A telecoms minister from a country of 100,000 people sits across from a satellite provider that serves half the planet. The provider has other markets to sell into. The minister does not have another provider to buy from. Both sides know it before the meeting starts, and the terms that come out of that room reflect it. That imbalance has a name now, and a prime minister said it out loud in front of five hundred people from his own podium.
In Nassau, from September 2 to 4, 2026, The Bahamas hosted the International Telecommunication Union's World Telecommunication/ICT Policy Forum, the first time in the event's three-decade history that a small island developing state has held it. Prime Minister Philip Davis used the platform to name a problem every Caribbean government negotiates around but rarely says plainly. He called it the sovereignty discount: the price a country pays in every negotiation where the other side knows it has no alternative.
The sovereignty discount is the term Bahamas Prime Minister Philip Davis used at WTPF-26 in Nassau to describe why small states accept weaker terms on cables, cloud contracts, and satellite deals. Because no single Caribbean island has the market size to negotiate as an equal, citizens' data gets generated and exported with no return, and the region pays a premium for infrastructure it has no real alternative source for.
Davis did not stop at naming it. He put numbers on the table and proposed a five-point framework for closing the gap. What he described is not an abstract policy grievance. It is the exact terrain where artificial intelligence, built and hosted on the Caribbean's own terms, could shift the balance of a negotiation the region has been losing by default for decades.
The Forum, the Framework, and the Numbers
WTPF-26 was not a small gathering. More than 500 participants from governments, industry, academia, civil society, and the technical community spent three days in Nassau producing five non-binding Opinions on digital divides, green digital transformation, telecommunication and ICT system resilience, space-based connectivity, and ICT-driven innovation and entrepreneurship. ITU Secretary-General Doreen Bogdan-Martin framed the goal as moving "from vision to action," positioning the Opinions as groundwork ahead of the ITU's Plenipotentiary Conference in November 2026, the meeting where binding global telecommunications decisions actually get made.
Bahamas Minister of Innovation and National Development Sebastian Bastian set the tone for what a small state wants out of that groundwork. "Small island developing states must not simply be affected by global digital policy, but be at the table helping to shape it," he told delegates, adding that the goal is to progress "from access to capability to opportunity." He also cautioned against treating technology as a solution on its own, telling the forum not to "fall in love with technology alone."
Davis's sovereignty discount framing gave that ambition a sharper edge. He cited figures that make the imbalance concrete. More than 99 percent of international data travels through undersea cables, which suffer 150 to 200 faults every year, most of them in waters the Caribbean does not control and cannot repair on its own timeline. Advanced economies export more than 60 percent of their services in digitally deliverable form. Least developed countries export roughly 15 percent. Around 80 percent of new digital infrastructure investment across the Global South concentrates in just ten countries, and none of them sit in the Caribbean Sea. Citizens in small states generate data continuously, in every transaction, every search, every mobile payment, and Davis put the region's position bluntly: "we supply it without an invoice."
His proposed fix runs to five points. Treat citizens' personal data as a protected national interest with negotiated terms for any cross-border transfer. Ensure the economic value that data generates returns to the country through local processing and local skills development, rather than leaving with the data itself. Exercise real telecommunications infrastructure authority through enforceable landing rights and open-access obligations, instead of accepting whatever terms a cable operator proposes. Designate telecommunications as critical national infrastructure with mandatory redundancy requirements. And adopt a unified satellite and spectrum policy backed by real technical capability, not a policy document with no engineering behind it.
The through-line across all five points is negotiating power, and Davis was explicit about where the region's strength actually sits. Individually, CARICOM states are small. Collectively, their combined territorial jurisdiction, much of it maritime, gives them a negotiating position no single island can claim alone. His closing instruction to the room was to stop showing up with fifteen separate grievances and start showing up with one position.
Why the Old Playbook Keeps Losing
None of this is new in substance. Caribbean governments have complained about correspondent banking de-risking, unfavourable cloud contracts, and cable landing terms for years. What changed in Nassau is that a sitting prime minister stood on an ITU stage and gave the pattern a name that ties it to a specific, fixable mechanism rather than a vague complaint about fairness.
