A fusion reactor is a machine designed to generate energy by forcing lightweight atomic nuclei together until they merge, releasing enormous amounts of energy in the process. This is the same reaction that powers the sun, and replicating it on Earth requires heating fuel to temperatures exceeding 100 million degrees Celsius. No commercial fusion reactor exists yet, but several experimental machines have demonstrated the physics, and a global race involving government programs and private startups is underway to build the first power-producing plant. The engineering challenges are staggering, and the timeline keeps slipping, but real progress in superconducting magnets, plasma control, and materials science has shifted the conversation from “if” to “when and how much will it cost.”
How Fusion Produces Energy
In a fusion reaction, two light nuclei collide at extreme speed and fuse into a heavier nucleus, releasing energy because the product weighs slightly less than the sum of its parts. That missing mass converts directly into kinetic energy, mostly carried away by the particles that fly out of the reaction. The most commonly pursued reaction combines deuterium and tritium, two heavy forms of hydrogen, producing a helium nucleus and a fast neutron. The neutron carries about 80 percent of the released energy and slams into a surrounding structure called a blanket, where its kinetic energy converts to heat. That heat, in turn, drives a conventional steam turbine to produce electricity.
Getting nuclei to fuse is the hard part. Because atomic nuclei are positively charged, they repel each other. To overcome that repulsion, the fuel must be heated until it becomes a plasma and the particles move fast enough to collide and stick. The threshold conditions a reactor must reach, balancing temperature, density, and the time the plasma stays hot, are collectively known as the Lawson criterion, a benchmark that fusion researchers treat as the minimum bar for a self-sustaining reaction.1Engineering: Open Access. A Review of Lawson Criterion for Nuclear Fusion Meeting that bar in a controlled, repeatable way is the central challenge every fusion reactor design is trying to solve.
Tokamaks and Stellarators
The leading approach to fusion uses powerful magnetic fields to confine the plasma in a doughnut-shaped chamber. The two main variants of this idea are the tokamak and the stellarator.
A tokamak uses a combination of external magnets and an electrical current running through the plasma itself to create the twisting magnetic field that keeps the hot gas from touching the walls. Most of the world’s fusion investment has gone into tokamaks. ITER, the multinational reactor under construction in southern France, is a tokamak, and so is the DIII-D experiment in the United States, where researchers have demonstrated techniques for suppressing dangerous bursts of heat at the plasma edge called edge localized modes, or ELMs. In those experiments, small magnetic perturbation coils eliminated the heat pulses for extended periods, a finding that feeds directly into the design of ITER’s plasma control systems.2Nuclear Fusion. Suppression of large edge localized modes with edge resonant magnetic fields in high confinement DIII-D plasmas
A stellarator takes a different route. Instead of relying on current flowing through the plasma, a stellarator creates the entire confining magnetic field with elaborately shaped external coils. This makes the plasma inherently steadier, since there is no current to disrupt. The downside has traditionally been that stellarators trap and lose energetic particles more easily than tokamaks. Recent optimization work has changed the picture substantially. Researchers have shown that stellarator configurations can be tuned to near-perfect magnetic symmetry, reducing particle losses to levels comparable to those in tokamaks, and that the coils needed to produce those fields are only moderately complex.3The Innovation. A promising approach to steady-state fusion: High-temperature superconducting strong-field stellarator with precise omnigenity Separate design work has produced stellarator configurations with a magnetic field spectrum close to the symmetry of a tokamak while keeping the steady-state advantage.4Physics of Plasmas. Stellarators with the magnetic symmetry of a tokamak The result is that stellarators, once considered a long-shot alternative, are now attracting serious investment from both public agencies and private companies.
Inertial Confinement
The other major strategy skips magnets entirely. In inertial confinement fusion, intense energy is dumped onto a tiny pellet of fuel so rapidly that the pellet implodes and, for a fleeting instant, reaches the density and temperature needed for fusion. The energy source is usually a battery of high-power lasers. The National Ignition Facility (NIF) at Lawrence Livermore National Laboratory in California is the best-known example. NIF’s 192 laser beams converge on a fuel capsule smaller than a pencil eraser, compressing it to conditions required for thermonuclear burn.5Nuclear Fusion. The National Ignition Facility: enabling fusion ignition for the 21st century The approach uses what is called central hot-spot ignition, where the capsule is simultaneously compressed and heated until the core ignites and a burn wave propagates outward through the fuel.6Nuclear Fusion. Ignition on the National Ignition Facility: a path towards inertial fusion energy
In December 2022, NIF achieved a landmark: the energy released by the fusion reactions exceeded the energy delivered by the lasers to the fuel capsule. That result proved the physics of ignition works, but it does not mean a power plant is around the corner. The lasers themselves consume far more electricity than they deliver to the target, and the facility fires only a handful of shots per day. A commercial inertial fusion plant would need to crush and ignite several fuel pellets per second, a manufacturing and targeting challenge that remains unsolved.
