For years, lidar has been synonymous with autonomous vehicles. Most conversations centered on helping cars navigate roads, detect obstacles, and improve safety. 

Those applications remain important, but they represent only a fraction of what is possible. As Physical AI expands into new industries, the same technology that helps a vehicle perceive a highway is being applied to defense platforms, rail corridors, airports, warehouses, and sports venues. 

Space mobility is the latest example, and it may be the most demanding one yet. 

AEye recently announced that its Apollo™ long-range lidar has been selected by Lunar Outpost for integration onto a crewed lunar terrain vehicle. It marks AEye’s entry into space mobility, and it raises an interesting question: what does perception have to look like when the environment is 384,000 kilometers away, there is no atmosphere, and no one can come fix the sensor if something goes wrong? 

Space Mobility Is Becoming a Physical AI Problem 

Lunar exploration is no longer only about getting to the Moon. It is increasingly about operating there, repeatedly and safely, over long periods of time. 

That shift is driving demand for surface mobility: vehicles that can carry crews, transport cargo, prospect for resources, and prepare sites for permanent infrastructure. The lunar rover market is among the fastest-growing segments of the space economy, driven by sustained, multi-year government investment and increasing commercial activity. 

Vehicles operating at that tempo cannot depend on a human operator watching every meter of terrain. They need to perceive, interpret, and respond to the environment around them in real time. This is the essence of Physical AI. 

Unlike traditional AI, which processes digital information, Physical AI enables machines to understand and interact with the physical world. Before a lunar vehicle can choose a route, avoid a crater rim, or safely deliver a crew to a work site, it must first perceive what is actually in front of it. That makes perception foundational. 

Why the Lunar South Pole Is Difficult to See 

The Moon’s South Pole is the focus of the next phase of human lunar exploration, and it presents lighting conditions unlike anything encountered in previous crewed missions. 

In the regions under consideration for landing, the Sun never climbs far above the horizon. Its highest possible elevation is roughly seven degrees. The result is permanent low-angle light: extremely long shadows, harsh glare, and deep shadowed areas immediately adjacent to brightly lit terrain. 

Human vision struggles in that regime. Crews have to see into near-total darkness while their eyes are adapted to brilliant sunlight, then back into that sunlight while still adapted to the dark. Camera-based systems face a related problem. They interpret a scene from brightness, contrast, and color, and that information becomes unreliable when a rock, a crater, or a slope sits inside a shadow that receives no scattered light at all. 

The Moon has no atmosphere to soften that transition. On Earth, air scatters sunlight and partially fills in shadows. On the Moon, a shadow can be effectively black, and the boundary between lit and unlit terrain can be abrupt. 

Lidar approaches the problem differently. Rather than inferring depth from appearance, it directly measures distance using laser pulses to generate a precise three-dimensional representation of the environment. It provides its own illumination, which means a shadowed boulder and a sunlit one are measured the same way. 

For a vehicle carrying astronauts across unmapped terrain, that distinction matters. 

Perception Where There Is No Second Chance 

Lunar hardware operates under constraints that have no real equivalent on Earth. 

There is vacuum, which eliminates convective cooling. There are thermal extremes across the day-night cycle and between sunlight and shadow. There is constant vibration during launch and traverse. There is abrasive regolith dust. And critically, there is no possibility of field repair once the vehicle is on the surface. 

Those conditions shaped why Apollo™ was selected. Its solid-state design offers the ruggedness required to withstand the lunar vacuum, thermal extremes, and constant vibration. Its compact form factor delivers that capability without competing for the mass and volume budget every lunar vehicle has to manage, where every kilogram carries a launch cost. 

“The lunar environment represents the most demanding proving ground any sensor can face: no atmosphere, no margin for error, and no opportunity for in-mission repair,” said Matt Fisch, AEye’s Chairman and CEO. 

Three Functions Perception Has to Support 

On a lunar vehicle, Apollo™ supports terrain perception, obstacle detection, and autonomous navigation. Each asks something slightly different of the sensor. 

Terrain perception is about understanding the surface itself: slope, roughness, and whether ground that looks traversable actually is. Loose regolith, small craters, and rock fields are hazards that depend as much on geometry as on appearance, and geometry is what a lidar point cloud measures directly. 

Obstacle detection is about what sits on top of that terrain, and about detecting it early enough to matter. The same principle that applies on a highway or a rail corridor applies here: greater range provides more time to react, whether the response comes from a driver, a remote operator, or the vehicle itself. 

Autonomous navigation ties the two together. A vehicle choosing its own path needs enough spatial understanding to identify where it can safely go, not just what is in front of it. On terrain that has never been driven and cannot be surveyed in advance at that resolution, that understanding has to be built in real time, on the move. 

Why Software-Defined Perception Matters 

Traditional sensing systems are often optimized for a single application. Software-defined lidar changes that equation. 

Instead of requiring new hardware for every environment, scan patterns and sensing performance can be configured through software and tuned to what a specific mission actually requires. A rover picking its way through a shadowed crater has different sensing needs than one traversing an open, sunlit plain, and those needs can shift over the life of a program. 

That adaptability is what allowed a sensing platform developed for terrestrial applications to be tuned for a lunar vehicle rather than redesigned from scratch. 

As Fisch put it, “Apollo’s™ software-defined architecture enabled us to tune its performance for exactly what this mission required, with the same adaptability and durability we have proven in defense, now validated in the harshest environment there is.” 

The Moon as a Proving Ground 

Space mobility may seem like an unexpected application for lidar, but it reflects the same broader trend visible in defense, rail, aviation, and sports analytics. 

As AI moves beyond software and into the physical world, machines across every one of those industries need the same underlying capability: the ability to perceive the world accurately, in real time, under conditions their designers cannot fully control. 

The lunar surface simply removes every safety net at once. Nothing about that environment is forgiving, which is exactly what makes it a meaningful test of what a perception system can do. 

At AEye, we believe long-range, software-defined lidar will play an increasingly important role in enabling the next generation of Physical AI, helping autonomous systems across transportation, infrastructure, robotics, defense, and now space better perceive, understand, and safely interact with the world around them. 

Learn more about our software-defined lidar here. 

This post contains forward-looking statements, including regarding expected product performance in future missions, the anticipated growth of the space mobility market, and the potential benefits of AEye’s selection by Lunar Outpost. Actual results may differ materially due to the risks described in AEye’s SEC filings, including the “Risk Factors” in its most recent SEC filings. AEye undertakes no obligation to update these statements except as required by law.