Lot of news here: AI-guided swarms, missile tracking and smart sensors

This week’s edition of SpaceNext: AI shifts from policy blueprints and platform-level autonomy to breakthroughs in AI-powered space hardware, defense systems, and next-gen satellite tracking. From neuromorphic chips to AI-guided missile interceptors, we explore how smart systems are moving off the whiteboards and into orbit.

Let’s get to it!

What if AI could identify satellites not by transponders, but by how they shine?

Slingshot Aerospace is doing just that. Under a new contract from the U.S. Air Force’s AFWERX program, the company is advancing its photometric fingerprinting capabilities — analyzing how satellites reflect sunlight over time to generate unique behavioral signatures.

Every night, its sensor network collects over 4.5 million light curve observations, which are processed by its Agatha AI. The goal? Detect unexpected maneuvers, reacquire lost objects, and spot deceptive spacecraft behavior in low Earth orbit — capabilities increasingly vital to U.S. Space Command.

This effort offers a non-traditional, stealth-resilient method to maintain orbital custody. It’s a step beyond cataloging — it’s cognitive classification in space.

GDMU3Gc65.tzd_{KVxWe1KJWidYik-kvjKTMUnLN6KKt_Ey.zedu1o2kjNM}k_Y3XHfI”>Read more: Go deeper with Sandra Erwin’s story in SpaceNews on Booz Allen’s Brilliant Swarms concept.

Cubesat propulsion is getting an AI makeover.

Miles Space has successfully demonstrated its Poseidon M1.5 plasma thruster, a water-vapor-based propulsion system engineered using a mix of genetic algorithms and neural networks. Fitting inside a 1U cubesat and running on just 1.5 watts, the thruster achieved 37.5 millinewtons of thrust at a specific impulse of 4,800 seconds.

What makes this stand out is the AI-driven design loop. CTO Wesley Faler fed simulation parameters into a neural net that learned his design preferences, refining the system even while he was offline. AI didn’t just assist; it co-created.

This is a story about AI as a hardware co-designer, optimizing propulsion performance for the next generation of ultra-efficient, mission-flexible spacecraft.

ESA’s AI digital twin takes root