Robotics AI News & Updates

FieldAI Secures $405M to Develop Physics-Based Universal Robot Brains for Cross-Platform Embodied AI

FieldAI raised $405 million to develop "foundational embodied AI models" - universal robot brains that can work across different robot types from humanoids to self-driving cars. The company's approach integrates physics into AI models to help robots safely adapt to new environments while managing risk, addressing traditional robotics limitations in generalization and safety.

Nvidia Launches Cosmos World Models and Infrastructure for Physical AI and Robotics Development

Nvidia unveiled new Cosmos world models including Cosmos Reason, a 7-billion-parameter vision language model designed for physical AI applications and robotics. The company also introduced neural reconstruction libraries, new servers, and cloud platforms to support robotics development workflows. These announcements represent Nvidia's strategic expansion into robotics as the next major application for AI GPUs beyond data centers.

AI Video Companies Luma and Runway Target Robotics and Autonomous Vehicles for Revenue Expansion

AI video-generating startups Luma and Runway are exploring partnerships with robotics and self-driving car companies as potential new revenue streams beyond their current focus on movie studios. Luma is particularly positioned for this expansion given their announced goal of building 3D AI world models that can understand and interact with physical environments.

Hugging Face Enters Robotics Market with $1M in Sales of Open-Source Reachy Mini Robot

Hugging Face, primarily known for open-source AI models, has entered the robotics market with its Reachy Mini robot, achieving $1 million in sales within five days of launch. The desk-sized robot features cameras, microphones, speakers, and is designed as a hackable entertainment device that runs open-source software and custom apps. The company positions this as an accessible entry point for consumers to become comfortable with AI-powered robots in their homes.

RealSense Spins Out from Intel with $50M to Scale 3D Vision Technology for Robotics

RealSense has spun out of Intel as an independent company after 14 years, raising $50 million in Series A funding to scale its stereoscopic imaging technology. The company's 3D perception cameras are used in robotics, autonomous vehicles, and drones to help machines understand their physical surroundings in real-time.

Amazon Reaches One Million Warehouse Robots and Launches DeepFleet AI Coordination System

Amazon has deployed one million robots across its warehouses after 13 years of automation efforts, with 75% of global deliveries now robot-assisted. The company also released DeepFleet, a generative AI model that coordinates robot routes and increases fleet speed by 10%.

Genesis AI Secures $105M to Develop General-Purpose AI Foundation Model for Robotics

Genesis AI emerged from stealth with $105 million in seed funding to build a foundational AI model that can power various types of robots for automating repetitive tasks. The startup uses proprietary synthetic data generation through a physics engine to train robotics models, avoiding the costly and time-consuming process of collecting real-world data. Genesis plans to release its model to the robotics community by the end of the year.

Google DeepMind Releases Gemini Robotics On-Device Model for Local Robot Control

Google DeepMind has released Gemini Robotics On-Device, a language model that can control robots locally without internet connectivity. The model can perform tasks like unzipping bags and folding clothes, and has been successfully adapted to work across different robot platforms including ALOHA, Franka FR3, and Apollo humanoid robots. Google is also releasing an SDK that allows developers to train robots on new tasks with just 50-100 demonstrations.

Meta Releases V-JEPA 2 World Model for Enhanced AI Physical Understanding

Meta unveiled V-JEPA 2, an advanced "world model" AI system trained on over one million hours of video to help AI agents understand and predict physical world interactions. The model enables robots to make common-sense predictions about physics and object interactions, such as predicting how a ball will bounce or what actions to take when cooking. Meta claims V-JEPA 2 is 30x faster than Nvidia's competing Cosmos model and could enable real-world AI agents to perform household tasks without requiring massive amounts of robotic training data.

Amazon Establishes Dedicated R&D Group for Agentic AI and Robotics Integration

Amazon announced the launch of a new research and development group within its consumer product division focused on agentic AI. The group will be based at Lab126, Amazon's hardware R&D division, and aims to develop agentic AI frameworks for robotics applications, particularly to enhance warehouse robot capabilities.