Three Refineries, Not One: The Case Against Guyana’s Binary Refinery Debate

ENERGY & ACCOUNTABILITY

Three Refineries, Not One: The Case Against Guyana’s Binary Refinery Debate

OPINION BY: Hem Kumar–September 2026

 A DEBATE BUILT ON A SHARED, UNEXAMINED PREMISE

Kaieteur News reported this week that energy expert Marta Jara;  a non-resident fellow at the Institute of the Americas, former president of Uruguay’s national oil company ANCAP, and a former Shell executive — has cautioned Guyana against President Irfaan Ali’s stated ambition for a 30,000 barrel-per-day refinery. Her advice, in essence: go big or go home. Jara argues that some studies place the economic threshold for an efficient refinery at around 100,000 barrels per day, and that if Guyana is serious about refining as a business, it should be evaluating a world-scale facility of 300,000 to 500,000 barrels per day — oriented toward export, not merely domestic supply.

Data from Guyana Energy Agency latest available report -2022

Jara is not wrong on the economics she is describing. Refining behaves like a commodity business; scale genuinely is a value driver; and a small, domestically-oriented facility built purely to hedge against import dependence will carry a very different cost structure than a export-scale plant competing on global product margins. Her caution is a legitimate contribution to a debate that, until now, has mostly consisted of the President restating his intentions at press conferences without ever publishing a feasibility case.

But the argument that has taken shape in public — small versus large, 30,000 versus 300,000 — accepts a premise that nobody has actually interrogated: that Guyana’s refining capacity, whatever its ultimate scale, should be built at a single site. That premise deserves the same scrutiny Jara has brought to scale. On both cost grounds and logistics grounds, it does not hold up as the obvious choice, and there is a substantial, well-documented alternative — modular, distributed refining capacity — that neither Ali’s framing nor Jara’s rebuttal has considered in public.

THE ONE DEAL GUYANA ACTUALLY SIGNED TELLS ITS OWN STORY

It is worth being precise about what currently exists on paper, because public reporting ; including this week’s Kaieteur piece — has repeatedly blurred the distinction between a refinery and what Guyana actually contracted for. In February 2025, President Ali announced a partnership with Arkansas-based Curlew Midstream. It was never a refinery deal. It was a bulk fuel storage and trading arrangement — a facility to hold 750,000 barrels of gasoline, diesel, jet fuel and heavy fuel oil, built around a “Fuel Exchange Agreement” for imported, U.S.-refined product, with an initial investment of at least US$300 million.

That deal has not closed. It missed its 2025 year-end target. Court filings in Florida, first reported by the Nashville Banner and later by Stabroek News, allege that former Tennessee Congressman Mark Green and lobbyist-attorney Marc Hebert; who had previously represented a Curlew-linked entity — used confidential information from that relationship to stand up a competing venture, Prosimos, in April 2025, around the same time the pair allegedly suggested in a Washington meeting that they could influence whether Curlew secured the agreement. A second, separate Florida suit, filed by IET Systems founder Donald Gantt against a former employee, also names Curlew Holdings and its CEO as defendants — but that case centres on a diverted Starlink communications proposal, not the fuel deal, and should not be conflated with it.

As of March 2026, President Ali told reporters discussions with Curlew were continuing through the government’s legal team; as of the most recent reporting, construction has not begun and no revised agreement has been finalized. Eighteen months after the announcement, Guyana has neither a refinery nor a functioning storage deal — and the single point of failure inherent in a one-project, one-partner strategy is no longer theoretical. It has already cost the country a year and a half.

 THE MODULAR CASE, BY THE NUMBERS

Modular refineries are an established, regulated category in the industry — not a theoretical alternative. Nigeria’s Federal Ministry of Petroleum Resources, which has built the most developed regulatory framework for this technology anywhere in the hemisphere, defines a modular refinery as a facility built on skid-mounted modules ranging from roughly 5,000 to 30,000 barrels per day; anything designed above that threshold is reclassified as a conventional refinery.                       The economics, drawn from peer-reviewed analysis of Nigeria’s downstream sector, are stark when set against a single large-scale facility.

Sources: Olaniyi, “Developing Modular Refineries in the Downstream Sub-Sector of the Nigerian Petroleum Industry,” International Journal of Engineering Research & Technology, Vol. 11, Issue 3 (2022); current 2026 modular refinery vendor pricing data.

The capital comparison is the headline, but the resilience argument matters just as much for a country with Guyana’s infrastructure record. Nigeria’s own literature on this makes the case plainly: rather than having one conventional refinery meet the demand of an entire country or region, in which any shutdown — mechanical failure, force majeure, sabotage — halts the whole supply, several modular units allow the others to keep operating when one goes down.

