In a humanoid surgical assistant robot demonstration hosted by Samsung Medical Center, the lead surgeon issued commands by voice only, and two humanoid robots, acting as an instrument-passing nurse and a camera-holding assistant respectively, collaborated to reproduce a cholecystectomy on a goat liver.
AgiBot has released AGILE 2.0, a second-generation perception-driven locomotion and whole-body control framework that brings visual perception into the motion control loop and introduces a unified whole-body motion representation, supporting manipulation while moving, real-time replanning, and cross-robot coordination. In the demo video Acrobat BOT, the AGIBOT X2 completes tasks such as ball balancing, three-robot jump rope, and collaborative carrying; the framework also supports risk perception and autonomous recovery from disturbances in human-robot shared environments.
PKU EPIC Lab proposes UniPart, a feed-forward cross-modal 3D Transformer conditioned on CLIP text embeddings that uses free-text phrases to select functional parts on point clouds, achieving zero-shot language-driven 3D part segmentation.
Why it matters: The paper presents a text-driven 3D part segmentation method and a self-built dataset, and validates transfer to real-world language-conditioned grasping.
PKU EPIC Lab has released FoldNet++, a large-scale synthetic dataset for robotic T-shirt folding and unfolding, covering 6 robotic embodiments, 1K T-shirts, 1K environmental assets, and 120K richly annotated demonstrations.
Why it matters: The source gives the dataset scale and the real-world deployment success rate, so readers can assess the feasibility of training garment manipulation policies purely on synthetic data.
MIT CSAIL, together with Google and Northeastern researchers, launched InstructMesh, which combines TRELLIS with GPT-4 to let users repair AI-generated 3D models using natural language before 3D printing them.
Why it matters: The research presents a new tool and user experiment data, letting readers learn how ordinary users can repair AI-generated 3D models and complete printing.
The Cell and BioRobotics Group (CBoRG lab) at the Technical University of Munich (TUM) Munich Institute of Robotics and Machine Intelligence (TUM MIRMI) proposes "Embodied Laboratory Intelligence," in which robotic assistants and virtual agents work together to explore the internal mechanisms of cancer cells.
NexArm ROS is an educational robotic arm designed for schools. It uses a dual-control architecture with an NVIDIA Jetson or Raspberry Pi host paired with an ESP32 controller, comes preinstalled with the OpenClaw agent, and supports voice and vision operations driven by large models such as Qwen and DeepSeek.
PKU EPIC Lab proposes GIF, an agentic generation framework for robot learning that reframes the problem as disentangled reconstruction followed by relative pose recovery. CoGen combines 2D and 3D generative models to produce instance-disentangled meshes and coarse initial poses; GPRM refines the relative poses under joint geometric and physical guidance; and a VLM verifier selects the candidate that best matches the structured specification.
Why it matters: The paper splits interactive and functional object composition generation into disentangled reconstruction and relative pose recovery, and provides a benchmark of eight contact-geometry classes along with collision-rate figures, which can help readers assess this route to simulation data generation.
At the 2026 XAIR Expo Embodied Intelligence Industry Expo, Zhido Future unveiled a technical preview of its differentiable world model "Zhizi World" (PGD World Model) and demonstrated validation progress on its embodied intelligent robot Xiaozhi S3. "Zhizi World" reconstructs particle-based scenes from short multi-view videos and can edit parameters such as object pose, material, mass, and friction to predict the physical outcomes of actions. S3 Xiaozhi features bionic dual arms, lidar-vision fusion navigation, and multimodal voice interaction; on-site demonstrations included scooping popcorn, filling cups and delivering them, and self-service retail shopping.
STIER has launched the PolishV series radial compliant spindle, featuring three main characteristics: high torque, precise force adjustment, and automatic tool change. The spindle delivers output torque of over 2.7 N.m and a floating torque of up to 24 N.m, providing 360° circumferential compliance and ±5° oscillating radial force, with floating accuracy of ±1 N, and supports automatic tool change in conjunction with a robot tool magazine. It has already been deployed in fields including liquid-cooling computing, chip manufacturing, aluminum alloy die casting, bearing manufacturing, medical devices, and 3C consumer electronics, and together with PolishX compliant force control and six-axis force sensors, forms its AI compliant grinding product system.
Stial has built a full-stack physical AI technology system spanning flexible force control, AI process decision-making, motion control, and robotic execution, focused on complex grinding, polishing, and deburring scenarios. Its PolishX flexible force control system can achieve contact force control accuracy of up to ±0.5% FS within the specified range and test conditions, and NextBrain™ AI Mushroom Cloud provides robot modeling, task planning, simulation, digital twin, and AI process decision-making capabilities. The related solutions have been applied in computing infrastructure, shipbuilding, auto parts, 3C electronics, and other fields, and have entered overseas markets such as Southeast Asia.
