For billions of years, the Sun has been running the natural fusion process scientists have spent decades trying to reproduce on Earth. That effort is now advancing on multiple fronts — from laser-driven ignition and ultra-long plasma runs to compact tokamaks and giant superconducting magnets.
Bengaluru-based Pranos Fusion has unveiled PRAGYA, a compact, low-aspect-ratio tokamak that the company describes as India’s first privately developed tokamak. Meanwhile, China has demonstrated the scale of its own fusion programme by completing a 582-tonne superconducting magnet for a future fusion reactor.
PRAGYA has produced plasma and is designed as an experimental platform to study and control the superheated state of matter needed for fusion. Its significance also lies in bringing private-sector participation to an Indian fusion programme that has historically been driven largely by government institutions.
Why fusion is so difficult
Fusion powers stars. For future fusion power plants, the most widely studied reaction combines two hydrogen isotopes — deuterium and tritium — to form helium and a neutron, releasing energy.
On Earth, this requires temperatures of roughly 100–150 million degrees Celsius. At those temperatures, fuel becomes plasma, an electrically charged state of matter. Since no conventional material can simply contain plasma at such temperatures, magnetic-confinement systems use powerful magnetic fields to keep it away from reactor walls.
The most established magnetic-confinement approach is the tokamak, a doughnut-shaped chamber that uses magnetic fields to confine plasma. But producing plasma is only the first step. It must remain sufficiently hot, dense and stable for long enough for fusion reactions to become useful.
The energy density of fusion fuel is enormous. The US Department of Energy estimates that the energy released by just 1 gram of deuterium-tritium fuel is equivalent to about 2,400 gallons (9,000 litres) of oil.
If completely fused, about 1 kg of a deuterium-tritium mixture could theoretically release roughly 340 terajoules, equivalent to nearly 95 gigawatt-hours of energy.
At power-plant scale, International Thermonuclear Experimental Reactor (ITER) estimates that a 1,000 MW fusion plant would require only about 250 kg of deuterium-tritium fuel a year, compared with roughly 2.7 million tonnes of coal for a coal-fired plant producing the same amount of power.
Fusion takes several routes
Fusion research is advancing along several fronts: producing more energy in short bursts, sustaining plasma for longer periods and developing the technologies needed for future reactors.
At the US National Ignition Facility (NIF), laser-based experiments are pushing inertial-confinement fusion. In June 2026, NIF achieved ignition for the 11th time, producing 7.9 megajoules of fusion yield from the target.
Earlier, in April 2025, it produced 8.6 megajoules from about 2.08 megajoules of laser energy delivered to the target. These are target-level figures — not net energy gains or electricity produced by the overall facility.
China is simultaneously tackling the engineering needed for larger fusion systems. In June 2026, the Institute of Plasma Physics under the Chinese Academy of Sciences completed and tested a 582-tonne toroidal-field superconducting magnet for its fusion reactor programme.
The magnet measures 21 metres long, 12 metres wide and 3.3 metres high and was described by its developers as the world's largest superconducting magnet built for a fusion reactor. It stores about 120 gigajoules of energy and is designed to generate the powerful magnetic field required to confine high-temperature plasma.
China's EAST tokamak sustained high-confinement plasma for 1,066 seconds in 2025, while its larger HL-3 tokamak achieved ion temperatures of about 120 million degrees Celsius and electron temperatures of about 160 million degrees Celsius.
France's WEST tokamak, meanwhile, sustained hydrogen plasma for 1,337 seconds, or more than 22 minutes, in 2025.
India’s fusion programme
Through the Department of Atomic Energy and the Institute for Plasma Research (IPR), India has built expertise in plasma confinement, superconducting magnets, heating systems, diagnostics and control.
IPR's programme includes the ADITYA-U and Steady-State Superconducting Tokamak (SST-1) tokamaks. India is also contributing major systems to ITER.
The country has recently expanded into spherical tokamak technology. IPR's first spherical tokamak achieved first plasma in December 2025.
A paper published by IPR researchers in August 2026 describes the machine's commissioning and first-plasma experiments, including the use of a 2.45-GHz microwave system to assist plasma initiation.
The machine was commissioned as a low-cost platform for studying low-aspect-ratio tokamak physics, including non-inductive start-up, current drive and shaped plasma operation.
Spherical tokamaks have a much lower aspect ratio than conventional tokamaks, giving them a more compact geometry. They are being studied as a possible route to smaller fusion systems, although the configuration also brings engineering challenges. That leaves India with two parallel developments: a publicly developed spherical tokamak at IPR and a privately developed low-aspect-ratio tokamak at Pranos.
From JET to ITER
Europe's Joint European Torus (JET), which completed its major experimental programme in 2023, demonstrated the potential of deuterium-tritium fuel. Its final campaign released 69.26 megajoules of heat in a single pulse, generating data for future fusion machines.
At the largest scale is ITER, the international tokamak being assembled in France with participation from China, the European Union, India, Japan, South Korea, Russia and the United States.
ITER is designed to demonstrate a burning plasma and produce 500 MW of fusion power from 50 MW of external heating. Under its revised schedule, deuterium-tritium operations are planned from 2039.
Taken together, these programmes show that fusion is no longer a single race for the highest temperature or biggest energy pulse.
NIF is pursuing ignition. EAST and WEST are tackling plasma stability and duration. JET has demonstrated deuterium-tritium operation. China is advancing the large-scale magnet technology needed to confine plasma. ITER is designed to bring these advances together in a much larger experimental system.
From plasma to engineering
Pranos is a deep-tech energy startup pursuing its own compact tokamak architecture, with PRAGYA serving as its experimental platform.
Founded in 2024, the Bengaluru startup is developing an integrated fusion technology stack spanning modelling and plasma-control software, high-temperature superconducting magnets and compact tokamak technology.
A March 2026 paper by researchers associated with Pranos detailed the design and mechanical analysis of PRAGYA's vacuum vessel.
The machine has a plasma major radius of about 0.4 metres, a minor radius greater than 0.18 metres, a planned plasma current of up to 25 kA and a toroidal magnetic field of 0.1 tesla.
The analysis found that the vessel design met the required safety margins under atmospheric pressure, self-weight and thermal stresses.
The problems that remain
Fusion still faces formidable scientific and engineering challenges. Plasma must remain stable while enormous heat loads are removed from the reactor. High-energy neutrons produced by deuterium-tritium fusion cannot be confined by magnetic fields and can damage surrounding materials.
A future reactor would also need a reliable tritium supply. Tritium is scarce in nature, so deuterium-tritium plants are expected to use lithium-containing breeding blankets to produce tritium from fusion neutrons while absorbing heat.
Then comes economics. A reactor could meet its scientific objectives and still fail commercially if it is too expensive, unreliable or difficult to maintain. Materials, magnets, heat removal, fuel cycles and remote maintenance will therefore matter as much as plasma physics.
That is what makes the emerging fusion landscape broader than a race to recreate the Sun.
Fusion is becoming an engineering challenge that spans superconductors, cryogenics, advanced materials, robotics, vacuum engineering, high-power electronics, computing and precision manufacturing.
Decades of public research have already built expertise across many of these areas in India. Private companies can now add another layer — developing hardware, testing alternative designs and building specialised capabilities.
The fusion race, in other words, is no longer just about making plasma. It is increasingly about learning how to control it, contain it and eventually build around it.
