Nuclear fusion versus fission: why the clean giant is always thirty years away

Every few years a fusion laboratory announces a breakthrough, the coverage declares limitless clean power is coming, and a decade later the grid looks exactly as it did. The cynic concludes fusion is a con; the enthusiast concludes it is imminent. Both are misreading the same number, because "net energy gain" means something far narrower in a physics paper than it does in a headline.

Start with what separates the two technologies. Fission splits heavy nuclei, usually uranium-235, and it is almost embarrassingly easy to start: a slow neutron strikes the nucleus, it splits, releases energy and more neutrons, and those neutrons split further nuclei. The chain reaction sustains itself at room temperature; the entire engineering problem is slowing it down safely. That is why Calder Hall was exporting fission electricity to the British grid in 1956 and why Sizewell C is, for all its cost overruns, a known quantity. Fusion runs the other way: it forces light nuclei, deuterium and tritium, to merge. Because both nuclei are positively charged, they repel each other ferociously, and overcoming that repulsion means heating the fuel to above 100 million degrees Celsius, roughly ten times hotter than the Sun's core, while confining it with magnetic fields or laser compression. Nothing about it is self-starting. Switch off the heating or the confinement and the reaction dies within milliseconds, which is also why a fusion plant cannot melt down.

The prize is real. A kilogram of fusion fuel yields around four times the energy of a kilogram of fissioned uranium, deuterium is extractable from seawater, there is no chain reaction to lose control of, and the waste problem shrinks from spent fuel that stays dangerous for tens of thousands of years to activated reactor components manageable within a century. This is why governments have funded it since the 1950s despite the running joke that commercial fusion is thirty years away and always will be.

Now look closely at the milestone that dominated coverage in December 2022. The National Ignition Facility in California fired 2.05 megajoules of laser light at a peppercorn-sized fuel capsule and got 3.15 megajoules of fusion energy back. That is genuine scientific ignition: the reaction released more energy than the light that struck the target. But the lasers that produced those 2.05 megajoules drew roughly 300 megajoules from the wall, because they are about one per cent efficient. Measured at the plug rather than the plasma, the facility consumed around a hundred times what the reaction yielded, and none of the output was captured as electricity. The Joint European Torus at Culham, before its 2023 shutdown, set the magnetic-confinement record of 69 megajoules over five seconds, again while drawing far more power than it produced. ITER, the international tokamak under assembly in southern France, targets Q of 10, meaning 500 megawatts of fusion power from 50 megawatts of plasma heating, but even that figure excludes the cryogenics, magnets and plant systems, and ITER will never generate a single watt for the grid. It is an experiment, deliberately so.

From the plasma to the plug

Converting a physics gain into a commercial one demands machinery nobody has yet built. Tritium barely exists in nature, so a power plant must breed its own by wrapping the reactor in a lithium blanket that captures neutrons and yields fresh fuel, a process demonstrated only in small samples. The 14 MeV neutrons that carry most of the energy out of the plasma also smash into the reactor's first wall, swelling and embrittling any known steel; qualifying a material that survives years of that bombardment is arguably a harder problem than the plasma itself. Add superconducting magnets that must run for months without a quench, and exhaust systems handling heat fluxes comparable to a rocket nozzle, and the distance between an ignition experiment and a power station comes into focus.

Why the honest reading matters for Britain

The UK has placed a serious bet on closing that gap: STEP, the Spherical Tokamak for Energy Production, is intended to put a prototype plant on the former coal site at West Burton in Nottinghamshire around 2040, and the UK Atomic Energy Authority's campus at Culham anchors a cluster of private ventures, including Tokamak Energy and First Light Fusion. None of this makes fusion electricity imminent, and none of it is undermined by the engineering gap. The sober position is that fission decarbonises grids now, fusion is a materials and systems programme with decades of hard graft left, and each new record should be read with one question: was that gain measured at the plasma, or at the plug?