The mechanism is straightforward. A negotiation only produces fair terms when both sides have a credible alternative if the deal falls through. A cable operator laying a line through the Eastern Caribbean has other routes and other markets competing for its capital. A finance ministry in an OECS state weighing whether to accept its terms has, in most cases, no second cable operator waiting in line and no realistic prospect of financing its own subsea infrastructure. That asymmetry is the sovereignty discount, and it repeats in cloud hosting contracts, in satellite spectrum allocation, and increasingly in the terms under which foreign AI platforms process Caribbean citizens' data.
Every Caribbean government that has tried to negotiate alone has run into the same ceiling. A market of 300,000 people, or even three million, is a rounding error to a global hyperscaler's enterprise sales team. It carries real weight only when combined with fourteen other markets asking for the same terms at the same time, which is precisely the coordination problem Davis's single-position call is aimed at solving.
The Caribbean Has Already Started Building the Alternative
The most useful evidence that this gap is closeable, not permanent, sits in Kingston. Maestro AI Labs, co-founded by two brothers with a core team of three people, built Project Maestro, a sovereign AI model trained on Jamaican data, by adapting and retraining existing open models rather than building from scratch. That choice compressed roughly a year of development into three months and avoided the US$5 million to US$15 million price tag typically attached to building a large language model from the ground up. Nvidia supplied the compute. Young Jamaican tech talent did much of the engineering work alongside the founding team.
The same lab built Credit Garden, a product that turns five years of remittance data across 18 Caribbean and Latin American economies into a credit signal that traditional credit bureaus never captured. That is the sovereignty discount inverted: data generated in the region, processed in the region, and turned into a financial product that pays a return to the region rather than exporting the raw signal to a foreign platform for someone else to monetise.
UWI's Sagicor AI and Financial Services Hub, backed by US$5 million and switched on across every campus in August 2026, points at the same target from the education and workforce side: building the skills base that Davis's second framework point, local processing and local skills development, actually depends on. None of these examples are hypothetical pilots. They are operating today, which is the strongest argument that the sovereignty discount is a solvable engineering and policy problem rather than a permanent condition of being small.
The Core Argument
The sovereignty discount is a negotiating problem, and negotiating problems respond to strength at the table. Sovereign AI infrastructure built on regional data, AI-driven tools that put a real number on what that data is worth, and AI-based monitoring that gives negotiators evidence instead of assumptions all shift power back toward the Caribbean side of the table. Jamaica's Project Maestro already proves the region can build this instead of only asking foreign platforms to be fairer.
What AI Could Actually Do to Close the Gap
Four applications map directly onto Davis's five-point framework, and none of them require a technology the Caribbean does not already have some working version of.
1. Sovereign AI Models That Keep Processing, and Its Value, Onshore
Project Maestro is the proof of concept. A model adapted from open weights and trained on regional data means citizen information, financial transactions, and government records get processed inside the jurisdiction that generated them, under that jurisdiction's own legal and regulatory regime, rather than shipped to a foreign data centre under a foreign country's law. Every CARICOM state does not need to build its own model from scratch. A shared regional model, built once and adapted per country, follows the same logic that makes CARPHA's public health data-sharing work: build the expensive core capability once, regionally, and let individual states plug into it.
2. AI-Driven Data-Value Analytics for Negotiators
Davis's complaint that citizens "supply data without an invoice" points to a specific, fixable gap. Governments entering a hosting or licensing negotiation with a foreign platform currently have no rigorous, defensible estimate of what their citizens' aggregate data is actually worth to that platform. AI-driven analytics, applied to anonymised transaction volumes, usage patterns, and comparable market pricing, can produce that number before the negotiation starts rather than after a bad deal has already been signed. A finance ministry walking into a cloud contract renewal with a modelled valuation in hand negotiates differently than one walking in with only a sense of grievance.
