CyberRobo

Humanoid Robots As intelligent automation systems, they are embedded within existing upstream and downstream processes, complementing and enhancing the capabilities of full-chain automation, such as production lines and warehouses. As auxiliary assistants, collaborating with humans and participating in human-led work and life, humanoid robots are learning and evolving from their masters, evolving from passive to active. As remote agents: Humans control robots through AR/VR and brain-computer interfaces, while humanoid robots act on their behalf, handling extreme tasks and meeting the human's needs. As complete agents: Humans empower the robots(AI), understanding their intentions, patterns, and preferences, autonomously planning and executing tasks, and establishing a symbiotic relationship with them. As human-machine symbionts: Humanoid robots' hands, legs, and other parts of the human body enhance the human being. As a replica of the human, embedding human memories, knowledge, experience, ego, and superego,a digital copy,within the human body,a kind of superhuman and real-world immortality.

@Ancient_Geo unclear,it has not yet been verified in reality.

@CyberRobooo Have they been demonstrated to operate without immediate failure in high radiation environments?

Let’s take a look at a Chinese manufacturer’s solution. Shanghai-based RunRobotics recently unveiled a centaur humanoid robot. 100–120 kg payload and peak torque of 830 N·m, it features tactile-sensing dexterous hands, end-to-end environmental perception, and an explosion-proof design. wheeled-legged hybrid configuration combines the speed of wheels with the strong terrain adaptability of legs. It’s clearly designed for demanding special environments such as steelmaking, nuclear facilities, oilfield inspections, and firefighting/emergency response.

@nillimtz They did not disclose exactly how it was implemented. and here is the tech report for their VLA model.robotics.xiaomi.com/xiaomi-robotic…

@CyberRobooo the videos are impressive, I wonder if they are also working on an orchestration algorithm to sync multiple robots and achieve more complex tasks

Xiaomi Humanoid Robot New Factory Training Since Q1, Xiaomi’s humanoid robots have been training on dual-side operations at the self-tapping screw workstation in its auto factory, achieving a 98% success rate. Today, Xiaomi shared that the CyberOne Gen2 autonomously performed two repetitive tasks in the final assembly logistics area: sorting center console side panels and folding material boxes --both reaching 90% success rates. They deployed both bipedal and wheeled models, working together with material transport robots as part of the broader intelligent automation system. I noticed the wheeled CyberOne was clearly more efficient and stable on the panel sorting task. In the first video, you can also see some noticeable jitter when the bipedal robot stops moving (though this should be relatively easy to fix). Xiaomi also candidly pointed out a current limitation in dexterity: when folding the second side of the material box, the robot still needs to adjust its orientation first to face the buckle, whereas skilled workers can pull the ring directly by feel without any micro-adjustments. Real-world scenarios are indeed the best benchmark.

@mikekalilmfg Pure editing

@CyberRobooo i can’t take it seriously with the edits

Crazy… Air humanoid Jordan Magic Lab has released a demo: The full-sized humanoid robot MagicBot X1 performs a flying slam dunk🤾‍♂️ Another kind of aesthetic of violence…

blog->anthropic.com/research/claud…

Anthropic is testing humanoid robots. Its latest research, Claude Plays Robotics shows Claude controlling real robots, including the Unitree G1 humanoid and the Unitree Go2 quadruped, across interfaces from low-level torque control to high-level VLA supervision. The results: >Go2 achieves stable balancing, walking, and basic navigation with pretrained policies. >G1 remains much harder, but balance has improved significantly despite the challenges of low-level control. >Newer Claude models (Opus 4.6/4.7, Mythos Preview) shine at high-level tasks like grasping and object manipulation. The gap in low-level control is narrowing fast. --- This builds on Project Fetch Phase 2 and signals that a safety-focused company like Anthropic actively benchmarking on real hardware signals that embodied AI is now core to their roadmap Meanwhile, nearly every major LLM player is moving into embodied AI: Google (Gemini robotics) OpenAI (humanoid robotics lab) Alibaba (Qwen-Robot Series) Mistral AI (Robostral Navigate-VLA) ByteDance (Seed-GR3/GR-RL,ByteDexter hand) … The era of large-model companies collectively racing toward embodied AI has truly arrived.

@PTrubey yeah. it even needs to be faster.

@CyberRobooo Nice to see. Still work to do, since 90% success isn’t good enough in manufacturing unless you have mitigation steps. This is the kind of useful work I expect to see all humanoids to start with.

@noah_schochet @kleinerperkins @NoahMcGuinn @SpaceX Infrastructure is the cornerstone of civilization. Humanoid robots are playing a part in this as well.

