ROBROS successfully demonstrated Dynamic Whole Body Motion with its humanoid robot IGRIS-C, including sit-to-stand, windmills, backflips, and kicks. The demonstration showcased dynamic balance and real-time whole-body coordination control amid rapidly shifting centers of mass. ROBROS calls it a step toward humanoid robots that can move, adapt, and work in real industrial environments.
On July 25, JAKA Robotics' self-developed JAKA π (π Zai) lightweight legged humanoid robot appeared at the Changzhou home match of the Suzhou Super League, where multiple robots danced in coordinated formation, hitting precise beats to the music. The robot is 1.2 meters tall, weighs 42 kg, and has 27 degrees-of-freedom joints. It uses a fused "brain + cerebellum" architecture, combining 1 ms EtherCAT high-speed servo communication with kilohertz-level joint feedback to achieve stable motion output. The JAKA Kargo wheeled humanoid robot project, built on the same underlying technology, was selected as a typical case among Jiangsu Province's top ten application scenarios for embodied intelligent robots.
A Carnegie Mellon University team brought 3 snake robots to Venezuela to search the rubble of collapsed buildings after the June 24 double earthquakes, working for three days from June 30 to July 3 without finding survivors. Each snake robot is 4 feet long and uses replaceable modular segments; operators control them with a gamepad and a tethered laptop, while software algorithms autonomously coordinate the body segments to move from point A to point B. The team previously deployed similar robots in the 2017 Mexico City earthquake.
The key to the NexArm robotic arm's smooth start and stop at a 500mm reach is its high-precision magnetic-encoder bus servos. The 12-bit magnetic angle sensors in the core joints provide 360° absolute angle detection and 4096-step position resolution, and their contactless measurement avoids wear.
The robot IGRIS-C completed its first public outing in Seongsu-dong, Seoul, performing dynamic moves such as backflips and windmills on site, then walked back to the office on its own after interacting and taking photos with the crowd. This is the first day out for IGRIS-C as documented by ROBROS, and the full version will be released next week.
Nock, aim, fire. 🏹 A bullseye at 20 meters. Watch PNDbotics Adam conquer the outdoors—from skipping stones to horseback archery, demonstrating extreme dynamic balance.
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.
Booster Robotics' humanoid robot ran the 21.0975 km Yizhuang Half Marathon in 1 hour 53 minutes, performing well in the small and medium-sized humanoid robot category. The course included ramps, sharp turns, and asphalt surfaces; a support tram losing power mid-race slowed the team, but the robot completed the race stably throughout. The team previously won a RoboCup championship and two titles at the German Open, and before the race the robot underwent a 48-hour continuous operation stress test.
Boston Dynamics' humanoid robot Atlas handed the match ball to the referee and performed a series of soccer celebration moves during halftime of the 2026 FIFA World Cup round-of-16 match between Norway and Brazil on July 5. This was the first public appearance of the production version of Atlas and the first robot halftime performance in World Cup history. The behavior is based on reinforcement-learning whole-body control training and was tested in both the lab and real environments; it was a live demonstration rather than an on-site autonomous operational deployment. The event is related to Hyundai's "Next Starts Now" project.
LimX Dynamics released V³-0 alongside COSA 0.5, a humanoid VLA technology stack in its VLA-tools layer that supports mobile whole-body manipulation, running on the 43-DOF LimX humanoid robot Oli.
Why it matters: The original text explains the architectural composition of V³-0 and its human-in-the-loop real-robot reinforcement learning loop, which lets readers understand where the mobile whole-body manipulation capability comes from.
RoboParty announced the establishment of RoboParty Lab and open-sourced three humanoid robot toolchains: the supervised learning training framework MimicLite, the unsupervised reinforcement learning control framework UFO, and the motion retargeting tool hhtools.
Why it matters: The three toolchains provide specific figures on training compute, time required, and the policies already integrated, which readers can use to assess their reuse value in their own R&D workflows.
