Guides › Market analysis

Domain Market Analysis: August and September 2026

Figures as of October 17, 2026

This analysis compares the domain sales DNScroll recorded in August and September 2026. It counts only sales of $100 or more, which leaves out the many small auction sales. Everything below describes recorded sales, which are a large part of the aftermarket but not all of it, and the figures can change slightly when records are corrected.

What happened to prices

Recorded sales of $100 or more. Source: DNScroll database.
Measure	August	September	Change
Recorded sales	17,158	20,564	+20%
Total value	$29,400,783	$22,915,107	-22%
Median price	$583	$462	-21%
Average price	$1,714	$1,114	-35%
90th percentile	$2,247	$1,600	-29%

The median recorded price fell from $583 to $462, about 21% lower, while the number of sales rose about 20%. The 90th percentile, the price that one sale in ten exceeded, fell by a similar share, from $2,247 to $1,600, so typical sales and the upper range both moved down.

Total value fell 22%, from $29.40 million to $22.92 million, and the average fell further, by 35%, because averages also reflect the largest sales. Those were smaller in September: the ten largest sales added up to $6.87 million in August and $2.61 million in September. That $4.25 million difference equals about two-thirds of the $6.49 million fall in total value.

The data show what changed, not why. A different mix of marketplaces or kinds of names can move a median as much as a change in buyer behaviour, so these figures do not show that domains became less valuable.

Many sales or a few large ones

Value is concentrated. In August the top 1% of sales (171 sales) made up 46.1% of the month's total value, and the ten largest made up 23.3%. In September those shares were 30.8% (205 sales) and 11.4%. When so much of the total comes from a small number of sales, one or two big deals can move a month's total value more than thousands of ordinary sales, which is why this analysis relies on the median for the typical sale.

Which extensions lead: by number of sales and by value

Selected extensions, recorded sales of $100 or more. Share of value is the extension's part of the month's total value.
Extension	Aug sales	Aug share of value	Sep sales	Sep share of value
.com	13,092	70.3%	15,524	70.7%
.org	959	3.5%	1,163	4.8%
.net	774	1.8%	787	2.4%
.ai	241	13.3%	405	3.9%
.top	325	0.5%	329	0.9%
.co.uk	168	0.9%	320	2.6%
.io	127	1.1%	252	1.2%

.com leads both rankings by a wide margin: 76.3% of sales and 70.3% of value in August, and 75.5% and 70.7% in September.

Other extensions rank differently depending on the measure. In August, .ai ranked sixth by number of sales (241) but second by value ($3.91 million, 13.3% of the total). In September it had more sales (405, fourth by count) but only about $896,000 of value (3.9%), third by value. The average recorded .ai sale was about $16,200 in August and $2,200 in September. The data show the gap, but not its cause.

.top ranked fourth by sales in August (325) and fifth in September (329), yet brought in under 1% of value in both months. .co.uk ranked sixth by sales in September (320) but fourth by value ($604,292, 2.6%). .xyz had just 74 sales that month but ranked seventh by value ($362,070, 1.6%), an average of about $4,900 per sale.

Sources of distortion in these two months

Dynadot. It supplied 39.7% of recorded sales of $100 or more in August and 36.5% in September, and 14.4% and 22.4% of the value. A source that supplies this much of the data can shift the totals when its coverage changes.
Coverage. Counting sales of every price, DNScroll recorded 30,799 sales in August and 57,957 in September, almost double, against a 20% rise in sales of $100 or more. A change in what DNScroll collects can produce a jump like that, so recorded sale counts measure DNScroll's coverage as well as market activity.
One new source. A marketplace that appeared only in September, Jinmi, contributed 33 sales worth $870,497, 3.8% of September's value. Leaving it out would make the fall in total value 25% instead of 22%, so it does not explain the decline.

What these figures cannot tell you

Private sales appear only when they are reported publicly, and some marketplaces are not covered.
For auctions, the recorded price may be the last price seen before the auction closed, which can differ from the amount paid. See the methodology.
Two months are not a trend, and a month-to-month change can reverse.
A median describes recorded sales, not what a particular domain would sell for. See the guide to reading sale comps.
This is a description of recorded data, not advice on buying or selling a domain.