Heating the Plasma
Reaching fusion temperatures requires multiple heating systems working together. No single method does the whole job.
Neutral beam injection, or NBI, shoots high-energy atoms into the plasma. Because the atoms carry no charge, they slip past the magnetic field; once inside the plasma they are ionized and transfer their energy through collisions. NBI remains an important tool for both heating the plasma and driving current through it. For future reactors, beam energies in the range of 100 to 200 keV produced from simpler positive-ion systems could be an attractive option during startup and the transition to a self-sustaining burn.7Nuclear Fusion. Neutral beam injection for fusion reactors: technological constraints versus functional requirements Design work for burning-plasma tokamaks is exploring NBI systems delivering around 15 megawatts of power.8Plasma Physics and Controlled Fusion. Beam transmission design for a 120 keV and 15 MW neutral beam injection system for burning plasma tokamak
Another method uses radio-frequency waves tuned to the natural spin frequency of electrons spiraling along magnetic field lines. This electron cyclotron resonance heating, or ECRH, can be aimed precisely at specific locations inside the plasma, making it a useful tool for controlling the plasma’s internal structure and suppressing instabilities.9IOPscience. Electron cyclotron resonance heating and current drive in toroidal fusion plasmas In stellarators, where no large plasma current is needed, ECRH plays an even bigger role because it can start up and sustain the plasma without any current drive at all.
Superconducting Magnets
The magnets that confine a fusion plasma are among the most demanding pieces of engineering in the entire machine. Conventional low-temperature superconductors, cooled with liquid helium to just a few degrees above absolute zero, have been the standard for decades. ITER’s magnets use this technology, and they are already among the most powerful superconducting magnets ever built.
High-temperature superconductors (HTS) are changing the game. HTS materials can carry current at higher temperatures and, crucially, in much stronger magnetic fields. Stronger magnets allow the plasma to be confined in a smaller volume, shrinking the entire reactor. Several private fusion companies are building their designs around HTS magnets, betting that a more compact machine will be cheaper and faster to construct. In tokamak fusion reactors specifically, HTS technology could be the factor that enables a sustained positive power output, where the electricity generated exceeds everything the plant consumes.10Nature Reviews Electrical Engineering. High-temperature superconductors and their large-scale applications Advances in HTS are equally relevant for stellarators, where the complex coil shapes have historically been expensive and difficult to manufacture. A stronger magnet made from HTS tape can be wound in tighter, more intricate shapes while remaining compact.
What the Reactor Walls Have to Survive
Even with magnetic confinement, the materials facing the plasma endure punishing conditions. The inner wall, often called the first wall or the divertor, must handle intense heat loads, a constant bombardment of fast neutrons, and erosion from plasma particles. Tungsten is the current front-runner for these plasma-facing components because of its extremely high melting point and resistance to erosion. Testing of full-scale tungsten components in the WEST tokamak in France has revealed the dominant failure modes: brittle cracking from sudden heat spikes and, under the most extreme steady-state loads, shallow surface melting with slow melt displacement.11Nuclear Materials and Energy. Testing of ITER-grade plasma facing units in the WEST tokamak: Progress in understanding heat loading and damage mechanisms
These components will need regular replacement. In a power plant running around the clock, the high-neutron-flux environment degrades materials faster than in today’s experimental machines, where shots last seconds to minutes. Developing steels and structural alloys that can withstand years of neutron bombardment without becoming too brittle or too radioactive is one of the field’s hardest unsolved problems. The frequency of these replacements directly affects how much electricity a fusion plant can sell, because every shutdown for maintenance is time the reactor is not producing power.
Fuel Options Beyond Deuterium-Tritium
Deuterium-tritium (D-T) is the easiest fuel to ignite, which is why every near-term reactor design uses it. But tritium is radioactive, scarce, and must be bred inside the reactor by bombarding lithium with the neutrons produced by the fusion reactions. Managing that tritium cycle adds engineering complexity and safety requirements.
