The AI TLD Land Rush: How To Bet On .CHATGPT, .AGENT, .MODEL And Other New AI Domains Before Pricing Explodes

If you looked at the ICANN 2026 Reveal Day list and felt your brain lock up, you are not alone. One minute, AI naming was mostly guesswork. The next, strings like .chatgpt, .gpt, .agent, .model, .claude and .agi were sitting on a real application list, with big companies attached and real money likely to follow. That is exciting, but also annoying, because this is how people end up panic-buying junk names or waiting so long that the smart options get snapped up first.

Here is the good news. You do not need to predict the entire future of AI branding. You just need a simple filter. Start with what is actually on the ICANN board, sort those strings into likely winners, risky but interesting bets, and probable dead ends, then build a short shopping list of second-level names before registrars and trademark teams flood the market. The goal is not to buy everything. The goal is to buy the few names that still make sense when the hype wears off.

What actually happened on ICANN Reveal Day

Reveal Day matters because it turned rumor into paperwork. We now know which AI-flavored top-level domains were actually applied for, and by whom. That changes the game.

Instead of guessing whether .gpt or .agent might someday exist, investors can now work from a real list. OpenAI reportedly showed up with a broad set of applications, including .chatgpt, .gpt and .agi. Google put real weight behind its own slate, helping push strings like .agent into the hottest part of the round.

That does not mean these extensions go live tomorrow. It means the market now has a map. And when a market gets a map, prices tend to wake up fast.

The first rule: not all AI domain extensions are equal

This is where people get tripped up. They see a flashy string and assume it is a great domain investment. Sometimes it is. Often it is not.

Category 1: Broad, usable terms

These are the names with the best chance of becoming real working neighborhoods on the internet. Think .agent and .model. They are generic. They are flexible. They can fit startups, tools, directories, media projects, consultants, and software products.

If one of these launches as an open registry with fair access, the second-level names could become very valuable. Examples might include support.agent, hire.agent, legal.agent, image.model, voice.model, or train.model.

Category 2: Big-brand adjacent terms

This is where .gpt sits. It is famous, yes. It is also messy. The term has become close to generic in daily speech, but it is still strongly tied to OpenAI in the public mind. That can create demand, but also policy risk, reserved-name risk, and legal headaches.

There may still be opportunities here, but you should assume tighter controls and less room for ordinary buyers.

Category 3: Pure brand TLDs

.chatgpt and .claude are the obvious examples. These sound exciting, but they may be the least useful for outside investors. A company may run them as closed spaces for its own products, support pages, developer tools, or marketing campaigns.

If that happens, you are not really shopping in a public neighborhood. You are peering through the fence at a private estate.

Your best bet is usually second-level strategy, not TLD fantasy

Most readers will make more money, or save more money, by planning for second-level names under the right extension than by daydreaming about owning the extension itself. Unless you are a registry operator with very deep pockets, the practical play is to identify which future TLDs are likely to be open and useful.

Think of it this way. The extension is the shopping mall. Your opportunity is the best storefront inside it.

How to rank AI domain extensions before you spend a dime

1. Check whether the term is generic or brand-specific

Generic wins more often. .agent and .model have broad commercial use. .chatgpt is highly specific. That usually means more restrictions and fewer chances for ordinary registrants.

2. Look at the applicant

If a company applies for a term closely tied to its own brand, assume it may keep a tight grip on it. If multiple large applicants are fighting over a broad string, that often signals stronger long-term value, even if launch timing gets delayed by contention.

3. Ask a boring question: who would actually buy names there?

This is the question that saves you from dumb purchases. Would agencies, SaaS tools, job boards, newsletters, APIs, consultants, marketplaces, and media brands want a name on that extension? If the answer is yes across several categories, that is a healthier market.

4. Think about speech

If the domain sounds natural when spoken aloud, it has a better chance. “support dot agent” or “studio dot model” feels cleaner than many novelty endings. Good radio test names tend to age better.

5. Price in launch friction

Even a great extension can be a bad buy if premium pricing gets silly. Some of the best opportunities disappear not because the name is gone, but because the registry turns every decent keyword into a four-figure annual bill.

The short shopping list: where I would focus first

.agent

This is the standout for many buyers. It is broad enough to cover AI agents, customer service agents, travel agents, talent agents, workflow agents, and software agents. That gives it more than one story, which is healthy.

