Tag: Nuclear Regulation

When the SHANTI Act took effect in December 2025, it ended a monopoly that had defined Indian nuclear policy since 1962. For the first time, an Indian private company can hold a licence to build and operate facilities that involve nuclear materials and ionising radiation, under the oversight of a regulator — the Atomic Energy Regulatory Board — that the same Act finally placed on a statutory footing. This is a structural reset, and it deserves the praise it has received.

But a law written principally for fission power generation cannot, on its own, settle how fusion should be governed. That is the next question, and it is a narrower and more technical one than the debate that produced SHANTI. It is also one the rest of the world has spent the last three years answering. India now has the rare advantage of being able to learn from settled precedent rather than improvise.

The argument begins with the physics

Those of us who have spent careers operating India’s experimental tokamaks know that a fusion device and a fission reactor are not variants of the same hazard. A fission reactor sustains a chain reaction in a large inventory of fissile material; its central safety problems are criticality control, the management of decay heat after shutdown, and a very large radioactive source term that must be contained in any conceivable accident. None of these apply to a magnetic-confinement fusion machine. There is no chain reaction to run away. The fuel present in the plasma at any instant is measured in fractions of a gram, and any disturbance — a loss of heating, a loss of vacuum, a magnet trip — causes the reaction to stop within seconds. The plasma cannot melt down because there is nothing sustaining it once conditions are lost.

This does not make fusion free of radiological concern, and no serious person claims it does. There are two real hazards, and they are bounded and well understood: the handling of tritium, a low-energy beta emitter used as fuel; and the neutron activation of the structure surrounding the plasma, which produces radioactive materials in the machine’s own components. These are radiation-safety problems. They are the kind of problem a regulator manages every day for medical accelerators, industrial radiography, and radioisotope production. They are not reactor-safety problems, and a regime designed for reactor safety is the wrong instrument for them.

Real hazards, of a different kind

To say that fusion is not a reactor is not to say it is benign. A fusion device is a large and demanding industrial plant, and several of its hazards are serious — they are simply the hazards that an occupational- and industrial-safety regime exists to manage, not nuclear-accident hazards. Its superconducting magnets store very large amounts of magnetic energy, and an unplanned quench — a sudden loss of superconductivity — must be engineered against so that the energy is released safely. Those magnets are held near absolute zero by substantial inventories of liquid helium and nitrogen, which in an enclosed space are asphyxiation hazards before they are anything else. The plasma heating and current-drive systems operate at extreme high voltages. None of this is unfamiliar to a competent regulator, but all of it is dangerous if managed poorly.

There is a materials hazard as well, and it is chemical rather than radiological. Many magnetic-confinement designs face the plasma with beryllium or tungsten, and the intense plasma–surface interaction generates fine dust inside the vacuum vessel. Beryllium dust is highly toxic: inhaled during maintenance, it can cause berylliosis, a chronic and irreversible lung disease. Controlling it — containment, ventilation, respiratory protection and disciplined access during maintenance — is an industrial-hygiene problem in its own right, entirely separate from radiation protection.

This is precisely why regulating fusion “for what it is” is not the same as regulating it lightly. What it calls for is the right combination: radiation safety for tritium and activated material, conventional industrial and occupational safety for the magnets, cryogenics and high-voltage plant, and hazardous-materials controls for beryllium and tungsten dust. The United Kingdom’s decision to place fusion under the Health and Safety Executive and the environmental regulators is, seen in this light, a choice to govern it through the industrial-safety system — not a choice to go easy on it. For India, the practical implication is that a workable regime must align AERB’s radiation oversight with the country’s conventional industrial-safety authorities, so that an operator meets one coherent set of expectations rather than a fragmented patchwork. The error the world’s leading jurisdictions have now moved to correct is the narrower one: applying the rulebook written for a reactor’s nuclear-accident hazards to a machine that has none of them.

What the rest of the world decided

The pattern across countries is strikingly consistent, and it amounts to a single principle: regulate fusion in proportion to the hazard it actually presents, which is the hazard of radioactive materials, and keep it legally distinct from fission.

In the United States, the Nuclear Regulatory Commission decided in 2023 to regulate fusion machines not as power reactors but under its byproduct-material framework — the same regime that governs radioisotopes and particle accelerators. The reasoning was explicit: fusion does not generate the decay heat that requires engineered emergency cooling, so the regulatory focus should sit on tritium, activation products, and activated dust, which existing materials licensing already handles well. Congress wrote that choice into law in the 2024 ADVANCE Act, and the NRC issued its proposed rule in February 2026. The resulting framework is described, in the agency’s own terms, as performance-based, technology-inclusive and risk-informed — deliberately less prescriptive than the rules for fission plants.

