Nvidia's Jetson GPUs are heading to the moon

Nvidia’s chips have conquered terrestrial data centers, and now they are heading somewhere with a lot less atmosphere.

Space robotics startup Lunar Outpost announced that its next lunar rover will use an Nvidia Jetson module to power its lidar navigation system. If the mission succeeds, it will likely be the very first GPU to ever touch down on the surface of the moon.

The rover is scheduled to launch before the end of the year on a SpaceX Falcon 9 rocket, riding aboard an Intuitive Machines lander. Its mission is to venture into deep lunar craters and eventually explore Reiner Gamma, a mysterious lunar region famous for its bizarre magnetic anomaly.

Bringing edge AI to the craters

Most spacecraft traditionally rely on decades-old computing architectures because ancient chips are hardened against cosmic radiation. But as space exploration shifts toward autonomous robotic fleets, those sluggish processors can no longer keep up with modern sensor data.

Lunar Outpost plans to run Jetson alongside its traditional flight compute platform to compare how modern edge AI stacks up against legacy space hardware. The goal is to allow the rover to process 3D lidar scans locally, making navigation decisions in real time instead of waiting for commands from mission control.

Nvidia is not stopping at the surface. The chipmaker recently partnered with Firefly Aerospace to put Jetson chips aboard a lunar-orbiting satellite to process reconnaissance imagery and track robots roaming on the ground below.

The lunar environment is notoriously brutal. Unlike satellites in low Earth orbit that get partial shielding from Earth's magnetic field, hardware on the moon faces raw cosmic rays and punishing temperature swings. The biggest engineering hurdle will be keeping the GPU alive through the freezing two-week lunar night using minimal battery power.

Why it matters

We think this mission represents a critical turning point in how space hardware gets built.

For decades, space agencies treated commercial off-the-shelf silicon like a liability. But if autonomous machines are going to build permanent outposts ahead of NASA’s planned astronaut return in 2028, robots need real brains, not just rudimentary controllers.

Putting an off-the-shelf Jetson chip onto a lunar rover is a clever move. It proves that space tech is finally adopting the same physical AI and robotics stack that terrestrial warehouses and autonomous vehicles use, slashing development time and cost.

That said, the risks are real. Consumer-grade silicon has not been battle-tested against long-term lunar radiation and brutal thermal cycling. If the chip freezes or suffers memory corruption during the lunar night, it could leave the rover blind. For now, high-compute space exploration still faces two major bottlenecks: cosmic physics and waiting on rockets to get there.