A desk-sized fusion reactor shrinks the laser, opening a tritium market to fund future plants

Research on nuclear fusion in the public mind is usually associated with massive facilities that span large complexes and are packed with rows of huge lasers or gigantic magnets, but a stealth startup near the Massachusetts Institute of Technology has decided to shrink that architecture to fit on a desk. In a small laboratory in Cambridge, USA, Yun Kai, co-founder and CEO of Cambridge Heavy Industries, unveiled a prototype of a miniature fusion unit that relies on integrated optics, arguing that the small size simplifies all the usual engineering complexities compared with the huge central projects.

The desk-size device allowed the team to carry out between fifty and eighty launch experiments daily.When we examined the experimental architecture of the project, we noted that this accelerated pace feeds its data to digital simulators that use artificial intelligence to improve the engineering design and tune the interaction physics, surpassing the slowness of large facilities that require long re-configuration periods between each experiment. This flexible approach attracted an early-stage financing round of $6.2 million led by Azula Ventures, with participation from Overlay, the Neta Foundation, B3A, the Massachusetts Clean Energy Council, and the Rothrock family.

The reactor uses a laser to compress a flat, circular fuel target, but it reduces the number of required pulses by employing only two main lasers to drive ignition. The first laser is split into multiple radiation lines that converge in concentric rings and strike the target with precise timing that starts at the edges and moves inward, generating a shock wave that propagates inward to enhance fuel compression. When the shock wave reaches the centre, the laser’s pulse hits the target to generate heat and produce a magnetic field, achieving fusion conditions with enormous pressure and relatively low laser power, according to a technique that originated in Keith Nelson’s chemistry laboratory at the Massachusetts Institute of Technology for studying materials under extreme conditions.

The sale of scarce tritium fuel precedes the construction of commercial power plants by a full decade.The company plans to develop a power plant within ten years that would occupy an area roughly the size of two basketball courts, yet it is deploying its miniature devices in the near term to produce the tritium isotope, a rare form of hydrogen needed by virtually all fusion plants under development. Tritium sells for about $30,000 per gram, and the current global supply does not exceed forty kilograms; a recent study projects a shortfall of roughly twenty-eight kilograms by 2054, valued at $840 million, which would provide the company with ongoing revenue while it builds its plant.

This shift forces a recalibration of infrastructure officials’ options and energy investments in the Gulf and Egypt, where planners are racing to secure sustainable, clean baseline power for the growing AI data-center clusters. The practical lesson for those monitoring the technical supply chains here lies in not tying financing solely to distant, massive fusion plants, but in exploring standardized test units that combine smart simulation with miniature hardware, and in leveraging early bottlenecks such as the supply of radioactive isotopes that could create independent markets before fusion power networks are born.