NVIDIA releases MotionBricks AI for Real-Time game and robot movement
NVIDIA unveiled MotionBricks, an AI system that can generate realistic movements for game characters and humanoid robots in real time. The technology was presented at SIGGRAPH 2026 and the company released source code and pre‑trained models in October 2026. MotionBricks uses a motion‑in‑betweening approach to interpolate between a starting pose and a target pose, producing motion clips that last…
Key points
- nvidia released motionbricks, an ai that generates real‑time movements for game characters and humanoid robots.
- motionbricks interpolates between a starting pose and a target pose, producing clips 0.6 to 1.9 seconds long.
- released models run at 30 fps on an rtx 4090 and are available on github and hugging face.
Key performance figures include a processing speed of 15,000 FPS for motion generation, while the released models support 30 FPS generation. The demo runs on an RTX 4090 and may work on other NVIDIA GPUs, though verification is needed. The open‑source implementation includes a Smart Primitive Editor and a Scene Editor for editing motion components and scenes. The system also powers NVIDIA’s GR00T Whole‑Body Control on the Unitree G1 humanoid robot, though additional balance and joint‑limit control is required for hardware execution.
MotionBricks demonstrates how AI can replace manual animation pipelines, potentially speeding up game development and robot control workflows.
Overview of NVIDIA MotionBricks - AI for Real-Time Generation of Game and Robot Movements
note.com · 30 October 2026
Loading the full article…
This text was published by note.com and written by npaka. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Anthropic removes Opus 4.8 fast mode on November 30 · 2 src
- Google announces Gemini 4 Carbon, its latest frontier AI model for cybersecurity · 7 src
- LiquidAI releases browser demo for LFM3-Omni voice AI · 1 src
- Datalab releases Surya 2.5 OCR model with 85.9% accuracy on OmniParseBench · 1 src
- Sensenova/SenseNova-Loop-9B-Preview · Hugging Face · 1 src
Comments
via GitHub Discussions