Are .io Domains Going Away? Is .io Safe and Worth It?

Short answer: No. As of July 2026 the .io domain is fully operational, the IO code is still in the ISO 3166-1 standard, and no retirement process has started. The uncertainty is real but it is slow-moving: the UK and Mauritius signed a treaty over the Chagos Archipelago on May 22, 2025, and if that eventually removes IO from ISO 3166-1, ICANN's ccTLD retirement policy gives a five-year runway from the notice of removal, extendable once to ten. Nothing about that timeline threatens a domain you buy this year. The better reasons to think twice about .io are the renewal price and whether your customers are technical.

This question gets asked in a panicked way and answered in an even more panicked way, usually by people selling you a .com. Here is what has actually happened, what would have to happen next, and what the one real historical precedent tells us.

What actually happened with the Chagos treaty

The .io extension belongs to the British Indian Ocean Territory, and the sovereignty of that territory has been disputed for decades. The relevant sequence is short:

Read that last row twice, because it is the row everyone skips. The trigger for a ccTLD retirement is not a treaty, a handover ceremony or a news cycle. It is the deletion of the two-letter code from ISO 3166-1. That has not happened.

What would have to happen for .io to be retired

Three separate things, in order, each of which takes time:

1. The sovereignty transfer completes and the ISO 3166 Maintenance Agency acts. The committee would have to decide that BIOT no longer warrants its own code element and delete IO from the standard. Territories can and do keep codes after political change.
2. ICANN issues a Notice of Removal. This starts the clock under the ccTLD retirement policy the ccNSO developed and ICANN's board approved in 2022.
3. The five-year transition runs. The retiring ccTLD comes out of the root zone five years from the notice, and the registry manager can request one extension of another five years. The maximum runway is therefore ten years.

So even in the worst realistic case, a name you register in 2026 has most of a decade before anything forces your hand, and you would have years of public notice to migrate. Compare that with the average startup's actual lifespan and the risk reprices itself.

The .su precedent nobody mentions

There is exactly one useful historical comparison, and it is reassuring in a slightly absurd way. The .su extension was assigned to the Soviet Union. The Soviet Union ceased to exist in December 1991. In March 2025, more than 33 years later, ICANN finally moved to begin retiring .su, with a target of 2030.

That is the closest thing to a benchmark we have: a country code whose country vanished outright still ran for three decades, and its wind-down was still measured in years rather than months. .io's situation is far less dramatic, since the territory continues to exist and only its administering state may change.

Is a .io domain safe to buy?

Yes, with a clear understanding of what you are buying. Every country code extension is ultimately a delegation that depends on a government and a registry, and that includes the ones people treat as safe. If you own a .ai name you are exposed to Anguilla; a .co owner is exposed to Colombia; a .io owner is exposed to whoever administers the Chagos Archipelago. None of these are new risks, they are just newly visible. The practical protections are unglamorous: keep the registration locked, keep multi-year renewals paid, keep your registrar account on two-factor auth, and own the matching .com if the brand ever becomes valuable enough to defend. If you are buying the name on the aftermarket rather than registering it fresh, the transfer itself is the moment worth protecting, which is what domain escrow is for: the name changes hands and the money releases together, or neither happens.

What .io is not is an SEO problem. Google treats .io as a generic top-level domain rather than a country-targeted one, so there is no geographic restriction and no ranking penalty baked into the extension. Rankings follow content, links and site quality, exactly as they do on a .com.

Why are .io domains so expensive?

Two separate costs get confused here. The first is the registration and renewal price: the registry sets a wholesale rate well above the .com level and registrars add margin, so roughly $30 to $80 a year is normal, every year, forever. The second is the aftermarket price of a good name, which has nothing to do with registry pricing and everything to do with scarcity. Short one-word .io names generally trade in the $5,000 to $50,000 band because a decade of developer-tool and API companies bought the good ones first. You can check that band yourself rather than taking it on trust: domain sales history covers where .io closes are published and how thin the record gets outside the top names.

The renewal number is the one investors underestimate. On a single brand domain, $60 a year is a rounding error. Across a 200-name speculative portfolio it is $12,000 annually before you sell anything, which is the difference between a book that compounds and one that quietly bleeds. That math is the whole subject of domain portfolio management, and it is why bulk renewal triage matters more than any single acquisition.

Are .io domains good for startups?

For developer tools, APIs, infrastructure and B2B SaaS, yes. The extension reads as native to that audience, and a short, clean .io is usually obtainable when the matching .com is a six-figure name or simply not for sale. For a local service business, a healthcare practice, a law firm or anything selling to non-technical consumers, .com remains the safer signal, because a meaningful share of those customers will type .com regardless of what your business card says.

The honest test is not "is .io respectable" but "will my customers mistype it". Ask five people in your target market to type your name from hearing it once. If they all reach for .com, buy the .com. If they are engineers who live in documentation and package registries, .io costs you nothing in credibility.

Is a .io domain worth it?

It is worth it when the .io is dramatically better as a name than the .com you could otherwise afford. A memorable, short .io usually beats a long, hyphenated or misspelled .com, and it beats waiting two years for a budget that never arrives. It is not worth it when you are buying .io purely because it is cheaper than the .com you actually want, since you will end up paying for the .com later anyway, at a price the owner sets once they see your traffic.

Price the decision instead of arguing about it. Run the name through an instant domain appraisal and read the comparable sales underneath the estimate: same length band, same extension, same keyword family. If the ask sits inside the comps corridor, the extension question is a preference. If it sits far above, the extension is not your problem, the price is. Current inventory sits on the .io domains for sale board with an appraisal attached to every line.

How to hedge without panicking

If you want insurance rather than reassurance, do these three things and stop thinking about it:

• Register the matching .com when it is affordable. Even parked and redirecting, it is the cheapest optionality in your brand.
• Keep the .io renewed several years out. Long registrations remove the failure mode that actually kills domains, which is a lapsed card on a forgotten account, not geopolitics.
• Know your migration path before you need it. If you ever do move a live product to a new name, the mechanics are permanent 301 redirects from every old URL to its exact new equivalent, updated canonical tags, both properties verified in Search Console, and a change-of-address request. The step teams forget is verification afterward: a redirect chain that silently breaks six weeks later undoes the link equity you just moved, so set up monitoring that pings you within minutes when a URL stops responding rather than discovering it in a traffic report a month late.

Migrations done this way typically recover most of their organic traffic within a few months. Migrations done by pointing the DNS somewhere new and hoping tend not to.

The bottom line

.io domains are not going away in any timeframe that should change a buying decision made in 2026. The code is still in the standard, no retirement clock has started, and the policy that would eventually govern one measures its transition in five-year blocks. The risks worth pricing are the annual renewal cost and whether your audience is technical enough for the extension to land. Both are knowable before you spend anything.

If the name matters to your brand, judge it the way you would judge any aftermarket purchase: against what comparable names actually sold for, not against a headline. The same discipline applies to every country-code extension in the market, including the one everyone is currently excited about, which is covered in .ai domains for sale and priced the same way, on comps.