Meet the Reactors Set to Upend Nuclear Fusion

Nuclear fusion has long been treated as the “holy grail” of clean, safe, near-limitless energy—but the central obstacle remains control. Hot plasma must be confined long enough for fusion to occur, and doing so reliably has proven brutally difficult. Recent momentum is shifting toward stellarators, a fusion approach that can run with greater stability than the tokamak method, and multiple companies have now published detailed plans showing how quickly that technology is converging toward practical power plants.

Tokamaks, the dominant and most studied fusion design, use a mix of magnetic and electric fields to keep plasma suspended. They often rely on a plasma current to help with confinement. That strategy has produced impressive control records without yet delivering sustained fusion energy: China’s East Tok achieved plasma confinement for over 1,000 seconds in January, while France’s West Tok later set a new record by maintaining plasma for over 1,300 seconds. But these milestones are still test runs—critical for learning control, not proof of net fusion power.

Fusion has been demonstrated in tokamaks, notably with the Joint European Torus (JET) in the UK, which produced a record-breaking 69 megajoules of energy in October 2023. Yet JET came close to break-even only for the plasma itself; it did not account for the extra energy required to run the reactor. More importantly, JET shut down at the end of 2023 after hundreds of plasma disruptions—sudden losses of control that can slam plasma into reactor walls and potentially damage the machine. The recurring theme is a tradeoff between the precision required by Maxwell’s equations and the reality of Murphy’s law.

Stellarators aim to sidestep tokamak’s reliance on plasma current. Instead, they generate the confining magnetic fields entirely through coils shaped into complex three-dimensional geometries. That design choice can make the plasma more stable and allows continuous operation, but stellarators were historically hard to build: coil geometry demanded heavy supercomputer simulation, and the resulting magnet structures were so intricate they sometimes had to be physically bent by hand.

That bottleneck appears to be easing. Three private companies have recently published stellarator roadmaps almost in parallel. In January, US-based TAE Energy released peer-reviewed papers and design plans for a “planar coil stellarator,” using simpler coils plus algorithms to steer the magnetic field—moving complexity from hardware into software. By March, it reported prototype magnetic fields above three Tesla, sufficient for a fusion-relevant confinement regime. In April, Germany’s Proxima Fusion published design plans for a commercial stellarator power plant called Stellaris and open-sourced the code used to calculate plasma behavior, targeting a demonstration device by 2031. Around the same time, Type One Energy released detailed modeling for its pilot plant, Infinity 2, projecting 800 megawatts of fusion power and aiming for an operational plant by the mid-2030s.

The clustering of announcements matters because it suggests stellarators are rapidly becoming a serious contender—possibly even the frontrunner—while also highlighting a cultural shift in fusion development. Private firms are publishing design details and, in some cases, code, making replication easier for the broader community. In fusion, that transparency is a major change: designs once required security clearance; now they can be shared over Wi‑Fi.