The mission reflects a growing interest in making space exploration less dependent on sending data back to Earth for analysis. By processing information directly on the Moon, future missions could make decisions faster and reduce the amount of data that needs to be transmitted to mission control.
A Computer Built for Lunar Conditions
Operating a computer on the Moon presents challenges that do not exist in a typical data centre. Equipment must cope with radiation, extreme temperature changes, limited power and the difficulty of carrying out repairs.
HPE says Spaceborne Computer-4 is designed to operate without cooling fans and withstand conditions near the lunar South Pole. Its hardware will include a low-power central processing unit (CPU), a graphics processing unit (GPU) and solid-state storage supplied by KIOXIA.
The system is expected to run large language models (LLMs) that can analyse visual data and generate useful insights. It will also process LiDAR measurements to support mapping, navigation and terrain analysis.
For a lunar rover, these capabilities could help identify obstacles, assess the surrounding landscape and highlight areas that deserve further investigation. Rather than transmitting every image or sensor reading to Earth, a computer could process some of the information locally and send back the most relevant results.
NASA Will Test How AI Models Work Together
NASA will contribute its Federated Learning for Unbounded and Intelligent Decentralization (FLUID) experiment. The test will examine coordination between three AI models trained on data from Earth, the Moon and the International Space Station.
Federated learning allows AI systems to learn collaboratively without necessarily gathering all their training data in one central location. Testing this approach across different environments could help researchers understand how distributed models exchange information when communication and computing resources are limited.
NASA will also test a large language model designed to search and interpret documents stored onboard. Such a system could make technical instructions, operating procedures and other mission documents easier to find when astronauts or mission operators need them.
Cybersecurity Takes a Place in the Mission
Deloitte will bring another capability to the experiment through Silent Shield, a cybersecurity system designed to detect potential intrusions using predictive analytics and anomaly detection.
As space missions become increasingly reliant on digital systems, protecting onboard computers and communications networks becomes more important. Detecting unusual activity early could help mission teams identify potential security problems before they affect operations.
The experiment will assess how this type of security technology can be used in a lunar computing environment. Its actual effectiveness will depend on the results of testing.
Linking the Moon With Earth-Based Computing
HPE will also support Astrolab’s mission control centre in Hawthorne, California, using HPE Private Cloud and HPE CloudOps Software to manage and automate computing workloads.
The setup combines local processing on the Moon with computing resources on Earth. This could help mission teams handle selected tasks closer to the rover while using terrestrial infrastructure for workloads that require additional computing capacity.
Scott Wiest, HPE fellow, global sales chief technologist and vice president, said the next stage of lunar exploration would depend on turning data into insights at the point of discovery. He said Spaceborne Computer-4 aims to bring AI and high-performance computing to the lunar edge.
The mission is expected to provide practical evidence of how these technologies perform beyond Earth. If the experiments succeed, they could help shape future systems for robotic exploration, scientific research and longer lunar missions. For now, the focus is on testing whether advanced computing can work reliably in an environment where maintenance, power and communication cannot be taken for granted.
