The Internet’s New Addresses: ICANN’s Latest Domain Batch Is Full of AI, Brands, and a Little Bit of “It Is What It Is”
For the first time in years, the Internet Corporation for Assigned Names and Numbers — the organization better known by its acronym, ICANN — is once again accepting applications for new top-level domains. For anyone who has ever typed a web address, these are the suffixes that come at the very end of every URL. They are the familiar .com, .org, and .net, but they can also be far more colorful and specific, like .pizza, .geek, or .photography. Top-level domains are the backbone of how we navigate the internet, and ICANN’s decision to open a new round of applications means that the web’s address book is about to get much bigger.
ICANN recently announced that 1,615 applications have already been submitted by 481 different applicants. The numbers alone show how much appetite there is for new digital real estate. But a closer look at the applications reveals a clear cultural and technological theme: artificial intelligence is everywhere. The trend is impossible to miss. Ten different companies, including both Meta and OpenAI, have applied for the .agent domain. Seven companies, including OpenAI, have applied for .agi, the acronym for artificial general intelligence. Six, including OpenAI, have applied for .asi, a clear attempt to ride the wave of President Trump’s push to replace the phrase “artificial intelligence” with “super intelligence.” Four companies have applied for .intelligence, and Meta is one of two applicants interested in the extraordinarily long .superintelligence. OpenAI has applied for a total of 15 top-level domains, while Meta wants an even more ambitious 21.
These applications are not just about staking a claim in the digital landscape. They reflect a deeper competition among tech giants to shape the language of the next era of computing. Companies are not only building AI products; they are trying to own the very words we will use to describe them. A domain like .agent could become the default home for autonomous AI assistants, while .agi and .asi evoke futuristic promises and anxieties. The fact that multiple companies are competing for the same strings suggests that confusion and legal battles may be ahead. ICANN will need to sort out who gets what, and why.
There are, broadly speaking, two types of domains in this new batch of applications. The first is the generic domain, the kind of string that many people and companies might want to use for a variety of purposes. There are hundreds of these in the current round. They range from the whimsical to the mundane: .geek, .haha, .anime, and .jackpot are all in the mix. One of the most delightful applications is .itiswhatitis, a domain that practically writes its own resignation letter. It is hard to imagine a more perfect internet address for the stoic acceptance of life’s absurdities. These generic domains are meant to be used by anyone who wants them, at least in theory. Businesses might use them to create memorable marketing campaigns, individuals might use them for personal websites, and developers might use them for creative projects.
The second type is the purely branded domain, which gives a single company its own exclusive top-level domain for its own internal and external purposes. There are 345 brand applications in this batch, and the list reads like a roll call of major corporations. Applebee’s has applied for .applebees. Bank of America has applied for .bankofamerica. Zillow has applied for .zillow, and Zoom has applied for .zoom. These branded domains are less about public adoption and more about control, security, and identity. A company that owns its own top-level domain can create deep, customized links, protect itself from typosquatting, and build a more cohesive online presence.
Meta, unsurprisingly, has gone all in on its brand architecture. The company has applied for .facebook, .instagram, .messenger, .meta, .frommeta, and .threads. Each of those could serve a different purpose. .facebook might eventually host official pages or user profiles in a way that makes the platform’s presence more distinct. .instagram could become the home of a new kind of social media link. .threads, the company’s Twitter rival, could be used to create memorable usernames or deep links. The applications suggest that Meta is thinking of its domains not just as marketing tools, but as part of the foundational infrastructure of its apps.
Other social and AI companies are doing the same. Bluesky, the decentralized social network that has grown in popularity over the past couple of years, has applied for .bsky. The apparent intention is to weave the domain directly into its usernames, giving people a way to identify themselves that is both elegant and independent from traditional platforms. If that happens, a username might look something like @name.bsky, which would give Bluesky’s users a distinct digital home.
OpenAI has also filed a wide-ranging set of applications. It wants .chatgpt, .codex, .openai, .gpt, and several others. These domains could be used for product pages, developer portals, or even future services. Anthropic, the company behind Claude, has applied for .anthropic and .claude. The race to control AI-related domains is not just about current products; it is about setting up the architecture for whatever comes next.
The history of alternate top-level domains, though, has been mixed. It has mostly been a tough road for these newer suffixes over the last several years. A few have become genuinely popular. The .ai domain, originally assigned to Anguilla, has surged in popularity because artificial intelligence startups love the acronym. The .app domain, promoted heavily by Google, has also gained traction, especially in the mobile app world. But most people have simply never gotten used to typing .restaurant or .babysitting when they are looking for a local business. The .com domain, and a handful of other legacy suffixes like .org and .net, have actually become more important than ever. When people are unsure of a website’s address, they almost always try the .com version first. This habit is so deeply ingrained that companies often pay enormous sums to acquire the .com version of their brand, even if they have perfectly good other domains.
It is therefore interesting to see so many major brands going after their own top-level domains. In the past, only Google and a few other companies have successfully operated branded domains at scale. Google has long used its own .google domain for a variety of purposes, including its search engine and internal pages. But the idea of hundreds of companies managing their own top-level domains is still relatively new. If these applications are approved, we could see a web where bankofamerica, applebees, zillow, and zoom all become their own top-level domains, functioning as self-contained corners of the internet.
This is only one step in the long process of creating a new top-level domain. The applications have been submitted, but nothing will be made official before sometime next year. ICANN has to review each application, evaluate possible conflicts, and decide who gets what. There will likely be disputes over the most valuable and contested strings. The .agent application, for example, has ten interested parties. The .agi application has seven. The competition for the .intelligence and .superintelligence domains could be fierce. Some of these disputes may end up in arbitration or private negotiations. The process is designed to be orderly, but it is also complex and slow.
Still, the sheer number of applications, and the prominence of the companies involved, makes one thing clear: even as the world becomes more filled with and mediated by artificial intelligence, great domain names continue to matter. In an age of apps, apps, and more apps, it might be tempting to think that the classic web address has lost its power. But the companies building the future of technology seem to believe otherwise. They are spending time, money, and legal resources to secure the digital strings that will define their brands for years to come. If the world’s biggest AI companies are fighting for the right to use .agent, .agi, and .asi, then the humble top-level domain is far from dead.
The internet is always changing, but the way we find things online still relies on names. ICANN’s new round of applications is a reminder that the web is not just made of content and code; it is also made of labels. Some of those labels will fail, as many have before. But some of them might become as familiar as .com is today. And someday, when we are all using AI agents to manage our daily lives, we may type a web address ending in .agent without even thinking about it. That future starts now, with 1,615 applications, 481 applicants, and a little bit of .itiswhatitis.
