• Mechonomics
  • Posts
  • China launches world's first robot that can run by itself 24/7 — watch it change its own batteries in unsettling new footage

China launches world's first robot that can run by itself 24/7 — watch it change its own batteries in unsettling new footage

+ Humanoid robot performs medical procedures via remote control

Author

Maddy Wong
July 25, 2025

🧠 Weekly Brief

China launches world's first robot that can run by itself 24/7 — watch it change its own batteries in unsettling new footage

China's UBTECH unveiled Walker S2, the world's first humanoid robot capable of autonomous battery swapping for 24/7 operation. The 5'3", 95-pound robot uses a dual-battery system, autonomously approaching charging stations to swap depleted batteries for fresh ones without human intervention, enabling continuous operation in industrial settings.

3 Key Takeaways:

  • Autonomous operation breakthrough: Walker S2's ability to independently swap batteries eliminates the fundamental limitation that has restricted humanoid robots to 2-4 hour operational windows, potentially enabling true 24/7 deployment in factories and customer service environments without human maintenance interruption

  • Intelligent power management: The robot doesn't just mechanically swap batteries—it intelligently decides whether to charge or swap based on remaining power levels and task priorities, demonstrating sophisticated decision-making capabilities that go beyond simple automation

  • Commercial viability milestone: By solving the uptime problem that has plagued industrial robotics, Walker S2 addresses a critical barrier to humanoid adoption, as continuous operation is essential for justifying the investment in manufacturing and customer service applications where downtime equals lost productivity

This represents a fundamental shift from human-dependent to truly autonomous robotics, potentially accelerating commercial deployment across industries where 24/7 operation is essential.

Humanoid robot performs medical procedures via remote control

UC San Diego researchers equipped a Unitree G1 humanoid robot with Inspire Gen4 hands and bimanual teleoperation system to perform seven medical procedures via remote control. Using foot pedals, HTC Vive trackers, and motion-capture cameras, the robot achieved 70% success rates in ultrasound-guided injections and consistent emergency ventilation.

3 Key Takeaways:

  • First clinical workflow integration: This marks the inaugural study adding humanoid robots to hospital work systems, moving beyond laboratory testing to actual medical procedure evaluation—a critical milestone as healthcare systems face unprecedented staffing shortages and patient overflow

  • Remote expertise multiplication: The teleoperation system enables non-clinicians to achieve 70% success rates in precision procedures like ultrasound-guided injections, potentially allowing expert medical knowledge to be transmitted remotely to underserved areas where specialist access is limited

  • Hybrid human-robot approach: Rather than replacing doctors, the system augments medical staff by handling routine and emergency tasks that currently burden overworked healthcare providers, addressing the root cause of physician burnout while maintaining human oversight for critical decisions

This research addresses the practical reality that healthcare systems need force multipliers, not replacements, as aging populations and labor shortages create unsustainable pressure on medical professionals worldwide.

Dobot unveils six-legged robotic dog suitable for multiple scenarios

Shenzhen-based Dobot Robotics launched the world's first six-legged biomimetic robotic dog, completing an embodied AI platform integrating robotic arms, humanoids, and six-legged robots. The inherently stable triangular support system can handle forces five times its body weight while operating quietly across uneven terrain without speed reduction.

3 Key Takeaways:

  • Third path robotics strategy: Dobot positions six-legged robots as the practical middle ground between humanoid versatility and industrial specialization, offering "greater commercial viability than humanoid robots while delivering broader environmental adaptability than traditional, task-specific machines"—potentially defining a new robotics category

  • Superior performance metrics: The six-legged design outperforms competitors across key dimensions—better than wheeled robots on complex terrain, lighter and quieter than tracked versions, more stable than quadrupeds, and less expensive than humanoids—solving multiple traditional robotics limitations simultaneously

  • Proven real-world deployment: Unlike many robotics demonstrations, Dobot's six-legged robots are already operational in demanding scenarios including explosion-proof inspections on offshore oil platforms, photovoltaic power station maintenance, and earthquake search-and-rescue missions, demonstrating immediate commercial viability

This launch completes Dobot's embodied AI ecosystem alongside their Atom humanoid, positioning the company as a comprehensive robotics platform provider rather than single-product manufacturer in the competitive Chinese robotics market.

