The Guinea Gap: Jamaica's $326 Million Bauxite Bet Has No AI In It

An open-pit mining operation. Jamaica's own bauxite pits in St. Ann and St. Elizabeth are being expanded in 2026, still without the sensors and autonomous equipment now standard at major mines elsewhere. Photo via Unsplash.

Jamaica's oldest export industry is having its busiest year in a decade, and almost nobody outside the mining ministry has noticed. Hurricane Melissa knocked bauxite output down 4.4% in 2025, to about 5.63 million tonnes from 5.9 million tonnes the year before. In the same stretch, investment in the sector jumped from US$251 million to US$326 million, a rise of roughly 30%, according to figures reported by the Jamaica Information Service. A sector that just took a hurricane to the face is spending more money than it did before the storm, not less. That only makes sense once you look past the weather and at Guinea, on the other side of the Atlantic, where a supply crunch is handing Jamaica an opening it has not seen in a generation.

None of the US$326 million is buying anything resembling artificial intelligence. It is buying roads, turbines, and trucks that still need a person in the cab. That is the story here: a real industrial boom, grounded in a real geopolitical opening, running entirely on the same equipment and the same methods Jamaica's bauxite parishes have used for decades, while competitors on other continents have already automated the exact same job.

A Rebound Built on a Storm and Somebody Else's Crisis

Hurricane Melissa made landfall near Black River, St. Elizabeth, on 28 October 2025 as a Category 5 storm, and its damage bill across the country ran past US$12 billion. Bauxite haulage roads, some of the same red-dirt tracks that move ore from pit to port, were among the infrastructure hit hardest. A 4.4% production decline in a year like that is, if anything, a sign the industry held up better than most of the country did.

What makes 2026 unusual is that the storm and the investment surge are happening at the same time, for almost unrelated reasons. Companies are not just repairing what Melissa broke. They are expanding capacity into a market that changed shape while Jamaica was busy recovering.

The Guinea Gap

Guinea is the world's largest bauxite exporter by a wide margin, and in 2026 its government tightened the national export cap to 150 million tonnes per year. Analysts tracking the sector put the resulting global supply deficit at roughly 33 million tonnes, a hole in the market that refiners in China, the Gulf, and elsewhere now have to fill from somewhere. Jamaica is one of the somewheres.

The catch is scale. Jamaica's realistic maximum output, even counting the idle Alpart refinery coming back online, is estimated at around 7.7 million tonnes, which covers only about a quarter of the Guinea shortfall. That is not a small number for an island economy. It is a meaningful, real opportunity. It is also nowhere near big enough to define the global market on its own, which is exactly why the smart move is capturing every tonne of extra yield and margin the existing footprint can produce, rather than only adding trucks and roads.

Where the $326 Million Is Actually Going

Walk through the individual projects and the pattern is consistent. Century Aluminum, which operates the Jamalco refinery, is spending about US$80 million on capital upgrades, including a new turbine generator meant to improve energy self-sufficiency, with a target of restoring the plant to its full nameplate capacity of 1.4 million tonnes by the end of May 2026. Windalco has committed roughly US$29 million to a haulage road project in St. Ann that opens access to higher-quality ore reserves. Discovery Bauxite is investing about US$45 million this year in a new haulage road, equipment purchases, and maintenance upgrades. Atlantic Alumina has struck a US$450 million strategic partnership with a facility in Gramercy, Louisiana, tying Jamaican alumina more tightly to a US processing customer.

Then there is Alpart, the largest open question in the sector. The plant, wholly owned by China's Jiuquan Iron and Steel Group (JISCO), has sat idle, and a proposed modernisation programme worth up to US$3 billion across three phases is on the table, with construction potentially starting before the end of 2026 and a relaunch targeted before June 2027. Hon. Floyd Green, Jamaica's Minister of Agriculture, Fisheries and Mining, said in May 2026 that government has to take a definitive decision about the plant's future this year. If Alpart restarts at anything close to its potential 1.7 million tonne addition, it alone could nearly double Jamaica's current bauxite output.

Every one of those projects is a road, a turbine, a generator, or a piece of conventional heavy equipment. Not one of the announced capital plans mentions sensors, predictive maintenance, autonomous haulage, or AI-assisted ore sorting. The industry is scaling up the exact operational model it has run for decades, just with more of it.

What Everyone Else Is Already Running

Rio Tinto's Gudai-Darri mine in Western Australia opened in 2022 as the first "greenfield" mine built from day one around a fully autonomous truck fleet, 42 vehicles with nobody in the cab. The company's own data from its wider Pilbara operations shows what happens when you remove human variability from haulage: eliminating the swings in operator speed, braking, and cornering technique cuts tire wear by up to 20% and fuel consumption by 10–15%. Fortescue Metals Group runs more than 175 autonomous trucks across its Chichester Hub and reports a 30% productivity improvement over manually operated equipment. Industry-wide, more than 1,000 autonomous haul trucks were in commercial operation globally as of 2023, and that number was projected to roughly double by 2026. Mining operators running autonomous haulage report safety incident reductions of up to 50%.

None of that exists in Jamaica's bauxite parishes. Not a pilot, not a single autonomous truck, not a predictive maintenance sensor bolted onto an existing haul fleet. The capital being spent right now at Jamalco, Windalco, and Discovery Bauxite is buying more of the same trucks that Rio Tinto and Fortescue have already started retiring from their own fleets.

