AlphaBot 2 is a semi-humanoid robot by Chinese company AI² Robotics featuring a humanoid upper body on a six-wheeled mobile base. Has over 34 degrees of freedom, adjustable working height up to 2,400mm, 6+ hours battery life, and runs the AlphaBrain embodied AI platform. Priced at approximately $15,000.
Primary use case: Industrial automation, quality inspection, logistics, and service operations in manufacturing and commercial environments
DOF
34
Battery
6 hrs
Price
$15K
Degrees of Freedom
34
Locomotion
wheeled
Balance
dynamic
Hand Type
unknown
Max Lift/Arm
20 kg
Battery Life
6 hrs
Autonomy Level
Semi-Autonomous
AI Capabilities
vision, speech, planning, manipulation, navigation, learning
Price
$15K
Availability
commercial
Primary Use Case
Industrial automation, quality inspection, logistics, and service operations in manufacturing and commercial environments
The 2026 World Robot Conference in Beijing showcased Chinese robotics startups demonstrating real-world capabilities like kickboxing and ice cream making, as companies race to go public and boost investor confidence. Shenzhen-based AI2 Robotics highlighted its AlphaBot powered by a proprietary foundation AI model at the event.
AI² Robotics has raised $735 million in a new financing round, pushing its valuation past 50 billion RMB (approximately $2.8 billion). The Shenzhen-based company differentiates itself by using wheeled bases instead of bipedal systems for its humanoid robots, trading stair-climbing capability for mechanical simplicity, durability, and lower production costs.
Crunchbase data shows global venture funding to robotics startups reached $9.9 billion in Q3 2026, more than double the $4.3 billion invested in Q3 2025, with Q1–Q3 2026 funding up roughly 185% year over year. China ($13.2 billion) and the U.S. ($12.1 billion) led investment, ahead of Germany ($1.7 billion), with billion-dollar rounds going to Xpeng Robotics ($900 million), AI² Robotics ($736 million Series C), Atoms, Skild AI ($1.4 billion) and Neura Robotics ($1.4 billion). Encord's co-CEO Ulrik Stig Hansen said physical AI still lacks a large corpus of real-world data, requiring expensive collection, curation and multimodal validation, and that the sector is moving from pre-training toward post-training.