3. AI-Based Monitoring for Undersea Cable and Satellite Resilience
With 150 to 200 cable faults occurring globally each year, and the Caribbean's connectivity running through a small number of physical routes, predictive monitoring matters as much as raw capacity. AI-based anomaly detection, already used by major cable operators elsewhere to flag stress and likely fault points before a full break occurs, gives Caribbean regulators independent data on how well a given cable route is actually being maintained. That data becomes an asset in Davis's fourth framework point, enforceable redundancy requirements, because a regulator citing its own monitoring evidence negotiates from a stronger position than one relying entirely on the operator's own reporting.
4. Pooled Regional Compute and Technical Capacity
Davis's fifth point, a unified satellite and spectrum policy backed by real technical capability, cannot be built by any single OECS state alone. It can be built by CARICOM and the Caribbean Telecommunications Union pooling compute, technical staff, and negotiating expertise the way CIMH and CARPHA already pool climate and health capability across the region. A shared regional AI capacity centre, jointly funded, gives a Grenada or a Dominica access to the same negotiating tools a Jamaica or a Trinidad and Tobago can more plausibly build alone, which is exactly the single-position strength Davis described.
The Economic Stakes
The sovereignty discount is not an abstract fairness argument. It shows up as a real number on every contract a Caribbean government signs for cloud hosting, satellite bandwidth, or cable capacity, and it compounds every renewal cycle a region negotiates from the same weak position. A government that pays retail rates for infrastructure a larger economy would negotiate at scale is a government with less left over for hospitals, schools, and the disaster response its geography makes non-optional.
The upside runs in the same direction. Every dataset processed locally instead of exported represents jobs, skills, and tax revenue that stay in the region rather than following the data offshore. Credit Garden's use of Caribbean remittance data to build a Caribbean credit product, instead of exporting that data to a foreign credit scoring platform, is a small but concrete illustration of what capturing that value actually looks like in practice, not in a policy paper.
What CARICOM and Regional Bodies Need to Do Next
Two months separate WTPF-26 from the ITU's Plenipotentiary Conference in November 2026, where the Opinions adopted in Nassau feed into decisions that carry actual legal weight. That window is short, and three actions need to move inside it.
First, CARICOM needs to convert Davis's single-position instruction into an actual negotiating mandate before Plenipotentiary, not after. A shared position drafted after binding decisions are made is a position that arrived too late to matter.
Second, the Caribbean Telecommunications Union should scope a shared regional AI and data-analytics capacity, built once and accessible to every member state, following the same cost logic that let Project Maestro compress a multi-million-dollar build into a three-month one by adapting existing tools rather than starting from zero.
Third, individual governments should start requiring, in every new infrastructure contract, the kind of independent monitoring data that AI-based cable and satellite anomaly detection can provide. A regulator with its own evidence negotiates the next renewal from a different position than one relying entirely on the vendor's own account of how well the infrastructure is performing.
A Discount That Was Never Permanent
What makes Davis's framing useful is that it treats the sovereignty discount as a negotiating outcome rather than a fact of geography. Small states have always had less market power than large ones. What they have not always had is a name for the specific mechanism translating that size gap into a price, and a five-point plan for closing it. Nassau produced both in the same week.
The Caribbean does not need to wait for that plan to be handed to it from outside the region. Project Maestro, Credit Garden, and UWI's Sagicor hub already show what building the alternative looks like when a small team decides to do it instead of asking a foreign platform to be more generous. The sovereignty discount holds for exactly as long as the region keeps showing up to negotiate without the tools, and the unified position, to make the other side's lack of an alternative matter as much as its own.
Frequently Asked Questions
What is the "sovereignty discount" in the Caribbean digital economy?
The sovereignty discount is a term Bahamas Prime Minister Philip Davis used at WTPF-26 in Nassau in September 2026 to describe the price a small state pays in every negotiation where the other side knows it has no alternative. In practice, it means Caribbean governments accept weaker terms on undersea cable landing rights, cloud contracts, satellite access, and data hosting deals because no single island has the market size or the technical leverage to negotiate as an equal. Citizens generate valuable data continuously, and Davis argued the region currently supplies it without capturing a fair return.
Does the sovereignty discount affect small Caribbean states differently to larger ones?
