Are all the good .com domains taken?

No. But the words you would reach for first are gone, and they are gone in the exact position you would use them. new in front of fifteen different nouns is registered fifteen times out of fifteen; the same word moved to the back is available twelve times out of fifteen. We checked every combination live against the registries on 8 September 2026 to find where that line falls.

Permutica generates domain names by crossing word lists and checking every combination against the registries, so we run a lot of these. The line between gone and available does not fall where you would expect. It falls between individual words, and then again between the two positions a single word can occupy.

Which .com words are actually taken — and where

Take new. In front of fifteen different nouns, so newledger.com and fourteen like it, every single one is registered. Not most of them. All fifteen.

Now flip it. Of ledgernew.com and its fourteen siblings, twelve are free.

The fifteen words come from across Permutica's lists — new, my, your and we from the prefix list, digital, modern and refined from adjectives, pixel and daemon from tech nouns, security and waterproofing from trades — while all fifteen partners come from one list, brandable nouns. That is deliberate, so that only the word being judged varies.

The flip is the finding. new is not a used-up word in general. It is used up in front, because in front is where everybody puts it, and ledgernew.com survives only because nobody wants to be called that.

One word is gone both ways. digital returned nothing in either position, zero out of thirty. Twenty-five years as the obvious modifier in every industry at once will do that.

Availability by category: the same split everywhere

Each word below was crossed with the same fifteen nouns: instinct, remix, anchor, cornerstone, vault, canvas, pivot, totem, dominion, chart, anvil, lens, registry, tome, relic.

Read a row across. securityinstinct.com and securityvault.com are registered; waterproofinginstinct.com and waterproofingvault.com are not. Same second word, same construction, same day. pixelinstinct.com is gone and daemoninstinct.com is free.

The lucrative trade is worked through and the unglamorous one is not. Nobody has been down the daemon row at all. None of these are names anyone was hunting for, which is the point. The only thing separating them is how many people got as far as that word.

Contested partners exist too. daemon misses one of its fifteen, and the one it misses is daemonvault.com, which is also the one vault takes from almost every other word we tried.

Is .io or .co any better? We measured that too

The standard advice when a .com is taken is to take the same name on an alternative TLD. So we ran the five most exhausted modifiers again — same fifteen partners, same day, leading position — against .io and .co.

modern is effectively unavailable in front on .com and completely open on .io. Across the five words, .com gives up two names out of seventy-five and .io gives up fifty-seven.

That answer comes with a catch. The gap reflects demand rather than registry policy: .io is open because far fewer people have gone looking there. The same measurement that says you can have the name says how little competition there was for it. Whether that trade is worth making is a business question rather than an availability one, and it is at least now a question with a number attached.

Why an available .com is not the same as a good one

Every domain here is an arbitrary combination. newledger.com is not a name anyone set out to own and neither is daemoninstinct.com. That is what makes the comparison work: same shape, same length, equally unwanted in the abstract, one registered and one not, with demand for the word as the only variable left. So read these as relative demand for a word, never as a shortlist. A free domain is not a good name, it just means nobody has been down that row.

This also sets a ceiling on what any of this can do for you. The names you would actually choose are not sitting in the registry waiting; they are registered, and where they are for sale at all they are on the aftermarket as premium domains, at prices no registrar sets. Availability tells you which words are uncontested. It cannot tell you which name is worth having.

The comparison breaks when the two sides are not the same shape. We also tested northdakotaactive.com, northdakotaelite.com and three like them: all five unregistered, against four of five taken for Georgia. Georgia is plainly more contested, but North Dakota is not "open" in any sense you can use. Nobody wants a 21-character state adjective domain, and that is most of why they are sitting there.

An earlier version of this piece tested five partners per word and reported seven words at zero. Three did not survive widening: vision went from 0/5 to 5/15 on the same test, which makes that zero a draw rather than a property. Everything above is fifteen partners in each position. Fifteen is still a sample, so read one row as a strong hint and the table as the finding.

How to find a good .com domain that isn't taken

Assume your first instinct is the exhausted one, and that the arrangement you put it in is exhausted too. The words that occur to you immediately occurred to everyone else immediately, in the same order.

That holds for whole vocabularies, not just single words. We crossed eight of Permutica's word lists against the same fifteen brandable nouns, sampling 120 combinations from each.

Mythology roots measured second-best here at 71% and was removed from the product on 10 September 2026 anyway. Availability was never the problem with it; the names it produced were. toneluna.com and wispcupid.com are free because they mean nothing, which is the same reason the longest bars above should be read as a map of where there is room rather than of where there is value.

The shortest bar is the list of words everyone reaches for: get, try, use, go, top, best, super, smart, my, your. Seven lists sit between 53% and 74% and one sits at 11%. Choosing a different register is a bigger lever than choosing a different word. It is a map of where there is still room to work, not a quality ranking: a word is free most often because fewer people wanted it, and that applies to a science term as much as to anything else.

We rebuilt that list on the strength of this, which is the honest disclosure: thirty-one replacement lead-ins were screened the same way and kept only at 10/15 or better, among them steer, scout, glean, pyro, thermo and winnow, which moved the list from 11% to 38% on a re-run the same day. Twenty-one candidates were measured and dropped, including ones that sounded obscure enough to be safe. terra came back 2/15 and omni 3/15.

Updated 10 September 2026. That fix was half right and we have since undone the other half. Screening on availability alone admitted words that are free because nobody wants them, and pyro, thermo and winnow were three of them: crossed with a noun they produce pyroledger.com and winnowinbox.com, which no one would ship. Twenty-four such entries have been removed. The screening rule went with them, because it was also excluding the words people actually reach for, and those are worth offering even when most combinations are gone. find scores 2/15 leading and open and play 0/15; all three are back in the list. Availability is decided per candidate against the live registries at search time, which is the only place it can be decided honestly, and a heavily-used word simply returns fewer hits. The 11% and 38% figures above stand as measurements made on the dates given.

None of that was predictable from the outside, which is the argument for the procedure rather than the instinct:

1. Start from the specific noun, not the generic modifier. The exhausted half of every pair we tested was the modifier. waterproofing scored 14/15 against the same fifteen nouns that security scored 6/15 against, so the noun was never the problem.
2. Move the word before you drop it. new, your and pixel all go from nothing available to most of them by changing position alone. It costs nothing to test, and no amount of thinking about the word would have told you.
3. Swap within the class rather than abandoning the idea. latent, refined and glacier are the same kind of word as modern and scored 10/15, 12/15 and 13/15 leading, where modern scored 1/15. You do not need a different concept, you need a less-used word for it.
4. Drop the category words outright. digital and my returned nothing usable in front. A word that has been the default modifier in every industry for twenty-five years has no space left in any of them.
5. Watch for the one partner that ruins a row. vault was taken against nearly every word we paired it with. A single contested second word can make a healthy first word look dead.
6. Check the combinations, don't guess them. Four thousand checks take about as long as one, and nothing in the tables above would have been predictable from the outside. A domain name generator that hands you a list without checking availability is showing you names that are mostly registered.
7. Re-check before you buy. These figures were true on 8 September 2026 and availability decays. The gap between deciding and registering is where names get lost.

If you want to run this rather than read it, the pattern behind every number here is below. Two of Permutica's word lists crossed on .com, checked live. Browse domain ideas by industry if you would rather start from a category than a word, or check a specific domain if you already have one in mind. If you are naming a local trade business rather than a brand, the city-plus-trade question has a much harsher answer. And if the length is what you are really after, we measured which lengths of .com are gone — every two-, three- and four-letter one is.