open-source

AGIBOT unveils a unified simulation infrastructure that uses LLM-driven spatial world models and massively parallel reinforcement learning to accelerate the path from digital training to physical deployment.

With a $6.5 million seed round and a $5,000 open-source arm, Anvil Robotics aims to provide the foundational hardware and software for the next wave of AI builders.

Menlo Research’s open-source project, Asimov, moves from GitHub repository to physical hardware with a bipedal kit designed for rapid iteration and "Processor-in-the-Loop" development.

NVIDIA researchers have revealed EgoScale, a framework that leverages a massive 20,854-hour egocentric human dataset to train robots in complex, fine-grained manipulation with minimal robot-in-the-loop data.

NVIDIA has released SONIC, a generalist humanoid controller trained on 100 million frames of motion data, aiming to replace manual reward engineering with a scalable "System 1" foundation for whole-body movement.