iRonCub3 Becomes World's First Jet-Powered Flying Humanoid Robot

🧠 Weekly Brief

iRonCub3 Becomes World's First Jet-Powered Flying Humanoid Robot

Summary (50 words): Italy's IIT achieved a robotics breakthrough with iRonCub3, the first jet-powered humanoid robot to complete controlled flight, hovering 50 centimeters off the ground. The 70kg robot uses four microjet turbines generating over 1,000 newtons of thrust, requiring a titanium spine to withstand 800°C exhaust temperatures.

3 Key Takeaways:

• Thermal engineering breakthrough: The team solved unprecedented challenges by developing a titanium spine and heat-shield covers to handle exhaust gases reaching 700°C and flowing at nearly the speed of sound—pushing the boundaries of materials science in robotics applications

• Advanced control system fusion: iRonCub3 integrates neural networks trained on simulated and experimental data with constrained Quadratic Programming algorithms, successfully balancing slow joint motors with high-speed jet turbines for stable flight control

• Asymmetric flight mastery: Unlike symmetrical drones, the robot's elongated humanoid frame and moving limbs create dynamic balance challenges that required entirely new control architectures, validated through wind tunnel testing at Polytechnic of Milan and AI-powered aerodynamic modeling from Stanford University

This achievement represents the convergence of advanced materials, AI control systems, and aerodynamics—setting the stage for future applications in disaster response and complex terrain navigation.

Source

Humanoid robots could soon start working at US Foxconn facility to build Nvidia AI servers

Summary (50 words): Foxconn and Nvidia plan to deploy humanoid robots at a new Houston facility to manufacture Nvidia's GB300 AI servers, targeting Q1 2026 operations. The collaboration includes testing UBTech robots and developing two Foxconn variants—one with legs, another with wheeled AMR base for cost-effectiveness.

3 Key Takeaways:

• First AI chip manufacturing deployment: This marks the inaugural use of humanoid robots in semiconductor production, potentially setting the industry standard for AI server manufacturing as companies seek to boost production speed and reduce costs in the rapidly expanding AI hardware market

• Dual-architecture robot strategy: Foxconn's two-variant approach—legged humanoids for complex manipulation and wheeled AMR-based units for cost-effective operations—reflects the industry's recognition that different factory tasks require different form factors rather than one-size-fits-all solutions

• Nvidia's foundation model integration: The deployment leverages Nvidia's Isaac GR00T N1 dual-system architecture (fast "System 1" reflexes plus deliberate "System 2" reasoning), demonstrating how AI foundation models are becoming the software backbone enabling humanoids to transition from demos to actual manufacturing workflows

This collaboration represents the convergence of AI hardware demand, manufacturing automation, and humanoid robotics reaching commercial viability in high-stakes production environments.

Source

Chinese humanoid robot cooks steak by remote control from 1,118 miles away

Summary (50 words): Dobot Robotics demonstrated its Atom humanoid robot cooking steak via VR teleoperation from 1,118 miles away, showcasing real-time precision control with 0.05-millimeter accuracy. The $27,500 robot has 28 degrees of freedom and five-fingered hands, proving long-distance teleoperation feasibility for dangerous or inaccessible environments.

3 Key Takeaways:

• Long-distance precision breakthrough: Dobot solved critical network latency challenges that have plagued remote robotics, enabling real-time teleoperation over 1,118 miles with sub-millimeter accuracy—a technical achievement that even NASA's Valkyrie program hasn't fully addressed for space applications

• Commercial teleoperation viability: At $27,500, Atom represents the first affordable humanoid capable of professional-grade remote operation, opening markets for dangerous environments like nuclear plants, offshore operations, and eventually space missions where human presence is impossible

• Market validation through demonstration: The cooking demo triggered a 3% stock price surge and prompted global shipments to Japan, proving that compelling real-world applications (rather than just walking demos) drive investor confidence and customer adoption in the humanoid robotics market

This milestone demonstrates that teleoperation, not full autonomy, may be the near-term path to humanoid commercialization, especially for high-stakes applications requiring human judgment with robotic precision.

Source

🤖 Startup Spotlight: China's $1.3B Humanoid Unicorn

Hugging Face: Democratizing Robotics with a $299 Desktop Revolution 📍 Location: New York, NY (HQ) / Paris, France

What they do: AI platform transitioning from "GitHub of machine learning" to open-source robotics with affordable, developer-focused hardware

Latest update: Launched Reachy Mini, a $299 desktop humanoid robot with native Hugging Face Hub integration, targeting millions of developers across their 10-million-user platform

Why watch: Unlike traditional robotics companies releasing one expensive product annually, Hugging Face plans to "release 100 prototypes a year" with full open-source designs. The 11-inch robot features 6 degrees of freedom, Raspberry Pi 5, and Python programming, positioning it as the antithesis to $20K-$70K industrial robots.

3 Key Takeaways:

• Open-source hardware disruption model: By releasing all hardware designs, software, and assembly instructions as open source, Hugging Face is applying successful software platform strategies to hardware—potentially creating the "Linux of robotics" that could accelerate innovation while undermining traditional proprietary approaches

• Platform ecosystem strategy: Native integration with Hugging Face Hub allows developers to access thousands of pre-built AI models and share applications through "Spaces," creating potential for "millions of apps" rather than single-purpose robots—fundamentally different from Tesla/Boston Dynamics' integrated approach

• Privacy-first positioning: CEO Delangue frames open-source robotics as fighting "concentration of power," arguing that transparent, locally-runnable AI models address privacy concerns better than "black box robots" controlled by a few companies—a compelling counter-narrative as robots enter homes

This represents the first major attempt to apply platform economics to robotics hardware, potentially reshaping industry dynamics from closed, expensive systems to open, affordable ecosystems.

