Summary
Andrew Reiter explained how Shield AI’s Hivemind autonomy stack pilots defense aircraft without GPS or persistent operator links in contested electronic warfare environments. He traced Hivemind from Nova indoor drones to VBAT, Ukrainian one-way attack drones, and the X-BAT fighter aircraft, emphasizing runtime assurance, rapid field iteration, and platform-agnostic AI sovereignty.
Key Topics
5 key topics from Andrew Reiter at HumanX Amsterdam 2026. Thicker branches were mentioned more often in the talk.
Making Military-Grade Robots Safe & Secure
Click a topic to open it
Key Takeaways
- Design autonomy for GPS-denied and communications-jammed environments.
- Separate machine-speed navigation from human decisions to employ force.
- Use runtime assurance to check autonomous commands before execution.
- Accelerate field-data loops to match adversary adaptation cycles.
- Build platform-agnostic autonomy so partners retain mission data and models.
“If your system's operational logic requires a Live Link or requires a GPS lock, it's simply no longer a warfighting system”
A
Andrew Reiter
Speaker
Share
Up Next
BRS
WatchNext in agenda
Physical AI Meets Real-World Autonomy
Bobby Healy, Radha Basu, Shawn Kerrigan, Éanna Kelly
physical aiautonomous systems
O
WatchAlso on defense tech
Opening remarks
OOliviana Bailey O'Brien
opening remarksai infrastructure
ISJ
WatchRelated
The Future of Software Architecture in the AI-Native Era
Ivan Burazin, Spencer Kimball, Jim Harris
ai nativesoftware architecture