Figure AI announced that the Figure 03 humanoid robot has entered Hall 52 at BMW Group Plant Spartanburg, shifting from the sheet metal loading and unloading work previously done by Figure 02 to a sequencing sorting use case; Figure 02 took part in assembling 30,000 vehicles last year.
Why it matters: The company has released details of Figure 03's deployment scenario at the BMW plant and of its Helix 02 drive, so readers can learn about progress in bringing humanoid robots from material handling to complex sorting.
STIER uses AI flexible grinding technology to address the pain points of easy scratching and insufficient consistency in weld seam grinding for battery box enclosures. Its solution uses PolishX flexible force control (accuracy ±0.1N) and millimeter-level visual positioning perception, combined with a process grinding knowledge base and the NextBrain™ AI cloud system to make grinding decisions, and has self-learning and adaptive capabilities. It also provides full-process support including customized laser scanning solutions, real-time data monitoring, and grinding head wear warnings.
STIER has launched an AI flexible grinding robot workstation that builds a "perception-analysis-decision-execution" physical AI closed loop, targeting the grinding of multi-curved-surface parts in high-end manufacturing such as aerospace and precision molds. The solution uses 3D vision inspection and positioning, a PolishX flexible force-control system belt grinder, and the self-developed NextBrain™ Mushroom Cloud system. In a wet grinding case involving silicon carbide composite workpieces, it achieved fully autonomous operation across the entire process, with contact force control accuracy of ±0.5% FS, per-piece processing time shortened by more than 22%, and product changeover completed within 10 minutes.
The world's first full-size humanoid robot to successfully climb a 1-meter-high box! Climbing is the ultimate test of full-body multi-joint coordination and core strength. PNDbotics' full-size Adam humanoid robot demonstrated the hardcore essence of a true "future warrior."
Professor Zhu Yanhe's team at Harbin Institute of Technology (HIT) published a paper in the International Journal of Extreme Manufacturing (IJEM), proposing an autonomous fault-tolerant repair strategy for modular robots based on cooperative reconfiguration, which restores joint fault damage by changing morphology without backup module replacement. The method combines topological description and kinematic self-modeling, particle swarm optimization to search for target configurations, and an improved RRT cooperative reconfiguration algorithm. In experiments, the fault-tolerance upper limit was 7 times higher than that of fixed-configuration redundancy methods. HIT is the sole completing institution, and doctoral student Qi Jian is the first author.
Why it matters: The paper presents a cooperative reconfiguration fault-tolerance strategy that requires no backup modules and data showing a 7-fold increase in the fault-tolerance upper limit, allowing readers to understand a concrete implementation path for self-repair in robot swarms under extreme environments.
CCTV News | 🤖 2026 WHRGF "Energy Relay": PNDbotics. Full-size humanoid robots take on traditional skills in an extreme test of force control and dynamic balance. 🔥
Galbot released AstraBrain-WBC 0.5 (paper titled Humanoid-GPT, accepted to CVPR 2026), training a Transformer whole-body controller on about 2 billion frames of retargeted motion data, achieving a 92.58% zero-shot motion tracking success rate with 0.39 ms inference latency; code and data are open-sourced.
Why it matters: This analysis clearly explains the data scale, architecture choices, and limitations of Galbot's new model, letting readers assess progress on the general motion control route.
In a payload demo, Duatic Alpha lifted 22.4 kg at full 1-meter extension, held it steady for 2 seconds, then safely set it back down, repeating this 10 times consecutively. The test was used to verify the platform's lifting capability, stability, and operational reliability, showcasing its mechanical design and motor control precision; the source did not provide information on speed, battery life, or mass-production delivery.
DEEP Robotics' small wheeled-legged robot, the Lynx S10 prototype, completed a 76-day, 11,852-nautical-mile Arctic Ocean research test aboard Sun Yat-sen University's Polar Research Vessel Zhongshan, marking the first time a quadruped robot has set foot on Arctic Ocean ice. The unit, weighing under 20 kg including its battery, was fitted with anti-slip bionic feet mimicking polar bear paws in collaboration with Sun Yat-sen University, Westlake University, and Hangzhou Dianzi University, and completed crawling and motion-control tests with crampons; the body uses an IP67 integrated sealed design, and after increasing the surface area of its limbs for paddling, it completed swimming tests in ice water.
