Shenzhen-based X Square Robot has raised RMB 1 billion (approximately $140 million) in a Series A++ funding round to accelerate development of embodied AI foundation models. The company, led by founder and CEO Wang Qian, aims to build a “robot brain” that gives machines a human-like understanding of the physical world.
While AI chatbots amaze us with conversation, getting a robot to reliably navigate a cluttered room or handle a delicate object remains a monumental challenge. X Square Robot is betting that the solution lies not in more code, but in a new kind of foundational intelligence built from real-world experience. The company’s recent funding haul, with participation from giants like ByteDance and HongShan, signals strong investor belief that the race to create general-purpose robots will be won by those who master the data-to-model “flywheel.”
“At X Square, we believe the key to enabling robots to truly master real-world tasks lies in the ‘robot brain’ — a foundation model for the physical world that parallels virtual LLMs to shatter generalization bottlenecks,” stated Wang Qian. This core philosophy drives their flagship system, WALL-A, which integrates vision-language-action (VLA) models with predictive world models. Think of it as giving a robot a physics-informed imagination; it can predict the outcome of its actions and use causal inference to understand why something succeeded or failed, drastically improving its ability to handle unseen situations.
READ ALSO: https://modernmechanics24.com/post/schaeffler-humanoid-robotics-alliance/
This isn’t just theoretical. X Square Robot recently demonstrated its Quanta X1 wheeled bimanual robot autonomously completing a complex food delivery in an open environment. Faced with strong winds, deformed packaging, and visual occlusions, the robot didn’t freeze. Using its model’s causal inference, it could “fill in the blanks” when objects were partially hidden and self-correct from operational stalls—all without human intervention. In another test, the system showed zero-shot generalization by identifying irregular items in a pile of parcels, a critical skill for logistics.
The company’s approach hinges on a closed-loop iteration of hardware, data, and models. X Square claims to be the first in China to scale up real-world data resources, using advanced capture tools like teleoperation rigs, exoskeletons, and its Universal Manipulation Interface (UMI). This massive, high-quality dataset fuels large-scale, real-robot reinforcement learning (RL), allowing their foundation models to learn through physical interaction, not just simulation. “The next phase of competition in embodied intelligence is essentially a battle of foundation models built on data closed loops and their capacity for model evolution,” noted Wang.
WATCH ALSO: https://modernmechanics24.com/post/us-hits-isis-terrorists-nigeria/
To complement its software, X Square is advancing full-stack hardware development. It has released two wheeled robot platforms—the Quanta X1 and the semi-humanoid Quanta X2—and manufactures its own core components like robotic arms and joint modules, readying for mass production. The company is already deploying its technology in advanced manufacturing, autonomous logistics, and senior healthcare services.
With this new capital injection, following a $100 million Series A round from backers like Alibaba and Meituan, X Square Robot is poised to scale its unique trifecta of models, data, and hardware. By building the “robot brain” from the ground up, they’re not just programming machines for tasks; they’re attempting to teach them to understand and adapt to our world, one real-world interaction at a time.
