US Air Force Seeks AI Solutions for Autonomous Cislunar Spacecraft

The U.S. Department of the Air Force has issued a request for information (RFI) to the defense and technology industry to explore the use of advanced artificial intelligence and machine learning algorithms for the autonomous operation of spacecraft in cislunar space.

The move is designed to support the administration’s policy of developing a long-term architecture beyond Earth orbit. The Department of the Air Force noted that the vast volume of cislunar space compared to near-Earth orbits creates communication delays, making it difficult for traditional ground-based command and control systems to operate within operational timelines.

The document emphasizes that the government does not intend to fund a new research and development program, but is instead seeking flight-proven, commercial off-the-shelf software that can be integrated with minimal adaptation onto government-furnished spacecraft hardware under a separate effort. The defined timeline requires delivery of initial software and its verification in a simulation environment within 90 days of authorization to proceed, and the completion of qualified, flight-ready software within 18 months. Prior flight experience in cislunar space is not a prerequisite, provided that respondents demonstrate the applicability of software flown in other orbits to cislunar dynamics.

The solicitation focuses on four key capability areas: autonomous mission management and spacecraft systems operations during extended periods without ground contact; predictive maintenance and real-time detection of hardware anomalies and faults; autonomous optimization of maneuvers and trajectories in complex gravity regimes without ground station intervention; and multimodal computer vision systems for image processing and the identification of celestial bodies and man-made objects.

Alongside software capabilities, companies must detail specific computing requirements for running the models, including processor architecture, memory capacity, power consumption, in-orbit verification and updating mechanisms, and safeguards to prevent erroneous decision-making during sensor failures or adversarial inputs. The deadline for responses is Dec. 4, 2026, with the collected data potentially serving as the basis for formulating future procurement requirements and contracts in the domain.