General Fusion heats plasma past 1 keV by squeezing it with metal

What happened

General Fusion announced that LM26 had passed 1 keV of electron temperature. The company had set this as the first major goal for the machine when it launched. LM26 was assembled in 2024 at the company's headquarters in Richmond, British Columbia. It made its first plasma, and its first plasma compression, in 2025. The company calls LM26 its largest demonstration machine so far.

Two instruments measured the result. A Thomson scattering system, built with the UK Atomic Energy Authority (UKAEA), measured about 1.1 keV just before peak compression. A second system, called absolute extreme ultraviolet (AXUV), gave about 1.2 keV at peak compression. According to the Nuclear Newswire report, the two readings followed the same trend. The company also reports ion heating of about 0.46 keV, based on neutron counts, density profiles and a reconstruction of the data over time.

Other private fusion companies have passed 1 keV before. Nuclear Newswire lists Tokamak Energy in 2023, Helion in 2023 and Zap Energy in 2024, each with a different machine design. The new point is that this is the first such result for General Fusion's method, which uses slow compression. Two papers on the result have been sent for peer review. They had not been published when the news came out.

The 1 keV mark has a long history. Nuclear Newswire notes that it was first reported on the Soviet T3 tokamak, announced in 1968. In 1969, physicists from the United Kingdom confirmed that result using Thomson scattering, the same kind of measurement used on LM26. The threshold has since become a common early test for any new fusion design. Reaching it does not mean a machine is close to producing power, but it shows the heating method works as planned.

The engineering behind it

LM26 uses an approach called magnetized target fusion. First, a plasma is formed with a magnetic field inside it. Then a solid lithium liner around the plasma is made to collapse inward by an electromagnetic theta-pinch, a strong pulse of current that creates a magnetic force on the liner. The squeeze takes milliseconds. Laser fusion compresses fuel in nanoseconds, which is about a million times faster. The company argues that slow compression needs less power and simpler equipment.

As a general rule in physics, when you compress a gas or plasma quickly enough that heat cannot escape, its temperature rises. The company says temperature, density and magnetic field all rose together during compression, as planned, and that the plasma stayed stable. Stability matters because a plasma that breaks apart during the squeeze loses its heat to the walls before the compression is finished.

Measuring the temperature was a hard problem in its own right. Thomson scattering works by firing a laser pulse through the plasma and studying the light that scatters off the electrons. The spread of that light's wavelength shows how fast the electrons move, and so how hot they are. LM26 has very little open space for instruments, so UKAEA designed a polychromator, the device that splits the scattered light into wavelength bands. Nuclear Newswire notes that UKAEA called this access problem a significant technical challenge.

The machine also depends on pulsed power. In general, a pinch like this is driven by large banks of capacitors that are charged slowly and then released in a very short burst through heavy switches. Timing must be exact, because the laser for the Thomson measurement has to fire at the right moment in a compression that lasts only milliseconds. That is a task for fast electronics, triggering circuits and high-speed data capture.

What it means in Nepal

This experiment has no direct link to Nepal's power system, and fusion power stations are not near. The company itself aims for net fusion energy, meaning more energy out than in, by the end of 2028, and a first commercial plant around 2035. These are company targets, not results. The result that can be checked today is the temperature, measured by two instruments with a national laboratory as partner. The planned commercial design would also add a liquid lithium wall to capture heat and breed fuel, and LM26 does not test that part.

The useful lesson for a student is about the type of work. Most of the engineering in a fusion experiment is not nuclear physics. It is high-voltage design, fast switching, sensor design, optics, signal processing and data handling. The same skills are needed in many other places, such as substation protection, medical imaging, radar, and industrial test labs. A student who can design a clean measurement chain, from a sensor to a digitiser to software, can work across all of them.

There is also a lesson in reading science news. A company announcement and an independent report are different things. In this case the measurement was made with UKAEA, a public research body, and the papers are still under review. A careful engineer reports the measured number, says who measured it, and treats future timelines with caution until outside experts confirm the work.