AI Domain Price Index 2026 — What AI Domains Cost
What AI and .ai domains actually cost in 2026: median $24,888 across 174 live asking prices, registration cost vs resale value, extension and length comparisons, open methodology.
AI Domain Price Index — 2026 edition. Data snapshot: 2026-08-10. Sample: 174 priced names. Methodology version 1.0. Published by BuyStartupDomains.
Most "domain price index" articles quote a precise median and never show you the underlying rows. This one publishes the dataset it is computed from, states plainly what the dataset is and is not, and recomputes every figure automatically from the live listings so the article can never drift away from the inventory it describes.
What this index measures — and what it does not
This is an index of asking prices, not sale prices. Every row is a name currently offered for sale on this site, with a published buy-it-now price you can open and verify. We do not have access to private or brokered sales data, and we do not estimate it.
That distinction matters more than any number below:
- Observed data — the asking prices, counts, medians and distributions in the tables. These are facts about a specific, disclosed inventory.
- Analysis — the comparisons between extensions, name lengths and name types. These are arithmetic on the observed data.
- Interpretation — the paragraphs explaining *why* a group prices where it does. This is judgement drawn from operating the portfolio, and reasonable people can disagree with it.
- Limitations — set out in full at the end. Read them before citing anything here.
Sample and date range
| Field | Value |
|---|---|
| Edition | AI Domain Price Index 2026 |
| Data snapshot | 2026-08-10 |
| Names in portfolio | 175 |
| Names with a published price (the sample) | 174 |
| Names excluded (price on request) | 1 |
| Price type | Buy-it-now asking price, USD |
| Methodology version | 1.0 |
The sample is the full priced inventory — there is no sub-selection, no trimming of outliers, and no weighting. Names priced on request are excluded because they have no published number to include; there is currently 1 of them (Sanctumary.com).
Headline figures
| Statistic | Value |
|---|---|
| Names in sample | 174 |
| Lowest asking price | $495 |
| 25th percentile | $9,888 |
| Median asking price | $24,888 |
| Mean asking price | $46,144 |
| 75th percentile | $80,000 |
| 90th percentile | $100,000 |
| Highest asking price | $280,000 |
The gap between the median ($24,888) and the mean ($46,144) is the single most useful thing on this page. It tells you the distribution is right-skewed: a small number of flagship names pull the average far above where the typical name actually sits. Any index that quotes only an average is misleading you. When you benchmark a name, benchmark against the median and the percentile band it belongs to.
Price distribution
| Asking price band | Names | Share of sample |
|---|---|---|
| Under $5,000 | 33 | 19.0% |
| $5,000 – $9,999 | 16 | 9.2% |
| $10,000 – $24,999 | 41 | 23.6% |
| $25,000 – $49,999 | 22 | 12.6% |
| $50,000 – $99,999 | 42 | 24.1% |
| $100,000 and above | 20 | 11.5% |
Read this as a shape rather than a set of numbers. There is a large entry band under $10,000, a dense working band between $10,000 and $50,000 where most funded startups actually transact, and a long upper tail. The upper tail is thin by count but holds most of the portfolio's total value.
By extension
| Extension | Names | Median | Mean | Range |
|---|---|---|---|---|
| .com | 139 | $24,888 | $47,756 | $988 – $280,000 |
| .bot | 30 | $17,555 | $36,868 | $495 – $150,000 |
| .ai | 4 | $50,000 | $70,000 | $50,000 – $100,000 |
*Interpretation.* .com carries the widest buyer pool and the deepest sample here, so its median is the most reliable figure in the table. The .ai and .bot samples are much smaller and are dominated by the specific names in them — treat those medians as descriptive of this portfolio, not as market-wide estimates. The broader point survives the small samples: .ai and .bot price at a premium *inside* their categories while narrowing the set of companies that would ever buy them.
### Registration cost vs resale price — two different numbers
People searching for what a .ai domain costs are usually asking one of two completely different questions, and conflating them produces wildly wrong budgets.
- Registration cost is what you pay a registrar for an *unregistered* name. The .ai registry sells in a mandatory two-year term, so the checkout figure is a multiple of the annual rate and lands materially above a .com registration. It is an annualised operating cost, not a capital purchase.
- Resale (aftermarket) price is what you pay the current owner for a name that is already taken. That is the number this index measures, and it is set by scarcity and buyer fit rather than by a registry price list.
Every figure on this page is a resale asking price. If a name you want is unregistered, register it and stop reading — no index applies. If it is taken, the registry price is irrelevant to what it will cost you, and the percentile bands below are the right benchmark. Short .ai inventory that is already registered and for sale is listed on the AI domains collection.