Today, we’re announcing that we’ve raised $115 million in funding, including a $100M Series A led by @kleinerperkins. America has lost the ability to build, and we’re here to restore it. My co-founder, @NoahMcGuinn, and I left our jobs at @SpaceX , where we worked on programs including Starship, Starshield, and @Starlink, to build a company that will solve construction’s greatest challenges. Infrastructure is the foundation of civilization, and construction is the precursor to innovation. If America wants to build a brighter future for the next generation, we have to make it faster, cheaper, and safer to build. That’s where @TerraFirma_Inc comes in. We’re a new type of company, a robotic construction company that builds the full technology stack needed to deliver an order-of-magnitude improvement in one of the world’s oldest, largest, most important, but least efficient industries. We are building technology that expands what’s possible in construction on Earth, and then we'll use that same technology to build megastructures and colonies on the Moon and Mars. We’ve made tremendous progress over the past year, growing the company more than 10x in the last 12 months. We are performing projects across the world. By the end of October 2026, we are on track to operate 3 of the top 3 largest robotic construction fleets in the world, each on a different continent, bringing unprecedented speed, scale, and efficiency to some of the world’s most complex critical infrastructure projects. This funding will allow us to step on the gas and scale our manufacturing, software, operations, and construction deployments, including work on massive commercial and government contracts. We’re building the future of construction right here in Austin, Texas, and scaling it globally. If you want to be part of the team changing the world, now and on Mars, join us. Our Series A was led by Kleiner Perkins, with participation from Bain Capital Ventures, Glade Brook Capital Partners, BANNER VC, Saga Ventures, Trust Ventures, Definition, PEAK6, Magnetar Capital, and Ravelin Capital. Huge thanks to all of our angel investors, friends, and family who have helped and supported us throughout this journey. Apply here: jobs.ashbyhq.com/TerraFirma-Inc…

>full video&home：robotics.xiaomi.com/xiaomi-robotic… >tech report:arxiv.org/pdf/2607.11643 >HF:huggingface.co/collections/Xi…

Xiaomi Robotics has introduced Xiaomi-Robotics-U0, a 38B-parameter autoregressive world foundation model. Generative models like this are increasingly becoming “data engines” for robots. U0 brings text-to-image generation, image editing, multi-view embodied scene generation, scene transfer, and embodied video generation into a single model. A single real-world robot trajectory can be expanded into variations with different backgrounds, lighting conditions, and object configurations, while preserving the original actions and task logic. This matters because real-world data can never cover every possible scenario. A robot may learn to pick up a cup in one kitchen, yet struggle when the background changes or the lighting shifts. According to the tech report, post-training with U0-generated data improved the robot’s average task success rate in unseen backgrounds and challenging lighting conditions from 36.9% to 63.2%. The value of world models may go beyond predicting the world. They can also generate new worlds for robots to train in.

Xiaomi quietly launched an autoregressive world model today called Xiaomi-Robotics-U0. One of its key capabilities is Real Robot Improvement: style-transferred demonstrations generated by U0 are combined with real-world demonstrations during policy post-training, allowing robot policies to generalize better to unseen backgrounds, changing lighting conditions, and other visual disturbances. robotics.xiaomi.com/xiaomi-robotic…

In the future, everyone might have their own humanoid replica… A way to continue their life, or to become the persona they wish to be.

@Scobleizer You will also witness the shift towards physical AI and robot.

I have watched every platform shift since the PC. This one does not ship as software. It gets hired. Meet the AI employee that 40,000+ teams already hired. He does not wait for prompts. He owns the outcome.

@bowtiedwhitebat “Almost any movable object can serve as a weapon.”

Italy's IIT really has so many humanoid gems. Came across their 2018 Centauro--a centaur-like disaster response robot that combines a quadruped lower body with a humanoid torso and arms. Built for harsh settings like rubble fields or disaster zones, it needs both stable movement over uneven ground and the strength to use human tools for forceful tasks such as breaching doors or clearing obstacles. The four legs deliver terrain-following locomotion and stair climbing, while the upper body handles dexterous, high-force manipulation. The robot blends full-body teleoperation--mapping an operator’s movements directly to the machine, with local autonomous modules for motion and impedance control, supporting supervised or higher-level autonomy when useful. At 1.5m tall and 93kg, it carries roughly 11kg payload per arm with impact-resilient actuators. Battery life is about 2.5 hours. Its hybrid wheeled-legged legs (6 DoF each, with wheels at the ankles acting like rolling hooves) provide better stability than pure bipeds while keeping tool use intuitive. This focus on matching robot morphology to real task demands, rather than defaulting to a standard humanoid form, remains a useful reference. Curious what others think:could hybrid centaur-style platforms play a larger role in future rescue or specialized operations?

Holiday Robotics (Korea) raised $105M Series A. Wheeled humanoid FRIDAY targets factory dexterity. Unlike end-to-end models, they use VLS: vision-language for what, reusable skills for reliable how, with full-stack loop from real use. In PoCs for manufacturing/auto/…

@USRoboRecruiter Gamified design, of course, does not aim to wash dishes.

@CyberRobooo I've not played yet, waiting to get home today, but I'm 1000% more likely to play a robot fighting sim over a dish washing or factory working sim

What if we expanded REK’s humanoid fighting simulator to real-life tasks? Build simulators(Simulated World)for housework, factory work, elder care, and multi-robot teamwork--all running on the same AI and physics as actual robots. Then anyone with a PC, controller, or headset could train humanoid robots just by playing games… Training humanoid robots through entertainment. I think someone is definitely designing this game.