Humanoid Robotics and Embodied AI Daily Wrap: October 9, 2025
The acceleration in humanoid robotics continues to dominate the AI landscape, with major announcements today focusing on design for mass manufacturing, autonomous skill acquisition, and the intensifying geopolitical race for market dominance. A notable critical voice, however, has tempered the excitement with a warning about fundamental dexterity challenges.
Figure AI Unveils Figure 03 Humanoid, Optimized for AI and High-Volume Manufacturing
Figure AI Inc. announced the release of its third-generation humanoid robot, the Figure 03, marking a significant redesign of both hardware and software to better suit mass production and eventual household use.
The company emphasized that the Figure 03 was built from the ground up for high-volume manufacturing, a departure from the traditional, time-consuming, and expensive prototype-focused designs of previous generations. This new approach involves a complete reinvention of the design and manufacturing process, the establishment of an entirely new supply chain, and the creation of a specialized high-volume manufacturing facility named BotQ.
A core focus of the redesign is the integration of artificial intelligence. Figure AI stated that scaling humanoid robots is impossible without advanced AI, and the Figure 03 is purpose-built to enable reasoning through its physical AI model, Helix.
- **AI-First Design:** The robot includes a fully redesigned sensory suite and hand system to bring its Helix physical AI model to life.
- **Enhanced Vision:** A new camera architecture delivers twice the frame rate, one-quarter the latency, and a 60% wider field of view per camera compared to its predecessor, engineered for high-frequency visuomotor control.
- **Commercial Ambition:** The company, which recently surpassed $1 billion in committed capital, aims to ship 100,000 humanoids over the next four years.
Figure 02 systems, the previous model, are already undergoing top-secret training and collecting data, with some units currently working 10-hour daily shifts lifting parts at a BMW factory.
Tesla Optimus Showcases Autonomous Kung Fu Mastery in New Demonstration
Tesla’s Optimus humanoid robot demonstrated a significant leap in AI-driven autonomy by flawlessly performing complex Kung Fu movements alongside a human trainer in a recently released video.
The demonstration was notable because Elon Musk confirmed that Optimus was operating completely on its own AI, without any human guidance or teleoperation. This is a stark contrast to earlier demonstrations, which sometimes required human intervention even for basic actions.
The ability to observe, analyze, and autonomously acquire complex physical skills—such as the balance and fluidity required for martial arts—signals a major advancement in the robot’s AI-driven learning capabilities and its ability to adapt to complex, non-repetitive tasks.
The continuous improvement in Optimus is part of Tesla’s ambitious plan to deploy robots in its factories by late this year, with a long-term goal of producing 1 million units annually by 2030.
Global Humanoid Robot Race Intensifies with Diverging National Strategies
The competition for dominance in the humanoid robotics market is escalating, with major global powers adopting distinct national strategies.
The United States, led by Big Tech companies, is focusing on the “brain” of the robots, leveraging its strengths in advanced AI. Companies like NVIDIA are concentrating on developing AI platforms, such as the humanoid-specific reasoning model GR00T, rather than manufacturing the robots directly. Google is similarly working on “Gemini Robotics” to bring its most intelligent AI models into the physical world.
Conversely, China is pushing for “China Speed,” building a comprehensive, vertically integrated industrial ecosystem from components to complete systems. This strategy, backed by government policy, has resulted in a rapid proliferation of new models and a focus on deployment on factory floors for immediate data collection. Chinese companies accounted for 61% of the 66 humanoid robot models unveiled between 2022 and the end of last year.
South Korea is also entering the fray, focusing on specialized humanoid robots that leverage its existing strengths in semiconductors and electronics manufacturing, with companies like Rainbow Robotics developing dual-armed humanoids for industrial settings.
Robotics Pioneer Rodney Brooks Issues Warning: Current Humanoids ‘Doomed to Fail’ Over Dexterity
Amidst the flurry of positive news and demonstrations, renowned robotics pioneer and co-founder of iRobot and Rethink Robotics, Rodney Brooks, offered a starkly contrasting view, predicting that the current wave of general-purpose humanoid robots—including Tesla’s Optimus and Figure 02—is “doomed to fail.”
Brooks argues that despite major advances in computing power and sensor technology, the fundamental challenge of dexterity remains the “wall that everyone runs into.” He asserts that without reliable hand control and sensitive touch sensors, current demonstrations are little more than a “show.”
His criticism centers on the lack of progress in fine-motor manipulation, which has plagued robotics development for decades. Brooks’ position suggests that the industry is prematurely scaling up without solving the core mechanical and sensory problems required for a robot to truly function as a general-purpose worker in an unpredictable human environment.
“Without sensitive touch sensors and reliable hand control, any demonstration remains little more than a show.” – Rodney Brooks on current humanoid robots.
Industrial Deployment Signals Shift to Commercialization
The humanoid robotics sector is transitioning from the lab to the factory floor, marking a critical step toward commercial viability.
Agility Robotics’ Digit humanoids are currently undergoing real-world trials in Amazon warehouses, where they are tasked with lifting boxes up to 16 kg, navigating narrow aisles, and loading items onto conveyor belts. The company’s RoboFab in Oregon, the world’s first mass-production factory for humanoids, has the capacity to produce 10,000 Digit robots annually.
In the automotive sector, Figure AI has deployed humanoids in BMW factories for autonomous tasks, while the Hyundai Motor Group is reportedly integrating Boston Dynamics’ Atlas robots into its Georgia factory. Unitree robots are also being utilized on production lines at Chinese electric vehicle manufacturers like BYD.
This rapid deployment in industrial settings is creating a massive learning engine, accelerating the training of AI models with real-world sensory data, which is essential for improving autonomy and reliability at scale.