NNVIDIA Latest coverage

Zero-Shot Real-World Deployment. RAI Institute Creates 5-Motor Bicycle Robot: 8 m/s, Half-Second Flips, 1-Meter Jump

The RAI Institute's UMV robot, a 23.5 kg bicycle-like robot with 5 motor-driven degrees of freedom, achieves high-speed movement (8 m/s), 1-meter jumps, and half-second aerial flips using reinforcement learning. All movement strategies are trained in simulation and directly deployed to the physical robot without additional tuning. The design philosophy focuses on reducing mechanical complexity and relying on algorithmic control for stability and dynamic capabilities.

Zero-Shot Real-World Deployment. RAI Institute Creates 5-Motor Bicycle Robot: 8 m/s, Half-Second Aerial Flips, 1-Meter Jump Platform

The RAI Institute's UMV robot, a bicycle-like machine with 5 motors, achieves 8 m/s speed, 1-meter jumps, and half-second aerial flips without specialized balancing hardware. All movement strategies are trained in simulation and directly deployed to the physical robot, demonstrating zero-shot transfer. The robot uses reinforcement learning to manage balance and dynamic maneuvers, with no springs, control moment gyroscopes, or dedicated stability mechanisms.

Jim Fan@DrJimFan· 16:53SignalEditorial score8585

Jim Fan's team at X Robotics open-sources ASPIRE V1.0, an agentic robotics framework with real robot skill discovery and self-improvement capabilities. The framework was tested on real robots and benchmarks, achieving ~99% performance overnight. The release includes updates on model performance and cost efficiency in robot tasks.

ETH Zurich scientists turn WUJI robotic hand into self-propelled mobile manipulator

ETH Zurich researchers have developed a system where a WUJI robotic hand can crawl, recover from falls, and perform manipulation tasks without a separate mobility system. Using reinforcement learning, the hand learns to support its own weight and move using its fingers. The work was presented in a paper titled 'Fingers as Legs: Learning Self-Supported Locomotion and Manipulation with an Anthropomorphic Hand.'

From AI Agents to Robot Action: Building the Edge Architecture for Physical AI

The article explores the integration of AI agents with robotics through edge computing, emphasizing the use of standardized protocols like MCP and the separation of cognitive and motor functions. It discusses the advantages of edge processing over cloud-based solutions, including reduced latency and cost predictability, and outlines a phased approach for deploying physical AI in industrial settings.

AWS and NVIDIA Launch Physical AI Toolchain for Robot Learning

AWS and NVIDIA have announced a new development platform that supports end-to-end robot AI training, from data generation to simulation and real-world deployment. The toolchain integrates NVIDIA's Isaac Sim and Isaac Lab with AWS's Cosmos to enable training in virtual environments, reducing the need for costly real-world experiments.

Embodied RSI Paradigm Breaks Through Zero-Shot Generalization Bottleneck for Robots: Recursia Outperforms NVIDIA ASPIRE by 10 Points in Long-Term Tasks

This article introduces the Recursia framework, which achieves embodied recursive self-improvement (RSI) by organizing the system as an internal and external dual-loop structure, enabling generalization in robot zero-shot long-sequence tasks. Recursia achieved success rates of 28.4% and 53.4% in LIBERO-Pro Long tasks, respectively 5.8 and 15.1 percentage points higher than NVIDIA ASPIRE. The study emphasizes the importance of integrating execution feedback into high-level decision-making and explores the challenges of maintaining task success across different execution contexts.

From farm to kitchen, Ficio is teaching robots about food

Ficio, a startup founded by Youhanna Meleka, is developing an AI platform for the food supply chain that can run on different machines and tackle tasks like sorting apples in orchards and preparing food in kitchens. FieldSorter uses computer vision to assess apple quality in the field, while FicioPrep automates food preparation in large-scale eateries. The startup aims to reduce waste and improve efficiency across the food supply chain.

Global First Robot Task Native 3D World Model: Cross-Dimensional UNITAS Enables Robots to Learn Physical Intuition, Sweeping Multiple Benchmarks

UNITAS is the world's first robot task-native 3D world model, achieving cross-perspective and cross-body generalization capabilities by unifying observation, action, and environmental changes within a three-dimensional physical space. The model performs exceptionally well in multiple benchmark tests, including LIBERO, LIBERO-Plus, and RoboTwin 2.0. With just 1.7B parameters, it outperforms larger parameter models such as Fast-WAM (6B) and Motus (8B).

