The new gTLD program explained: How ICANN created hundreds of domain extensions

What the new gTLD program is and why ICANN created it

The new gTLD program is ICANN’s initiative to expand the generic top-level domain space, adding hundreds of new extensions beyond the original handful to enhance competition, innovation, and consumer choice. Before it, the open generic space was tiny.

A generic top-level domain (gTLD) is an extension not tied to a country, like .com, .net, or .org. For nearly all of the internet’s history that list was short.

ICANN states the program’s goals directly: to enhance innovation, competition, and consumer choice, with new safeguards for a secure, stable, and resilient internet. The expansion was the largest systematic change to the namespace.

The original generic space was small, and the program opened it deliberately.

The starting point matters. The program did not invent the idea of a generic extension. It scaled it from a closed list into an open application process anyone with the fee and the technical capacity can enter.

That shift is the whole story. The string after the dot stopped being a fixed set ICANN handed out and became something an applicant can propose, operate, and run as a registry business.

What a top-level domain is, where it sits in a web address, and who governs each type is covered in the foundational sibling What is a TLD? The top-level domain explained, from the dot to the governance behind it.

The 2012 round that created hundreds of new domain extensions

The 2012 round is the first and largest application window of the new gTLD program, opening on 12 January 2012, drawing 1,930 applications, and ultimately delegating over 1,200 new gTLDs into the root zone. It ran on a USD 185,000 application fee.

The round was not ICANN’s first expansion attempt, but it was the one that changed the namespace at scale.

The 2012 round followed a decade of smaller steps:

• 2000 proof-of-concept round: a limited first test that added a small set of generic extensions.
• 2003-2004 sponsored round: a round of sponsored TLDs (sTLDs) for defined communities.
• 2005-2007 policy development: the GNSO ran the policy process that designed the open application model.

The 2012 round spanned four kinds of application, not one.

The applications were not all commercial land grabs. They sorted into four recognisable categories, and the type shapes how the resulting extension behaves today.

• Generic strings: open dictionary words like .app, .blog, and .shop, intended for public registration.
• Geographic strings: city and region names like .london, .paris, and .berlin, run for a place.
• Community strings: gated extensions like .bank, .law, and .pharmacy, restricted to a verified community.
• Brand TLDs: closed corporate extensions like .google, .apple, and .bmw, operated by a single company.

That spread is why the new gTLD label covers everything from a high-trust verified .bank to a bulk-registered commodity string. The category is broad enough that the label alone tells a buyer almost nothing.

How a new gTLD application becomes a delegated extension

A new gTLD application becomes a live extension through a fixed sequence: an applicant pays the evaluation fee, ICANN reviews the string and the operator, contention and objections are resolved, a registry agreement is signed, and the string is delegated into the root zone. The process gates on capability, not speed.

The path is deliberately slow and expensive, which is the point. A top-level domain is internet infrastructure, so the bar to operate one sits high.

The fee and the registry agreement are why the count stayed in the low thousands.

The economics explain the scale. A six-figure evaluation fee, multi-year processing, and an ongoing registry agreement price out casual applicants and bulk speculation at the top-level.

That filter is the reason the namespace gained roughly 1,200 extensions and not tens of thousands. Each one represents an operator that committed real capital and signed binding obligations.

The same process applies in the 2026 round, which is why the fee rose instead of falling. The barrier is a feature ICANN maintains on purpose.

What launched: the adoption reality behind the new gTLD count

The new gTLD program produced real adoption, but the headline registration totals overstate durable use, because renewal rates on the bulk of new gTLDs sit far below the legacy extensions. Volume and value are not the same number.

New gTLDs now hold roughly 48 million domains and added about 11 million in 2025, close to half of all global domain growth that year. The momentum reshaping the market is real, and the category is genuinely growing.

Global registrations reached 386.9 million in 2025, up 6.1% year over year, the first time growth exceeded 6% since 2014, when the first new gTLDs were arriving. The leading strings by registrations drove the bulk of that gain.

Renewal rate, not registration count, is the metric that survives scrutiny.

The distinction is the whole point of the section. A first-year registration on a discounted commodity string costs little and proves little. A renewal is a paid vote that the name earned its keep.

When the average new gTLD renews in the low-30s and an aged .com renews near 75%, the gap measures how much of the new gTLD volume is transient. Bulk registration inflates the headline while the held base stays thin.

The credible exceptions prove the rule. The strings that enforce a standard or serve a real audience, like .app and .dev, retain registrants at rates approaching the legacy extensions.

The 2026 next round explained: window, fee, and what it changes

The 2026 round is the second application window of the new gTLD program, opening 30 April 2026 and closing 12 August 2026, with a base evaluation fee of USD 227,000 per application. It is the first new round in over a decade.

The mechanics echo 2012 with a higher fee. The Applicant Guidebook and the program requirements were finalised by the Implementation Review Team and published in December 2025, and the submission process remains capability-gated, not first-come, first-served.

The opportunities sit mainly with established registries and large brands, the parties able to absorb a six-figure fee and a multi-year timeline. The window dates and the broad timeline are the facts a buyer needs, not the applicant detail.

The next round adds strings, and it does not reset the value of the ones already live.

The temptation is to treat a new round as a reason to chase the next string. The history argues for caution. The 2012 round produced over 1,200 extensions, and a short list of them carries durable trust and value.

A second round adds more strings to the same crowded space. It does not retroactively raise the value of a cheap commodity gTLD, and it does not lower the standing of an aged .com.

