EngineAI — Shenzhen humanoids known for the first front flip and a natural walking gait

• Oct 2023
  Founded
• Shenzhen, CN
  Headquarters
• T800 (2026)
  Latest flagship
• $200M Series B
  Funding raised
• China
  Country
• Private (HK IPO)
  Status

About EngineAI

Shenzhen EngineAI Robotics Technology Co., Ltd. — trading as EngineAI and known in Chinese as 深圳市众擎机器人科技有限公司 — was established in October 2023 in Shenzhen's Nanshan District by Zhao Tongyang and co-founder Ren Guowen. Zhao, who serves as CEO, is a veteran of China's first wave of legged-robotics work: he contributed to early humanoid research at Dogotix, launched a globally distributed quadruped in 2020, and co-founded XPeng's robotics program before striking out to build EngineAI. His stated ambition is a full-stack humanoid company — one that owns its actuators, its neural gait, and its embodied-intelligence models rather than assembling third-party subsystems.

The company's public trajectory has been unusually fast. EngineAI launched three humanoid models in 2024: the SA01 education bipedal (July 2024, ¥38,500), the full-size SE01 in October 2024 with a natural human-like walking gait, and the lightweight open-source PM01 in December 2024 at a launch promotion of ¥88,000. Within months the company had earned viral status: SE01 walking videos through Shenzhen streets drew tens of millions of views, and on February 23, 2025 EngineAI released a video showing the PM01 completing what the company presented — and Xinhua and multiple Western outlets subsequently reported — as the world's first humanoid robot front flip.

Commercial and financial milestones followed quickly. In April 2026 EngineAI closed a Series B round of approximately US$200 million at a valuation near US$1.5 billion and, in the same month, filed for an initial public offering on the Hong Kong Stock Exchange. On May 29, 2026 the company opened EngineAI Intelligent Manufacturing (Honghualing, Shenzhen) Base, a 12,000-square-meter facility with an integrated production line capable of rolling one new humanoid off the line every fifteen minutes, and it declared it had reached the threshold of 10,000-unit scalable delivery capability — a milestone very few humanoid programs anywhere in the world can claim.

Competitively, EngineAI operates in what is now the most crowded corner of the global humanoid robotics market. Its direct rivals include Unitree Robotics (whose H1 and G1 humanoids dominate the low-cost research segment), XPeng's Iron humanoid, UBTECH's Walker series, and Songyan Dongli's N2, along with Western competitors such as Tesla Optimus, Figure 03, 1X NEO, Apptronik Apollo, and Agility Digit. Against that field EngineAI's differentiators are a proprietary end-to-end neural gait, in-house actuators and joint modules, a rapidly maturing manufacturing base, and a track record of viral, physics-defining demonstrations that has repeatedly driven the humanoid category into mainstream news coverage.

Technology stack

EngineAI has framed itself from day one as a full-stack humanoid company: it designs its own joint modules, its own actuators, its own neural gait, and its own embodied-intelligence models. The company's core technical claim — the one that drove the SE01's viral street-walking videos and the PM01's front flip — is that walking, running, and acrobatic motion should be learned end-to-end rather than scripted joint by joint.

Perception

The SE01 platform uses EngineAI's Engine Sense LiDAR-assisted 360-degree perception suite, supplemented by a depth camera and a multi-microphone array for spatial audio and voice interaction. The PM01 combines the same multi-sensor fusion approach with an interactive chest-mounted core screen for direct user commands. Both platforms are designed to operate in unstructured environments — public streets, retail floors, and factory aisles — rather than the fenced, structured cells where industrial robots have historically lived.

Neural gait and motion control

EngineAI's central AI asset is an end-to-end artificial neural network that maps sensor input directly to actuator commands. Trained across Isaac, MuJoCo, ROS, and ONNX toolchains, the network produces the long-stride, upright walk that made the SE01 recognizable and the millisecond-level motion control that made the PM01 front flip possible. The same control stack has since supported walking, running, jumping, squatting, grasping, flying kicks, and capoeira spins across the product family — evidence, the company argues, that a single learned motion policy generalizes across an entire humanoid lineup.

Actuators, power, and compute

EngineAI builds its own BLDC joint modules in-house, an approach it considers essential to meeting the torque density and thermal envelope of an athletic humanoid — the T800 hits a peak joint torque of 450 N·m. The PM01 pairs an Intel N97 general-purpose CPU with an NVIDIA Jetson Orin AI accelerator, and the SE01 uses quick-swap 10,000 mAh battery packs in an aluminum chassis rated for a ten-year design life. On the manufacturing side, the new Shenzhen base uses automated locking, gluing, and laser welding to lift production efficiency by 40 percent, and every robot passes 79 full-dimensional quality inspections and 46 working-condition simulation tests before shipping.

Power management on a humanoid of this class is punishing. Every servo controller must deliver hundreds of watts of instantaneous torque without thermal throttling, which is why EngineAI-class platforms lean heavily on wide-bandgap SiC and GaN power semiconductors in the motor-driver stages, precision current-sense and gate-drive passives on every channel, and high-CV MLCC capacitor arrays across every DC rail. The onboard battery packs integrate cell-balancing, thermal-management, and multi-stage protection ICs whose reliability sets the practical ceiling on how long a humanoid can run between charges.

The robotics supply chain — where Hybrid Electronics fits in

A modern humanoid robot is one of the most component-dense products ever built. A single EngineAI-class platform depends on thousands of individual electronic components: precision BLDC motor drivers, high-torque servo modules, IMUs and 6-DoF force-torque sensors, LiDAR and time-of-flight modules, GPU compute boards such as the NVIDIA Jetson Orin family, battery-management ICs, wide-bandgap SiC/GaN power semiconductors, high-speed board-to-board connectors, precision resistors, and thousands of MLCC capacitors distributed across every power and signal rail.

Building that kind of hardware puts enormous pressure on the electronic-components supply chain. Prototype pilots and low-volume production runs frequently hit allocation, EOL notices, and long lead times on the exact parts they need, and R&D and field-service teams routinely have to support older platforms that use semiconductors no longer in mainstream distribution. During the current humanoid robotics and general-purpose robot AI boom, this pressure is amplified: dozens of well-funded teams — EngineAI, Unitree, XPeng, UBTECH, Songyan Dongli, Figure, Tesla, 1X, Apptronik, Agility, Boston Dynamics — are chasing the same short list of high-torque BLDC drivers, precision IMUs, GPU compute modules, and automotive-grade power semiconductors, and franchise stock on many key parts is now measured in months of lead time rather than weeks.

Hybrid Electronics has been sourcing electronic components for industrial and embedded systems since 1995. For robotics teams — humanoid, mobile, industrial, and research — we specialize in finding the parts you need when mainstream distribution can't:

Trademark and affiliation notice. EngineAI®, SA01, SE01, PM01, and T800 are trademarks of Shenzhen EngineAI Robotics Technology Co., Ltd. (深圳市众擎机器人科技有限公司). Hybrid Electronics is an independent electronic-components sourcing company and is not affiliated with, endorsed by, or authorized by Shenzhen EngineAI Robotics Technology Co., Ltd. All information on this page is compiled from public sources for educational and reference purposes.