SpatialGen Labs ships real-time 3D reconstruction for the U.S. military and the live pipeline for immersive sports at broadcast scale. Both delivered in 2025.
The pipeline from capture to action is fractured. Sensors outpace compute; standards lag behind hardware. Today, the layers of the spatial stack are built in isolation by teams and industries that rarely speak the same language.
SpatialGen Labs builds the distribution layer. We develop software that forces state-of-the-art research to actually function in high-stakes environments. There is no time for hesitation. Only time for delivery.
Legacy ISR delivers 2D orthomosaics that strip away the vertical dimension—the very data that determines fields of fire and structural masking. We deliver photorealistic, navigable 3D environments in seconds, not hours. If the situation is developing, your map should be, too.
We are moving past the era of static mapping. Our mission is to provide an persistent, living 3D layer over any Area of Operations—turning raw video feeds into high-fidelity spatial intelligence that can be measured, navigated, and acted upon in real-time.
Operationalizing 3D Gaussian Splatting for the field. We will accelerate the targeting cycle through real-time 3D reconstruction from the ISR drones that exist today.
Apple Immersive Video set the world standard for spatial quality. Yet no company threw their weight behind distributing in real-time. A recorded clip and a live game are not the same problem. They are four orders of magnitude apart in performance.
Live immersive is a battle. Capture, coloring, graphics, encoding, packaging, CDN, delivery. The process is a refinery, not a button on the remote. SpatialGen builds a box that holds it all together.
Operated with Spectrum executives to showcase real-time immersive streaming of Apple Immersive Video.
We're shortening the distance between a sensor reading and a decision.
RGB, thermal, IMU, GNSS, depth. One reconstruction. The sensors get cheaper and more numerous every year.
A splat scene that a targetter can measure against: distance, height, material, line-of-sight. Not just a point cloud that a human can play with.
Trained on the data we operate with. Tuned for the conditions we ship into. More on this later.
We don't care about the venue. Whether it's a courtside seat in LA or a target overseas. We're reconstructing the world fast enough to act on it.