By name length
| Name length (excl. extension) | Names | Median | Mean |
|---|---|---|---|
| 1–5 characters | 21 | $50,000 | $43,558 |
| 6–8 characters | 25 | $9,888 | $37,268 |
| 9–12 characters | 88 | $20,000 | $37,389 |
| 13–16 characters | 32 | $79,888 | $76,655 |
| 17+ characters | 8 | $49,888 | $54,916 |
*Interpretation.* Length alone does not drive price monotonically, and this table is a good illustration of why simplistic "shorter is worth more" rules fail. Very short names carry a scarcity premium; the longest band here is inflated by exact-match multi-word category names such as robotics subfield terms, which are long *and* expensive because they name a research field precisely. What actually moves price is a combination of scarcity, category legibility and phonetic strength — not character count.
Keyword vs brandable
| Name type | Names | Median | Mean |
|---|---|---|---|
| Keyword / descriptive | 96 | $32,000 | $57,093 |
| Brandable / coined | 78 | $20,000 | $32,668 |
*How the split is defined.* A name is classified as keyword/descriptive if its second-level label contains any of the following tokens: ai, robotics, robot, bot, neural, reasoning, inference, cognition, perception, planning, navigation, sensors, dexterity, alignment, embodied, world, model, sovereign, data, privacy, medical, bio, judgment, crypto, invest. Everything else is classified as brandable/coined. This is a mechanical rule published so the classification is reproducible; it will misfile the occasional edge case (a coined name that happens to contain a token), and that is a known limitation rather than a hidden one.
*Interpretation.* Keyword names price higher in this portfolio because they are bought by companies that want to own a category term outright, and there is exactly one exact-match name per category. Brandables compete against an effectively unlimited supply of alternatives, which caps what any single one can command.
By category
| Category tag | Names | Median asking price |
|---|---|---|
| Swiss | 11 | $100,000 |
| Agentic | 32 | $38,000 |
| Biotech | 9 | $30,000 |
| Medical | 15 | $30,000 |
| Japan | 28 | $29,888 |
| Robotics | 46 | $29,888 |
| Dubai/UAE | 19 | $25,000 |
| AI | 132 | $24,888 |
| Privacy | 42 | $20,000 |
| Luxury | 22 | $20,000 |
| Bot | 34 | $12,888 |
| Fintech | 17 | $9,500 |
| China | 7 | $7,888 |
Category tags are assigned per name and a name can carry several, so the counts sum to more than the sample size. Categories with fewer than five priced names are omitted because a median over four rows is not informative.
Notable trends in the 2026 edition
Stated as observations about this dataset, with the caveats attached:
- Robotics has become the densest premium category. Robotics-tagged names are the second-largest group in the sample and carry a median above the portfolio median. This tracks the capital going into humanoid and embodied-AI companies in 2025–2026, which is publicly reported and not something we are inferring from our own data.
- The exact-match subfield name is the strongest single pattern. Names describing a specific research capability — perception, navigation, planning, cognition — sit consistently in the upper bands.
- Geographic anchoring holds value where sovereign AI spending is concentrated. Japan-, Gulf- and Swiss-tagged names price above the portfolio median. The underlying spending programmes are public; the naming demand that follows them is our interpretation.
- The sub-$10,000 band has not disappeared. Roughly a fifth of the sample still sits under $5,000. Good names remain available to unfunded founders, which is worth stating because index articles tend to imply otherwise.
Methodology
Collection. The dataset is the complete set of names listed for sale on BuyStartupDomains.com with a published buy-it-now price, read directly from the site's own listing data at the snapshot date. No scraping, no third-party feeds, no survey.
Inclusion. Every priced listing is included. Exclusion: names offered at "price on request" only.
Calculation. All figures are computed programmatically from the raw price list. Median and percentiles use nearest-rank on the sorted price array. Mean is the arithmetic mean, rounded to the nearest dollar. Group statistics recompute the same functions over the subset. Distribution bands are half-open intervals (lower bound inclusive, upper bound exclusive). Name length counts characters in the second-level label only, excluding the extension.
Reproducibility. Because the figures are computed from the live listing data rather than typed in, this page updates itself when inventory or pricing changes. The snapshot date above reflects the current edition of the analysis.
Limitations and sources of bias
Please read these before citing this report.
1. Asking prices, not transactions. Nothing here evidences what a name sold for. Realised prices are typically below asking, by a margin that varies with category and buyer urgency. 2. Single-portfolio sample. This is one curated inventory, not the market. It reflects the acquisition preferences of one owner — heavily weighted toward AI, robotics, and geographically anchored names — and is not a random sample of AI domains. 3. Seller-set prices. The prices are set by the seller publishing this report. That is an obvious conflict of interest and the reason we publish the full method rather than a headline number. 4. Small sub-samples. Several extension and category groups contain a handful of names. Their medians describe those names and little else. 5. No time series yet. This is the first edition. It cannot show year-over-year movement; that becomes possible from the 2027 edition onward. 6. Survivorship within the portfolio. Names that sell leave the inventory, so a snapshot slightly over-represents names that have not yet found a buyer. 7. Mechanical classification. The keyword/brandable split uses a token rule and will misfile edge cases.
