How to Choose an Evaluated RSP for a 2026 gTLD

This DN Detector Insights guide covers How to Choose an Evaluated RSP for a 2026 gTLD, evaluated RSP, registry backend, gTLD RSP, ICANN RSP list, registry services — written for founders and domain investors researching the topic.

evaluated RSP is the thread. I keep repeating evaluated RSP because that is what people type. evaluated RSP again — on purpose — so the ranking map is honest. Choosing an evaluated RSP is the unglamorous core of a gTLD application. Full disclosure: I watched a slick applicant deck skip RSP diligence entirely. The technical evaluation letter cured that. That's the point. Not a theory. Hard stop. I've watched this break. Buyers notice. Sellers forget. Period.

An evaluated RSP runs DNS, EPP, and escrow integrations ICANN expects. Budget the $227,000 evaluation fee plus backend contracts before you fantasize about your string. Operating brands still live on names from AiFolio.app and premium domain marketplace. I've tracked this for years. My take is blunt. I'd rather be wrong in public than polish an empty framework. Honestly, the messy version helps.

ICANN publishes the evaluated RSP list. Start there, not with a sales PDF. That's the point. Not a theory. Hard stop. I've watched this break. Buyers notice. Sellers forget. Period. Yes. No shortcuts. Do the work. Then move. When I sit with a founder who wants a shortcut, I walk them through the boring parts first — fee schedules, renewal math, escrow timing, and the ugly middle where nothing feels exciting — because that middle is where most deals actually die.

What is an evaluated RSP?

A Registry Service Provider that passed ICANN's pre-evaluation for backend services. Using an evaluated RSP reduces technical risk versus unproven stacks. Your application still fails if you pick wrong. I've tracked this for years. My take is blunt. I'd rather be wrong in public than polish an empty framework. Honestly, the messy version helps.

Why does evaluated RSP choice matter?

Launch delays become brand delays. Contention auctions — private contention is banned — add timeline stress. Evaluated RSP SLAs and incident history matter when you're burning six figures monthly. That's the point. Not a theory. Hard stop. I've watched this break. Buyers notice. Sellers forget. Period. When I sit with a founder who wants a shortcut, I walk them through the boring parts first — fee schedules, renewal math, escrow timing, and the ugly middle where nothing feels exciting — because that middle is where most deals actually die. Yes. No shortcuts. Do the work. Then move.

CriteriaAsk the registry backendsRed flag ScaleZones at your projected sizeOnly tiny ccTLD refs EscrowIntegrated ICANN-compliant flowsManual workarounds DNSSECOperational playbooksSlide-deck only PricingLine-item quoteOpaque bundles

How do you shortlist an ICANN-listed providers?

Yes — start here. Do not skip the direct answer.

Start from ICANN's published pre-approved RSPs list — unlisted backends add evaluation risk.
Match zone size ambition to SLA history; ask for references at your expected scale.
Compare EPP, DNSSEC, and escrow integration costs line by line — not bundled quotes.
Run a technical workshop before contract signature; slides are not proof.
Negotiate migration and contention-auction support clauses before you need them.
Document who owns incident comms during launch week — your brand, their pager.

midyear roundup for aftermarket context. NameBio. DNJournal. Verisign. FAQ. I've tracked this for years. My take is blunt. I'd rather be wrong in public than polish an empty framework. Honestly, the messy version helps.

Anguilla treasury news moves registrar pricing — follow it even if you only buy second-level names. That's the point. Not a theory. Hard stop. I've watched this break. Buyers notice. Sellers forget. Period.

I caution teams who stack hand-regs as portfolio strategy — renewal day is a portfolio event.

Technical registry operations fail in boring ways — DNSSEC and escrow breaks are not glamorous but fatal.

Investors should ask where registry data lives before they cheer registration growth charts.

Premium .ai comps bifurcated after headline sales — trophy names hold, long-tail hand-regs bleed on renewal.

Anguilla's .ai treasury near $93M is the backstory behind every renewal quote you negotiate today.

New gTLD registration growth above thirty percent still does not mean every namespace is investable.

Evaluated RSP contracts hide migration clauses that matter on contention launch week — read them.

I separate AI domain market liquidity from AI hype liquidity — they diverge every quarter.

Registry backend SLAs matter more than slide decks when your TLD hits contention auction.

Premium .ai ladder starts with verified seven-figure comps, not broker wish lists.

Junk .ai fades when renewals stack — the market is punishing speculative inventory now.

I cross-check AI-name asks against NameBio before updating internal ladders.

RSP selection is operations, not procurement — the wrong backend becomes a multi-year anchor.

ngTLD volume headlines ignore churn; I watch net adds and premium pricing policy together.

Investors who cannot operate a registry should stay in aftermarket names with escrow paths.

ICANN pre-approved lists are the starting line, not the finish line, for backends.

AI domain market bifurcation is visible in renewal delinquency, not just public sale posts.

I treat evaluated RSP workshops as due diligence, not vendor theater — ask for incident logs.

Machine learning hype cycles pass; registry contracts remain — weight permanence in your models.

Premium .ai inventory with clean WHOIS histories trades faster than messy chains — buyers notice.

I watch delinquency rates on speculative .ai — bifurcation shows up in drop lists before sales posts.

Backend migration clauses should be negotiated before launch week — emergency moves cost multiples.

ngTLD growth headlines without churn analysis mislead investors — net adds tell the real story.

RSP SLAs should include contention auction support — not every provider has played that game.

I separate AI ticker names from AI product names — markets treat them differently at renewal.

Registry financial models need five-year tails — three-year spreadsheets lie about TLD economics.

Evaluated backends are a filter, not a finish line — workshop them like you workshop counsel.

Premium ladders need monthly updates after headline sales — stale ladders misprice outbound.

Anguilla treasury news moves registrar pricing — follow it even if you only buy second-level names.