The March 2026 global nuclear summit reinforced a voluntary declaration aimed at tripling global nuclear capacity by 2050, without adopting a universally binding objective. Several African, Latin American, and Asian countries participated, and the summit addressed the needs of countries wishing to develop new capacity. According to the IEA, total energy demand in Africa would increase by approximately 60% by 2040 under the stated policies scenario; electricity demand would more than double. Nuclear deployment in emerging countries requires sufficient institutional capacity to operate facilities over the long term.
The Essentials
- The March 2026 Nuclear Summit established the objective of tripling global capacity by 2050, positioning Africa, Latin America, and certain Asian countries as priority expansion zones (Crux Investor, IEA World Nuclear Outlook 2026); small modular reactors (SMRs), 25 GW planned, require the same institutional architectures as large reactors, only smaller.
- South Africa, host country of the only operational African reactor (Koeberg), displays 30 to 40% losses on its distribution network and a documented deficit of specialized operators.
- France provides a concrete warning: massive unexpected shutdowns in 2022-2023, linked to maintenance defects and erosion of operational expertise, put up to half the fleet offline simultaneously.
- The United Arab Emirates offer an instructive counter-example: by simultaneously building an independent safety authority, a national training pipeline, and a precise contractual framework for skills transfer, they succeeded in progressively nationalizing the operation of Barakah.
- The real inflection point occurs in the 2025-2035 decade: countries that build their regulatory frameworks, training pipelines, and operational data systems today will be able to operate a reactor in 2040; others will have built a building.
A Reactor Is Only as Good as Its Operator
Without an appropriate institutional framework, nuclear technology creates operational constraints.
Koeberg, in South Africa, has been operating since 1984. It is the only nuclear power plant in service on the African continent and has two reactors. Its longevity is real, but it masks a difficult operational reality: as of March 31, 2025, Eskom’s net debt was approximately 359 billion rands, or roughly 20 billion dollars at the exchange rate; primary sources consulted do not confirm losses of 30 to 40% for the South African network as a whole, and nuclear maintenance skills requirements must be assessed on a case-by-case basis. The reactor itself has experienced unexpected maintenance extensions in recent years, forcing scheduled shutdown extensions well beyond planned windows.
A reactor is only as good as the human organization that operates it. Technology can be entirely transferred; operational governance cannot be reduced to a contract, but contracts can impose and verify essential transfers of data, documents, skills, and responsibilities.
The French experience of 2022-2023 confirms this brutally. In July 2022, 29 of 56 reactors were shut down; this exceptional unavailability resulted notably from maintenance and inspections and repairs following the discovery of stress corrosion cracking, not from an established causality with skills erosion. The Court of Accounts notes fragilities in skills, reinternalization, and outsourcing; the exceptional technical cause of the 2022 shutdowns was stress corrosion cracking. If EDF stumbled over these issues after decades of mastery, the question arises even more forcefully for countries starting from scratch. This article on long-term nuclear operations management develops this thread in detail.
Three Unanswered Questions Before the First Concrete
The March 2026 summit reinforced a voluntary declaration aimed at tripling global nuclear capacity by 2050. It was far less precise about the operational conditions of deployment. Several institutional questions remain to be clarified for countries expressing interest in nuclear energy — in Africa, Latin America, the Middle East, or Southeast Asia.
The first concerns the ownership and governance of real-time operational data. A reactor continuously generates thousands of parameters: temperature, pressure, neutron flux, cooling circuit status. Managing this data requires robust computer systems, clear protocols on their storage sovereignty, and teams capable of interpreting them. In certain deployment models, access to operational data can remain in the hands of the supplier, which can create a dependency whose magnitude depends on the contracts.
The second concerns operator certification and standards applied. Certification of a reactor operator requires on average five to seven years of initial training and apprenticeship. Regulatory capacities vary significantly by country; several new countries must develop them, while others already have experience with nuclear operation and regulation. Without a competent and independent national safety authority, the country cannot fully exercise its regulatory control; however, operational qualification of technicians also falls under the operating organization, under regulatory supervision.
The third concerns financing of maintenance over forty years constrained by structural budgetary limitations. Plants are subject to periodic safety reviews, often decennial, but this does not systematically coincide with a ten-year maintenance shutdown. High debt and cost of capital complicate energy financing in many emerging countries; life-cycle financing needs must be documented project by project.