Names to watch: build.agent, sales.agent, home.agent, task.agent, research.agent, fix.agent, deploy.agent.

.model

This one is quietly strong. It works for AI models, 3D models, fashion, simulation, machine learning tooling, and marketplaces. It has range, and range matters.

Names to watch: vision.model, open.model, train.model, voice.model, data.model, local.model.

.gpt

Interesting, but more complicated. If this becomes open at all, strong descriptive names could carry value. But I would treat it as a selective, speculative basket, not the core of a portfolio.

Names to watch if available and compliant: support.gpt, legal.gpt, code.gpt, tutor.gpt. But go in expecting restrictions.

.agi

Potentially powerful as a prestige term, though narrower and more abstract. This feels more like a branding play for labs, think tanks, research groups, and ambitious startups than a mass-market namespace.

Names to watch: safety.agi, lab.agi, future.agi, ethics.agi.

.chatgpt and .claude

I would not make these your main hunt. They may end up valuable inside their own ecosystems, but they are not the cleanest target for ordinary domain buyers.

What to avoid right now

Do not buy based on hype words alone

“AI” plus a random noun is not a strategy. If the phrase would not make sense to a business buyer in two years, skip it now.

Do not assume every applied-for TLD will be open to the public

This is a huge mistake. Some strings may be tightly controlled, delayed, challenged, or never marketed broadly.

Do not ignore trademark risk

This is not just legal fine print. If a name depends on somebody else’s famous product mark, your resale options can shrink fast.

A simple buying framework for Domains Tip readers

Here is the plain-English version.

1. Make three lists: likely open generics, likely restricted names, and avoid-for-now strings.
2. Put .agent and .model at the top of your likely open generics watchlist.
3. Treat .gpt and .agi as selective speculation.
4. Treat .chatgpt and .claude as mostly brand-controlled unless proven otherwise.
5. Pre-build your second-level targets now, before landrush pricing and premium labels hit.
6. Focus on names with real end users. Think product categories, use cases, jobs, tools, and outcomes.

If you are stuck, use this test. Would a startup founder proudly print the name on a slide deck? Would a buyer understand it in two seconds? If yes, keep it on the list.

How pricing can explode, even before the public cares

People often imagine domain price jumps happen after TV ads or big launch events. In reality, the first pop often comes earlier. Registry insiders, brand protection teams, domain investors, and founders all start moving once the application list becomes public and the likely winners become clearer.

That means the narrow window is now. Not necessarily to register tomorrow, because some of these strings are still in process, but to decide what you want before registrars layer on premium pricing and before your best keyword targets become obvious to everyone else.

What I would do this week

Start a spreadsheet. Seriously. This is one of those moments where a boring spreadsheet beats gut instinct.

• Column 1: TLD string
• Column 2: Applicant
• Column 3: Open or likely restricted
• Column 4: Best second-level targets
• Column 5: End-user categories
• Column 6: Trademark or policy risk
• Column 7: Your maximum price

Then trim hard. If you begin with 100 ideas, cut to 20. The people who do best in these windows usually buy fewer names, not more.

At a Glance: Comparison

Feature/Aspect	Details	Verdict
Best broad-use AI extension	.agent has the widest commercial fit, with demand from AI tools and non-AI businesses alike.	Top watchlist candidate
Best under-the-radar option	.model works across AI, design, simulation, and fashion, which gives it several paths to adoption.	Strong practical bet
Highest hype, highest friction	.chatgpt and .gpt may draw huge attention, but access rules, premiums, and trademark concerns could limit real opportunity.	Proceed carefully

Conclusion

Reveal Day did not magically make every AI domain a winner. What it did do is give you a clean starting point. That is valuable. We now have a concrete public list, including OpenAI’s bids for strings like .chatgpt, .gpt and .agi, plus Google-backed competition that makes names like .agent especially important to watch. For the Domains Tip community, the edge is not in chasing every shiny new ending. It is in turning fresh ICANN data into a short, focused shopping list of AI domain extensions and second-level names that real buyers may want later. If you stay picky, respect trademark and access risk, and move before retail pricing catches up, you give yourself the best shot at finding value before the easy upside gets drained.