Source

📈 Investor Watch

2 Humanoid Robot ETFs Now Trading as AI Boom Expands

Two new ETFs targeting humanoid robotics launched within three weeks of each other, signaling mainstream investor appetite for the sector. The KraneShares Global Humanoid and Embodied Intelligence Index ETF (KOID) launched in early June with $3.2M AUM, followed by the actively managed Roundhill Humanoid Robotics ETF (HUMN) on Thursday.

3 Key Takeaways:

• Passive vs. active approaches emerge: KOID tracks 51 equally weighted stocks across the full humanoid ecosystem ("brain," "body," and systems integrators) with global diversification, while HUMN actively manages 30 holdings focused on direct manufacturers and critical enabling technologies like precision actuators and tactile sensors

• Tesla dominance across both funds: Tesla leads HUMN at 12.6% allocation due to Optimus development, while also featuring prominently in KOID alongside NVIDIA (8% in HUMN), demonstrating how established tech giants are becoming the primary investment vehicles for humanoid exposure

• Thematic ETF validation: Roundhill's entry (with $300M+ success in metaverse ETF METV) and KraneShares' participation (known for $6.5B China Internet ETF KWEB) suggest institutional confidence that humanoid robotics has moved beyond "sci-fi fantasy" to investable theme with genuine commercial potential

The rapid launch of competing ETFs indicates humanoid robotics is transitioning from niche research to mainstream investment category, offering retail investors structured exposure to this emerging sector.

Source

🧩 Pattern of the Week

The Emotional Intelligence Imperative

China's aging population crisis is driving a fascinating evolution in humanoid robotics: the transition from functional to emotional AI. Liu Lizheng from Fudan University's team developing the Guanghua No 1 humanoid for elderly care captures the shift perfectly—robots must not only "fulfill tasks" but be "interactive in emotional aspects, so elderly users feel warmth and companionship akin to that of a loved one." This represents a fundamental pattern emerging across the industry: as humanoids move into intimate human spaces, technical capability alone isn't enough. The Shanghai International Exhibition showcased this trend with "sympathetic" humanoids, intelligent wheelchairs, and automated exoskeletons all emphasizing human-machine integration over pure functional substitution. What's particularly striking is the timeline—Liu believes emotional AI functions will be achieved in 3-5 years, suggesting this isn't distant sci-fi but imminent commercial reality. This pattern extends beyond elderly care to hospitality, healthcare, and domestic applications where humanoids must navigate complex social dynamics. The companies that master emotional intelligence—not just physical dexterity or task execution—may ultimately win the humanoid adoption race as robots transition from tools to companions.

This emotional AI pattern represents the next competitive frontier as humanoids enter the most human-centric environments where acceptance depends as much on feelings as functionality.

Source

📚 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.

Source

🛠 Builder's Corner

Berkeley Humanoid Lite: The $5K Open-Source Revolution

UC Berkeley engineers led by PhD student Yufeng Chi developed Berkeley Humanoid Lite, a fully 3D-printable humanoid robot costing under $5,000. The 1-meter-tall, 16kg robot uses modular cycloidal gearboxes and commercially available components, taking about one week for novices to build. All hardware designs, embedded code, and training frameworks are completely open source.

What it does: The robot demonstrates grasping objects (including playing with Rubik's Cubes) via teleoperation joystick controls and bipedal walking through reinforcement learning controllers. The modular design allows builders to start with single actuators, progress to arms/legs, then complete humanoids. If components break, users simply 3D-print replacement gearboxes rather than buying expensive proprietary parts.

Why it matters: Berkeley Humanoid Lite directly addresses the customization barrier that has limited humanoid experimentation to well-funded labs. Unlike commercial robots with proprietary designs, this platform enables unlimited modification and learning. Chi's vision of an "ecosystem where people share ideas and knowledge" could accelerate humanoid development through distributed innovation rather than corporate R&D silos. The timing is perfect—as startups like Hugging Face pursue democratization from the software side, Berkeley tackles the hardware accessibility challenge. With active Discord communities already building and sharing assembly photos, this represents the emergence of true grassroots humanoid development that could produce breakthrough innovations from unexpected sources.

This open-source approach might prove that the next humanoid breakthrough comes from a garage, not a billion-dollar lab.

Source

💼 Jobs in Mechonomics

1. Technical Artist Co-op - Virtual Systems Amazon Robotics | Boston Area, MA | Co-op Position

Build virtual robotic worlds using Unity, Blender, and cutting-edge simulation technology. Create 2D/3D art assets for real-time graphics applications while liaising between programming and design teams. Work directly with robotics testing labs and manufacturing facilities to develop Digital Twin production workflows and maintain technical documentation for autonomous movement, AI/ML, and robotic management systems.

Source

2. Sr. Robotics Engineer, Mapping & Localization Serve Robotics | Los Angeles, CA / Remote | $119K - $160K

Design and implement next-generation mapping and localization engines for autonomous sidewalk delivery robots operating in urban environments. Build city-scale mapping systems, improve neural network-based navigation for safe traversable areas, and fuse motion/position estimates to reduce global localization uncertainty. Ship production-quality code while collaborating with autonomy teams to maximize mapping utility.

Source

3. Chief Unmanned Systems Robotics Engineer Leidos | Arlington, VA | $149K - $269K

Lead technology strategy and vision for unmanned maritime systems development, supporting Sea Systems Business Area's autonomous surface and sub-surface vessels. Work directly with customers, partners, and commercial companies to define cutting-edge solutions while collaborating with the Sea Systems CTO. Participate in program reviews, architect integrated autonomous systems, and provide technical expertise for proposal development.

Source

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.

Source

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.

Source