The Ambition of SMRs Does Not Simplify the Institutional Equation
Small modular reactors (SMRs) are often presented as the answer adapted to the context of developing countries: reduced initial cost, smaller physical footprint, possibility of connecting them to electrical grids less robust than those required by a large 1,000 MW reactor. Several countries are actively exploring these options — in Africa, Ghana and Egypt; in Latin America, Colombia, whose government planning documents provide for the deployment of a 300 MW SMR by 2035; in Asia, Indonesia and the Philippines.
But the promise of operational simplicity of SMRs deserves careful examination. An SMR remains a nuclear reactor. It requires a complete national safety framework, with regulatory adaptations and an approach proportionate to its design; its operators must be trained and authorized, and its maintenance as well as its aging management depend on its design. Certain microreactor concepts are factory-assembled, transportable, and designed to strongly reduce on-site maintenance requirements; however, the actual responsibilities for operation and maintenance depend on the regulatory model and contracts. For a country concerned about its energy sovereignty, this is a trade-off that deserves to be formulated explicitly, not slipped into the technical clauses of a contract.
Chile illustrates this regulatory paradox well. The country has been operating two research reactors since the 1970s. According to its National Nuclear Energy Commission, there are no regulatory barriers to nuclear installation, but development requires modifications to the legal framework and assurance that national safety authorities effectively exercise their functions independently. In other words: the technical door is open, the institutional house remains to be built.
South Africa as a Warning Model, the UAE as a Counter-Example
South Africa is the most instructive case on the African continent, precisely because it already has experience. Eskom has operated Koeberg for four decades. And yet, the country starts from a structurally difficult situation for any nuclear expansion project.
The IRP 2025 does not explicitly provide for a transition from 17 GW to more than 45 GW of renewables by 2030; its new capacity trajectories are mainly presented from 2031 onward. First, an verified loss rate must be established and technical losses, commercial losses, and unbilled electricity must be distinguished. Production deployment, including nuclear, must be coordinated with investments in networks and loss reduction; one does not replace the other. Grid quality is an important condition of the nuclear program, among several interdependent institutional and technical priorities.
At the opposite end of the spectrum are the United Arab Emirates — and the contrast merits serious study. Barakah, the first nuclear power plant in the Arab world, now operates with four reactors. This is not just a technological success: it is a success of operational governance. The regulatory authority (FANR) is now 70% composed of Emirati nationals. Emirati operators followed intensive training programs in South Korea before taking over operations. The transition was made possible by an explicit strategy: first use foreign expertise, then invest massively in training and developing national leadership, with contracted progression indicators from the outset. This model is not directly transferable — the UAE had considerable financial resources and a tradition of major infrastructure projects. But it demonstrates that a country without a nuclear past can build sovereign operational capacity, provided it addresses the institutional question before the technological question.
Skills requirements must be assessed according to countries and projects. Training a nuclear engineer takes ten years between initial study and full operational qualification. Governments that wait until after signing the construction contract to launch training pipelines risk having reactors without qualified operators at the time of commissioning.
What Institutional Architecture Enables Transition from a Failing Grid to Stable Nuclear Operation
The central prospective challenge for the 2035-2045 decade is to identify which countries will truly have the capacity to operate a reactor safely and sovereignly at that horizon, and what will distinguish them from others. The question is as relevant for Ghana or Egypt as for Colombia, the Philippines, or Jordan — which is refining in 2026 a call for tenders for an SMR for electricity production and seawater desalination.
A first scenario would be one of prolonged dependence. Reactors would be built and commissioned, but operation would remain under contractual control of the supplier for an extended period. This trajectory is technically viable and produces electricity.
Prolonged supplier control limits operational sovereignty in the absence of effective transfers of skills, data, and responsibilities to the national operator. It creates a dependency if national capacities and contractual transfers remain insufficient. The modes of operational control depend on contracts negotiated between states and suppliers.
A second scenario would be one of structured skills development. It would require several decisions made before any construction begins. First, the creation or strengthening of a competent national safety authority, sufficiently financed and legally independent in the exercise of its functions, particularly with regard to operators and promotion bodies; it should be able to authorize or certify personnel and regulated activities according to national law. Second, the launch of training pipelines for nuclear engineers, in partnership with universities and foreign operators. Third, the negotiation of contracts specifying ownership of operational data and making skills transfer a verifiable contractual deliverable.
