Our Investment in Overland AI
By Chris Morales and Ryan Solís
We’re proud to announce our investment in Overland AI to support their progress in building intelligent uncrewed ground systems for defense and national security.
Autonomy is hard, but ground autonomy can be especially difficult. Terrain introduces perception and navigation challenges not found in air or sea. Vegetation, ground conditions and weather can complicate the environment in which these systems operate. Additionally, autonomous ground systems must deal with a greater number of moving objects in close proximity compared to other domains.
For years, we’ve been hearing about a future filled with self-driving cars, but so far, developers of automotive autonomy have piloted a limited set of experiments in a few cities such as San Francisco or Phoenix. Their vehicles operate under strict conditions, typically in benign climates and rely on remote operations when their systems encounter edge cases.
We believe existing ground autonomy solutions such as these will not work for the Department of Defense (DoD). War is anything but predictable. Operations take place in the harshest conditions – on mountain ranges, dense forests, and expansive deserts, in snow, sleet, dust storms or driving rain, where the high-fidelity maps commercial autonomy providers rely on aren’t available. Roads and bridges are damaged or destroyed. There might not be roads to begin with. Radio signals and communications are jammed – making reliance on GPS and teleoperation not feasible.
This lack of flexibility and reliability is why we believe leaders at the DoD have expressed frustrations in ground autonomy projects to date. General James Rainey, head of Army Futures Command, has complained about the state of existing efforts, “Every time it’s like, go watch one [robotic tactical vehicle] follow another one around a parking lot and it runs over the curb and I’m like ‘Come on, we got to do better than this’.”
We believe that status quo vendors will continue to struggle to deliver operationally relevant ground autonomy. Historically, defense ground autonomy efforts have often focused on experimental leader-follower convoys and logistics use cases reminiscent of the conflicts in Iraq and Afghanistan. We don’t believe these fully reflect the dynamic battlefield of the future: the DoD needs tactically relevant autonomy built for use cases like off-road maneuver, mobile anti-ship platforms, long range artillery, or air defense systems.
Overland AI was built to help tackle these challenges. The Overland AI team first came together through the DARPA Robotics Autonomy in Complex Environments with Resilience (RACER) program, which was created to address gaps in existing ground autonomy efforts. Byron Boots, Overland AI’s co-founder and a robotics expert, recruited a team from leading tech companies to focus on this problem. They began with the hard part first: navigating completely off-road, with no maps or markers, in snow, sleet and rain.
In addition to working with DARPA, the Overland AI team is engaged on autonomy projects with the U.S. Army and Marine Corps. They were recently announced, along with Anduril and Palantir, as new participants in DIU’s GVAP project.
No human should be put at risk when a machine can do the job. We look forward to supporting Byron, Steph, Greg and the rest of the Overland AI team as they work to bring these critical capabilities online – helping keep servicemembers safe and out of harm’s way.