A single destination for timely, editor-curated robotics news from around the world.
On October 14, Google DeepMind officially launched the Gemini 4 Argon, a universal physical reasoning model. This model is designed to handle tasks such as robot control, virtual environment setup, training data quality inspection, and experimental replication. A demonstration with the Apptronik Apo...
Hesai has introduced Kosmo, a 3D world data platform that integrates the Kosmo One spatial camera and Kosmo World cloud service to transform real-world environments into digital formats for robot training and world modeling. An operator can scan a typical 100-square-meter indoor area in approximatel...
Hello Inc.'s robotaxi division has introduced HelloWorld, a generative driving world model designed to create multi-view camera video and lidar data. This innovative model can generate these outputs based on edited layouts or plain-language prompts, enhancing testing capabilities for autonomous vehi...
DexForce has introduced UNITAS, the world's first native 3D world model designed specifically for robotic tasks. This innovative model aims to unify the understanding of physical interactions by allowing robots to comprehend the three-dimensional physical world, which is crucial for advancing toward...
Odyssey, an AI research company, has launched Odyssey-3, its most advanced world model to date. This model represents a major leap towards general-purpose intelligence, enabling systems to understand, predict, and interact with both physical and virtual environments. Odyssey-3 can control robots, hu...
Odyssey has unveiled Odyssey-3, a versatile world model designed to control robots, humanoids, autonomous vehicles, and drones using a unified AI system. This autoregressive diffusion transformer is trained on a diverse set of visual observations, allowing it to adapt with minimal task-specific data...
On September 14, Shanghai Juna Technology Co., Ltd. officially launched Cog-WM 1.0, the world's first brain-inspired cognitive world model, at the 2026 Pujiang Innovation Forum. This model, based on systematic brain-like neural mechanisms, has demonstrated capabilities in autonomous navigation and m...
Shanghai Artificial Intelligence Laboratory has introduced the Intern physical world model W0, which features native force-tactile sensing and duplex collaboration capabilities. This model is designed to enhance robotics applications by integrating with Intern InkStone and the science model S2, faci...
On September 14, Digua Robotics and Giga Vision announced a strategic collaboration focused on integrating their respective technologies. Giga Vision will provide world models, embodied foundational models, and real-world application experience, while Digua Robotics will contribute an edge AI comput...
ACE Robotics and NTU S-Lab have announced the open-source release of Puffin-World, a multimodal world model that integrates physics, geometry, and appearance. This model utilizes the Puffin-16M dataset and achieves state-of-the-art camera absolute pose errors, reporting sub-degree accuracy across fo...
Scalabot has introduced the HERON-World Model, a new action-conditioned world model aimed at enhancing robotic understanding and future prediction capabilities. This model utilizes a three-tier data pyramid, incorporating real robot teleoperation data, simulation data, and first-person human operati...
From August 19 to 23, 2026, the World Robot Conference (WRC) took place at the Yichuang International Exhibition Center in Beijing. The event, themed 'Human-Machine Symbiosis, Integration of Production and Demand,' featured over 300 exhibitors and more than 2,000 exhibits. Keenon Robotics showcased...
Muka Robotics has achieved a remarkable feat with its LJM Latent Joint-conditional Model, securing the second position globally at WorldArena with a motion quality score of 89.17 using only 32 Zhenwu 810E GPUs. This accomplishment highlights the model's innovative dual-expert architecture, which emp...
Contactile has developed tactile sensors for robotic hands and grippers that enhance world models in robotics. These sensors address the limitations of current conditioning methods, which primarily rely on visual inputs and joint encoders, by incorporating friction as a critical input. This advancem...
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 r...
Runway AI Inc. has introduced Praxis-1, an open-weight world action model designed to enhance robotic control through video pretraining. The company is currently testing Praxis-1 across various embodiments and environments to identify and address potential gaps before its general release. The signi...
Xiaomi has open-sourced the Xiaomi-Robotics-U0, an autoregressive embodied world foundation model featuring approximately 4 billion parameters and full-scale weight lines of around 38 billion. This release includes training and inference tools designed to enhance robotic applications. The significa...
