Defense contractor Anduril Industries, Inc. and autonomy specialist Overland AI have successfully demonstrated a team of autonomous air and ground vehicles that can coordinate to neutralize threats, proving that the future of warfare involves less manual joystick-waggling and more algorithmic teamwork. The joint field test showed how connecting disparate systems with a shared AI brain can dramatically shorten response times in a chaotic battlespace.
The exercise was a textbook example of Manned-Unmanned Teaming (MUM-T), a concept the U.S. Army defines as the “synchronized employment of soldier, manned and unmanned air and ground vehicles, robotics, and sensors” to get an edge. The setup involved two of Overland’s ULTRA ground vehicles, running its OverDrive autonomy software, convoying with a crewed vehicle. Watching from above was an Anduril Ghost-X drone, all linked through Anduril’s Lattice software platform. When the Ghost-X spotted incoming enemy drones, the human operator simply tasked the two ULTRA vehicles to get a better look.
It’s the classic military maneuver: see a problem, then send the robots up the hill to get a better view. Except this time, the robots drove themselves. Overland AI’s software, honed in DARPA’s RACER program, enabled the 1,000-pound payload capacity ULTRAs to navigate the rough terrain and find elevated positions without human intervention. Once in place, Anduril sensors on the vehicles detected and tracked the drone activity, feeding the data across the network to allow operators to deploy electronic warfare countermeasures from a single, shared interface.
Why is this important?
The modern battlefield is a mess of incompatible systems that weren’t designed to talk to each other. This demonstration proves that a unified, AI-driven network can solve that integration nightmare. By having air and ground assets share sensor data and coordinate actions autonomously, you can significantly shrink the “sensor-to-shooter” timeline, extend situational awareness, and reduce the cognitive load on human operators. Every step—from spotting a threat with a drone to repositioning a ground vehicle and deploying a countermeasure—happened through one network without anyone needing to manually transfer data. It’s less about a single cool robot and more about the network that turns them into a cohesive, artificially intelligent pack.
