AI Domain: bubble or long-term asset?

.ai grew 47% YoY in volume, 18% in median sale price for Q1/2026. On one hand, this is healthy: AI investment is real, end-user willingness is high. On the other, the price curve is starting to resemble the .com cycle of 1998-2000 — sharp ascent, then 80% collapse. This brief presents 5 signals to distinguish smart AI investing from FOMO bubble.

.ai data Q1/2026

Public .ai volume Q1: $156M. Top sale: brand.ai $1.2M. 5-figure transactions: 312 (vs 187 in Q1/2025). Median sale: $9,800 (vs $8,300 Q1/2025).

Compared to .com cycle 1998-2000:

• .com 1998: $89M volume, $5,500 median. 1999: $312M, $11,200. 2000: $521M, $14,900. Peak.
• .com 2001: volume crashed to $186M (-64%), median fell to $7,800.

The .ai 2024-2026 curve is replicating the .com 1998-2000 pattern — different in absolute scale (.ai started smaller), same in shape.

Five signals of smart AI investing

Signal 1: Domain has "real-world utility". Brandable AI for clear niches — health.ai, tax.ai, legal.ai — have quantifiable end-user value. Avoid fictional names like "Quantum-Synth-Alpha.ai".

Signal 2: Multiple ≤ 50x annual cost. .ai registration ~$200/year. Buying at ≤ $10K (50x multiple) is the safe zone if holding 3-5 years. $50K+ purchases only when a real end-user buyer is identified.

Signal 3: 2-3 syllable, easy pronunciation. Brand recognition in AI sector requires short + memorable. 4+ syllable names with rare keywords still flop on liquidity.

Signal 4: Comparable transactions in past 12 months. No comparable = no price discovery. Buying on "broker says" is pure speculation.

Signal 5: Willingness to hold 5 years if needed. Names bought in FOMO waves without exit liquidity become "stuck inventory" when the wave passes.

Five FOMO signals to avoid

1. "I bought $30K, broker says I can sell $200K in 6 months". No real data — pure sales pitch.
2. 5+ syllable domains with awkward AI keyword stuffing. End-users do not pick these.
3. Buying on leverage / margin. Domains are illiquid — unsuitable for funded deals.
4. Putting majority of portfolio into .ai only. High concentration risk. Reference Brief #003 (Allocation Framework) — .ai should be at most 30% of total portfolio.
5. "This wave never stops". Every wave stops. The right question is when, not whether.

"Bubbles are not the enemy — they are an opportunity for those positioned beforehand. The enemy is believing you are at the start of the wave when you are already in the middle." — Howard Marks (paraphrase)

Conclusion: long-term, still an asset

.ai as a TLD will not "die" — unlike .nft or .crypto which burst entirely. Reason: AI is foundational technology (general-purpose), not application layer. Similar to .com surviving the 2001 crash and becoming internet plumbing.

But price level may correct 30-50% in the next 12-24 months. Standard strategy for individual investors: participate with maximum 30% NAV, prioritise 2-3 syllable brandable, ready to hold 5 years. No leverage. No all-in.

Brief #002 on price cycles. Brief #006 on the allocation framework with 3-scenario stress test — particularly relevant if cooldown materialises.