Artificial intelligence is ready to take over the operation of spacecraft, according to Martin Halliwell, a partner at NewSpace Capital and former chief technology officer of satellite operator SES. The open question, he says, is whether the space industry itself is prepared to let it.
Halliwell, who led global technology and research and development at SES from 2011 to 2019, is known for advancing flexible satellite architectures and for early work on reusable rocket launches with SpaceX. He holds an MBA and a BA in mechanical engineering and mathematics from the Open University. In his current role at NewSpace Capital, one of the first private equity firms devoted exclusively to growth-stage companies in the space technology sector, he is closely watching how new software and automation tools are adopted across the industry.
The idea of AI-driven spacecraft has moved from research concept to operational possibility. Satellite operators already use automation for routine tasks such as orbit keeping, power management, and constellation coordination. Advances in machine learning and onboard computing have made it possible for spacecraft to assess their own health, detect anomalies, and adjust their behavior without waiting for instructions from ground teams.
The potential benefits are significant. Autonomous spacecraft can respond to hazards more quickly, reduce the need for round-the-clock human monitoring, and enable missions in environments where communication delays make real-time control impractical. For large constellations with hundreds or thousands of satellites, an AI-based approach could make operations far more efficient than manual management.
But readiness involves more than technology. It requires trust among operators, regulators, insurers, and customers. Satellite operators must be confident that an AI system can make safe decisions in unexpected situations. Regulators will need to decide how to certify and oversee autonomous spacecraft. Insurers and financiers must understand the risks, and the workforce will have to adapt, with ground controllers becoming supervisors of AI systems rather than manual operators of individual satellites.
Halliwell's commentary frames these challenges as an opportunity. As an investor, he is focused on companies building the technology that could make autonomous operations mainstream. As a former satellite operator executive, he spent years on the other side, managing the technical and commercial demands of operating fleets in orbit.
The timing may be particularly important. The rapid growth of satellite constellations has created new pressure on operators to cut costs and improve efficiency. At the same time, the space sector has become more attractive to private capital, with investors looking for software-defined solutions that can scale. AI is well positioned to meet that demand, if the industry is willing to accept it.
The debate is no longer about whether AI can run a spacecraft. It is about whether the space industry will adapt its procedures, regulations, and culture quickly enough to make it possible. Halliwell's career has placed him on both sides of that transition. At SES, he helped shape the company's global technology strategy; with SpaceX, he worked on reusable rockets, one of the most disruptive developments in modern spaceflight. Through NewSpace Capital, he now backs startups that could define how spacecraft are operated in the future.
As constellations grow and missions travel farther from Earth, the question raised in the commentary is likely to become more urgent. Spacecraft that can think for themselves may soon become the norm. The industry's challenge is to decide how quickly to move and what safeguards to put in place.
