Nuclear power is not an energy source like any other. A nuclear plant can never be reduced to a unit of electricity generation, just as a nuclear programme cannot be viewed solely as an industrial decision. The atom sits at the intersection of energy sovereignty, scientific capability, financial power, technological diplomacy, national security and, at the far end of the spectrum, the possibility of mass destruction.

This singularity explains nuclear power’s enduring place in international relations. While some countries were closing reactors because of concerns over risk, cost or public opinion, others were consolidating their industries, training engineers, mastering enrichment, exporting power plants and modernising their arsenals. The atom had not disappeared. Its centre of gravity was shifting.

In the twenty-first century, its apparent return reflects less a technological breakthrough than a fundamental change in circumstances. Global electricity demand is rising, economies are attempting to reduce their emissions, power grids must integrate variable renewable generation, and governments are rediscovering the vulnerability created by energy dependence. At the same time, strategic rivalries are intensifying, arms-control regimes are weakening and the major powers are reinvesting in their nuclear forces.

Nuclear power has therefore returned to the centre of two fundamental questions: how can abundant energy be produced without creating excessive dependence, and how can the ultimate form of power be prevented from becoming an instrument of uncontrolled disorder?

The Return of the Atom

The nuclear revival is first and foremost part of a profound transformation of the global energy system. The electrification of transport, industry and heating, the expansion of digital infrastructure, growing demand for air conditioning in emerging economies and the development of artificial intelligence are all increasing electricity consumption. This demand cannot be met simply by adding intermittent generation capacity. It also requires dispatchable power, stronger grids, storage, interconnections and an institutional architecture capable of ensuring the continuity of the system.

Within this architecture, nuclear power has regained particular value. It produces large quantities of electricity with very low emissions across its full life cycle, while requiring relatively little land. Reactors can operate for several decades and provide stable output regardless of weather conditions. This combination of continuity, energy density and low carbon intensity explains why several governments have extended the lifetime of existing plants, revived suspended projects or restored nuclear power to their long-term planning.

According to the International Energy Agency, global nuclear generation reached a new record in 2025 and is expected to continue rising through 2030. At the end of 2024, 63 reactors representing approximately 71 gigawatts were under construction. Three-quarters were located in emerging economies and nearly half in China. The movement is therefore not simply a Western nuclear renaissance. It primarily reflects the migration of construction capacity towards Asia and towards states capable of sustaining an industrial policy over several decades. International Energy Agency

The picture nevertheless remains uneven. Some countries possess ageing fleets that continue to perform well but will be expensive to replace. Others are building rapidly through standardised programmes, integrated supply chains and financing largely supported by the state. Still others are announcing nuclear ambitions without having established the regulatory authority, technical expertise, grid infrastructure or financial capacity required to fulfil them.

The return of the atom cannot therefore be measured by the number of political declarations. It must be measured by the ability to transform ambition into reactors that are actually built, connected, operated and maintained.

An Industry of Sovereignty

Nuclear capability is one of the most demanding tests of a state’s ability to organise the long term. Nearly a century may pass between the decision to build a plant and the completion of its decommissioning. A nuclear programme outlives several governments, multiple economic cycles and sometimes several generations of political leaders. It requires a degree of institutional continuity that is increasingly rare in systems shaped by elections, budgetary constraints and pressure for immediate results.

Building a reactor involves far more than securing a site and selecting a supplier. Engineers must be trained, thousands of components qualified, subcontractors supervised, a rigorous safety culture maintained, grids prepared, financing secured, an independent regulator established and arrangements for waste management and decommissioning designed from the outset. Nuclear sovereignty therefore does not reside merely in owning a power plant. It rests on control of the entire system that makes its operation possible.

This industrial depth creates a hierarchy among states. A small group of powers can design reactors, manufacture critical components, enrich uranium, produce fuel, operate facilities, reprocess certain materials and export an almost complete nuclear package. Others remain dependent on a foreign supplier for components, fuel, maintenance, software, training or financing.

A nuclear contract consequently becomes an instrument of diplomacy. Selling a reactor establishes a relationship that may last sixty years or more. It ties the client country to a set of standards, a particular technology, engineering expertise, fuel supplies and maintenance networks. Russia has long exploited this model through an integrated offering combining construction, financing, training and fuel provision. China is gradually developing a comparable capacity. France, the United States and South Korea are also seeking to preserve or recover their positions in a market that concerns international influence as much as electricity generation.