Proton-boron fusion is often called the holy grail of alternative fuels because the primary reaction produces only helium nuclei and very few neutrons, meaning far less radioactive waste and no need for a tritium breeding blanket.12Physical Review Research. Proton-boron fusion scheme taking into account the effects of target degeneracy The catch is that boron has a much higher nuclear charge than hydrogen isotopes, which means the plasma radiates energy away faster and requires far higher temperatures to sustain. Researchers are exploring ways around this limitation, including firing intense laser beams or accelerated proton beams at a solid boron target instead of trying to maintain a burning plasma in equilibrium.
Helium-3 fusion is another tantalizing option, producing charged particles that could in principle be converted directly to electricity without a steam cycle. The practical problem is supply. Natural helium-3 is vanishingly rare on Earth, roughly 300 parts per million of all helium. The entire U.S. strategic reserve holds about 25 kilograms, and global production from the radioactive decay of tritium amounts to only about 18 kilograms per year.13ScienceDirect / Joule. Future Energy The helium bubble: Prospects for 3He-fuelled nuclear fusion Proposals to harvest helium-3 from the lunar surface, where it has accumulated over billions of years from the solar wind, come up repeatedly but remain firmly in the realm of speculation for now.
Safety and Radioactive Waste
Fusion’s safety profile differs from fission in a fundamental way: there is no chain reaction that could run away. The plasma in a fusion reactor contains only a few grams of fuel at any moment. If confinement fails, the plasma touches the wall, cools instantly, and the reactions stop. There is no mechanism for a meltdown or a Chernobyl-style accident.
That said, fusion is not waste-free. The neutrons produced by D-T reactions slam into the surrounding structure and make it radioactive over time. The dominant waste consists of structural materials such as steel and tungsten alloys.14Nuclear Fusion. Overview on the management of radioactive waste from fusion facilities: ITER, demonstration machines and power plants Crucially, this waste does not contain the transuranic elements like plutonium that make fission waste dangerous for hundreds of thousands of years. Fusion waste is dominated by shorter-lived activation products, plus tritium. A fusion plant actually requires more structural material by mass than a comparable fission plant, but the waste can be recycled or, after a period of cooling, disposed of as low-level waste.15Fusion Engineering and Design. Waste from fusion reactor: A comparison with other energy producing systems
The radiotoxicity comparison over time is striking. In the first few years after shutdown, a fusion reactor’s total radiotoxicity from short-lived activation products can actually exceed that of a fission reactor of similar power. But the decay is dramatically faster. After about a century of cooling, fusion waste radiotoxicity drops to more than 100 times lower than that of a pressurized water reactor. After 500 years, the gap widens further.16Fusion Engineering and Design. Fusion power plants, fission and conventional power plants. Radioactivity, radiotoxicity, radioactive waste That timeline matters enormously for waste management: instead of needing deep geological storage that must remain intact for millennia, fusion waste eventually reaches levels where surface storage or recycling become options.
Material choices matter, too. Designing structural components from reduced-activation alloys, steels that minimize the alloying elements most prone to long-lived radioactive isotopes, can further shorten the time before waste becomes safe to handle. This is an active area of research, with the goal of ensuring that a future fusion plant’s decommissioned components do not require the same disposal infrastructure as spent nuclear fuel.
How Fusion Is Regulated
Because fusion reactors carry none of the risks associated with fissile material, several countries have decided they should not be regulated like fission plants. In the United States and the United Kingdom, fusion machines fall under a different licensing framework than conventional nuclear power plants, one focused on byproduct materials rather than the containment and control of fissile material.17Fusion Engineering and Design. Fusion power plant radiation safety-by-design approach to licensing This distinction is significant for commercialization. Fission licensing is famously expensive and slow, sometimes adding a decade or more to a project timeline. A lighter regulatory path could shave years and billions of dollars from the cost of building and operating a fusion plant, which in turn affects how quickly the technology can compete in electricity markets.
Can Fusion Compete on Cost?