🤖 Startup Spotlight

EngineAI: China's $139M Bet on General-Purpose Humanoids

📍 Location: Shenzhen, China

What they do: Develops general-purpose humanoid robots powered by proprietary SEED multimodal large model integrating visual, auditory, and sensory data

Latest update: Raised approximately $139M across pre-Series A++ and Series A1 rounds, bringing total funding to over $166M since 2022 founding

Why watch: EngineAI secured backing from major Chinese tech players including XPeng (EV leader), JD.com (e-commerce giant), CATL Capital (battery leader), and Baidu Ventures, indicating broad ecosystem support. The company's SEED multimodal model aims to enable complex task execution in dynamic environments across manufacturing, logistics, and healthcare.

3 Key Takeaways:

  • All-star Chinese tech consortium: The investor lineup reads like a who's who of China's tech ecosystem—XPeng for automotive AI, JD.com for logistics automation, CATL for power systems, and Baidu for AI capabilities—suggesting coordinated effort to build a comprehensive humanoid robotics supply chain

  • Multimodal AI differentiation: EngineAI's proprietary SEED model integrates visual, auditory, and sensory data for "complex task execution in dynamic environments," positioning it as a competitor to OpenAI's foundation models but specifically optimized for embodied AI applications

  • Rapid scaling trajectory: From founding in 2022 to $166M total funding in three years demonstrates the accelerated timeline Chinese humanoid startups are following, with immediate focus on R&D, AI-hardware integration, and international expansion rather than extended prototype phases

This funding represents China's strategic push to dominate humanoid robotics through coordinated investment from automotive, e-commerce, battery, and AI leaders.

📈 Investor Watch

World to Have 300 Million Humanoid Robots by 2050, UBS Report Says

Swiss banking giant UBS projects the global humanoid robot population will exceed 300 million by 2050, with market value reaching $1.4-1.7 trillion. The comprehensive report, based on insights from over 30 analysts across 25 industries, forecasts 2 million androids working within a decade and 86 million units annually by 2050, reflecting 40% compound annual growth over 25 years.

3 Key Takeaways:

  • Investment timing strategy revealed: UBS identifies upstream players (component and material suppliers) as short-to-medium term winners, while midstream manufacturers face 5-year financial pressure from heavy R&D investments before long-term industry dominance shifts back to them—providing clear sector rotation guidance for investors

  • "EV moment" delayed but inevitable: UBS analyst Phyllis Wang suggests the technological breakthrough enabling million-to-tens-of-millions unit sales growth won't occur before 2030, but expects 70%+ price declines over two decades, indicating patient capital will be rewarded as the market matures

  • Geopolitical market divergence: China focuses on industrial manufacturing demand with government backing, while the US addresses labor shortages and reshoring through private sector initiatives—suggesting different investment themes and policy risks across the two largest humanoid markets

This institutional research validates the humanoid robotics investment thesis while providing specific timing and sector allocation guidance for navigating the 25-year growth trajectory.

🧩 Pattern of the Week

The Great Labor Liberation Thesis

A fascinating shift is emerging in humanoid robotics discourse: from job displacement anxiety to "human liberation" narrative. The latest industry analysis reveals telling statistics—8.1 million US job openings versus only 6.8 million unemployed workers, with 70% of openings in "essential roles" like warehouses and manufacturing. This gap, combined with global aging (16% of population will be 65+ by 2050), creates the perfect conditions for what industry leaders are calling "symbiotic collaboration" rather than replacement. The pattern goes deeper than labor economics: companies are repositioning humanoids as enablers of human creativity and fulfillment. The argument suggests that by handling "repetitive, dangerous, and undesirable tasks," robots free humans for roles requiring "creativity, emotional intelligence, and complex problem-solving." This isn't just marketing spin—it reflects a strategic pivot as the industry realizes adoption depends on worker acceptance, not resistance. The emergence of new job categories (robot trainers, human-robot interaction designers, ethics specialists) supports this narrative, suggesting the robotics revolution creates as many opportunities as it displaces. What's particularly striking is the connection to broader economic transformation: universal basic income becomes "feasible" when robots handle hazardous work, potentially offering an alternative to "disenchanted capitalism." This pattern represents the industry's attempt to reframe humanoids from competitive threat to collaborative tool for human flourishing.

📚 Resource / Reading

Human-centric Robots Boost Quality Professionals' Toolkits

This Quality Magazine piece by Mike DeGrace explores how collaborative robots (cobots) are transforming quality control and inspection workflows beyond traditional automation. Using real-world examples like Zippertubing Company's UR5 integration, the article demonstrates how cobots serve as flexible platforms for multiple quality applications rather than single-purpose tools.