An exported reactor is therefore never simply a piece of equipment. It creates a lasting strategic relationship between the state that controls the technology and the state that depends on it.

Fuel: The Hidden Geography of Power

Uranium is found in several parts of the world, but the availability of ore is not sufficient to guarantee nuclear autonomy. Between extraction and use inside a reactor lies a complex chain: conversion, enrichment, fuel fabrication, transportation, spent-fuel management and, in some countries, reprocessing.

The most sensitive dependencies are found within these intermediate stages. The International Energy Agency estimates that Kazakhstan accounts for approximately 43 per cent of global uranium production and that more than 99 per cent of enrichment capacity is concentrated among four major suppliers, with Russia alone representing around 40 per cent. This concentration turns the nuclear fuel cycle into an issue of economic security and geopolitical power. International Energy Agency

The European energy crisis that followed Russia’s invasion of Ukraine demonstrated that nuclear dependence does not take the same form as dependence on gas or oil. Fuel accounts for only a limited share of the total cost of nuclear electricity and can be stockpiled for several years, reducing immediate exposure to supply disruptions. Replacing a supplier, however, remains technically difficult. Fuel must be designed, licensed and tested for a particular reactor type. Diversification requires time, investment and close cooperation between industrial companies and safety authorities.

True autonomy does not therefore necessarily mean producing every component domestically. It rests on managed interdependence: several qualified suppliers, adequate reserves, long-term contracts, industrial capacity among allied countries and sufficient technical knowledge to avoid captivity to a single actor.

This distinction is essential. A country may operate several nuclear plants while remaining strategically dependent. Conversely, a state that does not control every stage of the fuel cycle may achieve a high degree of security by intelligently organising its alliances and supplies. In the nuclear sector, as elsewhere, absolute sovereignty is rare. Real sovereignty lies in the ability to withstand coercion.

The Economics of Patience

The principal obstacle to nuclear development is no longer necessarily technological acceptability. It is often financial. A large power plant requires substantial initial investment, immobilised during a construction period that may last a decade. Every delay increases interest costs, postpones revenue and undermines profitability. The final cost therefore depends as much on the price of capital, management quality and regulatory stability as it does on the underlying technology.

Recent Western projects have demonstrated the consequences of losing industrial continuity. When expertise has dispersed, suppliers are no longer qualified, designs change from one project to another and construction resumes after several decades of inactivity, the first reactor effectively becomes a prototype. Cost overruns do not necessarily prove that nuclear power is unviable. They often expose the weakness of the industrial ecosystem responsible for delivering it.

By contrast, standardisation, repeated construction and stable project teams can reduce both delays and risk. Nuclear power follows the economics of series production, but those benefits appear only when orders are sufficiently numerous and closely sequenced. A country that builds one isolated reactor every twenty years cannot preserve the same skills or achieve the same costs as a country committed to a continuous programme.

This reality explains the central role of the state. Few private companies can independently absorb the political, regulatory and financial risk attached to an asset of such duration. The models differ — public financing, government guarantees, long-term contracts, regulated returns or risk-sharing with consumers — but all implicitly acknowledge that short-term electricity markets cannot spontaneously finance infrastructure intended to serve several generations.

In 2025, global investment in new plants and the modernisation of existing fleets was expected to exceed $70 billion, an increase of approximately 50 per cent over five years. This recovery remains below the level required by scenarios involving substantial nuclear expansion. Above all, it shows that nuclear power is once again becoming an instrument of industrial policy rather than merely another category of energy asset. International Energy Agency

Nuclear and Renewables: A False Opposition

The energy debate remains trapped in a confrontation between nuclear power and renewable energy. This opposition is politically convenient but technically incomplete. An electricity system cannot be assessed simply by comparing the theoretical cost of a solar megawatt-hour with that of a nuclear megawatt-hour. It must deliver power when and where it is needed, despite changes in weather, peaks in demand, technical failures and grid constraints.

Solar and wind power can be deployed rapidly and have become highly competitive. Their output, however, varies according to time of day, season and weather conditions. Nuclear power requires more capital and time but provides dense, stable and low-carbon electricity. Hydropower, storage, flexible thermal plants, interconnections and demand management complete the system.

The relevant question is therefore not which technology should eliminate all the others. It is which combination can produce a system that is simultaneously affordable, resilient, dispatchable and compatible with climate objectives. The answer varies according to each country’s geography, resources, industrial structure and institutions.

