Why is nuclear fusion so hard?
Because you must force atomic nuclei — all positively charged and therefore violently repelling each other — close enough to fuse, which requires conditions comparable to the inside of a star, and then hold those conditions long enough to get more energy out than you put in.
The fundamental obstacle. Nuclei repel electrostatically, with the force rising steeply as they approach. Overcoming it requires enormous kinetic energy, meaning temperatures of 100 million degrees or more — considerably hotter than the Sun's core, because we cannot reproduce the Sun's crushing gravitational pressure and must compensate with temperature.
At those temperatures matter is a plasma, and the problem becomes containing something hotter than any material can touch.
The two approaches:
Magnetic confinement, holding the plasma in a magnetic field — the tokamak and stellarator. The difficulty is that plasma is unstable and finds ways to escape: turbulence, instabilities and disruptions have absorbed decades of research.
Inertial confinement, compressing a tiny fuel pellet with lasers so rapidly that it fuses before it can fly apart. This is where ignition — more fusion energy out than laser energy delivered to the target — was first demonstrated in 2022, a genuine milestone that is frequently misreported as net energy gain for the whole facility, which it was not.
The three conditions that must hold together — temperature, density, and confinement time — are captured by the Lawson criterion, and the difficulty is that improving one typically degrades another.
The engineering problems beyond the physics: materials that survive intense neutron bombardment for years; breeding tritium, since it does not occur naturally in useful quantities and a reactor must make its own; extracting heat; and superconducting magnets at scale.
Why it remains worth pursuing: abundant fuel, no long-lived high-level waste of the fission kind, no risk of runaway reaction, and no carbon emissions.
On the timeline joke. Fusion has been described as thirty years away for decades. Progress is real and the remaining problems are engineering as much as physics.