Yes. The discount scales inversely with market size and worsens the smaller a state is. A country like Jamaica or Trinidad and Tobago has enough population and commercial activity to attract some direct interest from cloud providers and cable operators. An OECS state with under 200,000 people has almost none, and typically inherits whatever terms a larger regional deal, or a larger non-Caribbean market, has already set. Davis's proposed fix, arriving at negotiations with a single CARICOM position rather than fifteen separate ones, is aimed squarely at correcting that imbalance for the smallest states.
How can Caribbean governments start closing the sovereignty discount with AI?
Four practical steps stand out. First, build or adapt sovereign AI models on regional data, as Jamaica's Project Maestro already has, so citizen data gets processed locally instead of exported to a foreign platform with no return. Second, use AI-driven data-value analytics to quantify what a country's data is actually worth before a licensing or hosting negotiation, replacing guesswork with a number. Third, deploy AI-based anomaly detection on undersea cable and satellite links to reduce outages and give negotiators evidence of the resilience investment they are, or are not, getting for their money. Fourth, pool compute and technical capacity regionally through bodies like CARICOM and the Caribbean Telecommunications Union so no single small state has to build AI infrastructure alone.
How much value does the Caribbean lose by not processing its own data?
There is no single published figure for the Caribbean specifically, but the scale of the gap is visible in the numbers Davis cited at WTPF-26. Advanced economies export more than 60 percent of their services in digitally deliverable form, against roughly 15 percent for least developed countries, and about 80 percent of new digital infrastructure investment across the Global South concentrates in just ten countries, none of them Caribbean. Every dataset generated in the region and processed on a foreign platform is a transaction where the processing, the jobs, and the resulting product sit somewhere else, and the region absorbs none of that value beyond the initial service fee, if there is one at all.
What is the difference between digital access and digital sovereignty?
Digital access means a country's citizens and businesses can connect to the internet and use digital services. Digital sovereignty means the country also controls where its data is processed, negotiates its own infrastructure terms, and captures a share of the economic value that data creates. Bahamas Innovation Minister Sebastian Bastian framed this at WTPF-26 as moving "from access to capability to opportunity." The Caribbean has largely solved access. It has not solved sovereignty, which is why a country can have widespread broadband and mobile coverage while still paying a sovereignty discount on every infrastructure deal behind that coverage.
What are the risks of Caribbean AI systems depending entirely on foreign cloud and data infrastructure?
Three risks compound each other. Continuity risk, because a foreign provider can change pricing, terms, or service availability with limited local recourse. Cost risk, because the region pays retail rates for compute and connectivity rather than negotiating from scale, which is the sovereignty discount in its purest form. And data risk, because citizen and government data processed abroad falls under a foreign jurisdiction's legal regime rather than the Caribbean state's own, complicating everything from law enforcement requests to control over how that data gets reused or resold. None of these risks require a crisis to matter. They are a standing cost the region pays every year it does not build alternatives.
What came out of the WTPF-26 forum in Nassau and does it bind CARICOM states?
WTPF-26 ran from September 2 to 4, 2026, in Nassau, organised by the International Telecommunication Union and attended by more than 500 participants from government, industry, academia, and civil society. It was the first time a small island developing state has hosted the forum. The gathering adopted five non-binding Opinions covering digital divides, green digital transformation, telecommunication and ICT system resilience, space-based connectivity, and ICT-driven innovation and entrepreneurship. The Opinions carry no legal force, but they are designed to build consensus ahead of the ITU's Plenipotentiary Conference in November 2026, where binding decisions on global telecommunications policy do get made.
Where is Caribbean AI and data policy headed after WTPF-26?
The immediate marker is the ITU's Plenipotentiary Conference in November 2026, where the Opinions adopted in Nassau will feed into decisions that do carry weight. Davis's call for CARICOM states to arrive at that conference and future negotiations with a single position, rather than fifteen separate ones, is the clearest near-term test of whether WTPF-26 changes anything or simply produced a well-attended forum. Regional institutions such as CARICOM, the Caribbean Telecommunications Union, and CARPHA-style data-sharing models already show the region can coordinate technically when it commits to it. Whether that coordination extends to a unified negotiating position on AI and data infrastructure is the open question the region now has two months to answer.