- Published on
Phybot M1 Debuts with a Backflip and a ‘Most Powerful’ Claim
One tap, and our reporting appears more often in your Google Top Stories. It makes a real difference to a small newsroom.


Just days after AgiBot set a world record for endurance on the streets of Shanghai, a new challenger from Beijing has emerged with a focus on raw explosive power. PHYBOT, a startup founded by Tsinghua University alumni, has unveiled the M1—a full-sized electric humanoid that announced its arrival by performing a standing backflip.
While acrobatic feats have become the standard marketing currency for humanoid robotics—pioneered by Boston Dynamics' Atlas —PHYBOT is positioning the stunt not just as a display of agility, but as proof of superior torque density. The company explicitly claims the M1 is the "most powerful humanoid robot ever created," a bold assertion in an increasingly crowded field.
The Engineering Behind the Flip
The footage shows the M1 performing a high-impact backflip from a standing position. According to data released by the company and analyzed by industry observers, the secret lies in the robot's proprietary actuation.
The M1 reportedly utilizes five-arc cycloidal drives developed entirely in-house. These joints are capable of a peak torque of 530 N·m, a figure that significantly outstrips many contemporary electric humanoids. Perhaps more critical for dynamic movement is the claimed torque density of 200 N·m/kg.
Stay ahead in humanoid robotics
One email a week: the launches, funding and research that mattered. Join the researchers, founders and technologists following the field.
"Higher torque translates to greater explosive power," noted industry commentator CyberRobo. "A higher peak torque density, much like denser muscles, allows for tremendous force within a smaller and lighter volume."
PHYBOT states that the robot can generate more than 10 kW of instantaneous power. The drivetrain is also rated to withstand shock loads of up to five times its overload capacity, a necessary durability feature for a machine designed to catch its own weight after airborne maneuvers.
Specs and Strategy: The "No Crutches" Approach
Standing 172cm (5'7") tall, the M1 is sized to navigate human environments. Under the hood, the robot's intelligence is powered by a hybrid compute architecture pairing an Nvidia Jetson Orin with an Intel Core i7 processor. For perception, it relies on a standard but robust suite of 3D LiDAR, stereo cameras, and an inertial measurement unit (IMU).
Notably, the company emphasizes that the M1 was developed without reliance on open-source frameworks or what it terms "technical crutches," signaling a vertical integration strategy similar to that of Tesla or AgiBot.
The founders, CEO Ren Shiau and co-founder Mao Shuhan, are former roommates from Tsinghua University’s mechanical engineering program. Ren brings specific domain expertise from stints at major Chinese robotics firms UBTech and Fourier Intelligence.
Their market entry strategy appears to be aggressive on price. PHYBOT is targeting a price point below $42,000.
Share this article
Read next
- Published on
- Reading time
- 9 min read
Beyond Mechanical Chatbots: Galbot Unveils RoboGesture to Tackle Real-Time Nonverbal Communication
- Published on
- Reading time
- 6 min read
OpenAI’s GPT-6 Astra Hits 95% on Physical Manipulation Benchmark, But Precision Tasks Expose Familiar Bottlenecks
- Published on
- Reading time
- 6 min read
Galbot Opens Pre-Orders for ET1 Bipedal Humanoid Aimed at Entertainment and Performance
- Published on
- Reading time
- 4 min read
Tuya Unveils Doova: An AI Companion Robot Targeting Elder Care and Smart Home Hubs
- Published on
- Reading time
- 5 min read
AGIBOT Deepens European Push at IFA 2026 with TÜV Rheinland Certification and Tekpoint Retail Alliance
- Published on
- Reading time
- 7 min read
NVIDIA Buys Hugging Face for $12.9B as Open-Source Robotics Scales Up
Stay ahead in humanoid robotics
One email a week: the launches, funding and research that mattered. Join the researchers, founders and technologists following the field.