In states that already possess a safe nuclear fleet, extending the operational life of existing reactors may be among the most effective ways to preserve low-carbon generation. In countries without the necessary expertise, suitable grids or financial capacity, launching a nuclear programme may instead create excessive dependence and divert resources from solutions that are more immediately accessible.

The atom is neither a universal answer nor a condemned relic. It is a strategic option whose value depends on the system in which it operates.

Small Reactors and the Battle for Standardisation

Small modular reactors, or SMRs, now occupy a central place in industrial strategies. Their promise rests on an inversion of the traditional model: build smaller units, manufacture more components in factories, reduce the investment required at each stage and add reactors gradually as demand grows.

They could serve smaller grids, replace certain coal-fired plants, provide industrial heat, support hydrogen production or supply remote sites. Their reduced capacity could also make it possible to finance a programme incrementally rather than immediately absorbing the cost of a large conventional plant.

That promise, however, remains to be demonstrated at commercial scale. A smaller reactor does not automatically produce cheaper electricity per unit of installed capacity. It must compensate for the loss of economies of scale through factory production, standardisation and shorter construction times. Yet mass production cannot emerge without orders, while customers are reluctant to order before costs have been proven. The technology therefore faces a familiar industrial dilemma: a market is required to achieve competitiveness, but competitiveness is required to create the market.

The OECD Nuclear Energy Agency’s dashboard now evaluates projects not only according to technical maturity, but also by their progress in licensing, financing, site selection, supply-chain development and fuel availability. This approach underlines that an SMR’s success depends less on the announcement of a new design than on its ability to complete every stage leading to actual operation. OECD Nuclear Energy Agency

The competition surrounding SMRs is therefore a battle to establish standards. The country that succeeds in licensing, manufacturing and exporting a design at scale may shape a significant part of the future nuclear market. Those that multiply concepts without moving into construction will remain trapped in an economy of demonstrators.

The Boundary Between Civilian and Military Power

Civilian and military nuclear programmes serve different purposes, but they rely on overlapping knowledge, materials and infrastructure. Most technologies used in an electricity-generating reactor do not lead directly to a weapon. Uranium enrichment and plutonium separation, however, are dual-use capabilities. They can support civilian fuel production while bringing a state closer to the technical threshold required for a military programme.

The Treaty on the Non-Proliferation of Nuclear Weapons is built on a compromise. Non-nuclear-weapon states renounce the bomb and accept international safeguards; in exchange, they retain the right to access peaceful uses of nuclear technology. The recognised nuclear powers simultaneously commit to pursuing disarmament. This arrangement has helped limit proliferation, but it remains profoundly asymmetrical. It institutionalises a difference in status between those that possess nuclear weapons and those required to demonstrate that they are not seeking them.

Inspections by the International Atomic Energy Agency constitute the principal verification mechanism. In 2025, the Agency applied safeguards in 190 states with agreements in force. Their effectiveness nevertheless depends on access to facilities, the quality of national declarations, the technical resources available and government cooperation. An inspection can reduce uncertainty; it cannot eliminate every clandestine intention or resolve a political conflict on its own. International Atomic Energy Agency

For many states, civilian nuclear capability consequently produces a form of latent power. It develops the scientific expertise, infrastructure and industrial base that could, under certain circumstances, reduce the time required for a change in strategy. This ambiguity fuels mistrust of national enrichment programmes even when their declared purpose is civilian.

The challenge of the twenty-first century will be to allow nuclear energy to expand without multiplying uncontrolled sensitive capabilities. Solutions exist: stronger safeguards, multinational consortia, international fuel banks, long-term supply contracts and mechanisms for taking back spent fuel. Their success will depend less on technology than on the level of trust among states.

The Return of Deterrence

The military dimension of the atom is regaining a centrality that it appeared to have partially lost after the Cold War. All nine nuclear-armed states are modernising their arsenals, delivery systems or supporting infrastructure. The United States and Russia are renewing their strategic triads. China is rapidly expanding its capabilities. India and Pakistan continue their programmes within an enduring regional rivalry. North Korea is consolidating its nuclear status, while France and the United Kingdom are maintaining their forces in a profoundly altered European security environment.