Why it matters: The original article provides the specific test conditions and hardware modifications for the Lynx S10's walking on Arctic Ocean ice and operation in ice water, helping readers understand how quadruped robots are adapted for polar scenarios.
Enchanted Tools' humanoid robot Mirokaï demonstrates multi-terrain navigation in a video, seamlessly switching between surfaces such as carpet, wood flooring, tile and concrete while continuously adjusting its behavior to maintain stability and smooth movement. This is a demo; the company teases further news on June 16, 2026, and invites people to join its Discord community.
PKU EPIC Lab released LIMMT (Less Is More for Motion Tracking), which it describes as the first data-centric study for physics-based simulated humanoid motion tracking. The method defines motion data quality along three dimensions: physics feasibility, diversity, and complexity. Training on less than 3% of AMASS achieves better tracking performance than training on the full dataset, and the team also cleaned web-sourced mocap data. The paper has been accepted to ICML 2026; see arXiv:2606.06953.
Why it matters: The paper presents a three-dimensional data quality framework and results from training on a small amount of data, offering a methodological reference for researchers working on humanoid motion tracking.
Boston Dynamics has partnered with Hyundai to train the Atlas® humanoid robot, using human demonstration references and reinforcement learning, to perform a Ghost Rabona kick it had never done before, including a feint step and a cross-legged shot.
STIER has built an integrated AI flexible grinding solution covering perception, analysis, decision-making, and execution, targeting the difficult grinding and polishing of special-shaped bent pipes such as liquid-cooling bent pipes and exhaust bent pipes. Its self-developed SeeV perception system combines 3D imaging to identify contours and deviations; the NextBrain™ AI Mushroom Cloud system plans trajectories and enables one-click changeover; and the six-axis AI flexible grinding robot is equipped with the PolishX flexible force control system, achieving contact force control accuracy of ±0.5% FS and supporting small-batch, multi-model customized production.
Enchanted Tools demonstrates in a video how the Mirokaï robot moves naturally in dynamic environments, relying on trajectory visualization, smooth navigation, and model predictive control (MPC) to continuously adjust its movements. The video targets hospitals, nursing homes, public spaces, and workplaces, emphasizing that movement is not just about avoiding obstacles but about making the people around it feel comfortable, able to predict its motion, and at ease intuitively. It ends with "WE'RE READY. ARE YOU?" and the date June 16, 2026.
ROBOTIS has integrated NVIDIA's Kimodo, a text-to-motion diffusion model, into the open humanoid robot AI Sapiens, allowing natural language prompts to directly generate full-body humanoid motions that are executed on real hardware. The motions are driven by DYNAMIXEL Q (QDD) actuators, with no motion capture suit and no frame-by-frame programming required. The full development source code will be released as open source, and the full video is available at https://youtu.be/2xAJzQ8V9p8 .
The CENTAURO project, in which PAL Robotics participates, redesigned the 7-DOF robotic arm for the TIAGo Pro mobile manipulator, optimizing the workspace to reach angles and positions that were previously unreachable. The arm integrates series elastic actuators (SEA) that can absorb impacts and measure contact force, and shifts to torque and admittance control, enabling the robot to sense applied pressure. The WebGUI v3 ecosystem adds a drag-and-drop visual programming interface, allowing non-experts to configure complex tasks in scenarios such as polishing and logistics.
LimX Dynamics has released LimX Luna, a full-size interactive humanoid robot for commercial services. It stands 160 cm tall and has 27 body degrees of freedom (excluding end effectors). Global preorders open today, with a domestic retail price of 298,000 yuan and a limited-time 40,000 yuan discount on the first 100 units.