PKU EPIC Lab published the ZETA study on arXiv (arXiv:2609.02546), proposing a unified definition of zero-shot cross-embodiment VLA transfer and a controlled benchmark covering 14 held-out target embodiments, distinguishing strict zero-shot transfer from pretrain-exposed transfer. Experiments show that local end-effector state-action representations, source embodiment diversity, and auxiliary co-training improve transfer by about 15, 18, and 7 percentage points, respectively, and that adding only 5% target-embodiment data during pretraining raises average target-embodiment progress by 13.4 percentage points.
Why it matters: The paper provides a unified definition of zero-shot cross-embodiment transfer and a controlled benchmark, allowing readers to understand the actual benefits of different representation and data strategies.
Huan'er Technology details the complete ROS 2 agent pipeline of the ROSOrin Pro robot: voice commands are recognized by WonderEcho Pro, then passed to the OpenClaw gateway for understanding and broken down into subtasks such as navigation, detection, and grasping.
Stial has released an AI flexible wet grinding solution for repairing scratches and pits in titanium alloy plates. The process includes the SeeV AI perception system to identify defects, the NextBrain™ AI brain to plan grinding paths and process parameters online, and PolishX flexible force control to stabilize contact force, with wet cooling simultaneously carrying away grinding heat. Under the specified range and test conditions, the system's contact force control accuracy can reach ±0.5% FS. The solution transforms experience-driven grinding into a controllable process driven by data and algorithms, and the related technology has been applied in multiple high-end manufacturing fields such as shipbuilding, automotive parts, and 3C electronics.
ZINOVA, using LimX Dynamics' TRON 2 dual-arm embodied robot, partnered with RIC Robotics to complete a POC demo of a scaled concrete slab construction scenario, covering processes such as formwork frame assembly, and multi-layer rebar laying and tying. In the demo, two TRON 2 robots respectively held a nail gun and wooden boards to collaboratively assemble and secure the concrete formwork frame. TRON 2 supports human-machine collaborative teleoperation, with operators intervening at key points while accumulating data to lay the foundation for future autonomous capabilities and model training.
RoboScience has launched REX G1, a wheeled humanoid general-purpose manipulation robot equipped with its self-developed Visics general embodied large model built on the VLOA architecture, targeting scenarios including logistics, factories, retail, and homes. REX G1 has 22 degrees of freedom, can carry 10 kg with both arms, supports ±0.1 mm repeat positioning accuracy, is only 540 mm wide and can enter 0.75-meter narrow passages, and relies on 2070 TFLOPS of on-device compute to complete the perception, decision, and control loop.
Lingchu Intelligence announced the completion of a new funding round of over $100 million. Investors include Tuopu Group, Ruicheng Fund under Chery Holding, Lens Technology, 37 Interactive Entertainment, Wuhu Municipal Investment Holding Group, and Fosun Chuangfu, while existing shareholder Zhuhai Technology Industry Group made an excess additional investment. The proceeds will focus on embodied world model R&D and commercial deployment in logistics and advanced manufacturing scenarios. The company's self-developed dual-system architecture comprises the Psi-R2 manipulation policy model and the Psi-W0 action-conditioned world model, and it plans to expand the scale of human data collection and accelerate the layout of data collection facilities.
Why it matters: Lingchu Intelligence officially announced the completion of the funding and stated that the funds will go toward world model R&D and commercialization in logistics and manufacturing scenarios, giving readers insight into its industry-finance resource layout.
Philips announced it has received up to $33.7 million in funding from the U.S. ARPA-H Autonomous Interventional and Robotics (AIR) program to advance automation of AI robotic surgery for stroke treatment.
Foundation Robotics released a demo video of a robotic hand catching a baseball. The post only provides a link to the full video and does not disclose the hand's degrees of freedom, payload, or whether it uses teleoperation or autonomous control.
At the 2026 World Robot Conference, Wiz Robotics showcased two product lines: humanoid robots and intelligent industrial 3D cameras. The humanoid robot "Xiaowei" can respond to voice commands to grasp and deliver drinks, while the inspection robot is equipped with an AGV chassis and supports 360-degree rotation; the intelligent dexterous hand has 16 degrees of freedom, fingertip pressure of 12N, and a maximum lifting load of 15kg. The industrial 3D camera matrix covers interpupillary distances from 90mm to 620mm, including the "Starlight Eye" welding series and the "Smart Light Eye" grasping series, and on-site demonstrations included teaching-free corrugated board welding and 3D point cloud dynamic imaging.
