Why .ai domains crossed median $50K aftermarket valuations in 2026: TLD economics breakdown
By Mustafa Bilgic, sole proprietor — domain investor since 2019, operator of names.center | Last reviewed: 2026-05-05 | 16 min read
The .ai market stopped behaving like a niche ccTLD when three things lined up at once: AI venture funding created end-user demand, Anguilla kept a premium renewal structure that filters low-quality speculation, and Identity Digital moved the registry onto institutional-scale infrastructure. The result is not that every .ai is worth five figures. It is that the investable top shelf of .ai now trades and gets valued in a range where $50,000 is no longer an outlier.
The .ai Market In Numbers
I pulled the market snapshot on 2026-05-05 from NameBio's public TLDStats API documentation and compared it with Sedo/InterNetX reporting and Anguilla registry economics. NameBio showed 21,795 public .ai sales totaling $54,552,341, with an all-market average of $2,502.98 and a maximum public sale of $1,200,000. The retail subset was much stronger: 1,171 .ai retail sales totaling $26,623,309, for an average of $22,735.53.
| Metric | Value | What it means for investors |
|---|---|---|
| NameBio .ai public sales, all time | 21,795 sales / $54.55M | Large enough sample to separate retail demand from wholesale churn. |
| NameBio .ai public sales, trailing 1 year | 7,486 sales / $25.79M | Nearly half of all public dollar volume arrived in the latest one-year window. |
| NameBio .ai retail subset | 1,171 sales / $26.62M / $22,735 average | Retail buyers pay a very different price curve than domainers trading inventory. |
| NameBio .ai 3-month average | $4,118.37 across 1,687 public sales | The broad market is liquid, but still not a universal $50K median. |
| Sedo/InterNetX 2026 average Sedo domain price | $2,753 | .ai retail comps sit far above the blended Sedo marketplace average. |
The key distinction is retail versus wholesale. The 20,624 wholesale .ai sales in NameBio's public TLDStats averaged $1,354.20, while the 1,171 retail .ai sales averaged $22,735.53. A professional portfolio model has to underwrite both curves. If you buy at wholesale, your pricing has to survive long hold times. If you pay retail, you need a specific end-user buyer thesis.
Why The $50K Premium-Comp Line Matters
The January 2025 InterNetX/Sedo snapshot listed Sedo's ten highest publicly reported .ai sales up to that point: you.ai at $107,293, agents.ai and growth.ai at $50,000 each, record.ai at $45,000, predictive.ai at $29,438, and several more in the $12,000 to $25,148 range. That table's median is $27,293, not $50,000. But the top half of the table already sat at or above $45,000, and later 2025/2026 public comps pushed the premium screen higher.
DNJournal's public sales charts and NameBio's current TLDStats show the step-change. The biggest public .ai comp in NameBio is now $1.2M, and DNJournal's 2025 ccTLD charts included multiple .ai names in the $50,000 to $125,000 zone. My premium-comp screen uses public names that meet three filters: one or two words, direct AI commercial use, and an end-user plausible price above renewal-carry cost. On that screen, the median has crossed $50K. On the entire reported .ai universe, it has not.
Registry Economics: Why Supply Is Not Cheap
.ai is the country-code TLD for Anguilla, delegated in the IANA root zone to the Government of Anguilla. It is not an ICANN-contracted new gTLD, which matters because ICANN's monthly registry reports cover contracted gTLDs and are useful as a baseline, but they do not explain .ai's registry revenue directly. For .ai, the stronger primary sources are IANA for delegation, the IMF for Anguilla fiscal data, and Identity Digital for registry platform migration.
The IMF 2024 Article IV report on Anguilla described the .ai boom in concrete fiscal terms: registrations rose from about 144,000 in 2022 to about 354,000 in 2023; auctions generated EC$87 million in 2023; the base registry fee was $140 for a two-year registration; and the report cited 90% renewal rates. That renewal rate is the hidden underwriting lever. A TLD with high churn can look hot for one year and collapse at renewal. A 90% renewal assumption means serious users and serious investors are keeping names.
Identity Digital's 2025 announcement that .ai migrated to its registry platform added the operating side: nearly 600,000 domains under management, 164 retailers at migration, and 90% of the top registrars by domains under management connected within six months. The same announcement said registry-run auctions delivered more than $600,000 in the first month and that daily auctions increased gross revenue by more than 20% week over week. That is not aftermarket hype; that is registry-side price discovery.
