embodied AI

Chinese state media reports that the country's embodied AI sector is shifting from research settings into practical deployment, with robots taking on actual jobs in industry and services. The report highlights China's push to turn embodied artificial intelligence from a laboratory technology into a working part of the economy, as companies and institutions expand real-world applications.

Experts unveiled what is described as the world's first embodied AI robot built on a fully domestic electronic architecture. Analysts say the milestone marks a shift in robotics from developing key components individually toward achieving underlying autonomy at the system level. The announcement is drawing attention for its potential implications for self-reliance in robotics and artificial intelligence technology.

Chinese state media reports that embodied AI robots developed in China are increasingly being deployed in real-world workplaces rather than remaining confined to research and classroom settings. The feature highlights how the country's robotics sector is pushing humanoid and intelligent machines into practical jobs, reflecting China's broader push to lead in embodied artificial intelligence and industrial automation.

A Chinese robotics startup says it could achieve a ChatGPT-style breakthrough in robot intelligence, often called a 'robot brain', as soon as next year. The claim suggests rapid progress toward general-purpose AI for robots, allowing machines to learn and adapt the way large language models do, and would mark a major milestone in China's push to compete in embodied AI.

South Korean robotics company Angel Robotics has launched PHAI-X1, a wearable robot platform aimed at physical AI research. The system is intended to give researchers a hardware base for developing and testing AI that interacts with the physical world, such as exoskeleton control and human-assistance applications. Details on pricing and availability were not included in the initial announcement.

Embodied AI company Knowin published a technical report on September 24 for GLOW, a framework designed to let household robots learn a new task from a single complete human demonstration. The approach aims to reduce the training data normally required for robot skills, drawing attention from robotics observers following the Pandaily coverage.

Chinese AI firm SenseTime has launched SenseMart OS, an operating system designed to bring embodied intelligence into retail environments. The system is aimed at powering robots and smart devices that can perceive, interact with, and assist customers in stores. Details about pricing, availability, and retail partners have not yet been widely reported.

A project called Roboharm is asking whether frontier robot policies actually refuse unsafe instructions. The work examines how AI systems controlling robots respond to harmful commands, a safety question increasingly urgent as AI models are deployed in physical robotics. Discussion is focused on how well current safeguards carry over from language models to embodied systems.

Black Forest Labs, the German startup best known for image generation, has taken the top spot on Nvidia's robotics leaderboard with FLUX 3 Action, a 7-billion-parameter model scoring 42.9%. The result is notable because it outperforms entries on Nvidia's own benchmark for robot foundation models, signaling that image-generation specialists are now competing directly in embodied AI and robot learning.

A British startup is turning video game controller inputs into training data for AI models designed to navigate the physical world. The idea is that human gameplay, including mistakes and clumsy manoeuvring, offers rich examples of how people learn to control complex systems, which could help robots and other embodied AI improve. The approach is drawing attention as a cheaper alternative to real-world data collection.

Figure AI says it has identified a scaling law for robotics, suggesting robot performance improves predictably as models and data grow. The claim comes with an important caveat: the relationship reportedly holds only about 56% of the time. The announcement is drawing attention in the robotics and AI community, where scaling laws have transformed language models and companies are racing to find an equivalent for embodied AI.

Industry commentary argues that physical AI — robotics and intelligent machines moving from pilots to real-world deployment — still lacks a repeatable, reliable path to production. Writers in the manufacturing press say companies need standardized processes to take AI-powered systems out of demos and into factories at scale, echoing broader debate about how fast embodied AI can be industrialized.