Do not read any figure here as a valuation of your own domain. For a range rather than a single number, run the name through the appraisal tool and read the valuation methodology, which explains exactly how that range is produced and where it is unreliable.
How to verify prices independently
Do this before trusting anyone's index, including this one:
- [NameBio](https://namebio.com) — searchable archive of publicly disclosed sales. Filter by keyword and extension to see comparable *sold* prices.
- [DNJournal](https://dnjournal.com) — weekly top-sales reports going back two decades; the best public source for six- and seven-figure comps.
- Marketplace buy-it-now pages — Afternic, Sedo, Dan and Atom show current asking prices across many sellers, which is the right comparison set for the data on this page.
A caveat that applies to all three: only a minority of sales are ever disclosed, and brokered private transactions skew high. Public sale medians therefore understate the top of the market.
What founders should budget in 2026
Anchored to the percentile bands above rather than to invented figures:
- Pre-seed / bootstrapped: below the 25th percentile ($9,888). Descriptive multi-word .coms and patient negotiation on brandables.
- Seed: the 25th–50th percentile band, up to roughly $24,888. A solid brandable .com — where most funded startups land.
- Series A: the 50th–75th percentile band, up to roughly $80,000. A category-legible two-word name. Frequently less than one month of paid acquisition.
- Series B and beyond: above the 90th percentile ($100,000). One-word or category-defining names that compound every marketing dollar for a decade.
The wrong move at every stage is buying the cheapest name you can tolerate. The right move is the most defensible name the current round can absorb without regret.
Using the index in a negotiation
1. Pull three to five genuinely comparable *disclosed sales* from NameBio — same length, same extension, same category. 2. Discard the outliers at both ends; they are stories, not comps. 3. Use the percentile band above to sanity-check where the ask sits relative to a live market of comparable inventory. 4. Anchor your first offer near the low end of what remains and expect to land between the middle and the ask. 5. Ask about lease-to-own. On five-figure names it closes deals that a cash offer cannot.
About this data / cite this report
Title: AI Domain Price Index 2026 Publisher: BuyStartupDomains Snapshot date: 2026-08-10 Sample: 174 priced domain listings URL: https://buystartupdomains.com/blog/2026-ai-domain-price-index Licence: free to quote and reference with attribution and a link.
Suggested citation:
- BuyStartupDomains, *AI Domain Price Index 2026*, 2026-08-10. https://buystartupdomains.com/blog/2026-ai-domain-price-index
This is intended as an annual series — the 2026 edition establishes the baseline, and later editions will reuse the same sample definition and calculations so year-over-year comparison is meaningful. Journalists and researchers who need a breakdown that is not published here can request it.
Frequently asked questions
What is the median price of an AI domain in 2026? In this dataset of 174 priced AI, robotics and adjacent listings, the median asking price is $24,888 and the mean is $46,144. Those are asking prices from a single curated portfolio, not market-wide sale prices.
Is a .ai domain worth more than the .com? Usually no. Inside AI research and infrastructure .ai carries genuine signal, and a strong short .ai can clear more than a mediocre .com. Across the whole market the matching .com still commands the higher price and the wider buyer pool. Extension medians in this report come from small samples — read them as descriptive, not predictive.
How much does a .ai domain cost in 2026? Two different numbers. Registering an unregistered .ai costs a registrar fee billed over the registry's mandatory two-year term — an operating cost, higher than a .com registration but not a capital purchase. Buying a .ai that someone already owns is an aftermarket transaction priced by scarcity and buyer fit; in this portfolio the priced sample has a median of $24,888 across all extensions. Do not use one number to budget for the other.
Where can I see top .ai domain sales? Not here — this index reports asking prices from one portfolio and explicitly does not estimate transaction values. Publicly disclosed sale prices, including the largest .ai sales, are archived at NameBio and reported weekly by DNJournal. Only a minority of sales are ever disclosed, so both sources understate the top of the market.
How do I value a specific AI domain name? This index gives you the distribution to benchmark against; it does not value an individual name. The step-by-step framework — comps, multipliers, and the attributes that actually move price — is in how to value a domain name in 2026, and the appraisal tool produces a range for a name you type in.
Why is one appraisal $4,000 and another $40,000 for the same name? Automated appraisals model liquidity — what a name fetches quickly to any buyer — while asking prices model the best-fit buyer. Both can be defensible at once. Full breakdown in why two appraisals can differ by 10x.
Can I use this data in an article? Yes. Quote any figure with attribution and a link to this page. Please state that the figures are asking prices rather than sale prices — that is the one thing we would like not to be misreported.
Related reading
- How to value a domain name in 2026 — the underlying valuation framework.
- .ai vs .com for startups — extension choice in depth.
- Best AI startup domains and best robotics domains — the inventory this dataset is drawn from.
- Robotics company names: a naming guide — naming strategy for the densest category in the sample.
- Appraisal methodology — how our per-name valuation range is calculated.