Atoms' $1.7B Debut, J&J's Surgical Breakthrough, and Humanoid Cage Fights
This week's digest covers Travis Kalanick's $1.7B industrial robotics venture Atoms, Johnson & Johnson's FDA approval for the Ottava surgical system, and the world's first humanoid cage fighting tournament. We also explore Sunday Robotics' laundry-folding AI, a new sensing skin for industrial humanoids, and BCI-controlled robots.
Podcast В· 3 min
Travis Kalanick's Atoms Raises $1.7B for Industrial AI
Former Uber CEO Travis Kalanick has unveiled his new robotics venture, Atoms, securing $1.7B in equity and debt to build specialized machines for industrial sectors. The company aims to develop 'atoms-based computers' for mining, construction, transport, and food production, explicitly avoiding the humanoid and warehouse robot trends. By integrating his previous ventures, including the acquisition of Pronto, Kalanick is positioning Atoms to tackle heavy industries that have historically resisted automation. This move represents a significant shift in Kalanick's 'bits to atoms' strategy, moving from digital platforms to physical infrastructure. The backing from a16z underscores the growing investor confidence in industrial AI, with the firm predicting that robots will soon handle the majority of physical production and logistics tasks. For the industry, this signals a potential pivot away from general-purpose humanoids toward purpose-built, ruggedized automation for high-stakes environments.
J&J's Ottava Breaks 25-Year Surgical Robot Monopoly
Johnson & Johnson has received FDA marketing authorization for Ottava, a surgical robotic system that features arms integrated directly into the operating table. This approval marks the first significant challenge to Intuitive Surgical's da Vinci system, which has held an estimated 80% market share for over two decades. Ottava is designed for common upper-abdomen procedures, including gallbladder removal and gastric bypass. The system's design allows for a 30–50% smaller footprint than existing surgical robots, addressing a major barrier to adoption in crowded operating rooms. By enabling the table and arms to move in sync, J&J aims to streamline surgical workflows. While robotic surgery adoption remains relatively low at 8% of global procedures, the entry of major players like J&J and Medtronic suggests a competitive shift that could accelerate hospital integration and lower costs.
Generative Bionics Unveils Gene.01 with Sensing Skin
Italy-based Generative Bionics has introduced Gene.01, a humanoid robot equipped with a full-body sensing skin capable of detecting touch, force, proximity, and temperature. The robot, developed in just six months, is currently being deployed for welding tasks at Fincantieri shipyards. The sensing skin serves a dual purpose: it acts as a safety layer to detect human workers before contact and functions as a teaching interface, allowing workers to guide the robot's movements by hand. This development highlights a critical trend in human-robot collaboration: the need for 'tactile awareness' in crowded industrial spaces. Unlike camera-based systems that may have blind spots, sensing skin provides a continuous safety buffer. As humanoids move from controlled labs to active factory floors, such safety-first hardware approaches are becoming a non-negotiable requirement for widespread adoption.
EngineAI Hosts World's First Humanoid Cage Fight
Shenzhen-based EngineAI has hosted the first freestyle humanoid fighting tournament, featuring 32 teams competing with standardized T800 robots. The event, held at the Nanshan Cultural and Sports Center, saw robots engaging in combat, with scoring based on effective strikes, stability, and durability rather than just knockouts. The tournament serves as a high-visibility stress test for hardware, pushing robots to maintain balance and functionality under extreme physical duress. Beyond the spectacle, EngineAI CEO Zhao Tongyang stated that the event is designed to gather real-world combat data to improve R&D. This follows a trend of using unconventional benchmarks—such as marathons and combat—to prove the durability and decision-making capabilities of Chinese humanoids. While the entertainment value is clear, the industry is watching to see if these 'battle-tested' capabilities can translate into reliable performance in factories and warehouses.
Rural NY School Adopts AI Humanoid Tutor
A rural New York school district has introduced Sally, a $58,000 AI-powered humanoid tutor from Realbotix, to assist students with STEM subjects. The robot, which features a silicone-skinned face and torso, is designed to guide students through coding, robotics, and math problems rather than providing direct answers. The district emphasized that Sally is an instructional tool intended to spark curiosity and will not replace human teachers. This deployment is a notable experiment in the efficacy of embodied AI in education. While chatbot tutors are ubiquitous, there is little empirical data on whether a physical presence enhances student engagement or learning outcomes. By investing in a high-cost humanoid, the school district is providing an early, real-world case study on whether the 'humanoid factor' makes a measurable difference in the classroom.
BrainCo Debuts BCI for Robot Control
At the World Artificial Intelligence Conference in Shanghai, BrainCo unveiled a Brain-Computer Interface (BCI) platform that allows users to control robotic arms and humanoids using only their thoughts. The system uses an EEG headset to decode neural intent, which is then translated into commands for the robot within 200ms. Unlike invasive BCI technologies, BrainCo's solution requires no surgery and is compatible with third-party hardware. The company is also using the platform to capture 'intent data'—a layer of information that records the neural signals behind human actions. BrainCo believes this data could be instrumental in training robots to better understand human tasks. While the practical application of brain-controlled robotics is still in its infancy, this development underscores the rapid convergence of BCI and physical AI.
Sunday Robotics' ACT-2 Model Masters Laundry
Sunday Robotics has previewed ACT-2, an AI model that enables its Memo home robot to fold laundry with a 99.1% success rate in unfamiliar environments. The system was tested across 785 folding attempts on various garment types, handling items in piles or baskets without requiring prior setup or demonstrations. The company plans to deploy Memo to homes via a beta program this fall. This achievement addresses the 'generalization gap' that has long plagued home robotics, where systems often fail when lighting or object placement changes. By pretraining on human-collected data, Sunday Robotics aims to create a system that learns techniques that carry over to unseen homes. If successful, this could mark a significant step toward truly useful, autonomous household robots.
Phantom Twist Drone Uses Motion Blur for Invisibility
Engineers at Northwestern University have developed Phantom Twist, a drone that rotates its body up to 25 times per second to exploit human motion blur. By spinning the frame and propeller in opposite directions, the drone eliminates stationary visual cues, making it 10 times harder to spot than conventional drones. The design was optimized by AI, which simulated 20,000 configurations to find the most effective layout. The drone is intended for wildlife monitoring and infrastructure inspection, where minimizing disturbance is critical. This project also highlights the growing role of AI in mechanical engineering, as the algorithm produced a configuration that researchers noted humans would likely not have conceived independently.