Niantic Spatial builds foundational artificial intelligence models designed for real-world operation. The company's work sits at the intersection of spatial intelligence, computer vision, and 3D mapping, with technical expertise spanning neural networks and geometric modeling. Its technology serves industries including robotics, the public sector, and energy and industrial applications.
Central to the company's platform is a proprietary spatial database containing over 30 billion posed images. This data underpins its Visual Positioning System (VPS), which enables precise localisation and orientation even in environments where GPS is unavailable. The VPS is powered by 50 million neural networks comprising 150 trillion parameters, converting aerial and ground capture devices into accurate positioning sensors.
Niantic Spatial also produces geometrically accurate digital twins from a range of input sources - smartphones, 360 cameras, drones, satellites, and LiDAR - outputting meshes and Gaussian splats. The combination of large-scale spatial data, advanced positioning technology, and digital twin creation positions the company as a significant player in applied spatial AI.
