Meituan exclusively invested in Series A. "Independent Variable Robotics" has successfully completed two consecutive rounds of financing worth hundreds of millions of yuan. | First reported by 36Kr

Text | Wang Fangyu

Editor | Su Jianxun

36Kr learned that the embodied intelligence company "X Square Robot" recently completed hundreds of millions of yuan in Series A financing, led by Meituan Strategic Investment and followed by Meituan Longzhu.

This round of financing will be used to continuously accelerate the synchronous iteration of the fully self-developed end-to-end general embodied intelligence large model and the robot body, as well as the cooperation and implementation of intelligent solutions for multiple future application scenarios.

In addition to the hundreds of millions of yuan in Pre-A++ round financing led by Lightspeed & China Partners and Legend Capital disclosed in February this year, before the Series A financing, X Square Robot also completed hundreds of millions of yuan in Pre-A+++ round financing invested by China Media Capital, Yunqi Partners, and GF Xinde Investment.

Within less than a year and a half since its establishment, X Square Robot has currently completed seven rounds of financing, with a cumulative financing amount exceeding one billion yuan.

X Square Robot was founded in December 2023, focusing on the research and development of the "general embodied large model". It uses real-world data as the main data source to build general robots with fine operation capabilities.

As one of the earliest companies in China to adopt a completely end-to-end approach to realize the general embodied intelligence large model, X Square Robot is committed to building an integrated system architecture for unified perception, planning, and control.

The "WALL - A" model of the "Great Wall" operation large model series independently developed by the company has the ability to fuse multi-modal information. It can integrate different types of perception data, natural language instructions, and motion control signals to achieve end-to-end mapping from input to output.

Based on large-scale general knowledge pre-training and multi-task learning mechanisms, the "WALL - A" model has demonstrated zero-shot generalization ability in some unseen new task scenarios - zero-shot generalization in a wide range of scenarios is one of the key indicators for realizing general robots.

At the same time, X Square Robot has built a model-driven data closed-loop system. The company has independently developed dozens of data processing models and multiple generations of data collection equipment to achieve automatic control of data quality and comprehensively improve data collection efficiency.

At the hardware level, the company independently develops and continuously optimizes the robot body adapted to the control of multi-modal large models to better meet the needs of fine operations and stable operation in an open environment. At present, X Square's robot body has been applied in multi-step complex task scenarios.

It is worth noting that since the end of last year, X Square Robot has made breakthroughs in multi-modal output and embodied thought chains, earlier than the time when Gemini Robotics announced the research results of thought chains in March this year.

The model architecture developed by the company focuses on multi-modal output, realizing end-to-end information fusion, enabling the robot to achieve efficient alignment among multiple information channels such as vision, language, and action, and significantly improving the model's context reasoning and self-feedback capabilities in ultra-long sequence tasks.

The company has also continuously made new progress in dynamic environment perception, real-time task planning, and ultra-long-range tasks. The above breakthroughs and progress have further improved the robot's autonomous decision-making ability and execution efficiency in complex and open environments.

Wang Qian, the founder and CEO of X Square Robot, believes that the development of embodied intelligence depends on the advantages of model algorithms in the short term, data advantages in the medium term, and product advantages in the long term. X Square has always deepened its accumulation around these three major directions and was the first to implement practices in domestic open service scenarios to explore service closed-loops in different scenarios.

The company's core team brings together experts from world-renowned artificial intelligence and robotics laboratories and scholars from top universities at home and abroad, committed to promoting the innovation and development of embodied intelligence and robotics technology. The core team members include researchers who were among the first in the world to propose the Attention mechanism (the core of the Transformer structure), experts from international top robotics laboratories, technical leaders of China's first batch of 100-billion-parameter large models, and experts in the field of robot hardware.