Robot Manufacturer Mobile Manipulator Human-Machine Interaction Education & Research
Stardust AI Launches Lumo-2: Innovative Robot Action Model for Home Automation
Original from leaderobot.com: Stardust AI Unveils Lumo-2: Advanced Robot Action Model for Household Tasks

Stardust AI Launches Lumo-2: Innovative Robot Action Model for Home Automation

On July 15, Stardust AI introduced its second-generation embodied base model, Lumo-2, which is the industry's first household latent world-action model. This launch includes the physical AI symbiotic agent, Agent Philia, enhancing their full-stack architecture of AI models, embodied operating systems, and rope-driven entities. The company will showcase its 'trinity' multi-scenario implementation solutions at the World Artificial Intelligence Conference in Shanghai from July 17 to 20.

Lumo-2 autonomously performs 22 complex household tasks, demonstrating industry-leading capabilities in task range and complexity. This model addresses the challenges faced by robots in open environments, such as the inability to explain actions and the high costs of training complex skills. By predicting future scenarios before generating actions, Lumo-2 aims to overcome these bottlenecks and improve the practical execution of robotic tasks.

Looking ahead, Stardust AI plans to enhance the scalability of Lumo-2 by expanding training data diversity and exploring efficient data engineering paradigms. The team is also focused on advancing real-world interactive learning to enable robots to adapt and evolve autonomously in dynamic environments. No further timeline was disclosed at the time of publication.

Editor's Note

The introduction of Lumo-2 marks a significant advancement in the robotics sector, particularly in household automation. As robots increasingly enter open environments, addressing the challenges of action prediction and skill acquisition becomes crucial for widespread adoption. The focus on scalable learning and diverse data integration could reshape the competitive landscape in robotic applications.

RobotToday Initiative

Robotics needs a service framework.

RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.

Share

Related Articles/News

Robotic Reaction Mapping Uncovers Hidden Reactivity in Biginelli Reaction

Robotic Reaction Mapping Uncovers Hidden Reactivity in Biginelli Reaction

Recent advancements in robotic reaction mapping and chemical AI have unveiled a surprising pseudo-seven-component transformation in the Biginelli reaction, a process that has been studied for over a century. This discovery is significant as it enhances the understanding of the Biginelli reaction, potentially leading to new applications in organic synthesi...

Nvidia Launches Cosmos 3 Edge AI Model and Expands Physical AI Ecosystem in Japan

Nvidia Launches Cosmos 3 Edge AI Model and Expands Physical AI Ecosystem in Japan

Nvidia has introduced its new AI model, Cosmos 3 Edge, aimed at enhancing physical AI applications in Japan. This model is designed to help systems perceive and navigate real-world environments, marking a significant step in Nvidia's strategy to penetrate the physical AI market. The expansion is part of CEO Jensen Huang's visit to Japan, where Nvidia is f...

1X Unveils NEO Robot Hand with Unmatched Precision for Household Tasks

1X Unveils NEO Robot Hand with Unmatched Precision for Household Tasks

1X has introduced the NEO robot hand, showcasing its ability to perform intricate tasks with remarkable precision. The hand features 25 degrees of freedom and utilizes a tendon-driven structure, mimicking human muscle and tendon systems, allowing for smooth and natural movements. The NEO hand can perform tasks such as separating grapes, twisting light bul...

Robotics Dexterous Manipulation AI Technology
LimX Dynamics' Oli Humanoid Completes Household Tasks with Advanced COSA 0.5 Brain System

LimX Dynamics' Oli Humanoid Completes Household Tasks with Advanced COSA 0.5 Brain System

LimX Dynamics has announced that its Oli humanoid robot has successfully autonomously completed a range of household chores. This achievement utilizes the COSA 0.5 brain system, which features a three-tier S2-S1-S0 architecture, marking a significant advancement in humanoid robotics. This development is noteworthy as it positions LimX Dynamics' Oli as the...

