SPARO: AI Drone Navigation Without GPS

A new Chinese venture, SPARO (硅羽科技), is developing advanced AI for autonomous drone navigation, even in environments devoid of GPS signals. The company recently announced a new funding round from prominent investors including Seas Capital (雲時資本) and Hidden Hill Capital (普洛斯隠山資本). This marks the fourth financing round for SPARO, which was established less than six months ago, bringing its total funding to hundreds of millions of yuan (billions of Japanese yen). The capital infusion will primarily fuel talent acquisition, commercialization efforts, and ongoing technological development.

Founded in Shenzhen in February 2026, SPARO is not focused on manufacturing drones themselves, but rather on creating the sophisticated AI platforms that enable true autonomous flight. The drone industry, while rapidly expanding into areas like aerial photography, infrastructure inspection, and surveying, still faces significant limitations. Current systems heavily rely on GPS, pre-programmed flight paths, or remote pilot control, rendering them less effective in subterranean, indoor, low-light, or smoke-filled conditions where positional accuracy degrades and adaptive task execution becomes challenging.

SPARO aims to overcome these hurdles by integrating core functionalities—environmental perception, decision-making, path planning, and task execution—directly onto the drone. This approach seeks to create what the company terms "flying AI agents" capable of independent judgment and task completion. The company highlights its full-stack proprietary technology, encompassing sensor-based perception, a "cerebellum" for control, a "cerebrum" for decision-making, and even multi-drone coordination.

By fusing data from multiple sensors like cameras, LiDAR, and inertial measurement units, SPARO claims centimeter-level positioning accuracy is achievable even without GPS or in adverse conditions. Their autonomous navigation model goes beyond traditional SLAM (Simultaneous Localization and Mapping) by understanding spatial structures and environmental changes, allowing for dynamic flight path planning. Furthermore, the system boasts obstacle avoidance response times under 5 milliseconds, enabling drones flying at speeds up to 20 meters per second to react instantly to moving obstacles.

The company's vision extends beyond mere observation; SPARO is developing capabilities for drones to perform intricate tasks in the air, moving them from passive data collectors to active agents. Their technology also supports coordinated operations involving up to a thousand drones, promising enhanced automation for logistics, industrial inspections, and emergency response scenarios. The founding team's pedigree is a significant draw for investors. Zhang Fu, the founder, is a tenured associate professor at the University of Hong Kong and recognized as a top 1% global researcher by Clarivate. He previously served as a technical advisor to DJI for eight years.

Despite its short operational history, SPARO has already secured dozens of early customers, including drone manufacturers and logistics firms. The company plans to develop both complete drone systems for real-world validation and modular AI components for integration into various aerial platforms, including multi-rotors, fixed-wing aircraft, and eVTOLs. Zhang emphasizes a strategy focused on creating a "new market" by providing the essential "brains" for drones, rather than directly competing with established platform providers like DJI.