This second scenario is slower. It can delay the commissioning date compared to immediate deployment. But it can help create a sustainable energy asset and strengthen national operational capacities. Commissioning between 2040 and 2045 may be envisageable in certain cases, but no general timeline can be deduced from the IAEA framework.
Between these two scenarios, the signals to watch are few but determinative. The creation of a safety authority with its own budget and independent status is a solid indicator of institutional seriousness. The launch of training partnerships with reference operators, whether EDF, Korea Hydro & Nuclear Power, or the IAEA safety agency, is a second signal. And the nature of data ownership clauses in pre-contracts currently being negotiated is probably the most informative and least visible signal. Colombia, by adopting in 2024 its first comprehensive nuclear safety bill, laid a first stone in this direction. Most of its neighbors and counterparts are not there yet.
Comparison with other institutionally intensive sectors is useful here. Mobile telephony deployment in Africa succeeded because telecommunications regulators existed, frequencies were allocated transparently, and local operators progressively built competence. Failures of major infrastructure projects, meanwhile, often followed the same pattern: imported technology, operation delegated, local expertise never built. The dynamics of skills transfer in other industrial contexts illustrate the importance of the institutional environment in the outcomes of technology projects.
The Actors Building the Foundations
Institutions are already working on these questions, without making headlines. The IAEA works with more than twenty African countries interested in nuclear energy; in March 2025, it had conducted eleven INIR missions in nine African countries as part of its Milestones program, which identifies institutional prerequisites before any construction decision. The Milestones program and INIR missions help countries identify regulatory and infrastructural gaps; the timeline must be documented separately with each commercial agreement. The IAEA also launched in 2026 technical cooperation on SMR planning with several Latin American countries.
On the financing side, the landscape has notably evolved since 2025. The World Bank ended decades of exclusion of nuclear from its financing, partnering with the IAEA to support SMRs, network upgrades, and extensions of existing reactors. The Asian Development Bank updated its energy policy to include nuclear and forged a partnership with the IAEA for countries in Asia and the Pacific exploring this option. The Development Bank of Latin America and the Caribbean (CAF) also concluded an agreement with the IAEA to the same effect. These developments are notable because the absence of complete life-cycle financing is precisely the blind spot that transforms good projects into unanticipated burdens.
The determining question is the following: Are governments that today sign nuclear framework agreements with foreign suppliers including a binding component for skills transfer, complete with verifiable indicators and penalties for non-compliance. The presence of such a provision can shed light on the conditions for nuclear development in emerging countries. UAE experience shows that it is possible. It also shows that it takes time, costs money upfront, and requires sustained political will over two decades — not over a single term.
Sources
- Crux Investor, Nuclear Energy Summit Signals Long-Term Demand Growth, Global Nuclear Policy Coordination and Uranium Market Expectations, March 2026, https://www.cruxinvestor.com/posts/nuclear-energy-summit-signals-long-term-demand-growth-global-nuclear-policy-coordination-and-uranium-market-expectations
- International Energy Agency, World Nuclear Outlook 2026
- Court of Accounts (France), Report on the Management of EDF’s Nuclear Fleet, 2025
- International Atomic Energy Agency, Milestones in the Development of a National Infrastructure for Nuclear Power (ongoing program)
- South Africa, Integrated Resource Plan 2025 (Department of Mineral Resources and Energy)
- Energy for Growth Hub, 2026 Update: Who in Latin America is Ready for Nuclear Power?, July 2026, https://energyforgrowth.org/article/2026-update-who-in-latin-america-is-ready-for-nuclear-power/
- Mondaq / Gowling WLG, The UAE’s Nuclear Transformation: Lessons from Barakah and Beyond, February 2026
- Élysée / Presidency of the French Republic, Statement on Nuclear Energy Financing, March 10, 2026, https://www.elysee.fr/en/emmanuel-macron/2026/03/10/statement-on-nuclear-financing
- RSIS, Nuclear Energy’s New Momentum: Can World Capacity Triple by 2050?, March 2026
- IAEA, Comprehensive Report on Nuclear Technology 2025, GC(69)/INF/9
