Nuclear Fusion

The Sun gets away with fusion because it has 333,000 Earth masses of gravity doing the confinement for free. Earth has to replace that with magnets, lasers, vacuum vessels, superconductors, neutron-resistant walls, and a tritium supply chain that does not yet exist at power-plant scale. The physics works. The unresolved question is whether the machine can be made cheaper than the electricity it sells.

The reaction

The workhorse reaction is deuterium-tritium fusion:

D + T → He-4 + neutron + 17.6 MeV

Deuterium is hydrogen with one neutron. Tritium is hydrogen with two neutrons and a 12.3-year half-life. When they fuse, most of the energy leaves as a 14.1 MeV neutron, which is useful for heating a blanket but brutal for materials.

The temperature target is roughly 100 million °C, several times hotter than the Sun's core. That number is less magical than it sounds: the plasma is thin, so the total heat content is manageable, but the charged particles must collide often enough, stay confined long enough, and be hot enough to beat the Coulomb barrier. Fusion is not one problem. It is density, temperature, and confinement time all refusing to be solved alone.

The main machines

Approach	Basic idea	Named benchmark	The catch
Tokamak	Magnetic donut traps plasma with external fields plus plasma current	ITER in France, aiming for Q = 10	Disruptions, cost, pulsed operation
Stellarator	Twisted magnetic geometry traps plasma without large plasma current	Wendelstein 7-X in Germany	Hard geometry, hard construction
Laser inertial fusion	Lasers crush a tiny fuel pellet for billionths of a second	NIF ignition, December 2022	Lasers and targets are far from power-plant economics
Pulsed / alternative concepts	Compress plasma in bursts, sometimes with direct energy conversion	Helion, TAE, First Light	Less proven physics at reactor scale

The National Ignition Facility reported fusion ignition in December 2022: more fusion energy came out of the target than laser energy entered it. That was a real physics milestone, not a grid prototype. The facility used far more electrical energy to power the lasers than the lasers delivered to the pellet, and a power plant would need to repeat the shot many times per second.

ITER is the opposite bet: not tiny pellets, but a 23,000-ton tokamak built to test burning plasma. Its first plasma schedule has slipped, with deuterium-tritium operation pushed into the 2030s. The machine may teach the world how reactor-scale plasma behaves, even if it never sells a kilowatt-hour.

What actually blocks the power plant

The hardest part is not making fusion happen. JET, NIF, and many smaller machines have done that. The hard part is making a device that survives its own success.

The 14.1 MeV neutrons from D-T fusion slam into the first wall and blanket. They displace atoms, create helium bubbles, weaken materials, and activate steel. A reactor that works for minutes but needs its wall replaced every few months is not an energy system.

Tritium is the second bottleneck. Civilian stockpiles are tiny, measured in kilograms, much of it sourced from heavy-water fission reactors. A fusion plant must breed tritium by letting neutrons hit lithium in a surrounding blanket. That full loop, burn tritium, breed tritium, extract it, purify it, and feed it back, has not been demonstrated in a commercial reactor.

What's contested

Fusion people disagree less about the plasma physics than about the economics. One side sees high-temperature superconductors, better simulation, and private capital as a real break from the old “30 years away” joke. The other side sees the same missing pieces as before: tritium breeding, neutron damage, remote maintenance, and capital cost.

There is also a timing problem. Solar, wind, batteries, and grid software are not standing still while fusion matures. A fusion plant that arrives in 2045 has to beat the energy system of 2045, not coal plants from 1985.

Why this has to do with other realms

Fusion is a physics page that becomes a space page the moment the reactor shrinks. Chemical rockets are energy-poor: Voyager 1 has spent almost half a century reaching only a tiny fraction of the distance to dest proxima centauri. A real fusion drive would not make stars nearby, but it could move interplanetary travel from months and years toward weeks and months. That is why tech fusion drive is not science fiction in the same way warp drive is; the fuel is known, the reaction is known, and the missing part is engineering violence.

Fusion also rhymes with concept nuclear fission in an uncomfortable way. Both promise dense energy. Both turn into materials science, regulation, supply chains, and public trust once the equations leave the blackboard.

An open question

If the first commercial fusion plant produces expensive electricity but proves tritium breeding and wall survival, does that count as success, or does fusion only matter if it beats solar-plus-storage on price?

Key Sources

J. D. Lawson, “Some Criteria for a Power Producing Thermonuclear Reactor” (1957) — the classic confinement condition behind the Lawson criterion.
ITER Organization technical materials — baseline reference for the tokamak path, Q targets, and schedule.
National Ignition Facility, Lawrence Livermore National Laboratory ignition reports (2022-2024) — primary source for laser fusion milestones.
EUROfusion / JET deuterium-tritium campaign reports (2021-2024) — record-setting magnetic fusion experiments before JET shutdown.
National Academies, Bringing Fusion to the U.S. Grid (2021) — sober view of what a pilot plant must prove.

Abhishek