TL;DR
- Caribbean AI startups grew from roughly 12 in 2023 to about 47 by mid-2026. Seed money is following; money for Series A and later is not.
- Jamaica has a draft AI Act in progress, Trinidad and Tobago runs AI in its energy sector, Barbados is building a fintech cluster, and Guyana is deciding what to do with the data its oil sector produces.
- The Caribbean AI Association and the Caribbean AI Risk Management Council are operating institutions now.
- UWI's Faculty of Engineering reported 380 students in the first cohort of its AI graduate programme in 2025.
- Still thin: Caribbean training data, local compute, production experience and binding regional rules.
Three years ago I could name every Caribbean AI startup from memory and have fingers left over. Today I cannot keep up with the list.
The count went from roughly 12 in 2023 to about 47 in mid-2026. By Silicon Valley standards that is small. By Caribbean standards it means founders now learn from each other, investors can compare companies, and universities can point students to real employers.
I have spent fifteen years watching the region catch technology waves late and on other people's terms. Since about 2022, a small group of people across the Caribbean have been building the institutions to avoid that with AI. This is my mid-2026 survey of where things stand, written as someone building in the field rather than studying it from outside.
Four countries at a glance
| Country | Main driver | Strongest asset | Main gap |
|---|---|---|---|
| Jamaica | Policy and startups | Data Protection Act 2020, National AI Task Force, draft AI Act | Few AI systems in production inside local firms |
| Trinidad and Tobago | Energy sector | Production AI for maintenance and operations | Spreading that skill beyond energy |
| Barbados | Financial services | A central bank willing to work with fintech firms | Scale of the local market |
| Guyana | Oil revenue | Capital and large operational datasets | Keeping data processing and skills in the country |
Jamaica: policy lead, thin production base
Jamaica carries most of the region's AI policy work. The National AI Task Force, on which I serve, has produced detailed recommendations, a draft AI Act is in progress, and the Data Protection Act 2020 gives a legal base that several CARICOM peers still lack.
Kingston's tech cluster around New Kingston and Half Way Tree has grown. StarApple AI, which I founded as the Caribbean's first AI company, now works alongside younger firms building for logistics, health and financial services. Maestro AI Labs plans its public launch in August 2026 and a JSE Junior Market listing in 2027.
AI job postings in Jamaica rose 280% year on year in 2025, from a small base. UTech students are asking for AI coursework they did not ask for two years ago. The weak point is depth: plenty of products are being built, and few AI systems run at production scale inside Jamaican companies.
Trinidad and Tobago: energy pays for AI first
In Trinidad and Tobago the energy sector supplies the data, the money and the reason to invest. bpTT and the National Gas Company run AI for predictive maintenance and operational efficiency as production systems, not pilots. The Trinidad and Tobago AI Society has built a professional community around that work, which gives engineers a way to carry what they learn into other sectors.
The spread into the wider economy is slow. The more useful signal is capital: energy-sector AI budgets are far larger than anything a consumer startup can raise, and some of that money has started going to Caribbean-founded AI firms instead of only global vendors.
Barbados: fintech built on a cooperative regulator
Barbados chose financial services as its niche, and AI sits inside that choice. A cluster of Bridgetown firms use AI for compliance checks, fraud detection and credit risk. The enabling factor is the Central Bank of Barbados, which has worked with fintech operators so that new financial products can be built and tested legally. That openness is not standard across the region.
Financial firms in Barbados face real compliance obligations when they deploy AI, and demand for AI risk frameworks there has been strong. The Caribbean AI Risk Management Council has seen much of its uptake from that sector.
Guyana: who processes the oil data
Stabroek Block revenue has given Guyana capital no CARICOM country had five years ago. The decision in front of it is whether to build local AI and analytics capability or buy services from abroad as part of its infrastructure build-out.
The conversation has started. The University of Guyana is developing data science courses, the Natural Resources Ministry is working with analytics firms on production and environmental monitoring, and ExxonMobil's operations generate data that local analysts could work on. If the processing happens abroad, Guyana keeps the oil revenue and loses the skills and value in the information. People I have spoken with in Georgetown see this clearly. I have not yet seen a funded programme to address it.
The institutions now in place
In 2023 Caribbean AI had builders and no institutions. The Caribbean AI Association now has members in several jurisdictions, a seat in regional policy discussions and a voice in international governance meetings. The Caribbean AI Risk Management Council gives banks, hospitals and procurement teams risk frameworks written for Caribbean conditions rather than adapted from US or EU templates.
On the education side, UWI's Faculty of Engineering reported 380 students in the first cohort of its AI graduate programme in 2025, which means a pipeline that does not depend on people leaving to study. UTech covers the undergraduate level, and Be A Genius reaches secondary and early tertiary students through AI Jamaica programmes.
