Naming a tech company used to come with a familiar set of advice.
Keep it short. Make it brandable. Avoid anything too generic. Invent a word if you have to. And if the domain you want is gone, add “get,” “try,” “app,” “hq,” or another small linguistic spare part until something becomes available.
But what are people actually doing now?
We analyzed 6,420 .tech domains registered between January and July 2026 to find out.
The dataset reveal a remarkably consistent naming culture.
Clean names dominate. Descriptive names don’t. AI companies largely avoid putting AI in their names. The average name sits inside a surprisingly narrow length range.
And the most popular solution to the naming problem, inventing a brand-new word, appears to be creating a new problem of its own.
Five in six founders got a clean domain name on .tech
The first thing that jumps out is how little scaffolding these names need.
83.4% of the domains we analyzed have no modifier at all.
No “get” or “try” at the beginning. No “app,” “hq,” “labs,” “online,” “ai” or similar addition at the end.
95.3% are also a single unbroken token. Only 4.7% contain a hyphen and 2.8% contain a digit.
That’s notable because modifiers have traditionally been the escape hatch of domain naming.
If the perfect domain isn’t available, Acme becomes GetAcme, AcmeApp, or AcmeHQ. The brand gets the name it wants, while the domain carries the compromise. That pattern is largely absent here.
There is an important caveat. We can see the resulting domains, not the decision that produced them. A company that wanted “Harlow” and settled on “HarlowWorks” would still appear as a clean name in this analysis.
But the broader pattern is hard to miss: the overwhelming majority of these domains don’t look patched together to make the registration work.
For a founder currently staring at get{startupname} because {startupname} isn’t available, that’s worth knowing.
A modifier might solve the registration problem today. You still have to introduce yourself with it tomorrow.
With .tech, the domain name doesn’t have to explain the product; tech says it all!
If the names aren’t full of domain workarounds, perhaps they’re being used to explain what the company does.
Mostly, they aren’t.
84.5% of the domains contain no obvious category word such as health, pay, data, secure, robot, learn, or ship.
Only 15.5% do.
Even “tech” itself is rare. Just 85 of the 6,420 domains, or 1.3%, include “tech” in the name before .tech.
There are several possible explanations for the broader pattern.
Companies may want room to move into adjacent products later. They may be naming themselves before the product is fully settled. Or descriptive names may simply be harder to secure.
Our data can’t tell those explanations apart.
What it can tell us is that descriptiveness isn’t the default.
That matters if you’re currently trying to make a company name carry the entire pitch deck.
Your name doesn’t necessarily have to explain what you do. That’s what positioning, copy, design and the product itself are for.
And when the domain extension already supplies context, repeating that context in the name can become redundant. somethingtech.tech is doing twice the work to say the same thing.
Nearly a third are building with AI. But most don’t put AI in the name.
This is where the pattern gets especially interesting.
1,865 domains in the dataset described work involving AI.
But among those AI-related sites, 84.2% did not have “ai” in their name.
In other words, nearly three in ten of the sites we analyzed talked about AI in their positioning, while only a minority of that group chose to bake AI into the brand itself.
The pattern also isn’t confined to one corner of tech. AI-related work appeared across developer tools, crypto, healthtech, marketing and other categories.
There are several plausible reasons for leaving AI out.
One is longevity.
Tech startups have been through enough naming waves to know that today’s defining technology can become tomorrow’s invisible infrastructure. “Internet company” once told you something meaningful. Eventually, almost every company became an internet company.
AI may be following its own version of that trajectory.
We can’t know whether that’s why any individual company in the dataset chose its name. We can only observe the result: most of the people building with AI aren’t naming themselves after AI.
So if you’re putting “AI” into a company name simply because AI is central to the product today, it’s worth asking a second question:
Will you still want it there when AI is simply part of how software gets built?
60% of the names are coined or invented words
If companies aren’t using category words, what are they using?
Mostly, words that didn’t exist before.
60.0% of the names we analyzed are coined words. Another 32.4% are compounds. Real words with an affix account for 6.5%.
A single standalone real word?
Just 1.1%.
And coining isn’t being driven by one especially inventive category.
It appears everywhere.
Coined names account for 71% of edtech names, 69% of developer tools, 66% of cybersecurity and robotics, 64% of business software, 58% of fintech and 53% of hardware and IoT.
Inventing a word makes intuitive sense. You get something ownable, searchable, and potentially easier to trademark. You’re also not competing with every existing use of an ordinary dictionary word.
There is just one problem.
Everyone else had the same idea.
24.5% of the names in our dataset have another name within two character edits.
Almost one in four.
3.1% are within a single edit.
The dataset contains pairs such as revyn/nevyn, erivo/nerivo, viddy/vidsy, deepdock/deeplock, keelstone/keepstone, forgeos/forgeops, and cdash/ddash.
None of these necessarily resulted from copying. In fact, the more interesting possibility is the opposite.
Give thousands of people the same naming brief:
Short. Pronounceable. Modern. Brandable. Invented.
And independent naming processes can begin producing remarkably similar answers.
The crowding hasn’t disappeared.
It has moved.
The vocabulary changes depending on what’s being built on .tech
The naming patterns may be broadly similar across tech, but the vocabulary isn’t.
Certain words and fragments show up more prominently depending on what a company is building.
Crypto and Web3 names lean toward sol, token, chain, and node. In fintech, pay, trade, and flow appear repeatedly. Healthtech gravitates toward health, med, and care, while developer tools use agent, dev, and api.
Marketing has its own cluster around seo, rank and digital. Data and analytics names bring up data, dash and path. In robotics, it’s aero, drone, and robotics.