The United Kingdom went further in statute. Its Energy Act 2023 confirmed that fusion energy facilities fall outside the Nuclear Installations Act 1965, so they do not require a nuclear site licence and are not overseen by the Office for Nuclear Regulation. Instead they are regulated by the Health and Safety Executive and the environmental regulators, under rules judged proportionate to fusion’s lower hazard. The government’s stated motive was as much economic as scientific: early regulatory clarity, it found, was a decisive factor in where private fusion companies chose to locate.

Japan is moving the same way, with its expert bodies recommending that fusion be regulated under the radioisotope law rather than the law governing fission reactors, expressly to avoid over-regulation. China oversees fusion devices through its radiation-protection and radioisotope-device regulations rather than its reactor regime. Four very different systems, four different legal mechanisms, one shared conclusion.

What none of these countries did is lower their safety standards. They removed a category error. A technology whose dominant risk is contained radioactive material should be regulated as such — rigorously, but not under a rulebook written for runaway chain reactions and molten cores.

What India should do next

India does not need to repeat this debate from first principles. It needs to make a small number of specific, well-precedented choices within the framework SHANTI has already given it.

First, define fusion machines and accelerator-based neutron sources distinctly in AERB’s subordinate regulations, and place them on the radiation-safety licensing track rather than the power-reactor track. AERB already operates a mature apparatus for licensing medical accelerators, industrial sources and isotope facilities. The earliest commercial fusion-adjacent devices in India — accelerator-driven neutron sources for medical and industrial use — belong on that track by their physics. The Board should say so explicitly, so that applicants are not left guessing which regime applies.

Second, adopt a graded, risk-informed pathway that scales requirements to the device. A research-scale machine, a demonstration plant, and an eventual power plant present escalating but still bounded hazards. The licensing burden should escalate with them rather than being fixed at the level appropriate to a gigawatt fission station. This is the performance-based logic the United States has now codified, and it is the right fit for a sector that will move through several device generations.

Third, provide predictability. Investors fund certainty at least as much as they fund physics. Published licensing pathways, structured pre-application engagement with AERB, and defined timelines for review do more to enable a private sector than any subsidy. The UK experience is unambiguous on this point: clarity itself was the incentive.

Fourth, make liability proportionate. SHANTI already replaced a single statutory cap with a graded liability framework linked to plant capacity, and it gives the Central Government power to relieve facilities of liability obligations where the risk is judged insignificant. Fusion’s small source term is exactly the case that provision was written for. A fusion liability tier set well below that of a fission station would reflect the real risk and remove a needless barrier to insurance and finance.

Fifth, align with the emerging international consensus. As AERB develops fusion-specific guidance, it should draw on the IAEA’s work and on the converging US, UK and Japanese frameworks, so that a design licensed in India can be recognised abroad and Indian regulators can benefit from the operating experience accumulating elsewhere. For a country that hopes to export this technology, harmonisation is not a courtesy — it is market access.

The honest boundary: where the lighter touch must not apply

A credible case for proportionate regulation has to be equally clear about where proportion runs the other way. Not everything that travels under the banner of “fusion” carries fusion’s benign hazard profile.

The clearest example is the fusion–fission hybrid: a device that surrounds a fusion neutron source with a sub-critical fission blanket. Whatever its merits as a route to early net power, such a machine holds bulk fissionable material and breeds fission products. Its hazards are genuinely fission-grade, and it should be regulated as such. The United Kingdom recognised this and deliberately kept hybrids inside its nuclear site licensing regime even as it freed pure fusion from it. India should do the same. SHANTI already reserves the most sensitive parts of the fuel cycle — enrichment, reprocessing, heavy-water production and the management of spent fuel beyond on-site storage — to the Central Government. Any hybrid programme therefore belongs inside the national framework operated by the Department of Atomic Energy and its institutions, in partnership with them, not as a freelance enterprise around them.

This is not a concession reluctantly made. It is the discipline that makes the rest of the argument credible. A sector that asks for proportionate treatment of its low-hazard activities must visibly accept full treatment of its high-hazard ones. That is how a regulator, and a public, come to trust a new industry.

A proportionate, not a permissive, ask

None of this is a request to regulate fusion lightly. It is a request to regulate it accurately. The countries that have separated fusion from fission did not weaken their safety standards; they matched the standard to the science and removed a mismatch that would otherwise have strangled a low-hazard technology under rules built for a high-hazard one.

India has just taken the hardest step. Repealing a six-decade monopoly and giving its nuclear regulator statutory independence was the difficult, structural reform. What remains for fusion is smaller, more technical, and extensively precedented: a clear definition, a proportionate licensing track, predictable timelines, fair liability, and an honest line drawn at the hybrid. Get these right, and India can grow a private fusion sector that does not compete with the national programme that DAE, BARC and IPR have built since the 1950s, but complements it — adding a commercial route, on Indian soil, to capabilities the country has spent two generations learning to master.

Continue reading “Regulating Fusion for What It Is”