The Magnetic Cathedral
In a valley north of Marseille, on a 180-hectare site carved from scrubland and vineyard, the largest scientific instrument ever built is taking shape . The central component — a magnetic coil called the central solenoid — will generate a field 280,000 times stronger than Earth's own . When energised, it will induce and sustain a current of 15 million amperes through a volume of superheated gas for up to 500 seconds at a time . The goal is modest only by comparison to its ambition: to prove that humans can replicate the physics of stellar cores, fusing hydrogen into helium and releasing energy in the process, not once or twice in a laboratory flash, but continuously, controllably, at industrial scale.
This is ITER — the International Thermonuclear Experimental Reactor, though the acronym, borrowed from Latin, means simply "the way." It is the culmination of magnetic fusion research stretching back more than half a century, the hinge between today's physics experiments and tomorrow's power stations . Seven members — the European Union, the United States, Russia, China, Japan, South Korea, and India — signed the formal agreement in 2006 to share the cost and the risk of building it . The project has survived the financial crisis, the fraying of the liberal international order, and Brexit (the United Kingdom, no longer an EU member, opted to remain a participant ). Construction of the tokamak complex basemat began in August 2014 . The timeline has slipped repeatedly; at one point the start date was pushed to 2019 . Yet the work continues, a kind of slow-motion moonshot, because the prize — if it can be won — is civilisational.
"ITER is one of the most ambitious energy projects in the world today."
The reactor will not generate electricity. It will not, in itself, solve climate change. It is, explicitly, an experiment: the penultimate step before a demonstration fusion power plant can be designed . What it will do, if successful, is produce a "burning plasma" — a self-sustaining fusion reaction in which the majority of the heat comes not from external sources but from the alpha particles born in the fusion process itself . This has been the holy grail of magnetic fusion research for decades. No machine has achieved it. ITER's designers believe their creation will.
The Physics of Confinement
Fusion, in principle, is simple. Take two isotopes of hydrogen — deuterium and tritium — and collide them with enough violence that their nuclei overcome electromagnetic repulsion and merge . The result is a helium nucleus, a spare neutron, and a flood of energy, released according to Einstein's equation as a tiny amount of mass converts to kinetic energy. This is what powers the sun. The difficulty is that the conditions required — temperatures exceeding 100 million degrees Celsius — vaporise any container. The fuel must be held not by walls but by magnetic fields, in a configuration called a tokamak, a doughnut-shaped chamber where the plasma is confined by coils arrayed around and through its core .
The central solenoid is the backbone of this system . It runs vertically through the centre of the tokamak, and when pulsed, it induces the plasma current that heats the fuel and shapes the magnetic cage. ITER's solenoid will reach a field strength of 13 Tesla , an intensity that requires superconducting magnets cooled to within a few degrees of absolute zero. The engineering is formidable: the machine must sustain these conditions for minutes at a time, long enough for the plasma to stabilise and the fusion reactions to dominate the energy balance.
The prize is a plasma volume of 830 cubic metres, more than eight times larger than the biggest tokamak operating today . Size matters in fusion. The energy lost from the plasma scales with its surface area; the energy produced scales with its volume. Make the reactor large enough, and the ratio tips in your favour. ITER is designed to cross that threshold, producing ten times more fusion energy than it consumes to heat the plasma — a milestone no magnetic fusion device has reached .
In parallel, researchers pursuing inertial confinement fusion — a different approach using lasers to compress fuel pellets — recently reported that their experiments achieved alpha-particle self-heating, with fusion yield exceeding the energy delivered to the fuel by a factor of two or more . It is a proof of concept for the underlying physics, though the path from laboratory pulse to power plant remains unclear. ITER represents the magnetic fusion community's answer: not a single flash, but a sustained burn.
The Fuel Question
Deuterium is abundant. It can be extracted from seawater at modest cost. Tritium is not. It is radioactive, with a half-life of 12.3 years, and it does not occur naturally in useful quantities . Today's supply comes largely from the cooling systems of heavy-water nuclear reactors, a flow that is both limited and geopolitically concentrated. If fusion is to become a practical energy source, it must produce its own tritium.
This is where the breeding blanket enters the picture. In a deuterium-tritium fusion reaction, the neutron carries away 80 per cent of the energy released. It is not confined by the magnetic field; it slams into the reactor wall. Surround that wall with a blanket containing lithium, and the neutrons will transmute the lithium into tritium, which can be harvested and fed back into the plasma. The reactor becomes self-sustaining, at least in principle .
ITER will not operate with a full breeding blanket — that is a task for the demonstration plants that will follow. Instead, it will host a Test Blanket Module programme, installing small sections of prototype blanket designs from different member states into its chamber walls . The goal is to prove the concept, to measure the tritium production rates, to validate the materials under neutron bombardment. It is a critical link in the chain. A fusion reactor that cannot breed its own fuel is a laboratory curiosity, not an energy source.
The tritium question also shapes ITER's operational profile. The machine will begin with hydrogen and helium plasmas, used to commission the systems and refine control techniques. Only later will it transition to deuterium-tritium reactions, when the physics is well understood and the neutron flux can be managed. The tritium itself is hazardous — not in the quantities that will be present in the plasma at any moment (on the order of grammes), but cumulatively, and in the activation of the reactor structure. The engineering must account for this: remote handling, shielding, containment. Fusion is often called clean, and relative to fission, it is. But it is not without radiological consequence.
