that has taken years from application to launch for strings applied for in the 2012 round, not months. ICANN’s next concrete milestone here is November 17, 2026, when it discloses whether any applicants dropped or switched their applications to a different string, which is itself just a procedural checkpoint, not a resolution of who wins.
Anyone encountering a site offering “free pre-registration” for .agent domains right now is encountering speculation about a namespace that doesn’t exist yet, run by an entity with no guarantee it will ever control that namespace at all. That’s worth treating with real skepticism rather than urgency; a pre-registration list for a TLD years away from delegation, run by a party that may never win the string, has no guaranteed path to becoming an actual working domain.
What This Actually Means for a Small Business Right Now
The realistic takeaway is that .agent isn’t a decision point for a small business today, and won’t be for a while even after the auction resolves. What the story is genuinely useful for is as a reminder of how new TLD land grabs tend to play out: a handful of deep-pocketed companies fight expensively over the generic, category-defining string, a registry eventually launches it years later, and a long tail of speculative registrations follows, most of which never develop into anything a customer actually searches for or trusts the way they trust a .com.
For an actual domain decision happening today, the more useful question is which existing, already-established TLD best matches the business and the audience actually being served, not which speculative new string might matter someday. A .com remains the default a typical customer types without thinking, and for anyone weighing it against alternatives, registering a domain in an established extension that customers already recognize and trust is a far more reliable move than waiting years for a category-defining TLD to finish being fought over by some of the world’s biggest tech companies, and then hoping it catches on with actual buyers once it finally exists.
Watching Without Waiting
None of this means the .agent story is irrelevant to watch, particularly for anyone building software that genuinely operates in the AI agent space rather than just selling to small businesses that might someday use one. The eventual outcome will say something real about how the biggest platforms expect agent-related naming and trust infrastructure to work. But watching a multi-year ICANN process unfold is a different activity than acting on it today, and the two are easy to conflate when a headline makes a years-out auction sound like next quarter’s opportunity.