Reading the .AI Bifurcation: What 3,651 Sales Actually Show

Sarah Chen leads quantitative research at MainSearches, maintaining one of the most complete transaction datasets in the aftermarket. We asked her to walk through what the numbers actually say about the .AI market after the 2026 registry price increase.

That the market isn't one market anymore — it's two, moving in opposite directions. Below roughly $2,500, volume collapsed after the price increase. Above $25,000, floor prices climbed around 22 percent year over year. The middle is thinning. Anyone holding a portfolio needs to understand which tier each of their assets actually lives in, because the strategies are completely different.

Pure arithmetic. When renewal cost jumps from $92 to $114, that's a 22 percent increase in carry. For someone trying to flip a $500 registration for $1,500, that carry cost is decisive — it eats the margin and extends the break-even horizon. Speculators operate on thin margins and short clocks. Raise the floor and you remove the weak hands almost immediately. What remains at the low end is increasingly genuine end-user demand, which is actually healthier.

It's a luxury-goods dynamic. Higher barriers reduce noise and concentrate attention on quality. A buyer evaluating a $50,000 .AI is indifferent to a $114 renewal — it's a rounding error. But that same cost structure scares off the churn registrants who inflated supply. So premium inventory faces less competition, more focused buyer attention, and improved perceived scarcity. When renewal costs eliminate weak hands, strong hands gain pricing power. That's the entire bifurcation in one sentence.

Comparable sales within a recency window, character length, dictionary status, pronounceability, and vertical commercial activity. But the single most underweighted variable is liquidity evidence — whether comparable names have actually transacted, not just been listed. I'd rather anchor to three real sales than thirty asking prices. Listing prices are opinions. Clearing prices are facts.

Triage, immediately. Score every name against the premium criteria. The top 10 to 15 percent get held with patience. The clearly bottom tier should be liquidated before the next renewal cycle, even at a loss — paying $114 to keep a lottery ticket with negative expected value is a slow bleed. The middle is where people get stuck, and that's where you make hard, name-by-name decisions rather than portfolio-level hope.

Time to sale. Everyone fixates on price, but the median time-to-sale for realistically priced domains is around 14 months, while aspirationally priced names exceed 47 months and most never sell at all. Liquidity has a cost. If you price for a faster exit, you're often net-ahead versus holding years for a price that never arrives. Capital has a time value, and most domainers never account for it.