At the beginning of 2026, SIPRI estimated the global inventory at approximately 12,187 warheads, of which 9,745 were held in military stockpiles and considered potentially available for use. Around 4,012 were deployed, while between 2,100 and 2,200 were maintained at a high level of operational alert on ballistic missiles. The decline in total weapon numbers since the Cold War therefore conceals a more disturbing development: active arsenals are being modernised while transparency is diminishing. SIPRI Yearbook 2026

The expiry of New START in February 2026, without a successor treaty, removed the final major bilateral agreement limiting American and Russian strategic nuclear forces. Its disappearance does not automatically trigger an unlimited quantitative arms race, but it reduces verification mechanisms, increases uncertainty and weakens predictability between the two largest nuclear powers. SIPRI

Deterrence rests on a permanent contradiction. To prevent war, each power must convince its adversary that it would be prepared to use a weapon whose employment could lead to its own destruction. This logic can produce cautious stability when capabilities are known, communications remain open and red lines are understood. It becomes more fragile when doctrines are ambiguous, decision times shorten and several nuclear powers interact simultaneously.

Hypersonic weapons, anti-satellite capabilities, cyberattacks, artificial intelligence and long-range conventional systems further complicate the equation. A cyberattack against a command network may be interpreted as preparation for a strike. A dual-capable missile may be impossible to identify before impact. The destruction of an early-warning satellite may appear to signal an attempt at decapitation. The danger no longer lies solely in a rational decision to initiate nuclear war, but in a sequence of errors, misunderstood signals and algorithms accelerating the pace of decision-making.

Nuclear stability in the twenty-first century will therefore depend as much on the governance of emerging technologies as on the number of warheads.

Risk, Safety and Trust

The history of nuclear power has been permanently shaped by Three Mile Island, Chernobyl and Fukushima. These accidents differed in their causes, scale and consequences, but they demonstrated that a nuclear event immediately extends beyond the boundaries of industry. It affects public trust, national finances, energy policy and sometimes relations between generations.

Nuclear safety is based on preventing rare but potentially severe events. It requires an institutional culture capable of resisting complacency, economic pressure and concealment. A reactor may be technologically advanced yet politically vulnerable if the regulator lacks independence, the operator minimises incidents or the state treats transparency as a threat.

New risks have been added to traditional safety concerns. Facilities must now be protected against cyberattacks, sabotage, drones and the effects of armed conflict. The war in Ukraine demonstrated that a nuclear plant could find itself at the centre of a military theatre, transforming civilian infrastructure into an instrument of strategic pressure. Protecting nuclear sites has consequently become a matter of international law, physical security and crisis management.

Waste introduces another dimension of the long term. The volume of highly radioactive waste is limited compared with the enormous quantities produced by fossil-fuel systems, but its hazards persist across periods far exceeding the normal lifespan of political institutions. Deep geological disposal rests on solid scientific foundations, yet implementation remains slow and politically sensitive. The difficulty is not simply to isolate the material. It is to ensure that decisions made today will remain financed, documented and respected over several generations.

Nuclear power therefore requires a particular form of trust: trust in science, in the state, in the operator and in the continuity of institutions. When any one of these elements collapses, the legitimacy of the entire programme may be called into question.

The Conditions of Power

The twenty-first century will probably be neither entirely nuclear nor free from the atom. Civilian nuclear power will remain a minority share of global energy production, but its strategic importance will extend far beyond its statistical weight. It will contribute to electricity security in certain countries, the decarbonisation of difficult sectors, industrial heat production and, potentially, the emergence of new energy systems built around small reactors.

Its development will nevertheless widen the gap between states capable of sustaining a complete nuclear industry and those able only to import a foreign solution. The former will possess an industrial, diplomatic and scientific instrument. The latter must ensure that energy diversification does not become a form of technological dependence lasting for generations.

In the military sphere, the atom will continue to structure the international hierarchy. Nuclear weapons guarantee neither prosperity, conventional victory nor freedom from vulnerability. They do, however, protect the core of a political regime against certain forms of existential threat and radically alter an adversary’s calculations. It is precisely this political effectiveness that makes disarmament so difficult: states ask others to abandon an insurance policy they themselves consider indispensable.

The nuclear question of this century is therefore not simply how many reactors will be built or how many warheads dismantled. It concerns the ability of societies to govern a form of power whose benefits, costs and risks extend far beyond the present.

Mastering the atom means mastering time: the industrial time of construction, the political time of decision-making, the strategic time of deterrence and the almost motionless time of radioactive waste. Few technologies demand such continuity, expertise and responsibility.

Nuclear capability will consequently remain a measure of real power. Not the power proclaimed in speeches, but the kind that enables a state to finance, build, regulate, protect and transmit an infrastructure across several generations. The atom is not returning because it ever ceased to exist. It is becoming central again because the world is rediscovering the price of continuity — and the fragility of nations that can no longer organise it.

Main Sources