How Hesai’s Kosmo captures 3D scenes for Pixar workflows and physical AI
乔布斯的皮克斯，狂赞中国物理AI数据终端
- Who
- QbitAI
- What
- analyzed Hesai's Kosmo platform and its role in physical AI
- Quote
Kosmo的特性，实际上已经远远超越智能汽车，甚至是具身智能应用场景。
What happened
Hesai’s Kosmo spatial capture system features in a Pixar RenderMan tutorial demonstrating workflows with real-world 3D Gaussian splatting data. The article explains how Kosmo One synchronizes lidar, imagery, and IMU readings before cloud reconstruction through Kosmo Spatial Core, which outputs 3DGS, mesh, and depth data. Hesai says scanning a roughly 100-square-meter space takes 15–20 minutes; the article also describes semantic segmentation and AI completion of unscanned areas for robot training data.
Read next
See the full storyHesai’s Kosmo Demonstrates Real-World 3D Capture in Pixar RenderMan TutorialThe story so far
Hesai’s Kosmo spatial capture system appears in a Pixar RenderMan tutorial demonstrating how real-world 3D Gaussian splatting data can be used in rendering workflows. Kosmo One synchronizes lidar, imagery and inertial measurement unit (IMU) readings, then sends them for cloud reconstruction through Kosmo Spatial Core, which outputs 3D Gaussian splatting, mesh and depth data. Hesai says scanning a roughly 100-square-meter space takes 15–20 minutes. QbitAI also describes semantic segmentation and AI completion of unscanned areas as capabilities for producing robot training data, connecting the platform’s scene-capture workflow to physical AI applications.
Keep up with AI, one email a day
Model releases, price changes and compute developments, with original sources. Free, confirmed by email, and easy to unsubscribe. Read a past issue · 2026-10-27