Technology
Xiaomi Launches Robotics-U0: A Unified 38B-Parameter Generative Model for Robotics

Xiaomi Launches Robotics-U0: A Unified 38B-Parameter Generative Model for Robotics

Xiaomi has unveiled Robotics-U0, an advanced embodied generative model featuring 38 billion parameters. This innovative model is designed to perform multiple tasks, including scene generation, embodied transfer, video generation, and text-to-image capabilities, all within a single unified architecture. The introduction of Robotics-U0 is significant as it ...

Technology
MIT and Toyota Research Institute Unveil SceneSmith for Robot Household Training

MIT and Toyota Research Institute Unveil SceneSmith for Robot Household Training

MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Toyota Research Institute have developed SceneSmith, an AI-powered system that allows robots to practice household tasks in a virtual environment. This system utilizes three visual language models to collaboratively create realistic 3D scenes, enabling robots to learn complex sk...

AI Robotics Virtual Reality
Robbyant Launches LingBot-VA 2.0, the First Embodied-Native AI Model for Robotics

Robbyant Launches LingBot-VA 2.0, the First Embodied-Native AI Model for Robotics

Robbyant, a company under China's Ant Group, has introduced LingBot-VA 2.0, touted as the first embodied-native video-action world model specifically designed for robotics. Unlike traditional models adapted from digital content, LingBot-VA 2.0 is built from the ground up for physical-world tasks, enhancing physical accuracy and execution efficiency throug...

AI and Robotics
Flexible Soft Sensor Enables Robotic Action Without Traditional Electronics

Flexible Soft Sensor Enables Robotic Action Without Traditional Electronics

Researchers have developed a new soft sensor that allows soft robots to convert touch into action without relying on traditional electronic components. This innovation addresses a critical limitation in soft robotics, which often depend on bulky electronic sensors and actuators. The sensor is made from flexible materials, enhancing the robot's adaptabilit...

Robotics
PHANES AI Launches TouchWorld Tactile Model for Enhanced Robot Dexterity

PHANES AI Launches TouchWorld Tactile Model for Enhanced Robot Dexterity

PHANES AI, established by 28-year-old Yang Shuo from HIT, has introduced TouchWorld, a tactile foundation model designed to enhance robotic manipulation capabilities. This model enables robots to perform precise physical tasks by incorporating a sense of touch, marking a significant advancement in robotic dexterity and interaction with their environment. ...

Technology
Thirteen New Embodied AI Models Released in June 2026 by BAAI and Alibaba

Thirteen New Embodied AI Models Released in June 2026 by BAAI and Alibaba

In June 2026, the landscape of embodied AI saw the introduction of 13 new models, including significant contributions from BAAI and Alibaba's Qwen-Robot. This rapid development indicates a shift from traditional hardware benchmarks to a focus on software intelligence, highlighting the competitive nature of the field. The emergence of these models undersco...

Technology
ROBOTTODAY Weekly July 6 – 10, 2026

ROBOTTODAY Weekly July 6 – 10, 2026

This week in robotics: AI² Robotics raises $735M for wheeled humanoids, Unitree G1 performs first live surgery, Dexmal launches DexOS platform, Agility Robotics announces SPAC, and Ondas acquires DZYNE for $875.8M. July 6–10, 2026.

RobotToday Weekly
Physical AI Landscape: From Digital Intelligence to the Embodied Physical World

Physical AI Landscape: From Digital Intelligence to the Embodied Physical World

A comprehensive industry analysis of Physical AI — covering Embodied AI, Field AI, Factory AI, VLM, World Model, VLA, and the VLA vs WAM architecture debate. Includes 20+ companies, 28 references, and the RobotToday Physical AI Landscape diagram. July 2026.

Artificial Intelligence

Related Suppliers

LiOps

The Spatial Intelligence Agent leverages advanced 3D segmentation and real-time object recognition to enhance robotic automation in logistics and manufacturing. Utilizing hybrid mapping techniques, it integrates spatial intelligence for dynamic environment understanding. The system enables immediate deployment without custom development, significantly reducing integration costs and improving operational efficiency through precise path calculation and obstacle avoidance capabilities.

