Shield AI is an American robotics and defense technology company based in San Diego, California. It addresses a specific operational challenge: enabling drones to perceive their surroundings, move around, and complete a mission without constantly relying on a remote operator or a data link. Its business combines the development of autonomy software, Hivemind, with that of aerial platforms. It primarily serves the armed forces, with applications in reconnaissance, intelligence, and surveillance.
A company born from needs in the field
Shield AI was founded in 2015 by brothers Brandon and Ryan Tseng, along with Andrew Reiter. A former Navy SEAL, Brandon Tseng brings knowledge of military operations, while his co-founders contribute to the technological side of the project. The initial ambition notably involves sending robots to explore dangerous places before deploying people there.
This focus takes shape with Nova, a small drone designed to operate inside buildings without GPS. The company then expands its scope to aircraft capable of covering larger areas. In 2021, the acquisition of Martin UAV brings it V-BAT, a vertical takeoff and landing drone. This acquisition strengthens its position at the intersection of software, aircraft integration, and robotic systems manufacturing.
Hivemind and V-BAT, two complementary components
Hivemind forms Shield AI’s software foundation. This autonomy system uses sensor information to perform perception, navigation, and planning functions. It is designed to operate onboard aircraft, including when communications are degraded. The aim is to reduce reliance on remote piloting and enable missions to be carried out in contested environments.
Shield AI seeks to integrate this technology into various platforms beyond its own drones. This strategy involves adapting it to the sensors, flight controls, and constraints of each aircraft. Simulation and real-world testing therefore play an important role in developing and validating autonomous behaviors.
V-BAT represents the hardware component of this offering. Its architecture allows it to take off and land vertically, then fly like an airplane. It can therefore be deployed from sites without runways, particularly for surveillance missions. The combination of an aircraft and autonomy software distinguishes Shield AI from a software-only provider.
What next?
Shield AI’s development will depend on its ability to turn technical demonstrations into reliable systems that can be produced and maintained at scale. Integration across multiple aircraft types opens up commercial opportunities but also increases validation requirements. In defense, autonomy must come with assurances regarding cybersecurity, control over system behavior, and human oversight. These issues, along with export regulations, will influence the international deployment of its technologies.