At the 2026 World Robot Conference (WRC 2026) in Beijing, held August 19–23, Elite Robots presented three AI+robotics application scenarios: Lingju model random bin picking (CS66 collaborative robot + Lingju Zhizao end-to-end VLA model + SMC electric gripper), a dual-arm coffee robot co-exhibited with Heshisiwei (supporting 50+ flavors), and a fully automated ice cream robot co-exhibited with Beijing Rural Commercial Bank.
Tokyo Robotics demonstrated its dual-arm robot Torobo2 autonomously clearing a table (bussing): it collects dishes and a cloth from the tabletop into a tray, completing the task autonomously in real time (1x speed) throughout. Trajectories are generated by a fine-tuned VLA policy, with visual information from head and wrist cameras. The demonstration verified stable execution under drastic changes in ambient light, as well as the adaptability to re-recognize and continue completing the task after a human intervenes and moves target objects or the tray.
Tokyo Robotics demonstrated its dual-arm robot Torobo2 seamlessly transitioning from "Laundry Folding" to "Table Bussing" in a continuous operation, completed autonomously in real time at 1x speed throughout. Trajectories are generated using a fine-tuned VLA policy, relying on visual information from head and wrist cameras to judge environmental changes. Each task has its own individual control policy (action model), and the policies are switched at appropriate timing to achieve uninterrupted continuous execution between tasks.
The dual-arm robot Torobo2 demonstrates autonomous clothes folding, using visual information from head and wrist cameras to generate trajectories in real time via a fine-tuned VLA policy, executing the entire process autonomously. The first half of the video completes the folding task at 1x speed, while the second half (played at 1.5x speed) shows its flexible adaptation and correct folding of clothing with different shapes and properties, such as shorts, a towel, and a T-shirt.
At the Beijing World Robot Conference (WRC), held August 19–23, X Square Robot showcased three exhibition areas: home services, logistics sorting, and bodyless data collection. Its fully self-developed end-to-end unified world model WALL-B drives robots to complete dozens of household chores. Logistics sorting reaches an efficiency of 1,816 items per hour with 98% accuracy, at a cost 70% lower than the "humanoid robot + five-fingered dexterous hand" approach, and has already entered routine operation on real production lines. The QUANXTA Zero data collection solution achieves a data-ingestion validity rate above 85% and can reduce training-data costs by 90%.
In a StridingAI demo, a robot autonomously cleared a cluttered tabletop: it understood the semantics of each object and distinguished trash from reusable tableware. The task was completed through coordinated bimanual manipulation and object handovers, and is a long-horizon, multi-step operation. This content is a demo clip released by StridingAI; the original text did not provide a specific robot model or success rate metrics.
Skild AI released the robot foundation model S1, which uses video demonstrations as task prompts and can execute manipulation tasks not seen during training without fine-tuning, with long-horizon tasks lasting up to about 10 minutes. At a pretraining data scale of 100,000 hours, the ICL model achieved a 66% success rate on unseen tasks, compared with 9% for the language-prompted VLA; a single demonstration was equivalent to about 380 post-training demonstrations, and the planting task took 11 minutes from recording the demonstration to autonomous execution. S1 was trained on NVIDIA AI infrastructure and has already been put into use with commercial partners.
Why it matters: The original article provides comparison figures for ICL and VLA at the same data scale and a deployment timeline, allowing readers to assess the feasibility of the in-context learning approach for manipulation tasks.
In August 2026, the 2026 Embodied Intelligence Summer Special Training and Study Camp, hosted by Zhejiang Lingqiao Intelligent Technology Co., Ltd. and co-hosted by Shanghai Artificial Intelligence Research Institute and Zongshen Chuangke (Shanghai) Technology Co., Ltd., concluded. Over the 5-day, 40-class-hour program, all trainees from more than ten institutions and dozens of enterprises and public institutions passed the assessment, with an overall satisfaction score of 96.8 points.
The embodied dual-arm loading and unloading robot jointly developed by Elite and Maixing Robotics has achieved 24/7 unmanned mass production in the drilling workshop of a leading Apple supply chain factory, with a single-cycle takt time of 52 seconds, a loading and unloading yield of 99.9%, and 5 units stably online. The solution uses hybrid force-position control and dynamic trajectory planning to achieve dual-arm coordination, is equipped with the self-developed Yuanqi Primo embodied large model, achieves loading accuracy within ±0.2 mm, and uses a wireless IO four-step handshake scheme to eliminate the need for modifications by the original machine tool manufacturer, shortening the deployment period from 10 days for the first unit to 7 days for the fifth unit.
Why it matters: The original text provides actual performance figures such as line takt time, yield, and deployment period, allowing readers to understand the implementation path of the embodied dual-arm solution in a mass-production workshop.
