ByteDance and HSG Lead a 1 Billion Yuan Investment in the Largest Embodied AI Financing Deal at the Beginning of the Year

Embodied intelligence remains hot at the beginning of 2026.

Just now, X Square Robot officially announced the completion of a 1 billion yuan Series A++ financing round.

This financing round was jointly led by ByteDance and Sequoia China, with participation from top institutions and diverse local platforms such as Beijing Information Industry Development Fund, Shenzhen Capital Group, Nanshan Strategic Emerging Industry Investment Fund, and Wuxi Venture Capital Group.

Among them, Sequoia China chose to increase its investment after the Series A+ round in September last year; ByteDance, on the other hand, rarely made a direct investment in the field of embodied intelligence.

It is worth mentioning that, in addition to ByteDance, X Square Robot had previously received investments from Meituan and Alibaba, becoming the only domestic embodied intelligence company backed by these three major Internet giants.

From Series A+ to Pre-Series A++: Continuous Investment within a Year

Before this Series A++ financing round, X Square Robot had completed multiple financing rounds in 2025, covering Series A+, Series A, Pre-Series A+++, and Pre-Series A++.

From a chronological perspective, this is a clear financing curve that expands as technology and products progress.

On September 8, 2025, X Square Robot completed a nearly 1 billion yuan Series A+ financing round.

This round was led by Alibaba Cloud and Guoke Investment, with participation from China Development Financial, Sequoia China, Yingce Capital, and Meituan Strategic Investment. As an old shareholder, Meituan Strategic Investment chose to increase its investment, and Legend Star and Junlian Capital continued to follow up.

The financing funds were mainly used for the continuous training of the fully self-developed general embodied intelligence foundation model, the R & D and iteration of hardware products, and the advancement of the open-source large embodied model WALL - A.

This financing round also sent two significant signals: one is that Alibaba Cloud clearly entered the embodied intelligence track for the first time; the other is that this was the fifth time X Square Robot completed a financing round worth over 100 million yuan.

Looking further back, on May 12, 2025, X Square Robot completed a financing of hundreds of millions of yuan in the Series A round, led by Meituan Strategic Investment and followed by Meituan Longzhu.

The funds from this round were mainly used for the synchronous iteration of the end - to - end general embodied intelligence large model and the robot body, and to accelerate the actual implementation of application scenarios.

Before the Series A round, X Square Robot also completed a financing of hundreds of millions of yuan in the Pre-Series A+++ round, led by Huaying Capital and followed by Yunqi Capital and GF Xinde Investment.

By this stage, the company had been established for less than a year and a half and had completed a total of 7 financing rounds, with a total financing amount exceeding 1 billion yuan.

Even earlier, on February 17, 2025, X Square Robot completed a financing of hundreds of millions of yuan in the Pre-Series A++ round, led by Lightspeed China Partners and Junlian Capital, and followed by Beijing Robot Industry Investment Fund and Shenqi Capital.

The funds were used for the training and scenario implementation of the next - generation unified general large embodied intelligence model.

As of now, since its establishment more than two years ago, X Square Robot has completed 9 financing rounds, with a cumulative financing of over 3 billion yuan, reflecting the high recognition of the capital market for its technical route of the "independent foundation model for embodied intelligence".

Who is X Square Robot?

X Square Robot was established in December 2023, focusing on the self - developed "general large embodied intelligence model".

The founder and CEO, Wang Qian, graduated from Tsinghua University with a bachelor's and a master's degree, and then pursued a doctorate at the University of Southern California, with a research focus on Robotics Learning. He was one of the early explorers in the research field related to the neural network attention mechanism and one of the first to introduce the Attention concept into the neural network system.

The co - founder and CTO, Wang Hao, is a doctor in computational physics from Peking University. He used to work at the Institute for Digital Economy in the Guangdong - Hong Kong - Macao Greater Bay Area (IDEA Institute) as the head of the large model team, leading the release of several open - source large models and having a profound foundation in the field of foundation models and system engineering.

In terms of the technical path, X Square Robot made a relatively clear judgment from the beginning. Wang Qian has repeatedly emphasized that the embodied intelligence model is an independent foundation model parallel to the language model.

The so - called independence is not just a distinction at the application level. It means that embodied intelligence needs to directly face the continuous states, causal relationships, and action feedback in the real physical environment. Its modeling goals, training methods, and evaluation criteria are fundamentally different from those of language models and multimodal models serving the virtual world.

Based on the view that "embodied intelligence is an independent foundation model for the physical world", X Square Robot self - developed the "WALL - A" series of VLA operation large models.

This series integrates perception, understanding, decision - making, and action output into an end - to - end model, aiming to solve the reasoning fault and generalization bottlenecks of traditional modular architectures in long - sequence complex operation tasks.

In October 2024, X Square Robot released the WALL - A model, which at that time became one of the largest end - to - end unified large embodied intelligence models in terms of parameter scale globally.

In September 2025, X Square Robot further open - sourced the embodied foundation model WALL - OSS, which ranked third globally in the recent RoboChallenge list, demonstrating the model's generalization ability and stability under a unified evaluation framework.

On the hardware side, X Square Robot is simultaneously promoting two generations of embodied robots, Quantum No. 1 and Quantum No. 2.

Among them, Quantum No. 1 is a wheeled dual - arm robot ("Xiaoliang"), equipped with the end - to - end embodied base model WALL - A, supporting multimodal remote operation, and mainly used for data collection in high - frequency operation tasks, model verification, and scientific research and teaching scenarios.

Quantum No. 2 is upgraded to a wheeled humanoid structure, with higher degrees of freedom and fine force - control ability, used for collecting complex operation and high - quality physical interaction data.

Overall, X Square Robot is not in a hurry to provide the "final form of the robot answer", but instead chooses to first build a sustainable and evolving embodied intelligence base: the model learns in the real physical world, the hardware serves the model, and the data feeds back to the model for iteration, forming a closed - loop.

It is precisely under this technical and engineering logic that X Square Robot has gradually attracted continuous attention and investment from the capital market and the industrial side.

Reference Links

[1]https://www.x2robot.com

[2]https://mp.weixin.qq.com/s/92U2n8wyDxzXCxz8JHGoFQ

This article is from the WeChat official account "QbitAI". Author: henry. Republished by 36Kr with permission.