How .ai Compares With .com And .io
NameBio's public TLDStats API is useful because it lets us compare extensions with the same yardstick. On 2026-05-05, .com showed 1,397,893 public sales totaling $2.47B, with a retail average of $8,500.28. .io showed 19,995 public sales totaling $22.93M, with a retail average of $8,757.16. .ai showed fewer total public sales than .com but a stronger retail average than .io and a much higher maximum public sale.
| TLD | Public sales | Retail average | Maximum public sale | Investor interpretation |
|---|---|---|---|---|
| .com | 1,397,893 | $8,500.28 | $70,000,000 | Deepest market, broadest buyer pool, lower renewal drag. |
| .ai | 21,795 | $22,735.53 | $1,200,000 | Smaller market, stronger retail AI-specific pricing, higher carrying cost. |
| .io | 19,995 | $8,757.16 | $230,000 | Mature tech ccTLD, but no longer the default AI-native signal. |
This comparison explains why .ai can support aggressive asking prices without implying that .ai is safer than .com. The .com buyer pool is horizontal. .ai is vertical: the name either fits an AI company, model layer, data workflow, agent product, automation platform, chip/cloud stack, or it does not. That creates bigger price dispersion. The best names are obvious. The middle of the market is dangerous.
A Premium-Comp Screen I Would Actually Use
My premium screen is intentionally narrow. A name needs to pass at least four tests before I let a $50,000 comp influence pricing: it must be short, it must describe a funded AI use case, it must have no obvious trademark dependency, and it must be useful as a product or company name without explanation. That excludes most "keyword + ai" registrations and most long prompts-as-domains. It includes names tied to agents, models, copilots, inference, robotics, data, security, chips, workflow automation and enterprise AI operations.
| Premium-comp test | Pass example | Fail example | Why it matters |
|---|---|---|---|
| Buyer category | Clear AI product category | Generic adjective with no buyer | Funded buyers pay for category control, not decoration. |
| Length | One word or tight two-word phrase | Three-plus-word prompt phrase | Short names survive pivots and investor conversations. |
| Comparable sales | At least two public comps in adjacent categories | No direct or adjacent comps | Without comps, the price is mostly hope. |
| Renewal math | Expected ask is 500x+ annual registry carry | Expected ask is 20x annual carry | .ai renewal cost punishes marginal inventory. |
| Legal safety | Dictionary or generic technical term | Famous model, product or company mark | Trademark risk can erase liquidity. |
That screen is why I do not apply the same target to every .ai. A mediocre two-word .ai might be worth $500 to $2,500 wholesale or low retail, even during a hot cycle. A one-word infrastructure term might deserve a five- or six-figure ask because the buyer pool contains companies for whom the domain is a fundraising, recruiting and trust asset. The extension is only one variable; the buyer's economic use is the main variable.
The Renewal Math Behind A $50K Ask
At the official registry-level price noted by the IMF, .ai costs $140 for two years, or $70 per domain-year before registrar markup. A 200-name .ai portfolio therefore carries roughly $14,000 per year at the registry-fee level alone. Add acquisition cost, marketplace commissions, payment processing and time, and a low-quality .ai portfolio can become a cash drain quickly.
Here is the break-even logic I use. Suppose 200 .ai names cost $14,000 per year to renew and 15% commission applies on sale. If the portfolio sells one name per year at $20,000, net proceeds are about $17,000 before tax, barely covering renewals. If the same portfolio sells one name every two years at $50,000, net proceeds average $21,250 per year after commission, which covers registry-level renewals and leaves room for acquisition amortization. That is why a $50K premium-comp median matters: it can make the carry model work at low sale-through rates.
| Portfolio size | Annual registry-level carry | One sale at $25K after 15% commission | One sale at $50K after 15% commission | Read-through |
|---|---|---|---|---|
| 50 .ai names | $3,500 | $21,250 net | $42,500 net | A small curated book can survive one retail sale every several years. |
| 200 .ai names | $14,000 | $21,250 net | $42,500 net | One $25K sale covers carry but leaves limited acquisition recovery. |
| 1,000 .ai names | $70,000 | $21,250 net | $42,500 net | Scale requires multiple retail exits or much higher ASP. |
This table is the reason I prefer concentrated .ai ownership over broad hand-registration. A 50-name book of real AI category names can be rational even if annual sale-through is low. A 1,000-name book of average names needs enough sales velocity to behave like a business. Without buyer flow, it becomes a renewal liability.
How I Treat Registry Auctions Versus Aftermarket Sales
Registry auctions and aftermarket sales should not be merged blindly. Registry auctions can surface names that never had a previous private owner; aftermarket sales transfer names between registrants. The buyer psychology is different. Registry auctions often attract investors because the inventory is clean and directly released. Aftermarket sales often indicate an end user or another investor outbid the current owner.
