Runway releases Praxis‑1 world action model with open weights
Runway has made Praxis‑1, a 27-billion-parameter model designed to predict robot video and actions jointly, generally available with Apache 2.0 open weights.
1 min readRead story
CHOOSE YOUR FOCUS
What do you create?
NEWS FEED
Robotics
TopicRoboticsClear topic
Sorted by AI Radar publication time · Times shown in Istanbul (UTC+3).Clear filters
NTT’s physics-based world model tests robot actions before movement
NTT developed a physics-based world model that predicts candidate robot actions as video before execution and evaluates risks such as dropped objects or collapsing loads.
2 min readRead story
Hedra Model Factory puts physical AI models into a continuous improvement loop
Hedra launched Model Factory for robotics, manufacturing, biology and defense teams. The platform combines GPU sandboxes, post-training, low-latency deployment and retraining from production evidence in one loop.
1 min readRead story
MemryX MX3+ targets more edge AI capacity within the same power envelope
MemryX is targeting up to 10 times the peak AI compute and nearly 10 times the on-chip model memory of MX3 with the MX3+ accelerator, which is planned for 2027.
1 min readRead story
Gemini Robotics 2.5 advances precise manipulation and multi-step robot tasks
Google DeepMind's new vision-language-action model reports higher success rates than Gemini Robotics 2 across precise manipulation, contact-rich tasks and reliability on multiple robot embodiments.
1 min readRead story
Boston Dynamics uses a wheeled Atlas to test sim-to-real robustness
The wheeled Atlas is not a product; it is a research testbed for stressing actuators at high torque and speed and transferring a whole-body reinforcement-learning controller to real hardware.
2 min readRead story
NVIDIA GR00T 2 learns robot skills from human video and transfers them across embodiments
NVIDIA’s open world action model GR00T 2 learns dexterous manipulation from human and robot data and can transfer the same skills across different robot bodies.
2 min readRead story
Physicl Studio turns robot failure scenes into controlled training datasets
Physicl opened Studio in private beta, allowing physical AI teams to describe difficult scenes, generate controlled variations and receive exact labels for every frame.
1 min readRead story
Teradyne and Sereact bring Cortex physical AI to UR robots
The partnership scales Sereact’s Cortex platform on Universal Robots arms for unstructured logistics tasks such as returns handling and piece-picking.
2 min readRead story
CoreWeave launches one-command Physical AI templates for robotics and autonomous-vehicle workloads
Physical AI Launchables provides a catalog of tested simulation, data-generation, training, and evaluation templates for Isaac Sim, CARLA, MuJoCo, AlpaSim, and Cosmos 3.
2 min readRead story
Microduck carries an Isaac Lab-trained walking policy onto a real robot
NVIDIA and Hugging Face published an open workflow that trains Microduck’s walking policy in parallel simulation with Isaac Lab 3.0 and exports it through ONNX for deployment on the physical robot.
2 min readRead story
Laplace stress-tests robots and autonomous vehicles in simulation before the real world
Laplace Intelligence is building interactive simulations and automated red-team agents to identify dangerous edge cases in physical AI systems before deployment.
2 min readRead story
BEYOND THE HEADLINE
What changed?
What can it bring to your work?
Every story includes a concise summary, its creative workflow implications and original sources. Explore official demos and examples alongside the news.
Our editorial approach