On September 10, at the Bund Conference, Luo Yihang, co-founder and CEO of Shengshu Technology, unveiled Motus2, a self-evolving general world model aimed at robotic dexterous manipulation. Unlike traditional robots that rely on explicit instructions, Motus2 enables robots to learn autonomously from...
At the 2026 World Robot Conference, a clear trend emerged indicating that the embodied intelligence industry is shifting focus from creating general-purpose robots to developing specialized solutions that deliver real value in specific scenarios. Innovations in areas such as garment sewing, biomedic...
Samsung Electronics is accelerating the development of its proprietary humanoid robot by integrating internal manufacturing data, custom hardware, and foundation AI models into a centralized initiative. Reports indicate that this effort operates independently from Rainbow Robotics, where Samsung hol...
A startup named 'Xindu Qiyuan' has successfully utilized a 'white box world model' based on 'uncertain differential geometry' to enable a humanoid robot to perform a complete task from voice command to cup retrieval without prior calibration or pre-training. This achievement is particularly signific...
At the WAIC, RoboScience showcased a groundbreaking cloud-based embodied large model named Visics, capable of controlling multiple robotic hands. This innovation allows for seamless switching between different robotic hands while maintaining operational efficiency, demonstrating the ability to recog...
On July 17, Extreme Intelligence unveiled Gravity 4D WAM, the first module of its embodied intelligence framework, at the 2026 World Artificial Intelligence Conference (WAIC). This launch addresses a critical issue in current visual language models: the discrepancy between visually plausible predict...
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, en...
At COMPUTEX in Taipei, NVIDIA unveiled Cosmos 3, a groundbreaking open world foundation model designed to integrate vision reasoning, physical simulation, and action prediction. This launch represents a significant shift in the robotics industry, moving from language-centric frameworks to video-firs...
Physical Intelligence has introduced a beta version of its Pi API, enabling robotics developers to access its foundation models for fine-tuning and inference. This platform allows developers to upload robot data and fine-tune models on company-managed compute resources, streamlining the adaptation o...
Neura Robotics has unveiled Neuron AI, a multimodal foundation model designed to enhance robots' understanding of the physical world. This model integrates vision, touch, audio, robot-state data, language, and memory, utilizing a Perception-Language-Action architecture to maintain context and track...
Physical Intelligence has introduced a beta API designed for training and running its robot models, marking a significant advancement in providing intelligence for various machines. Announced on October 13, this service allows teams to upload demonstrations, train policies, and test them on hardware...
On October 14, 2026, NEURA Robotics launched Neuron AI, a multimodal foundation model designed for Physical AI. This innovative model is trained using real-world robot data, simulation-enhanced data, and human demonstrations, allowing it to operate across five different NEURA robot embodiments, incl...
From October 12 to 16, 2026, the 26th China International Industry Fair took place at the National Exhibition and Convention Center in Shanghai. Elite Robots showcased its 'Yuanqi Primo' industrial-grade embodied large model at booth 8.1H-F008, emphasizing the transition of embodied intelligence fro...
Recent advancements in frontier AI models, particularly Astra and Opus, are significantly expediting the development processes in robotics. These models streamline labor-intensive tasks such as building simulated environments, creating tasks, and coding for testing. Demonstrations of Astra and Claud...
In 2026, China's humanoid robot shipments surpassed 40,000 units, capturing a 97% global market share. The Ministry of Industry and Information Technology forecasts that the total output for the year may exceed 100,000 units. However, a significant barrier remains in the form of data, as training th...
AMD has signed a definitive agreement to acquire World Labs in an all-stock transaction valued at approximately $8.2 billion, expected to close by the end of 2026. This acquisition marks a significant investment in the 4D world modeling sector, with World Labs founder Fei-Fei Li joining AMD as Execu...
On October 10, Luming Robotics announced a strategic partnership with ModelBest, focusing on data collection, model training, and scenario applications for embodied intelligence. This collaboration aims to advance embodied intelligence from technical validation to large-scale industrial deployment b...