For a country whose most recent domestic transport disaster is the still-unresolved MV Barima ferry sinking, and whose varied  construction has spent 2026 alternating between  delays  and an outright  suspension, that is not an abstract engineering preference. It is the single most concrete argument in this entire debate, and it is the one argument Jara’s centralization case cannot answer.

SIZING THE REAL QUESTION: WHAT DOES GUYANA ACTUALLY CONSUME?

Any serious siting debate has to start with demand, and here the numbers cut against both existing proposals more than either side has acknowledged. Guyana consumed an estimated 26,760 barrels per day of oil and petroleum products in 2024 — essentially flat against 2023 — split roughly across distillate fuel oil (48.2 percent), motor gasoline (26.2 percent), and residual fuel oil (18.7 percent). Gasoil made up 23 percent of 2023 petroleum imports and finished gasoline 15 percent, confirming diesel as the dominant product the country actually needs refined.

That figure is worth sitting with. Even a single 30,000 bpd facility, run at full utilization, would already exceed the entirety of Guyana’s domestic consumption. Jara’s own cited economic threshold — 100,000 bpd — is nearly four times national demand. Her recommendation, understood honestly, is not a domestic energy security proposal at all; it is an export strategy, dependent on regional and international product markets Guyana has no existing footprint in. That is a legitimate ambition, and this piece does not dismiss it. But it answers a different question than the one Ali’s 30,000 bpd figure was originally posed to answer, and the public debate has conflated the two.

Three 30,000 bpd modular units sited across the country’s three historic counties — Demerara, Berbice and Essequibo — would together total 90,000 bpd: comfortably above domestic demand, built for security and price insulation rather than export volume, distributed against the country’s actual distribution infrastructure, and, per the cost table above, achievable for roughly half the capital of a single plant at even a third of that combined capacity.

THREE SITES, WEIGHED HONESTLY

Demerara — the anchor.

A Demerara-sited unit serves the country’s largest population and industrial demand center and plugs directly into the existing retail distribution backbone. Guyana’s three major fuel distributors; Guyana Oil Company, Sol Petroleum and Rubis — already operate nationally, with GuyOil alone running more than 50 stations. This is the one leg of the model that requires no new infrastructure bet of any kind; it rides entirely on capacity that already exists.

Berbice — real domestic case, conditional cross-border upside.

A Berbice unit would serve Region 6 demand directly and speak to Region Ten’s documented and repeated power shortfalls in Linden — the same shortfall that has, this year, required the Bosai Mineral Groups Company Inc.  to donate generators to fill a gap the state itself has not closed. On its own domestic terms, this leg needs no further justification.

The cross-border trade case for this leg; the one raised in this debate as a possible extension into Suriname — is real but considerably more conditional than the Essequibo leg below, and it should be reported as such rather than assumed. Guyana currently imports US$96.2 million in refined mineral fuels and oils from Suriname, meaning the existing trade flow across the Corentyne runs in the opposite direction from what a Berbice refinery’s export ambitions would require; any near-term case here is import-substitution, not new market capture.

The physical link that would eventually make a larger cross-border case real — the proposed 3.1-kilometre Corentyne River Bridge between Moleson Creek and South Drain, a roughly US$236 million, 100-year-design-life crossing — remains unbuilt. As of mid-2026, Suriname’s government announced it would finance and build the bridge unilaterally, a claim Guyana’s Foreign Ministry formally and publicly rejected; construction has not started and no financing agreement has been finalized. In the meantime, the ferry alternative, the MV Canawaima, has been suspended since August 22, 2026 over safety deficiencies. A Berbice modular refinery is fully justified on Region 6 and Region 10 grounds alone. Its Suriname trade upside should be understood as a future option contingent on a bilateral dispute this newsroom cannot predict the resolution of — not as a pillar of the present case.

Essequibo / Rupununi — the strongest and most underreported leg.

This is the site the current public debate has entirely missed, and it is the one with the clearest evidence behind it. Guyana is building, with its own oil revenue and largely on its own initiative, a 680-kilometre highway from Georgetown to Lethem on the Brazilian border — a project industry estimates could cut freight time to the Panama Canal route by eight to ten days by giving Brazil’s northern state of Roraima a shorter path to the Atlantic than the existing Amazon River route through Manaus. Roraima’s exports to Guyana have grown from roughly US$1.6 million in 2018 to somewhere between US$36 million and US$50 million by 2024–2026, depending on the source consulted; a more than twenty-two-fold increase in six years.

