JPL's 120-kilowatt lithium thruster marks step toward megawatt electric propulsion

On a February afternoon in 2026, inside a water-cooled vacuum chamber eight metres long at NASA's Jet Propulsion Laboratory, a tungsten electrode began to glow. The metal reached 2,800 degrees Celsius. Around it, a red plume spread from the mouth of an experimental engine as electrical current surged through lithium plasma. Five times that day, engineers fired the machine. At its peak, the prototype drew 120 kilowatts of electrical power—a number that matters less than what it represents and far more than what it does not yet prove.

To understand the scale, consider what is already flying. NASA's Psyche spacecraft, currently cruising toward a metal-rich asteroid, carries Hall-effect thrusters that draw about 4.5 kilowatts. The JPL prototype operated at more than 25 times that power. The comparison is immediate and striking. It is also incomplete. Power and thrust are not the same thing. A kilowatt measures how quickly electrical energy enters a system. Thrust measures force. Two engines can draw identical power while producing vastly different combinations of force and exhaust velocity, depending on their efficiency, design, and propellant. JPL did not publish the prototype's measured thrust, efficiency, or specific impulse. Those numbers will be essential before anyone can translate the power level into a useful spacecraft design. Without them, the 120-kilowatt figure is a starting point, not a destination.

The engine itself is a magnetoplasmadynamic thruster, or MPD—a technology studied since the 1960s but never flown operationally. Unlike Psyche's Hall thrusters, which ionize xenon gas and accelerate it through electric and magnetic fields, the MPD turns lithium metal into vapor and then plasma. Very high electric currents pass through that plasma. Their interaction with a magnetic field creates an electromagnetic force that accelerates the plasma out of the engine. The appeal is straightforward: at high power, an MPD might process far more electricity through a single thruster than today's flight systems while still using propellant efficiently. A vehicle could obtain substantially more electric thrust without assembling an impractically large bank of small engines. That is potential. It is not yet performance.

The test took place in JPL's condensable metal propellant facility, known as CoMeT. The chamber itself is part of the result. High-power electric propulsion cannot be tested in an ordinary room. The vacuum must be sufficiently thin, the plume must be contained, diagnostic equipment must be protected, and enormous amounts of heat must be removed without distorting the measurements. The facility was designed to eventually support tests reaching the megawatt class. Five ignitions brought the prototype to 120 kilowatts. James Polk, a senior research scientist at JPL, said the firing showed both that the thruster worked and that the testbed could support the scaling work ahead. Five starts are not 23,000 hours of operation. The test did not show that the engine could run through years of cumulative use, survive thousands of thermal cycles, or restart after a long coast in deep space. It established the first point on an engineering curve whose difficult end is endurance.

Moving from 120 kilowatts to one megawatt means putting 8.3 times as much input power through a single thruster. In a simple drawing, the engine only becomes larger. In hardware, every current path, magnetic field, propellant channel, seal, electrode, insulator, and cooling boundary must cope with a harsher operating environment. The scale-up may not be linear. A component that remains cool enough during a brief 120-kilowatt pulse can overheat during a longer firing. Plasma may distribute itself differently at higher current. Erosion that looks trivial across five starts can remove unacceptable amounts of material over tens of thousands of hours. The central tungsten electrode exceeded 5,000 degrees Fahrenheit. Tungsten is used because ordinary metals would fail far sooner, but even refractory materials erode, crack, or change properties under prolonged heat and plasma exposure. In space, where a vacuum removes convective cooling, heat must move through the structure and eventually leave as infrared radiation from large surfaces. At megawatt scale, radiators, pumps, and plumbing can become a substantial part of the vehicle.

A one-megawatt thruster is inseparable from the machine that produces and manages that electricity. JPL placed the MPD work within NASA's Space Nuclear Propulsion project because crew-scale electric propulsion will probably require a compact, solar-independent source. A fission reactor produces heat. A conversion system turns some of that heat into electricity. The electricity then powers the thruster, while radiators dispose of the heat that was not converted into useful electrical output. JPL estimates that a human Mars vehicle might need two to four megawatts in total, using several MPD thrusters. Even if one-megawatt engines become available, the spacecraft would still require multiple units for total power, control, and redundancy. It would also carry a reactor, shielding, turbines or another conversion system, radiators, lithium stores, and a structure capable of keeping the crew away from the reactor's radiation. NASA's planned Space Reactor-1 Freedom pathfinder is intended to test a 20-kilowatt-electric reactor system with a 12-kilowatt Hall thruster. A two-megawatt Mars vehicle would demand one hundred times that reactor output.

Electric propulsion can use far less propellant than a chemical stage for the same broad mission task. The trade is low thrust, which means the vehicle accelerates over a long arc rather than through a brief burn. Raising electrical power offers a route to more thrust while retaining much of electric propulsion's propellant advantage. That combination is attractive for Mars because crewed vehicles and their supplies are heavy. Reducing propellant mass can free launch capacity for habitats, shielding, consumables, and abort margins. Greater thrust can also make an efficient trajectory less slow than it would be with today's kilowatt-class engines. None of this supplies a transit duration from the February firing. JPL said the technology could support human missions, not that the prototype had completed a simulated voyage or proven a particular schedule. Travel time will emerge only when engineers can place measured thruster performance inside a complete vehicle model with realistic reactor, radiator, and payload masses.

The next tests can answer questions the 120-kilowatt headline cannot. How much thrust did the engine produce? What was its efficiency? How steadily did it feed lithium? Where did material erode or condense? Did repeated starts change performance? Longer firings at the existing level would establish a baseline before power rises. Tests at 500 kilowatts and one megawatt would then show whether performance scales as hoped. Component inspections could reveal which parts set the lifetime limit, while increasingly flight-like electronics would expose integration problems early. After that would come a full subsystem demonstration and, eventually, a flight experiment. The thruster would need to operate with its actual power processor, propellant feed, and thermal hardware in a package light enough to launch. Qualification would have to address vibration, radiation, vacuum, cold starts, fault recovery, and the impossibility of routine maintenance. The February test deserves attention precisely because it made one part of this chain physical. A lithium-fed MPD thruster was built, fired, and taken to 120 kilowatts in a facility prepared for higher power. The remaining chain is still long.