On April 23, Beyond Imagination brought SMASH, billed as the world's first outdoor autonomous humanoid ping pong robot system, to the Shandong City Table Tennis Super League (Zibo stop). World champion Chen Meng, CCTV host Ju Ping, commentator Gao Han, actor Zhang Xiaolong and other guests took the stage to play against the robot live, producing real rallies. This was a live demonstration and interactive challenge; the original article did not provide metrics such as the autonomous execution success rate.
Noble Machines published an article introducing its three-layer robot learning stack for industrial autonomy, along with on-site test results from GTC 2026: over four days, the same model simultaneously drove multiple Moby robots to complete two autonomous tasks at the NVIDIA booth and its own booth.
Figure AI demonstrated two humanoid robots equipped with Helix-02 collaborating to tidy a bedroom within two minutes, completing tasks such as opening a door, hanging clothes, putting away headphones, closing a book, taking out trash, pushing a chair under a desk, and making a bed together. The two robots run the same learned vision-language-action (VLA) policy, with no shared planner, message passing, or central coordinator; each infers the other's intent from its actions via its own cameras. The company says that to its knowledge this is the first demonstration of multi-humanoid collaborative mobile manipulation completed by a single learned neural network directly from pixels to actions.
Why it matters: The official demonstration shows two humanoid robots driven by the same VLA system collaborating to tidy a bedroom, giving readers insight into an approach to multi-robot collaboration without a shared planner.
The Kepler K2 humanoid robot is being marketed in a "fitness beast" mode for extreme endurance, strength, and precision control, with claims that it can easily handle a 30KG payload, deliver continuous output on a single charge, and show no performance degradation. The source does not provide specific figures such as battery life, degrees of freedom, test conditions, or success rates; the performance claims are the company's marketing statements and are not labeled as verified results.
Team Equilibrio, this year's Focus project at ETH Zurich's Robotics Systems Lab, has built Wheeliam, a single-wheel robot with highly dynamic legs and a single-wheel design that can perform agile movement and jumps.
Researchers at EPFL published a method in Science Robotics that lets multiple robots with different designs complete the same task from a single human demonstration. The method does not rely on AI; instead, it analyzes the physical properties of three-revolute-joint robotic arms, marks safe regions within each robot's range of motion, and classifies three-joint robots into six categories based on shared physical constraints, writing those constraints into each robot's program. The authors call this kinematic intelligence.
A robot from Chinese smartphone maker Honor surpassed the men's half-marathon record on April 19, and robot maker Unitree announced that its bipedal H1 reached 10.1 meters per second, while the average speed required for Bolt's men's 100-meter record is 10.44 meters per second.
An EPFL research team published the Kinematic Intelligence framework in Science Robotics, enabling robots to safely transfer skills to robots with different configurations from a single human demonstration. The paper is available at http://dx.doi.org/10.1126/scirobotics.aea1995.
Sony AI's autonomous table tennis robot Ace competed under the rules of Japan's professional table tennis league, defeating multiple elite and professional players; the research was published in Nature (DOI: 10.1038/s41586-026-10338-5).
Why it matters: The article explains Ace's competitive record and three key technologies, and places it alongside the milestones of Deep Blue and AlphaGo, helping readers judge where physical sports competition currently stands.
PNDbotics Adam robot: stair climbing
Tokyo Robotics demonstrated human-like walking, whole-body teleoperation, and motion tracking on its custom-built Torobo humanoid. The control policies are trained through large-scale parallel reinforcement learning, and policies learned in physics simulation, once deployed to the real hardware, achieved dynamically stable whole-body motion. The company says it will continue improving motion stability in both software and hardware, and plans to introduce AI for autonomous task execution.
Noble Machines and Schaeffler have partnered to bring high-fidelity digital twins of high-performance actuators into the open-source physics engine Newton, for developing whole-body control policies before hardware deployment. The models can capture manufacturing variations and predict aging and energy consumption, with energy consumption predictions up to 10 times higher than idealized models, and feedforward compensation can reduce tracking error by up to 68%. Robots can also use control signals as identification signals to enable online self-diagnosis and compensate in real time for wear or environmental drift.