Anguilla Made $85 Million Selling .ai Web Addresses. Only 13% of Caribbean Adults Use the Technology

Pieter van der Heyden, The Battle about Money, after 1570. The Metropolitan Museum of Art, public domain.

Anguilla is a British Overseas Territory of 91 square kilometres. In 2025 it earned $85.3 million from registrations of its country-code domain, .ai, according to PYMNTS.com reporting on the territory's 2026 budget. The money helped put the budget into surplus.

Across the wider Caribbean, only about 13 percent of adults aged 18 to 65 use generative AI tools, according to survey findings reported by the Jamaica Observer in May 2026. The global estimate is roughly 55 percent. The region is being paid for the name of the technology while most of its people are not yet using it.

How Anguilla earns from .ai

The .ai suffix is Anguilla's country code, assigned under the same convention that gave Tuvalu .tv and Colombia .co. For years it was registered mainly by companies with "AI" in their name by coincidence. The generative AI boom made it a label for a whole industry: start-ups, research labs and large technology firms now want .ai addresses, and Anguilla collects a fee on each registration through its registry operator.

For a territory of Anguilla's size, $85.3 million is revenue on the scale of a mid-sized export industry, earned without a factory or a port. It has nothing to do with Anguillians using or building AI. The domain income and the technology economy are linked only by two letters.

Spanish or French, Coffer, 16th or 17th century. The Metropolitan Museum of Art, public domain.

What the 13 percent figure measures

The same regional reporting put generative AI use among micro, small and medium enterprises at about 19 percent. That is higher than the figure for individuals but well below markets with similar internet access.

Connectivity is not the main barrier. Mobile broadband covers most of the region, and the tools run on an ordinary phone. The barriers I see are these:

• No structured first session. Most people have never had thirty minutes to try a tool on their own work with someone who can answer their first questions.
• Unclear permission. Staff do not know whether their employer allows it, so they either avoid it or use personal accounts without telling anyone.
• Trust. People have heard that the tools make things up, and nobody has shown them how to check the output.
• Not knowing where to start. Small business owners in particular cannot see which task to try first.

In StarApple AI's training sessions, from bank boardrooms to community centres, people who get that first structured session usually keep using the tools and pass them on to colleagues and family. So I read the 13 percent as a count of how many people have had that exposure so far, not as a verdict on Caribbean readiness.

A first session that works

The format we use most is simple enough for any employer or trainer to copy:

1. Each person brings one real task from their week: a letter, a report, a spreadsheet, a customer complaint.
2. They do it with an AI tool while a trainer watches and answers questions.
3. They check the output against what they know, and mark every error.
4. They write down one task they will use the tool for next week.

It takes 30 to 45 minutes. Step three matters most, because it builds the habit of checking that makes the tools safe to use at work.

The governance progress in July 2026

The Caribbean AI Forum 2026 ran on 23 and 24 July at UWI St Augustine, hosted by the Caribbean Telecommunications Union with the university and the Artificial Intelligence Innovation Centre, under the theme AI for Caribbean Transformation. The CTU launched the Caribbean AI Task Force's Final Report, the result of nine months of work by more than 35 experts from governments, universities, civil society and business. It sets out a roadmap for harmonised AI governance, shared regional infrastructure and implementation steps for CARICOM member states.

The same month, the UNESCO Office for the Caribbean presented its Caribbean Artificial Intelligence Policy Roadmap to the 126th Special Meeting of CARICOM's Council for Trade and Economic Development on Information and Communication Technologies, which endorsed it. A region long described as fragmented on AI policy now has two frameworks, from a UN agency and from its own telecommunications body, agreed within weeks of each other.

What national strategies and data centres will not do on their own

Jamaica and Trinidad and Tobago are expected to publish national AI strategies before the end of 2026. Both are also where AI infrastructure is arriving: Huawei Cloud's Model as a Service has reached both countries, according to Jamaica Observer reporting in May 2026, and Trinidad and Tobago has signed memorandums of understanding for hundreds of megawatts of AI data centre capacity.

A national strategy directs government procurement and public investment. It does not teach a shop owner in Montego Bay or a nurse in Port of Spain to use the tools. The Task Force report names AI literacy and capacity building as one of its five pillars, alongside governance, infrastructure and support for sector-specific innovation. That pillar will be judged on how many people are trained and using the tools, and nobody has yet set a regional target for it.

The weak point in my argument is the data. The 13 percent figure comes from one set of surveys reported in one newspaper, and I have not seen the method or the country breakdown. A properly sampled regional adoption survey, repeated each year, would tell us whether training is working.

What to do next

1. Employers with 20 or more staff: run the four-step first session above for every team this quarter, and send a one-page note saying which tools staff may use and what data they must never paste in.
2. HR managers: run the three-question adoption survey in the FAQ now and again in three months, so you can see whether the sessions changed anything.
3. Small business owners: pick one weekly task, such as social posts or supplier emails, and do it with an AI tool for four weeks before deciding whether to pay for a subscription.
4. Ministries responsible for ICT: set a numeric adoption target under the Task Force's literacy pillar and fund an annual survey to measure it.
5. The CTU: publish the Task Force report's recommendations as a tracker, with an owner and a date for each, before the next forum.