SE3 Labs builds SpatialGPT, a system that converts unstructured visual data into queryable insights by combining 3D computer vision with large language models. The core idea is to bridge the gap between raw sensor input - particularly from event cameras and conventional imaging pipelines - and structured, searchable spatial information that can be consumed by downstream AI or human operators. The technology targets the challenge of making physical-world data as accessible and queryable as digital text.
The company's Spatial AI platform is deployed across defense, smart infrastructure, construction, and smart factory verticals. Use cases include real-time situational awareness, process-level insights from 3D scene understanding, and autonomous systems capable of navigating and reasoning about three-dimensional space. The platform is designed to handle the constraints of real-world environments where unstructured visual input is the norm rather than the exception.
SE3 Labs was founded by researchers and engineers with backgrounds at Tesla Autopilot, NVIDIA Research, Bosch, and BMW Autonomous Driving. The team's work spans 3D computer vision, large language models, event cameras, robotics, and spatial AI. The company operates in the European Union, where it positions its technology against the challenge that roughly 45% of work in the EU remains non-digital, suggesting a large addressable surface for physical-world AI systems.