Notable Quotes:

"I've stood beside operators and inspection workers when a collaborative robot (or 'cobot') cell is unveiled for the first time and witnessed that 'click' of recognition when they realize that cobots are not replacements for human labor."

"What people are recognizing is that cobots offer a paradigm shift from traditional automation used in quality and inspection applications. Put simply, cobots offer capabilities that traditional robots don't, especially when it comes to flexibility, programmability, and reprogrammability."

"When you buy a cobot, you may have just one application in mind, but leading cobots are platforms for almost any number of quality applications - and not just quality applications."

This resource highlights the critical distinction between traditional industrial automation and human-centric robotics, emphasizing how cobots enhance rather than replace human capabilities in quality control environments—a key insight for understanding the broader humanoid robotics adoption curve.

🛠 Builder's Corner

Intro robotics students build AI-powered robot dogs from scratch

Stanford's CS 123 course teaches students to build "Pupper" robots from basic hardware kits over 10 weeks, combining engineering fundamentals with cutting-edge AI. Students start with motor control and locomotion, then add neural networks for improved walking, vision, and environmental response. The course culminates in a "Dog and Pony Show" where teams demonstrate creative applications like maze navigation and firefighting, attended by industry leaders from NVIDIA and Google.

What it does: The course evolved from Stanford's Doggo robot project to create an accessible platform requiring only basic programming skills. Students build hardware, code movement systems, and train neural networks while working through playfully-named labs like "Wiggle Your Big Toe" and "Do What I Say." Teams customize their Puppers for specialized tasks, with some even 3D-printing custom parts for final projects.

Why it matters: This represents the democratization of advanced robotics education—making "state-of-the-art physical intelligence" accessible to undergraduate students rather than restricting it to graduate research labs. By combining AI and robotics from day one, CS 123 reflects the industry reality where foundation models are transforming embodied AI. The course's evolution from student club project to formal curriculum taught by Karen Liu (Stanford), Jie Tan (Google DeepMind), and Stuart Bowers (Apple) demonstrates how quickly robotics education is adapting to industry needs. Most importantly, the "low barrier to entry" philosophy proves that complex robotics concepts can be taught through hands-on building rather than pure theory, potentially inspiring a new generation of robotics innovators during what the instructors call the "robotics boom."

💼 Jobs in Mechonomics

1. Staff Reinforcement Learning Engineer - Robotics XPENG | Santa Clara, CA | $215,280 - $364,320

Lead cutting-edge research in deep-learning methods for legged locomotion and whole-body control in humanoid robots. Develop end-to-end motion controllers using reinforcement learning and imitation learning while tackling critical sim-to-real transfer challenges. Work with state-of-the-art algorithms like PPO, DQN, and SAC on real-world robotics problems.

2. Robot Service Triage Specialist- 1X Technologies AS| Sunnyvale, CA | $80,000 - $120,000 (estimated)

Join the forefront of household humanoid robotics by ensuring safe, reliable operation of AI-powered robots in customer environments. Rapidly diagnose complex issues across mechanical assemblies, sensors, actuators, and firmware using proprietary 1X tooling. Execute field-level electrical debugging, coordinate repairs, and maintain detailed service analytics to support engineering feedback loops.

3. Robotics Engineer Rediantt LLC | Remote | $100K - $120K

Design and develop robotic systems across multiple industries including manufacturing, healthcare, aerospace, logistics, and consumer electronics. Handle full lifecycle from blueprints and control systems to software development, testing, and maintenance. Write code for control algorithms, path planning, and user interfaces while managing robotics projects including planning, budgeting, and team coordination.

4. Robotics Engineer Rediantt LLC | Remote | $100K - $120K

Design and develop robotic systems across multiple industries including manufacturing, healthcare, aerospace, logistics, and consumer electronics. Handle full lifecycle from blueprints and control systems to software development, testing, and maintenance. Write code for control algorithms, path planning, and user interfaces while managing robotics projects including planning, budgeting, and team coordination.

5. Sr. Machine Learning Engineer (Autonomous Vehicles) General Motors | Remote | $158K - $220K

Develop and implement machine learning techniques for driverless technology, transforming raw data into meaningful information for autonomous vehicle systems. Lead technical efforts to optimize, supervise, and refine on-road performance for production models. Guide technology choices as a technical leader while enabling team effectiveness through extensible design and code contributions.