Guyana's Grid Forecast Jumps 5x: Who's Buying, What's Contracted, and Why Ratepayers Risk the Overbuilt Capacity

Kaieteur News reported in August 2026 that Guyana Power and Light's Development and Expansion Plan projects total electricity demand to surge more than fivefold, from 2,000 GWh in 2026 to 10,441 GWh by 2030, with peak demand climbing from 318 MW to 1,671 MW.^[1] Industrial sales alone are forecast to reach 2,783 GWh by 2030, and GPL management has stated publicly that it is designing the grid explicitly around "large-scale digital infrastructure," including data centres.^[8] The scale of these projections is the claim on which GPL will justify billions in generation and transmission investment. But the projections also hide the most critical question any ratepayer and regulator should ask: how much of this forecast is contracted, and who bears the cost if it does not materialize?

The danger is structural. A forecast is not a contract. GPL has not disclosed whether any portion of the projected data-centre load is backed by signed agreements with demand ratchets, minimum-take provisions, or collateral; nor has it disclosed the cost-allocation mechanism for the assets being built to serve this load. Under standard utility practice in jurisdictions with special contracts, a large anchor load (say, 50 MW or more) triggers negotiation of a custom tariff with the anchor customer. That tariff typically specifies: a demand charge applied to reserved capacity (so the customer pays whether or not it uses the power), a minimum-take ratchet (often 70 to 90 percent of contracted capacity over the life of the contract), and long-term commitment matching the life of the assets (typically 10 to 25 years). The protective point is that the customer, not ratepayers, bears the risk of underutilization. GPL's published plan contains no such disclosure. Either no such contract exists, in which case GPL is building rate-based capacity on a forecast alone, or it does exist and is confidential, in which case GPL and the anchor customer are hiding the term and the ratchet from public scrutiny.

The forecast itself should be interrogated. GPL projects industrial demand of 2,783 GWh by 2030, excluding Linden, and 2,846.6 GWh including Linden's integration by 2029.^[1] The document names data centres as a driver but provides no breakdown of contracted versus announced versus speculative megawatts. Have multiple data-centre developers submitted interconnection requests for the same sites, creating phantom load that GPL is counting once in the official forecast but which may represent duplicate applications shopped to multiple utilities (a documented practice in North America)? What is the historical realization rate of GPL's own past load forecasts? Without these answers, the plan reads as aspirational demand dressed in planning language, serving to justify capex that GPL wanted regardless.

The political economy is also visible. Linden, Guyana's former bauxite hub, is being integrated into the national grid by 2029, and the Mayor has already flagged anxiety about whether electricity costs will rise for residents as a result of the integration and the new industrial loads.^[6] A former Linden Mayor and current Member of Parliament noted that the town once powered a 32 MW steel plant and 16 MW aluminum plant but now struggles with 15 MW total capacity, and that poor planning, not just generation shortage, is the root cause.^[7] This is the classic pattern: GPL projects enormous industrial growth, builds generation and transmission to serve it (claiming ratepayer dollars as investment), and if the industrial load underperforms or delays, the cost burden falls on residential and small-commercial customers who bear the tariff structure. Residential demand is projected to reach 4,467 GWh by 2030, and commercial 1,249.6 GWh.^[1] If industrial demand trails forecast by even 10 to 20 percent, those customers will be paying for stranded capacity.

Guyana's regulatory environment, unlike that of many jurisdictions with mature utility oversight, does not yet have a standing framework for special contracts with demand ratchets, collateral requirements, or cost-isolating large-load tariffs. The immediate intervention available is a formal docket on the Development and Expansion Plan itself, requiring GPL to file: (1) a detailed breakdown of contracted, committed, and announced data-centre and industrial load, with the underlying customer names and signed interconnection/offtake agreements; (2) the minimum-take ratchet and term of any special contract already executed or proposed; (3) a realization-rate audit of prior GPL load forecasts versus actual demand; (4) a least-cost analysis comparing the cost of GPL-built generation with bring-your-own-generation and flexible-curtailment tariffs available to the anchor loads; and (5) a cost-allocation schedule showing which customer classes bear which portions of the new assets. Without these filings and public scrutiny, GPL's plan amounts to a vote of confidence in every forecast the utility has produced, with ratepayers holding all the downside risk.