In a StridingAI demo, the STEAM model from Professor Chao Yu's team at Tsinghua University drives a robot to autonomously complete a multi-stage unboxing task, achieving millimeter-level bimanual coordination through tool use. STEAM models task progress and action quality frame by frame, extracting learning signals from mixed-quality data including expert demonstrations, model rollouts, and human takeovers to improve task success rates and execution speed.
In its home breakfast service scenario, UniX AI's Panther humanoid robot autonomously plans and completes the full workflow, from preparing ingredients and heating food to making drinks and serving the meal at the table. UniCortex handles long-horizon task planning, UniFlex supports rapid learning of new household skills from a few demonstrations, and UniTouch combines visual and tactile perception to improve manipulation reliability for food, kitchen tools, and everyday objects. More information at www.unix-group.ai.
UniX AI released a continuous, uncut video in which the humanoid robot Wanda autonomously completes a full set of service tasks in a real hotel environment, including room organization, table and bathroom cleaning, bed making, and everyday item handling, demonstrating smooth and stable autonomous operation.
The Hongxing Technology Xiaomei M9 humanoid service robot focuses on continuous service off-grid and targets elderly care, medical, government, and commercial service scenarios. Its perception, modeling, navigation, and decision-making are 100% completed on local edge computing, enabling it to keep working for more than 6 hours under disconnected or weak network conditions, with zero data leakage. Its millimeter-wave radar fall prediction has an accuracy rate of 99% and a false alarm rate below 1%, and its robotic arm completes the entire process from perception to intervention in just 10 seconds.
At its WAIC 2026 booth, Keenon Robotics demonstrated autonomous humanoid robot operation across four scenarios: coffee making, hotel laundry, a dessert shop, and retail pickup, all without teleoperation. The XMAN-R1 independently picked up a cup, made coffee, and delivered it to a latte art machine for printing.
Stial focuses on the vertical scenario of complex grinding, building a closed-loop physical AI technology around "perception—analysis—decision—execution," and recently received the "2025 Science and Technology Innovation Leading Enterprise" honor from Ningbo Cixi. Its AI flexible grinding solution has been applied in fields such as aerospace, new energy vehicles, chips and semiconductors, and energy storage, in cooperation with a number of leading enterprises. The company is also one of the drafting units of the group standards "Technical Requirements for Industrial Grinding Robots" and "Technical Requirements for Multimodal Perception Fusion of Humanoid Robots," and operates the AI brain NextBrain™ Mogu Cloud system.
FedEx and Dexterity announced an expansion of their partnership to deploy Dexterity's Foresight world model and Mech dual-arm trailer-loading system at scale at the Hagerstown hub in Maryland, after the technology completed years of joint development, testing, and operational validation at pilot sites.
Why it matters: The original article provides details on the production-scale deployment from pilot to a large hub, helping readers understand how physical AI is actually being implemented in logistics loading and unloading scenarios.
Stial recently won a project to upgrade the intelligent flexible polishing production line of a leading domestic precision metal composite materials company, with an order value in the millions of yuan. The solution deploys multiple AI flexible polishing workstations, uses the NextBrain™ Mogu Cloud system for multi-task intelligent decision-making, and is equipped with the self-developed PolishX flexible force control system at the end effector. Through hybrid force/position control and real-time gravity compensation, it adapts to irregular curved surface profiles and dynamically adjusts contact force, serving the AI computing center's liquid cooling pipelines to improve mass-production quality and efficiency.
Hongxing Technology showcased self-developed robots including Dazhuang Z6 and Xiaomei M9 at IME 2026 in Bangkok, Thailand, winning intent for cooperation from many overseas customers. The Dazhuang Z6 industrial humanoid robot is equipped with an edge computing engine; when offline, it processes 100% of data locally and can continue working for 10 hours, with repeat positioning accuracy of ±0.05mm. Xiaomei M9 is equipped with the self-developed local R1 large model and features multimodal human-like interaction and scenario-adaptive capabilities.
During the 2026 World AI Conference, Fudan University's Institute of Intelligent Robotics and Advanced Manufacturing Innovation presented achievements in embodied intelligence, physical intelligence, and related areas. Professor Su Hao delivered a keynote speech on physical intelligence at the main forum; Sudo Technology made its first public appearance in China and demonstrated a four-robot collaborative battery module assembly production line; and Zhang Lihua's team's intelligent inspection robot dog appeared on a CCTV program.
At 2026 WAIC, JAKA Robotics' full-stack, in-house developed embodied intelligence products were showcased together. The JAKA K1-25 humanoid robot weighs under 25 kg, has a maximum payload of up to 30 kg, and features 1 kHz force-control response and ±0.1 mm repeat positioning accuracy; the new energy motor collaborative assembly system jointly built by JAKA Kargo and Hasco Automotive has a recognition accuracy of 99% and can complete 0.5 mm fine-thread screw fastening in 10 seconds.