This is where optimism meets hard math. For fusion to be competitive with other energy sources after 2040, generation costs will likely need to come in at or below roughly $80 to $100 per megawatt-hour at 2020 prices. Modeling of early magnetically confined designs, whether large or small, suggests that first-generation plants will struggle to hit that mark. Energy costs are projected to exceed $150 per megawatt-hour, driven by low power availability from pulsed operation, the need to frequently replace internal components, and relatively low thermal efficiency.18Energy Policy. Can fusion energy be cost-competitive and commercially viable? An analysis of magnetically confined reactors
That does not mean the economics are hopeless. The same analysis identifies a path forward through technology improvements: higher-efficiency power cycles, longer-lasting components that reduce downtime, and production standardization as multiple plants are built from a common design. Separate economic modeling of a demonstration-scale fusion plant found levelized costs of electricity competitive with current renewable energy sources when external costs like pollution and climate damage are factored in.19Energy. Approximation of the economy of fusion energy The honest picture is that fusion’s first plants will be expensive, but the technology has a credible scaling story if the engineering problems can be solved.
An important nuance: fusion’s economic case may not depend solely on producing the cheapest electricity in a given market. A dispatchable, carbon-free power source that generates no long-lived waste and requires no geological disposal could command a premium, particularly in regions that lack good renewable resources or face political resistance to fission. Whether that premium is large enough to sustain a fusion industry through its expensive early years is an open question.
Alternative Reactor Concepts
Tokamaks, stellarators, and laser-driven inertial fusion get most of the attention, but they are not the only ideas on the table. A field reversed configuration, or FRC, is a compact magnetic confinement device that operates at high beta, meaning the plasma’s own pressure nearly balances the confining magnetic field. FRC devices have demonstrated that they can reach thermonuclear fusion conditions through magnetic compression, and theoretical models suggest the concept could scale to energy-producing generators.20Journal of Fusion Energy. Fundamental Scaling of Adiabatic Compression of Field Reversed Configuration Thermonuclear Fusion Plasmas At least one well-funded private company, TAE Technologies, is pursuing an FRC-based reactor, betting that its simplicity and compactness could offer a faster route to commercialization than the massive tokamak approach.
Magnetized target fusion, Z-pinch devices, and various hybrid schemes also populate the landscape. The diversity is partly a reflection of how much room there still is for a breakthrough: no one has yet demonstrated a commercially viable design, which leaves the field open to ideas that conventional wisdom once dismissed. The risk, of course, is that resources get scattered across too many concepts without any single one reaching the finish line.
Where the Private Sector Fits In
For most of fusion’s history, the field was almost entirely government-funded, with timelines measured in decades. That has changed dramatically since roughly 2020. Dozens of private fusion companies have collectively raised billions of dollars in venture capital and corporate investment, each betting on a particular combination of physics concept, magnet technology, and reactor scale. Some are pursuing compact tokamaks with HTS magnets. Others are building stellarators, FRC devices, or laser-based systems. A few are chasing proton-boron fusion despite the physics being far more challenging than D-T.
The private sector’s entrance has compressed timelines, at least on paper. Several companies have announced plans to demonstrate net energy or even connect a pilot plant to the grid before 2035. Whether those schedules hold will depend on how quickly engineering problems in materials, tritium handling, and sustained plasma operation can be solved. What the private companies have already accomplished, however, is a shift in perception: fusion is no longer seen exclusively as a government science project. It is increasingly treated as an emerging energy technology with real commercial potential, subject to the same investor pressure and milestone-driven development as any other startup sector.
The Tritium Supply Problem
Even if the physics and engineering come together, a D-T fusion industry faces a practical bottleneck: there is almost no tritium available. Global inventory is estimated at only about 25 kilograms, most of it produced as a byproduct in heavy-water fission reactors in Canada. Those reactors are aging, and no large new source of tritium is on the horizon outside of fusion reactors themselves.
The plan has always been for fusion reactors to breed their own tritium by surrounding the plasma with a blanket containing lithium. When a fusion neutron strikes a lithium atom, it splits into tritium and helium. In theory, a well-designed blanket can produce slightly more tritium than the plasma consumes, allowing the reactor to be self-sufficient and even build up a small surplus to start up the next plant. In practice, no one has demonstrated a complete tritium breeding cycle at reactor scale. The blanket must capture nearly every neutron, the tritium must be extracted efficiently from the blanket material, and losses throughout the fuel cycle must be kept to an absolute minimum. If the breeding ratio falls even slightly below 1.0, the reactor starves itself of fuel. Getting this right is widely considered one of the most critical milestones on the path to commercial fusion.