NASA to spend $1.3 million on physical AI to search for water on the Moon

NASA has awarded a $1.3 million contract to startup AstraQua, specializing in physical artificial intelligence algorithms, to build advanced autonomous navigation systems. The technology will allow future lunar rovers to land and maneuver near permanently shadowed lunar craters without GPS or pre-existing maps. This initiative is part of NASA's broader effort to integrate commercial robotics into Artemis missions.

• Contract value totals $1.3 million
• AstraQua startup focuses on physical AI algorithms
• Primary objective is finding water ice in shadowed craters
• System operates without GPS and maps in real time

Navigating extreme lighting conditions on the Moon

The primary challenge of landing near lunar craters is that their floors never receive sunlight, rendering standard optical cameras useless. Furthermore, radio commands from Earth suffer from noticeable transmission delays. AstraQua plans to overcome these obstacles using a combination of infrared and thermal sensors alongside tactile surface sensing during descent.

The role of physical AI in autonomous rovers

Physical AI technology will enable rovers to instantly adjust their descent trajectories upon detecting rocks or loose soil without waiting for operator instructions. By processing landscape data on the spot, the AI makes split-second navigation decisions suited for navigating rugged and treacherous extraterrestrial terrain.

Significance for the Artemis program and lunar exploration

Successfully operating within the eternal shadows of cold-trap craters is considered the key to discovering water ice reserves, which are vital for long-term human presence on the Moon. The current contract represents another step in NASA's strategy to harness commercial robotics innovations, with initial simulations slated to test the AI in lunar-analogue environments over the coming year.