In February 2026, Roraima’s state government, Guyana’s national government, and TriStar — which now manages Guyana’s new, Panamax-capable deep-water port — signed a memorandum of understanding to formalize the relationship.

A modular refinery sited to serve this corridor would not be betting on speculative demand. It would be positioning to supply the return leg of a trade relationship that is already growing rapidly in one direction — Brazilian goods flowing into Guyana — with refined fuel product as the commodity flowing back. That is a materially stronger evidentiary case than anything offered for either the 30,000 bpd or 300,000 bpd proposal currently on the table, and no public official on either side of this debate has raised it. The honest caveats: no cross-border cargo treaty yet exists to govern trade of any kind along this corridor, fuel included, and a stretch of the Guyanese side of the highway remains unpaved. Both are solvable in the same window the refinery itself would take to build.

THE HINTERLAND LEG: A PLAUSIBLE CASE THE STATE HAS NO DATA TO SUPPORT

A fourth possibility raised in the course of this newsroom’s reporting deserves an honest accounting rather than a confident claim either way. Region Seven — Cuyuni-Mazaruni — produces an estimated 70 percent of Guyana’s gold output, and Bartica, at the confluence of the Essequibo, Mazaruni and Cuyuni rivers, is the established fuel and supply chokepoint for the country’s interior mining sector, a role it has held for more than a century. The qualitative case for a hinterland-sited modular unit; proximity to an existing high-diesel-consumption industry, an established river-and-road logistics network already centered on Bartica — is genuinely plausible.

It is also, at present, entirely unquantifiable. No published data exists; not from the Guyana Geology and Mines Commission, not from the Guyana Energy Agency, not from any other public source this new-media could locate — breaking down diesel or fuel consumption by the mining industry. The Guyana Energy Agency’s Energy and Energy Statistics Division tracks petroleum import volumes at the national level only. There is no regional or district disaggregation of any kind.

This news-media is not prepared to cost a fourth refinery site on a plausibility argument alone, and readers should treat this section as exactly that — a plausibility argument, not a proposal. But the absence of the data itself is worth stating in its own right: a government that has spent eighteen months unable to close a single fuel storage deal, and that has offered no feasibility study for either refinery proposal now before the public, also does not track — or at least does not publish — the basic consumption data that would let anyone outside the Ministry of Natural Resources site a fourth facility responsibly. That is not a footnote. It is the same failure of institutional capacity running underneath the entire debate.

WHAT THIS ARGUMENT IS NOT

This is not a rebuttal of Marta Jara’s underlying economics. A world-scale, export-oriented refinery of 300,000 to 500,000 barrels per day, competing on Caribbean and northern South American product markets, may well be a sound long-term ambition for a country producing close to 900,000 barrels of crude per day and rising. Nothing in the distributed model argued for here forecloses that possibility, and the two are not, in principle, mutually exclusive — a country could pursue domestic-security modular capacity now while separately evaluating an export-scale mega-project on its own timeline, its own financing, and its own feasibility study.

What this argument does reject is the framing that has dominated the public conversation since Kaieteur’s report — that Guyana’s only live choice is between one small plant and one large plant, both sited wherever the government’s negotiating partner of the moment happens to propose. On cost, on build time, on resilience, and on the actual geography of Guyana’s demand and trade relationships, that framing has never been defended. It has only been assumed.

 WHAT ACCOUNTABILITY REQUIRES HERE

• The Ministry of Natural Resources and the Guyana Energy Agency should publish a regional and district-level breakdown of petroleum product consumption and imports — the absence of this data is not a technical oversight; it is a precondition failure for any rational siting decision, distributed or centralized.
• Any renewed negotiation with Curlew Midstream, or any successor arrangement, should be conducted and disclosed with the transparency the current 18-month record has conspicuously lacked, given that the collapse of that single deal is now the subject of active federal litigation in the United States.
• Before either the 30,000 bpd or the 300,000–500,000 bpd proposal advances further, the government owes the public a comparative feasibility study that at minimum weighs the distributed modular alternative on its own terms — not as an afterthought to a decision effectively already made in press-conference remarks.
• The Guyana Geology and Mines Commission should be pressed, specifically, on whether it tracks — even internally — fuel consumption data for the country’s mining districts, and if so, why that data has not been made public.

None of this requires Guyana to choose smallness. It requires Guyana to choose deliberately — on the basis of where its people, its industries, and its fastest-growing trade relationships actually are, rather than on the basis of whichever number was most recently floated at a conference podium.

— The Board