Search results for Allex
AI & Enterprise
RLWRLD unveils \'RLDX-1\' robotics foundation model
Physical AI company RLWRLD said on Wednesday it unveiled its in-house robotics foundation model, RLDX-1. The company released three 8.1B-parameter variants, including a pre-training checkpoint and two platform-specific mid-training checkpoints, and made model weights, training code and technical documents available to external researchers via GitHub and Hugging Face. RLWRLD said the model processes vision, language, torque, touch and task memory, and highlighted its Multi-Stream Action Transformer architecture.
AI & Enterprise
Manufacturing convergence puts \'physical AI\' in focus
Companies are accelerating moves to develop physical AI that blends AI with the real world. At CES 2026, firms showcased technologies ranging from Nvidia’s autonomous driving lineup to robotics training based on Nvidia’s open humanoid model. Partnerships between manufacturers and AI companies are expanding, while Mobileye plans a $900 million acquisition in the sector. The article also notes signs that applying AI to manufacturing can be difficult, citing limited progress in Nvidia’s Omniverse monetisation efforts.
AI & Enterprise
RLWRLD uses open robot AI GR00T N1.5 to advance robot hand manipulation training
RLWRLD said on Jan. 6 it is enhancing training for its Allex robot and a five-finger robotic hand to manipulate objects stably, based on Nvidia Research’s open humanoid robot AI model NVIDIA Isaac GR00T N1.5. The company said grasp-and-move tasks require perception, understanding and fine finger control, and that small errors can lead to failure. RLWRLD said it is optimizing and fine-tuning the pre-trained model and building a structure for rapid, repeatable experiments. It also plans to unveil a proprietary VLA model early this year.