metricname

Guidesby the numbers

Choosing .com vs .io vs .ai for a Startup — What the Sales Data Says

Most TLD advice is branding vibes. This is the other half: measured resale medians per extension (the same word sells at ~4× more on .com than .io), the ccTLD risks nobody mentions, and a decision procedure that uses data where data exists.

Start with what the extensions actually are

Most naming advice treats .com, .io, and .ai as three flavors of the same product. Administratively they aren’t. .com is a generic TLD operating under ICANN’s gTLD contracts. .io is the country-code TLD of the British Indian Ocean Territory and .ai is the country-code TLD of Anguilla — both adopted generically by the tech world, but each governed by its own registry’s rules and pricing rather than by gTLD contracts. Most of the practical differences below — renewal prices, policy stability, even whether the extension exists in twenty years — flow from that distinction.

The measured resale gap nobody quotes

A startup’s domain is also an asset it may one day sell, and here the data is unambiguous and unfashionable. Holding name quality fixed — the same dictionary-word class, measured across reported NameBio sales — the median one-word .com sale is $1,711. The same class on .ai has a median of $6990.41), and on .io it’s $4040.24) — roughly a quarter of the .com figure. Our TLD comparison tool shows the full table with sample sizes, and .com’s dominance runs deeper than medians: it carries roughly 72–74% of all reported aftermarket dollar volume (NamePros’ NameBio-based market analysis).

Two honest footnotes on that data. First, medians, not averages — the “.ai averages six figures” headlines are single mega-sales dragging a mean; the median .ai sale is $699, and any average quoted without a sample size deserves distrust (more on that habit here). Second, .ai’s sample is the largest in our table — 3,923 reported sales against .com’s 1,163 in the same class — so the boom is real in volume. Volume just isn’t price: more .ai names are changing hands than ever, at medians well under half of .com’s.

The branding factors — judgment, labeled as judgment

None of the following is measurable the way sale prices are, which is exactly why most TLD advice consists of it. Labeled accordingly:

  • The type-in default. When someone hears your name spoken, many will type .com first. That’s the strongest practical argument for .com, it compounds with every ad and podcast mention, and it’s why “the .com of your name is a competitor or a parked page” is a real cost even when your .io is the brand.
  • Audience signaling. To developer and AI audiences, .io and .ai read as native; to the general public they read as a typo risk. If your buyer is technical, the signal is fine. If your buyer is everyone, the default wins.
  • Search ranking is not a factor. Google has stated directly that keywords in a TLD confer no ranking advantage, and that generically-used ccTLDs like .io are treated as generic rather than geo-targeted. Anyone selling you a TLD for SEO is selling something else.

The structural costs of the ccTLDs

  • Renewal prices run at multiples of .com’s, set by each registry and subject to change — .ai in particular registers in multi-year minimums at rates several times a .com renewal. Read the actual figure off the registrar checkout, and if the name is speculative inventory rather than your brand, run that figure through our renewal calculator before committing to years of it.
  • A ccTLD’s existence is tied to its country code. Codes have been retired before when territories changed status, on multi-year transitions. .io’s territory has been the subject of a widely covered sovereignty handover; no change to the TLD has been announced, and none may ever come — but it is a category of risk .com structurally cannot have, and a fair thing to know before you put a brand on it.
  • Registry policy is the registry’s. Pricing, dispute rules, and transfer mechanics on a ccTLD follow the registry’s own terms rather than the gTLD baseline — usually invisible day to day, occasionally the whole story.

A decision procedure that uses data where data exists

  1. Run your name through the TLD comparer — it appraises the identical name across all six measured extensions with the same engine, so you see the asymmetries honestly: strong names hold value everywhere but concentrate it on .com; weak names floor out on the low-multiplier TLDs.
  2. If a clean, pronounceable .com of your name is gettable, the data says take it — highest medians, the type-in default, no registry risk. This is the one point where the measured and the judged agree completely.
  3. If it isn’t, decide what you’re buying. A .io or .ai as your operating brand is a rational purchase priced mostly in the judgment factors above. The same name as an investment is priced by the resale table — and the table says a quarter to two-fifths of .com, before renewal costs that run higher. Buying the ccTLD and believing the .com multiple is how naming decisions quietly become bad trades.

What no table can tell you

Whether your company’s name on .ai will matter more than the median, whether the .com owner will sell to you in two years for less than the rebrand would cost, whether the audience you have in year five looks like the one you’re imagining now. The medians describe what happened to thousands of names that weren’t yours. They’re the right starting point precisely because they’re nobody’s story — and the decision, as usual, is the part we don’t do for you.

This guide is for informational purposes only. It is not financial, legal, or investment advice, and it is not a certified appraisal. A domain’s real price is set by what a specific buyer actually pays — no article or model can know that in advance, and we say so instead of pretending otherwise.

Last reviewed: July 2026 · Against the NameBio-measured TLD medians behind our comparison tool (retrieved July 2026) and the sources linked in the body.