Robotera's open-source VPP2 has surpassed RoboDojo's dual-arm benchmark by achieving a 58.5% zero-shot success rate on Real ALOHA Arms. This success was made possible by training the model to predict video sequences before executing robot actions, demonstrating a significant advancement in robotic c...
SafeWorld has emerged from stealth mode, securing $12.2 million in seed funding to develop its robot safety simulation platform. The technology aims to simplify safety testing for AI-powered robots across various embodiments, such as robotic arms, humanoids, and mobile robots. The significance of S...
SafeWorld has officially launched after securing $12.2 million in seed funding, co-led by Shine Capital and a16z Speedrun. The Palo Alto-based startup aims to create simulation software that tests the safety of autonomous robots prior to their deployment in real-world environments. This platform all...
Moxin Technology has introduced its MoWorld application, which transforms pet photos into animated 4D digital twins. This innovative feature is available on Huawei's Mate 90 smartphone, utilizing Ascend cloud generation technology alongside a sophisticated 600M-parameter model powered by Kirin. The...
Safeworld, founded by Dr. Ding Zhao and his team, aims to tackle the unpredictability of generative AI in robotics. With over $12 million in seed funding, the company focuses on creating safety standards for robots operating in unstructured environments. Their approach involves simulating real-world...
SafeWorld, an AI lab focused on robot safety simulation technologies, has officially launched after securing $12.2 million in seed funding. The funding was co-led by Shine Capital and a16z Speedrun, with participation from Box Group, Carnegie Mellon University Endowment, Innovation Endeavors, SV Ang...
Dalian University of Technology has introduced VA-Bench, a testing framework that evaluates 12 multimodal model setups on various robot-arm tasks. Among these models, Qwen3.8-max achieved the highest success rate, completing 53.93% of the tasks. However, none of the models managed to finish a strict...
SafeWorld has emerged from stealth mode with $12.2 million in seed funding, co-led by Shine Capital and a16z Speedrun, to develop robot safety simulation technologies. This funding will support the creation of a platform that provides safety testing and simulation software, enabling enterprises to e...
MindOn has launched the Mind-1, a new AI model aimed at enabling robots to execute everyday manipulation tasks at human-level speeds. Demonstrated on September 30, the Mind-1 showcased robots handling parcels, folding shirts, and collecting laundry, all marked as 'Fully Autonomous' and operating at...
AIsphere has launched PixVerse R2, a groundbreaking real-time world model capable of generating explorable audiovisual environments. This innovative platform takes various inputs, including text, images, audio, and keyboard commands, while maintaining state throughout a session. The introduction of...
Runway, a prominent developer of generative AI video technology, is venturing into robotics with the introduction of Praxis-1, an open-weight AI model. This model is designed to leverage knowledge gained from video to control physical robots, allowing developers to download and adapt it independentl...
DexRobot is presenting its advanced dexterous manipulation technologies at IROS 2026, highlighting the DexGEM platform and the DexHand021 Pro. The company is showcasing its capabilities through three accepted research papers, including one focused on dexterous robot manipulation from human demonstra...
Sales of humanoid robots are experiencing a significant surge, primarily driven by high-performance models, according to a recent report by Frost & Sullivan. The report, released on September 28, indicates a shift in the industry from technology validation and small-scale deliveries to large-scale p...
The world's first embodied intelligent robot, featuring fully domestically produced electronic architecture, made its debut in Yichang. This milestone represents a significant advancement in robotics technology, showcasing the capabilities of local manufacturing and innovation. The introduction of...
On September 18, 2026, independent testing organization Robocurve released a benchmark report titled RoboHarm. The report addresses a critical question regarding AI: can large language models effectively recognize and reject dangerous commands when controlling real robotic arms? The study tested thr...
Light Origins has launched Light-O1, its first general-purpose embodied foundation model, which utilizes human actions extracted from internet videos to provide a reusable framework for robots to learn various tasks. The company trained six versions of the 4-billion-parameter model on multimodal tok...
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