The Disruption Problem
Plasmas are unstable. The forces that confine them are delicate, and if the configuration is perturbed — by impurities in the fuel, by fluctuations in the magnetic field, by the plasma's own turbulent dynamics — the result can be a disruption: a sudden collapse in which the plasma's energy is dumped into the surrounding structure in milliseconds . In a machine the size of ITER, with a plasma current of 15 million amperes, a disruption is not a minor event. The magnetic forces generated can be large enough to shift components weighing tonnes. The heat load can exceed the tolerance of the wall materials.
Understanding and mitigating disruptions is one of the central challenges for ITER . Smaller tokamaks have developed techniques to detect the precursors of instability and either suppress them or trigger a controlled shutdown. ITER will inherit and extend these methods, but the scale is unproven. The machine is designed to tolerate a certain number of disruptions over its lifetime, but minimising their frequency is essential both for operational reliability and for the longevity of the reactor components.
This is the kind of problem that cannot be fully solved on paper. ITER is, in this sense, a test bed not only for the physics of burning plasma but for the engineering of disruption control, remote maintenance, and long-pulse operation. The data it generates will inform the design of the demonstration plants that follow, refining the architecture, the materials, and the control systems.
The Politics of Patience
The project has always been as much a geopolitical exercise as a scientific one. The idea of an international fusion reactor emerged during the Reykjavik summit in 1986, a moment of thaw in the Cold War . The formal agreement came two decades later, in 2006, after years of negotiation over cost-sharing, site selection, and governance. Cadarache, in southern France, was chosen as the European candidate, and ultimately as the host site, in a decision that balanced technical suitability with political compromise .
The cost has been a persistent source of tension. Early estimates proved optimistic; the budget has grown substantially, and timelines have stretched . For a project of this scale — described as the most expensive science experiment ever attempted — such overruns are perhaps inevitable, but they test the patience of member states and the goodwill of domestic publics. The European Union, which bears the largest share of the cost, has defended its commitment, framing fusion as a long-term investment in energy security and climate mitigation . Other members have periodically grumbled but remained aboard, bound by the recognition that no single nation could afford to build ITER alone.
Brexit introduced a wrinkle. The United Kingdom, a major contributor to European fusion research, faced the prospect of exclusion from ITER as it left the EU. After negotiation, a pathway was found for the UK to remain a participant , preserving continuity for British researchers and industry. It was a rare moment of pragmatism in a fractious divorce.
The Horizon Beyond
ITER is not the endgame. It is, by design, the experimental step between today's plasma physics machines and tomorrow's demonstration power plants . If it succeeds — if it produces a burning plasma, if it validates the breeding blanket concept, if it demonstrates long-pulse operation and disruption control — then the next step becomes feasible: a fusion reactor designed not to produce data but to produce electricity, to feed the grid, to prove that fusion can be economically viable.
That demonstration plant does not yet exist, even on paper. It will take the lessons of ITER and incorporate advances in materials, superconductors, and plasma control that are still in development. The timeline is uncertain. Optimists speak of the 2040s; sceptics note that fusion has been "thirty years away" for the past sixty years. The difference now, proponents argue, is that the science is settled. The physics of fusion is not in doubt. What remains is engineering: can we build machines robust enough, efficient enough, and cheap enough to compete with other low-carbon energy sources?
The answer will be written, in part, in the data that flows from ITER's first plasma campaigns. The machine is expected to achieve plasma energy breakeven — more fusion energy out than heating energy in — but not engineering breakeven, which would account for the energy needed to run the entire facility . That is a task for the demonstration plants. ITER's role is to prove that a burning plasma is achievable, controllable, and scalable. Everything else follows from that.
The Weight of the Wager
On the construction site in Cadarache, work proceeds with a kind of methodical urgency . The tokamak complex rises in stages, a cathedral of steel and concrete and superconducting coil. The technical challenges are immense: components manufactured on different continents must fit together with sub-millimetre precision; systems that have never been tested at this scale must work in concert; a machine designed to contain stellar temperatures must do so reliably, repeatedly, for years.
The project is, in one sense, a bet: that the physics of burning plasma will yield insights and techniques that justify the investment; that the international collaboration will hold together long enough to see the work through; that fusion, which has eluded practical realisation for so long, will finally cross the threshold from laboratory curiosity to engineering reality. It is a bet underwritten by seven governments, representing half the world's population, at a cost measured in tens of billions.
But it is not a blind bet. The science is grounded in decades of experiment and theory. The design is informed by the hard-won experience of earlier machines. And the need is undeniable. Humanity requires low-carbon, baseload energy at a scale that renewables alone cannot easily provide. Fusion, if it can be made to work, offers that: fuel abundant in seawater, no long-lived radioactive waste, no risk of meltdown, no carbon emissions. The prize is large enough to justify the patience.
ITER will not deliver that prize. It is the penultimate step, the proof of principle, the machine that opens the way . If it succeeds, the demonstration plants that follow will carry the work forward. If it fails, or if the costs prove insurmountable, the dream of fusion power will recede again, perhaps indefinitely. The next decade will tell. In a valley in southern France, the magnets are being wound, the chamber is being assembled, and the world is waiting to see if the sun can be brought to earth.