WAKU Care (WAKU Robotics)

The Intelligent Maintenance Agent (IMA) leverages AI for operational efficiency, integrating Augmented and Mixed Reality (AR/MR) technologies. It supports Enterprise Asset Management (EAM) and Computerized Maintenance Management Systems (CMMS) for optimized asset management and maintenance processes. Key features include predictive maintenance, data insights, and a ticketing system, facilitating effective field service management and spare part optimization across various industries.

UC Berkeley Robotics & Intelligent Machines Lab

The BAIR and RAL labs at UC Berkeley conduct foundational research in robot learning, dexterous manipulation, and reinforcement learning. Berkeley has produced landmark work on deep RL for robotics, imitation learning, and vision-language-action models.

Education & Research Software & Algorithm Provider

Physical Intelligence

Physical Intelligence focuses on integrating general-purpose AI with physical systems. The team employs foundation models and reinforcement learning algorithms to enhance robotic manipulation and task execution. Recent advancements include real-time action chunking for large vision-language-action (VLA) models, enabling improved generalization and efficiency. Notable projects address Moravec's Paradox and explore human-to-robot transfer in scalable robotic frameworks.

7Xrobotics

7Xrobotics specializes in the development of autonomous robots designed for household tasks, specifically dishwashing and laundry folding. Utilizing advanced machine learning algorithms and robotic process automation (RPA), these robots integrate computer vision for object recognition and manipulation. The systems are engineered for efficiency, employing sensors for navigation and task execution within domestic environments, optimizing user convenience and operational performance.

System Integrator

Weave Robotics

Weave Robotics specializes in the development of domestic robotics, focusing on autonomous systems designed for household tasks. The flagship product, Isaac, integrates advanced AI algorithms and machine learning for efficient task management, including laundry folding and home organization. Scheduled for consumer deployment in 2026, Isaac aims to enhance home automation and personal assistance capabilities within residential environments.

Generalist AI

Generalist focuses on advancing robotics through the development of embodied foundation models that enhance physical interaction capabilities. Recent innovations include the integration of physical commonsense and reflexive intelligence, enabling robots to perform complex tasks like one-shot assembly. Ongoing research evaluates dexterous sensorimotor policies across diverse environments, aiming to bridge the gap between AI models and real-world applications in robotics.

Advex AI Inc.

Advex AI develops AI-powered software solutions for automation and robotics. Emphasizes scalable AI models for industrial applications, including dexterous manipulation, process optimization, and integration in manufacturing environments.

Software & Algorithm Provider

The Robot Learning Company

The TRLC-DK1 is an open-source development kit designed for AI-native robotics, featuring a single-arm configuration with a 700mm reach and 1kg payload capacity. It supports multiple operating systems (Linux, MacOS, Windows) and utilizes USB-C for data transfer. The advanced TRLC-DK1-X variant enables bimanual learning with dual leader-follower setups and enhanced sensory input through three cameras.

U Leuven

KU Leuven is a Belgian research university with active robotics research groups covering robot learning, manipulation, autonomous vehicles, and human-robot interaction. The university collaborates extensively with industrial partners on applied robotics projects.

Education & Research Healthcare & Senior Care Research Institute

Impaqt Robotics Private Limited

Based in India and built for the world, we design pneumatic and vacuum interfaces, modular grippers, case forming systems and palletizing platform grounded in real integration experience to reduce complexity and help teams grow their automation with confidence.

Robot Manufacturer Software & Algorithm Provider System Integrator

Planys Tech

Planystech is a pioneering company at the forefront of robotics and smart device innovation. Specializing in cutting-edge automation solutions, they offer a diverse range of products designed to enhance efficiency and streamline operations across various industries. Their services include tailored robotics integration, smart device development, and comprehensive support, ensuring clients can leverage the latest technology to optimize their workflows. With a commitment to quality and innovation, Planystech is transforming the landscape of automation, making it accessible and effective for businesses of all sizes.