Who Pays for the Energy Transition? A Decision Framework for Fair Costs, Reliable Power, Fiscal Durability and Measurable Results

Executive summary from the original paper

The energy transition is often presented as a choice between technologies. For decision makers, the more consequential choice is how to provide reliable energy service, who finances the infrastructure, who bears the downside risks and how the benefits are distributed. A low price for generation does not establish a low delivered cost. A public subsidy does not eliminate a cost. An emissions reduction claim does not establish that households have dependable access. The 2026 evidence strengthens the case for a service-first approach. Global energy investment is estimated at USD 3.4 trillion in 2026, with about USD 2.2 trillion directed to renewables, nuclear, grids, storage, low-emissions fuels, efficiency and electrification [1]. Yet the capital mix is only one part of the transition. IEA estimates that annual grid investment must rise by roughly 50% from around USD 400 billion by 2030 to meet expected electricity demand [4]. The gap between generation spending and network spending explains why low-cost generation does not automatically produce low-cost, reliable electricity at the meter. The distribution problem is equally material. Tracking SDG7 reports 655 million people without electricity access, including more than 560 millio

n in Sub-Saharan Africa [6]. At the same time, the World Bank reports that only about 40% of utilities in its developing-country sample cover operating and debt-service costs [14]. A transition that expands generation while weakening the utility, raising arrears or leaving vulnerable customers unable to use the service is incomplete. The approval question therefore needs to cover access, bill burden, utility cash flow, fiscal exposure and reliability together. Technology economics have improved, but headline generation costs remain an incomplete decision metric. IRENA reports 2025 global average levelised costs of USD 44/MWh for solar PV and USD 33/MWh for onshore wind, while offshore wind averaged USD 78/MWh [3]. Those values are useful benchmarks, not retail tariffs. They exclude many location-specific costs and values such as network reinforcement, congestion, reserves, curtailment, taxes, financing conditions, backup, land and the timing of generation. The paper therefore distinguishes plant economics from delivered service economics. A second change is the speed of demand and system transformation. Renewables are forecast to add about 4,600 GW of capacity from 2025 to 2030, with solar PV accounting for close to four-fifths of

the increase [25]. Data-centre electricity consumption is projected to more than double from 2024 to around 945 TWh by 2030 in the IEA base case [24]. Battery storage added 108 GW globally in 2025 [11]. These trends increase the value of coordinated planning. Generation, grids, storage, demand response, firm capacity and customer protection must be evaluated as a portfolio rather than as isolated assets. Five mandatory approval tests 1. Affordability across households and firms. 2. Reliability through stressed hours. 3. Effective access and service quality. 4. Fiscal and utility durability. 5. Verified climate performance against a documented counterfactual. Figure 2. Global average renewable generation costs for projects commissioned in 2025. Source: IRENA [3]. These values are not delivered retail tariffs. The International Energy Agency estimates total energy investment at USD 3.4 trillion in 2026, of which about USD 2.2 trillion goes to clean energy categories [1]. Its electricity analysis also identifies grid investment as a constraint, with RAGUNAUTH RAMSAROOP | WHO PAYS FOR THE ENERGY TRANSITION? annual expenditure needing to rise by about 50% from roughly USD 400 billion by 2030 under its demand outlook [4]. These are glo

bal estimates, not a project appraisal for any country. IRENA reports global average generation costs of USD 44/MWh for solar PV and USD 33/MWh for onshore wind in 2025 [3]. The cost of financing and integrating these technologies varies materially by location. This paper proposes five tests for each pathway: affordability across households and firms; reliability through stressed hours; effective access; fiscal durability; and verified climate performance. It joins these tests to a full cost ledger and a scenario register. Every material project should disclose who pays if construction is late, demand disappoints, a currency falls, fuel prices spike or an extreme event damages assets. The central recommendation is to approve complete service packages on the basis of comparable, independently reviewed outcomes. Contracting should allocate risks to those able to manage them. Protection for vulnerable households should exist before prices change. Utilities must fund maintenance. Public reports should retain the original forecast alongside observed performance and revisions. Test Decision evidence Affordability Bills and total energy spending by income group and firm type. Reliability Hourly adequacy, outages, resilience and restorati

on. Access Sustained service quality, connection and use. Fiscal durability Cash outlays, utility gap, debt and contingent liabilities. Climate Measured system emissions against a documented counterfactual. This is a decision framework, not a claim that one technology or policy suits every country. The report sets out methods and stress tests. Local approval still requires local data, legal review, engineering studies and meaningful engagement with affected people. Research question and scope The question is: how should a country allocate the costs and risks of an energy transition while delivering dependable, affordable service and measurable emissions reductions? The unit of analysis is the complete pathway from supply through network and end use. The relevant payers are households, businesses, utilities, taxpayers, investors, international partners and future users. The paper uses a broad definition of payment. Payment includes direct bills and taxes, but also foregone public revenue, utility losses, guarantees, foreign-exchange exposure, land and livelihood costs, pollution damages, transition support, maintenance backlogs and risks shifted to future users. The purpose is not to force every impact into a single monetary number

. It is to prevent important burdens from disappearing because they sit outside the project company or outside the first year of a tariff. The relevant comparison is between complete pathways that deliver a stated service. A generation option with low energy cost but insufficient output in stressed hours should not be compared directly with a dispatchable option without adding the resources needed to reach the same reliability standard. The same principle applies to access. A connection programme should be judged against sustained, affordable service, not connection count alone. This equal-service rule is the foundation for the cost, risk and distribution analysis that follows. This paper addresses electricity and connected uses in cooking and transport. It treats climate mitigation, adaptation, energy security and distribution as related decision criteria. It does not assume that every energy technology belongs in every system. The analysis is global and comparative. No country-specific financial return, tariff or avoided-emissions figure is calculated. Method First, distinguish observed values, institutional estimates, modeled scenarios and the author's recommendations. Second, establish a common service standard, counterfactual

, time horizon and currency for comparing options. Third, draw a payer ledger across investment, operations, social protection and contingent obligations. Fourth, stress the plan under plausible disruptions. Fifth, verify distributional and environmental outcomes with independent data. The method draws on IEA, IRENA, IMF and IPCC publications listed in the references [1-9, 11]. Evidence is treated in four classes. Observed data describe measured outcomes. Institutional estimates use defined methods to infer values not directly observed. Scenarios show conditional futures under stated assumptions.