TechCapsules

Robots pour cocktails and run marathons, but still can’t multitask

Humanoid robots showcased at the Robotics Summit in Boston demonstrated capabilities such as mixing cocktails, running marathons, and folding laundry. However, despite marketing claims, these robots are often limited to performing a single task or require human control. Experts attribute this limitation to the complexity of multitasking and the need for more advanced artificial intelligence. Robots are making progress, with advancements in hand dexterity and AI-powered vision-language-action models. However, creating a general-purpose robot that can perform various tasks independently is still years away.

💡 The stakes are higher for robots than for chatbots, as their physical actions can cause harm. To bridge this gap, companies are gathering vast amounts of data on human movement to improve their robots' abilities.

Chang Robotics and Wheel.me Partner to Make Any Cart, Shelf, or Fixture…

Chang Robotics and Wheel.me have expanded their partnership to bring autonomous mobility to a wider range of industries and facility types. The collaboration enables Chang Robotics to serve clients with compact, flexible automation that works with their existing infrastructure. Wheel.me's Genius platform uses autonomous wheels with lidar and 3D vision to transform standard carts, shelves, and fixtures into self-driving units, operating in environments with limited space and flexibility. The partnership builds on a relationship that began in 2023 and will integrate Wheel.me's technology into client solutions.

Embodied AI Robotics Startup Astrall Dynamics Gains Investment from SAIC and Others

Astrall Dynamics, a Shenzhen-based embodied AI robotics startup, has secured investment from SAIC Motor and Dongfang Precision, increasing its registered capital to 67.33 million yuan. Founded in June 2024, the company specializes in R&D, manufacturing, and sales of smart robots and motor control systems. Astrall Dynamics has also partnered with China Life Property & Casualty Insurance to develop dedicated insurance products for quadruped robots, aiming to drive innovation in smart equipment and establish a new model for risk coverage.

💡 The investment and partnership signify a growing recognition of the potential for embodied AI in China's robotics industry, with Astrall Dynamics poised to play a key role in driving innovation and adoption.

China’s Alibaba unveils AI brains designed to power the next generation of…

Chinese firm Alibaba has unveiled its first embodied AI model family, the Qwen-Robot suite, which links large language models with real-world robotic actions. The suite comprises three models focused on navigation, manipulation, and world modeling for robots operating in physical environments. Developed by Alibaba's Tongyi Lab, the Qwen-Robot suite enables machines to perceive, reason, and interact with the real world, joining a growing global push to advance embodied AI beyond traditional chatbot applications. The models are currently undergoing pilot testing with selected Alibaba Cloud enterprise clients.

💡 The Qwen-Robot suite's success could accelerate the development of robots capable of performing complex tasks in real-world environments, potentially transforming industries such as manufacturing, logistics, and healthcare.

Leju Robot and Huaxi Intelligence Form Strategic Partnership

Leju Robot, a major player in the humanoid robotics sector, has formed a strategic partnership with Huaxi Intelligence to collaborate on high-computing main control platforms, edge-side model deployment, industry solutions, and benchmark projects. The partnership focuses on four key areas, including co-developing domestic high-computing main control solutions, driving efficient deployment of embodied AI large models, and creating benchmark solutions for the industry. Leju Robot and Huaxi Intelligence aim to build a developer open-source ecosystem to boost scenario adaptation for domestic chips and bridge the full chain of "data-model-chip."

💡 This partnership has significant implications for the development of domestic chips in the robotics industry, as Leju Robot and Huaxi Intelligence aim to bridge the full chain of "data-model-chip" through open-source data and ecosystem development.

Evaluating humanoids for surface finishing applications

Researchers evaluated humanoids for surface finishing applications, examining features such as legs, hands, and heads. Legs are inefficient for factory floor locomotion and add complexity, while multi-fingered hands are expensive and may not provide a strong grasp. Heads offer limited value due to restricted camera placement. Dual arms are useful, but a configuration with general-purpose robot arms, rails, and cameras is more effective. This configuration can operate faster and apply higher forces, making humanoids unlikely to be a winning solution in manufacturing applications.

💡 Humanoids' limitations in manufacturing processes underscore the importance of tailored robotic solutions. By leveraging general-purpose industrial robots, manufacturers can optimize production speed and precision.