For an acquirer, the round is context, not a buy signal. The names that hold value before the round are the names that hold value after it.

Which new gTLDs carry real value versus hype

The new gTLDs that carry durable value are the ones that enforce a standard or serve a defined audience, while the hype concentrates on cheap, open, high-volume strings that trade on first-year price and discount at renewal. The label is not the signal.

Reading value means reading the specific extension, not the new-gTLD category. A verified .bank and a bulk-registered .xyz sit at opposite ends of the same program.

The comparison with the legacy space is stark. Security researchers find phishers love the cheap new strings, and the abuse concentration on a short list of them is what drags the category’s reputation. New gTLDs hold about an 11% market share yet carry roughly a 37% share of reported cybercrime domains.

Three patterns separate value from hype:

• Enforced-standard strings: .app and .dev enforce HTTPS at the TLD level, and .bank verifies every registrant, so the extension itself signals a standard.
• Descriptive open strings: .shop, .store, .tech, and .online read as on-topic but carry lower recall and resale depth than an aged gTLD.
• Cheap commodity strings: .xyz, .top, and .icu are low-cost and high-volume, the condition under which abuse and churn concentrate.

Value tracks the standard a string enforces, not the program it came from.

The working rule reverses the marketing. A new gTLD earns value by enforcing something, a security standard, a verification gate, a real community, that a buyer or a registrant cannot fake.

.app and .dev are credible because HTTPS is mandatory at the TLD level. .bank is credible because every registrant is verified. The extension carries a standard the registrar cannot waive.

The full abuse and trust data behind which extensions get flagged is set out in the sibling guide New gTLD trust and spam risk: Which domain extension gets flagged, and whether the extension itself moves rankings is answered in Does your domain extension affect SEO? The data, the myths, and the aged-domain nuance.

What a new gTLD signals for resale liquidity and aged-domain value

On the aftermarket a new gTLD signals thin resale liquidity, because the deep buyer pools sit on the established gTLDs and the aged vanity ccTLDs, not on the post-2012 strings. This is the lens the registry marketing omits.

The investor read inverts the registration question. A fresh registration weighs branding fit and first-year price. An acquisition weighs inherited equity and how fast the name converts to cash.

The extension prices the liquidity, and the domain’s history prices the equity.

The distinction that matters on an acquisition is between equity and liquidity. The inherited backlink profile and topical history decide the SEO value. The extension decides how deep the resale pool around it runs.

An aged .com with a clean, relevant inherited profile combines both. A new gTLD with identical metrics carries the same on-page equity and a thinner exit. The program created the strings; it did not create resale depth.

How extensions sort into resale tiers, from the .com gold standard down to speculative new gTLDs, is ranked in TLD value and resale tiers: which domain extension holds value.

The way an extension sits alongside the other signals on an acquired name is set out in the foundational list sibling Domain extensions list: Every TLD type and what each one signals.

How a screened catalogue reads a new gTLD inside the program

The whole program resolves to one operational point: a new gTLD is one signal, and the inherited history, the renewal-backed demand, the abuse exposure, and the resale liquidity are what decide value. The history, the adoption reality, and the value matrix all point the same way.

SEO Domains reads a new gTLD exactly that way at intake. The curated catalogue screens each aged domain in a fixed order:

• Its inherited backlink profile, read for relevance and cleanness.
• Its topical history, read against the buyer’s intended use.
• Its abuse and trademark exposure, surfaced before acquisition.
• The extension and its program type last, read as a trust-and-liquidity modifier on top of that screen.

Domain Authority, Domain Rating, Trust Flow, and Citation Flow are reported alongside the inheritance read, so a buyer sources a name selected on the history that decides the SEO and the extension that prices the resale liquidity around it.

The catalogue reads inherited equity before the extension, which is the order the program demands.

The discipline SEO Domains applies is to invert the order the registry marketing encodes. A raw listing prices a name on its extension and its registration count, and leaves the history unread.

The catalogue reverses that. It reads the inherited backlink profile for relevance and cleanness, reads the topical history against the buyer’s intended use, reads the abuse and trademark exposure, and only then reads the extension and its program type as the trust-and-liquidity modifier they are.

An aged domain on an established gTLD combines the inherited SEO equity that carries the ranking with the deepest resale pool the extension prices. ICANN-accredited transfer applies to every acquisition regardless of extension, and the underlying diligence runs on RDAP after the WHOIS sunset of 28 January 2025.

A buyer who reads the program by governance and renewal-backed value, not by the new-gTLD label alone, is the buyer best served by inventory screened on the history that is the larger lever.

A screened catalogue raises confidence in the history, and it guarantees no outcome.

The honest takeaway is two-sided. New gTLD hype, a cheap commodity string sold as in-demand, a new round read as a buy signal, an abuse-associated extension passed off as clean, and a history ignored in favour of the suffix are real ways a domain decision goes wrong.

They concentrate in unscreened pools where a name reaches a buyer priced on its extension and its registration count with its history unread.

A screened catalogue does not write the content, earn the new links, or run the conversion the inherited equity rewards. It does not provide financial, investment, or legal advice, and it promises no ranking or sale outcome on any name or extension.

What it does is read the inherited equity, the topical history, the abuse exposure, and the resale liquidity that decide the outcome, and place the extension where this program places it, as one signal on top of the asset.

A reader who finishes this explainer is equipped to stop reading the new-gTLD label in isolation and start reading the domain.