Niantic Spatial

Niantic Spatial builds foundational models and mapping technology intended to let AI operate in the physical world. The company develops spatial intelligence systems that allow humans and machines to understand, represent, navigate and engage with their surroundings. Its stated focus is the roughly 80% of economic activity that occurs beyond screens, with applications in robotics, the public sector, energy and industrial settings.

Technical work spans computer vision, 3D mapping, geospatial AI, digital twins, Gaussian splats, LiDAR processing, mesh generation, neural networks and sensor fusion. The company generates geometrically accurate digital twins from smartphones, 360-degree cameras, drones, satellites and LiDAR, producing meshes and Gaussian splats as output.

Underlying these systems is a proprietary database of more than 30 billion posed images. Its Visual Positioning System turns aerial and ground capture devices into positioning and orientation sensors, delivering localisation in places where GPS fails; the company states it is powered by 50 million neural networks and 150 trillion parameters.