Surface mining equipment at work. Globally, machines like this are increasingly paired with predictive maintenance sensors and autonomous haulage systems that Jamaican operations have not yet adopted. Photo via Unsplash.

The Squatting Problem AI Could Actually Solve This Year

Not every fix requires a US$3 billion capital programme. The Jamaica Bauxite Institute, the government agency that manages bauxite lands, has flagged a much smaller and more solvable problem: encroachment. Kemoy Lindsay, the JBI's Director of Bauxite Lands, has described squatting on bauxite lands as a growing issue, with enforcement expected to intensify in the coming financial year. That is a monitoring problem, and monitoring problems are exactly what cheap satellite and drone-based computer vision already handles well in other mining jurisdictions. A JBI that could flag new structures or clearing on its land within days, instead of catching it after a settlement is established, would save money and legal headaches far beyond what the technology costs. It is the kind of AI project that does not require a hurricane-sized capital plan, just the decision to run one.

Why the Gap Persists

The honest answer is not that nobody in Jamaica's bauxite sector has heard of autonomous mining. It is that the capital and the incentives point the other way. Haulage jobs support parish economies in St. Ann, St. Elizabeth, Manchester, and Trelawny in ways that are politically and socially real, and a truck driver's job is not an abstraction to be optimised away without a plan for what replaces the income. Foreign owners like JISCO and Century Aluminum answer to head offices weighing Jamaica against other assets in their global portfolios, and a modernisation decision on the scale of Alpart gets made in board rooms far from St. Elizabeth. And the current investment wave is explicitly reactive: rebuild what Melissa broke, expand fast enough to catch the Guinea window before the price moves elsewhere. Reactive capital tends to buy what it already knows how to buy.

As one haulage contractor working the St. Ann roads put it recently, "Wi jus a try get back to normal, and now dem waan we automate before we done rebuild? One ting a time." That is a fair complaint about sequencing, not about the destination. The point is not that Jamaica should have skipped rebuilding the roads to install robot trucks first. It is that a US$3 billion Alpart modernisation and a new Jamalco turbine are both once-in-a-generation capital decisions, and once-in-a-generation capital decisions are exactly the moment to design in the sensors and the autonomous-ready groundwork, even if full autonomy arrives years later. Retrofitting later always costs more than building it in from the start.

What Vision 2030 Actually Requires Here

Jamaica's Vision 2030 development plan describes a knowledge-driven economy in policy language, and the bauxite sector is the test of whether that language means anything in the industries that already exist, not just the ones being invented from scratch. This is where StarApple AI, the Caribbean's first AI company, founded in Jamaica by Adrian Dunkley, the region's recognised AI leader, fits into the picture. The gap in bauxite is not unique. It is the same gap StarApple AI Jamaica's research has found across the wider economy: businesses and agencies that own the problem and could afford a solution, but have not yet built the readiness to run one. A capital-intensive, foreign-owned, export-driven sector like bauxite is a harder place to close that gap than a small business adopting a chatbot, but it is also where the money and the motive already exist.

The Guinea window will not stay open indefinitely. Prices respond to supply shocks, and other bauxite-producing nations, from Guyana to West African rivals, are watching the same gap Jamaica is chasing. The tonnes Jamaica ships in the next two years while the gap is wide will earn more than the same tonnes shipped once global supply rebalances. Every dollar spent on capacity that also builds in monitoring, predictive maintenance, and a path toward autonomy captures more of that window than a dollar spent on capacity alone.

A Practical Starting Point, Not a Moonshot

Nobody serious is arguing Jamaica should try to build a Gudai-Darri overnight. A realistic sequence looks more modest and more achievable within the current investment cycle. Start with drone and satellite monitoring of bauxite lands, the JBI's own stated priority, which is cheap relative to the enforcement costs it replaces. Add predictive maintenance sensors to the haul truck fleets already being expanded at Jamalco and Windalco, since the trucks are being bought anyway and the sensors are a rounding error against an US$80 million capital programme. Pilot AI-assisted ore grade sorting at the port, where consistency in what gets shipped affects the price Jamaica's alumina fetches on arrival. None of that requires JISCO to resolve Alpart's fate first, and all of it compounds: land monitoring protects the asset, maintenance sensors protect the equipment, and grade sorting protects the revenue.

Full autonomous haulage, the approach Rio Tinto and Fortescue have already proven at scale, is a legitimate longer-term target, but it is also the piece that should be designed into the next major capital decision rather than bolted on after the fact. If the Alpart modernisation goes ahead this year, as Minister Green has said it must be decided, that is the moment to write autonomy-ready infrastructure into the plant's specification, not the year after it reopens.

The Bottom Line

Jamaica's bauxite industry is doing something genuinely difficult right now: rebuilding from a Category 5 hurricane and expanding into a supply gap opened by a crisis on another continent, at the same time, with real money. That is not a small achievement. But every dollar of the US$326 million now flowing into the sector is buying more of the same equipment that got Jamaica to 5.63 million tonnes, not the equipment that would let it produce more from the same footprint, more safely, at lower cost per tonne. The Guinea window rewards whoever gets the most value out of every tonne shipped while it stays open. Right now, that is not the plan.

Frequently Asked Questions

A bulk cargo vessel at dock. Jamaican bauxite and alumina ship out through ports like Discovery Bay and Rocky Point, where AI-assisted grade sorting could protect the price the ore fetches on arrival. Photo via Unsplash.