That makes intuitive sense. Even when a company doesn’t use an explicitly descriptive name, the language of its category can still find its way into the brand.
But another interesting part of the category data is that the naming structure changes much less than the vocabulary does.
Coined names lead in every major category we analyzed, from 53% of hardware and IoT names to 71% in edtech. The median name is nine characters in most categories, dipping to eight in crypto, data and analytics, and hardware and IoT.
Different parts of tech seem to develop their own naming dialects without creating completely different naming rules.
| Category | Words that show up more often |
|---|---|
| Crypto & Web3 | sol, token, chain, node |
| Fintech | pay, trade, flow |
| Healthtech | health, med, care |
| Developer tools | agent, dev, api |
| Marketing | seo, rank, digital |
| Data & analytics | data, dash, path |
| Robotics | aero, drone, robotics |
The real naming white space might be real words
There is another curious gap in the dataset.
Standalone real-word names make up just 1.1% of all 6,420 domains.
They remain below 2% in fourteen of the sixteen categories we examined. Hardware and IoT, creator tools, and ecommerce had none. Crypto was the highest at 3.1%. Fintech was the lowest at just 0.4%.
That doesn’t automatically make real-word names better.
They come with obvious tradeoffs. They can be difficult to trademark. Search intent may already belong to something else. The matching domain may be unavailable. A familiar word can also arrive carrying meanings you didn’t ask for.
But one common objection deserves another look: that real words are simply too generic to stand out.
Our dataset suggests the naming crowd has gone somewhere else.
Even when real words appear inside compounds, they tend to be used literally. A name such as Classview describes something related to classrooms. Agentstatus monitors agents. Wirekit works with wiring. Supportgraph deals with support knowledge.
The much rarer move is the evocative one: taking an existing word and allowing the company to give it a new association.
Think less “word that describes the software” and more “word the software eventually owns in your head.”
We haven’t quantified that distinction because deciding whether a name is literal or evocative requires your taste and judgment. But qualitatively, reading through the dataset, the territory looks remarkably empty.
For founders willing to accept the trademark, search, and availability tradeoffs, that makes it worth investigating.
Nine characters, three syllables: tech names have found their shape
For all the creativity involved in inventing names, their physical dimensions are remarkably consistent.
The median name in the dataset is nine characters and three syllables long.
62.9% fall between seven and twelve characters.
Another 21.2% are six characters or fewer, while 15.9% run beyond twelve.
There are some category differences. Crypto, data and analytics, and hardware and IoT run about one character shorter than the rest.
But overall, the naming silhouette barely moves.
Short is still desirable. Ultra-short is another matter.
Some of the shortest names in the dataset are strings such as aemz, cgqa, btpx and ayvu. At a certain point, removing another character stops making a name more memorable and starts turning it into a Wi-Fi password.
Brevity has a floor.
The morphology of invented names also shows how naming fashion has moved.
Among the 3,849 coined names, 32.1% end in a vowel and 19.8% contain a doubled letter. “ai” is both the most common two-letter ending and the most common two-letter beginning.
One naming trick, however, has almost vanished.
The Flickr/Trackr-era habit of dropping a vowel appears in just 0.7% of coined names.
Some naming trends die loudly. This one appears to have quietly packed its bags.
What should a founder naming a company in 2026 take from this?
The data doesn’t reveal the perfect company name. It does something more useful: it shows you what thousands of other people naming things in tech are already doing.
And that gives you a better set of questions to ask.
Can you get a clean domain name?
83.4% of the .tech domains registered between January and July this year did so without the usual modifier words. Which may not be possible with .com.
So before settling for GetSomething or SomethingHQ, see if the extension you’re trying to force is the problem and needs changing.
Does your name need to explain your category?
Most don’t. 84.5% contain no category word. Clarity matters, but the name doesn’t have to carry it all. A domain extension like .tech, with a category cue, can often do the trick. As is evident in the dataset.
Are you naming the company after a technology that may become infrastructure?
Nearly a third of the dataset describes AI work, yet 84.2% of those names don’t say AI. Keep that in mind as you start naming.
Is your invented word actually distinctive?
Don’t stop at an exact-match search. Look for names one or two letters away, names that sound similar when spoken, and names that blur together on a search-results page.
One in four names in this dataset already has a near-neighbor within two character edits.
Does your name sound like everything else in your category?
Different parts of tech develop their own vocabulary. Crypto leans toward words like “sol,” “token,” “chain” and “node,” while developer tools use “agent,” “dev” and “api.”
Knowing your category’s language can help you decide whether you want your name to signal where you belong or deliberately step outside it.
And perhaps most importantly:
Don’t confuse following the naming playbook with being distinctive.
The dominant 2026 tech name is clean, compact, non-descriptive, and invented.
That may be exactly the right answer for your company.
But once 60% of everyone else reaches for the same answer, breaking one of those rules becomes important to stay distinctive.
Methodology
We analyzed 6,420 .tech domains registered between January and July 2026, drawn from commercial web data and cleaned of any parked pages, unedited templates, and domains listed for sale.
The dataset includes personal sites, consultancies, and small local businesses alongside companies, so findings describe naming patterns across the analyzed .tech domains rather than tech startups specifically.
AI classification is based on how sites described their own products and work, not independently verified technical capabilities. Category analysis was possible only where site descriptions provided enough information to classify the domain; 41% had descriptions too thin to classify.
Smaller categories should be interpreted cautiously. Rows with fewer than 150 domains carry margins of approximately ±9 percentage points or wider, and the robotics category, with 97 domains, should be treated as directional.
.Tech Domains and the .tech domain extension are managed and operated by Radix Registry.