Talent: enough builders, too few who have run systems in production
The skills gap for building AI demos has narrowed. Tools are cheaper and easier, the UWI programme adds graduates, and remote work lets Caribbean practitioners take international projects from home.
The gap that remains is production experience. Far fewer Caribbean practitioners have kept an AI system running for eighteen months than have built a prototype. Production is where people learn that models drift, pipelines break and rare cases turn up every week. Courses do not teach this. Funded deployment projects that place junior practitioners inside live systems with a senior mentor do, and the region has only a handful.
Capital: seed exists, growth money comes from abroad
Early money is arriving. The 14 West Fund, which I helped set up, has made early-stage commitments to Caribbean AI firms. The Caribbean Development Bank has started adding AI criteria to its technology financing, and diaspora professionals in New York, Toronto and London are writing angel cheques.
A $500,000 seed round is news in the Caribbean and routine in San Francisco, so local founders have to reach revenue early. Several have told me that constraint helped them. The gap is at $2 million and above: no Caribbean fund writes cheques that size into AI companies consistently, so firms that grow raise from US or European funds and take on those investors' priorities.
What is still missing
- Caribbean training data. Global models do adequate generic work and poor work on Caribbean law, economics, English and Creole. Building those datasets needs funding and a CARICOM-level data governance framework that does not exist yet.
- Local compute. Connectivity is uneven, and processing data in the region rather than in Miami or Atlanta needs data centres. Proposals have circulated for years; none operates at scale.
- Binding rules. Data protection laws vary by country, and there is no CARICOM-wide AI framework with enforcement. The region can still design its own rules instead of inheriting them.
- Production experience, covered above, which no amount of new coursework fixes on its own.
The diaspora is sending back experience, not only money
Caribbean engineers in Silicon Valley, London and Toronto have started investing, advising and in some cases moving back part-time to Jamaica, Barbados or Trinidad. Their motive is partly commercial: a market they understand, with less competition than they face abroad.
The bigger contribution is experience. An engineer who spent five years running production AI at a large US company knows how mature teams handle drift, outages and audits. Through mentoring and hiring, that shortens the production gap faster than any curriculum. The Caribbean AI Association has made diaspora engagement a priority, and the technical level at regional AI events has risen as a result.
The next phase is about depth: AI systems running inside Caribbean enterprises, Caribbean-owned datasets, rules that protect people without blocking new products, and ownership structures that keep companies partly Caribbean as they grow. I do not yet know where the Series A money will come from.
What to do next
- Founders: before raising, get one AI system into production with a paying customer and keep it running for six months. That record matters more to investors than a demo.
- Enterprise leaders: pick one internal process and fund a production deployment with a named owner and a 12-month budget, and staff it with at least one junior Caribbean practitioner.
- Governments: publish which public datasets can be opened for AI use and on what terms, and put the CARICOM data governance framework on the next ministerial agenda.
- Universities: add a required production placement to AI graduate programmes, with an industry host and a senior supervisor.
- Diaspora engineers: commit a fixed amount of time, such as two hours a month, to mentoring a Caribbean team through the Caribbean AI Association.
Frequently Asked Questions
What counts as a Caribbean AI startup in the count of 47?
The figure covers firms headquartered in the region whose main product uses AI, including companies applying AI to an existing business problem as well as those selling AI products. There is no official registry, so different counts will vary depending on where that line is drawn. If you need a figure for a grant or investment case, state your definition next to the number.
Which Caribbean countries already have data protection laws?
Several do, with different ages and coverage. The Bahamas passed its Data Protection Act in 2003, Barbados in 2019 and Jamaica in 2020, with Jamaica's Act fully in force from December 2023. Trinidad and Tobago's 2011 Act has only been partly brought into force. Check the current status with each country's data protection office before relying on it, because amendments and commencement orders change the picture.
How can a startup test an AI financial product legally in Barbados?
Start with the regulatory sandbox run by the Central Bank of Barbados with the Financial Services Commission. It lets a firm test a product with real customers under agreed limits and supervision. Contact the regulator before building anything that takes deposits, lends or moves money, and bring a written description of the model, its data and how a customer can appeal a decision.
Does the EU AI Act apply to Caribbean companies?
It can. The Act entered into force in August 2024, with obligations phasing in over the following years, and it applies to providers outside the EU whose AI systems are sold in the EU or whose outputs are used there. A Caribbean firm serving European clients or tourists should map its systems against the Act's risk categories now, because clients will ask.
What should a Caribbean AI founder have ready before approaching investors?
Paying customers, evidence the system works in production, and written rights to the data the model uses. Regional investors also ask about unit costs, since model API fees can eat margins at small scale. A clear plan for where later funding will come from helps, because local funds rarely write cheques above $2 million.