For valuation, I discount registry-auction comps unless the buyer can be identified as an end user or the term is so strong that the source does not matter. I give more weight to public marketplace sales that show the name was already privately held and still attracted a strong bid. That is where WHOIS-history tools help: nameserver changes, registrar moves and ownership patterns can show whether a price was followed by a real operating buyer or parked again with another investor.
The strongest .ai underwriting stack therefore has five layers: NameBio and DNJournal for public sales; Sedo/InterNetX for marketplace and extension context; IMF and Identity Digital for registry economics; IANA/ICANN for delegation and baseline reporting context; and WHOIS-history tools for ownership verification. Any one layer alone is weaker than the combined picture.
Where Whoisology Fits
Whoisology is useful for reverse-WHOIS and ownership-history checks, not as a public price tape. For this article I treat Whoisology-style WHOIS history as verification support: did nameservers change after a reported sale, did registrant patterns consolidate under a known investor, and was the asset likely registry-auction inventory or aftermarket inventory? Price evidence comes from NameBio, Sedo/InterNetX and DNJournal; ownership-history evidence comes from WHOIS-history tools such as Whoisology.
That distinction matters for E-E-A-T. A public sale price without ownership context can overstate liquidity if it was a registry auction. Ownership history without sale price can prove movement but not value. A strong comp has both: public price reporting and a plausible post-sale ownership or nameserver transition.
Practical Underwriting Rules For .ai In 2026
- One-word commercial nouns: price from retail comps, not wholesale averages.
- AI infrastructure terms: require evidence of funded buyers, open-source category growth, or enterprise spend.
- Two-word .ai: underwrite against $2,000 to $10,000 unless the phrase maps to a real SaaS category.
- Novel brandables: avoid paying retail unless you can name five likely buyers.
- Renewal discipline: if a name cannot plausibly sell for 30x to 50x annual carry, it is not a portfolio hold.
The best .ai names now behave like scarce product-category assets. The rest behave like expensive lottery tickets. The job of the investor is to know which side of that line a name sits on before renewal season forces the decision.
Replication Notes For This Analysis
To recreate the numbers, start with NameBio's TLDStats endpoint and query extension=.ai, extension=.com and extension=.io. Record all-market, retail, wholesale, one-year and three-month fields separately; do not mix them. Then pull the Sedo/InterNetX 2026 report for average marketplace price and the InterNetX .ai snapshot for the Sedo top-sale table. For registry economics, use the IMF Anguilla report and Identity Digital migration release instead of generic ICANN gTLD files, because .ai is a ccTLD.
My final valuation screen then removes three classes of misleading comps: registry-auction names with no end-user evidence, sales where the buyer category is not repeatable, and outlier names so strong that they do not help price ordinary inventory. That is why the article can say the premium-comp median crossed $50K while also saying the broad .ai market did not. Both statements can be true only if the segment definition is explicit.
Frequently Asked Questions
Did the entire .ai market reach a $50,000 median sale price?
No. NameBio's broad .ai market averages are much lower. The $50K point applies to a screened premium-comp segment, not to every public .ai sale.
Why is .ai different from a normal new gTLD?
.ai is a ccTLD delegated to Anguilla, not a new gTLD under an ICANN registry agreement. That changes the source set: IANA, IMF Anguilla data and Identity Digital .ai platform statements matter more than ICANN monthly gTLD reports for direct registry economics.
What is the strongest public .ai data source?
For reported sale prices, NameBio and DNJournal are the strongest public sources. For marketplace pricing and annual market context, Sedo/InterNetX reports are useful. For registry economics, use IMF Anguilla reporting and Identity Digital's .ai migration release.
Is .ai safer than .com for investors?
No. .ai has strong vertical demand but higher carrying cost and a narrower buyer pool. .com remains broader and more liquid.
What .ai names should investors avoid?
Avoid weak two-word names, hyphenated names, trademark-risk names, and brandables that are not recognizably tied to AI workflows or funded buyer categories.
Sources And Verification Notes
- NameBio API documentation - TLDStats queried 2026-05-05 for .ai, .com and .io.
- Sedo/InterNetX Global Domain Report 2026 newsroom summary.
- InterNetX Snapshot: .ai deep dive - Sedo .ai sales table and registration context.
- IMF Anguilla 2024 Article IV report - registration growth, registry fee, renewal and auction revenue.
- Identity Digital .ai migration release - platform migration, registrar coverage and auction revenue.
- IANA root database: .ai delegation record.
- ICANN monthly registry reports - gTLD baseline reporting context.
- DNJournal domain sales charts - public high-value .ai comp verification.
- Whoisology - WHOIS-history/reverse-WHOIS verification context, not used as a public price tape.