Black Forest Labs' FLUX 3 Action Applies Video Expertise to Robotics

Black Forest Labs (BFL) has released FLUX 3 Action, a 7 billion-parameter world-action model (WAM) that converts camera feeds and text instructions into robot movements. It ranks first on the RoboLab-120 benchmark, achieving 42.9% success in 1,200 trials. The model is trained on video data and fine-tuned for specific robot arms, with open weights available under a restrictive license. While it shows promise in simulated environments, its performance on real-world benchmarks like RoboArena remains untested.

Tsinghua's Embodied Model Tops Global Rankings. Outperforming GPT-6 and NVIDIA Without Hacks or Extra Data

Xingdong Jiyuan's self-developed world action model VPP2 has topped the RoboDojo simulation leaderboard with an average success rate of 32.26% and a score of 39.26, outperforming models like GPT-6-Astra. VPP2 improves prediction and action generalization by decoupling video prediction from action learning through a phased training strategy.

OpenWAM: An Open Framework for Composable World-Action Models

This paper presents OpenWAM, an open framework for world-action models (WAM) that aims to standardize the evaluation of different WAM designs. The framework includes a three-stage training process, a shared MoT architecture, and four configurable interaction modes. It introduces modular components like IDM and FDM, trained on a large-scale 'counterfactual' dataset, leading to improved task success rates in both simulation and real-world robotic tasks.

Into the Omniverse: How Developers Turn Ideas Into Simulations With Frontier AI Agents

NVIDIA's blog explains how developers use AI agents and Omniverse libraries to build simulation environments for robotics and autonomous systems. Examples include humanoid robot simulation, autonomous driving testing, and digital twin creation, showing the integration of AI with physics, rendering, and sensor simulation.

Amazon Launches Physical AI Toolchain on AWS to Help Companies Build Machines That Think

Amazon has launched the Physical AI Toolchain on AWS, an open-source platform that combines architecture guidance, deployment automation, and ready-to-use code. It integrates with NVIDIA's Physical AI stack to support the full development lifecycle of physical AI, enabling machines to perceive, understand, and act in the real world. The toolchain includes synthetic data generation, model training, simulation, edge deployment, and continuous improvement, powered by AWS services and NVIDIA's physical AI ecosystem.

The Machines that Make the Machines

This paper describes the development of intelligent robots for assembling GB300 tester trays, a complex manufacturing task requiring high precision and adaptability. The authors explore challenges such as uncertainty in part geometry and performance requirements, and present a technical strategy combining classical robotics, imitation learning, and simulation-based reinforcement learning. The work emphasizes the importance of real-world data and infrastructure for achieving reliable automation in industrial settings.

Half a Year of Winning Millions in Orders, Shipyard Production Scheduled Until 2030: Shipbuilding Welding Robots Enter the 'Golden Age'?

The article discusses the surge in welding robot orders in the shipbuilding industry, driven by the need for automation and efficiency. It highlights the growth of the market, key players like Fao Robot and Zhongxun Technology, and the technological advancements in AI-driven systems and collaborative robots. The text also addresses challenges in complex welding scenarios and the importance of adaptability and reliability in robotic solutions.

Animesh Garg@animesh_garg· 20:59SignalEditorial score8585

FLUX 3 Action is an open-source 7B parameter world action model that achieves first place on the RoboLab benchmark. It outperforms previous models by 6.1 percentage points with 56% fewer parameters and runs 3.95x faster. The model supports fine-tuning for specific robots and tasks and is integrated into LeRobot with deployment on NVIDIA Jetson.

Robotiq Releases New 2F-85 Isaac Sim Asset on Newton

Robotiq has released updated Isaac Sim assets for the 2F-85 gripper, addressing simulation inaccuracies by supporting both PhysX and Newton physics backends. Newton enables precise closed-loop kinematics, crucial for sim-to-real fidelity, while Phys, and is part of Robotiq's open-source developer tools.

Into the Omniverse: How Open World Models Push the Frontier of Physical AI

NVIDIA introduces Cosmos 3, an open-world model family for physical AI, combining vision reasoning, world generation, and action prediction. Available under the Linux Foundation’s OpenMDW 1.1 license, it enables post-training and adaptation for robotics, autonomous vehicles, and vision AI systems. The model ranks highly in benchmark evaluations and is part of NVIDIA’s broader physical AI stack.

NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics

NVIDIA introduces Cosmos-H-Dreams, a real-time, action-conditioned generative simulator for surgical robotics. It distills the capabilities of Cosmos-H-Surgical-Simulator into a causal, few-step student model, enabling interactive environments for policy evaluation and synthetic data generation. The system runs on a single NVIDIA RTX PRO 6000 GPU and supports integration with surgical platforms like Versius.