Yunqi Capital Joins "Intelligent Square" in Hundreds-of-Millions-Yuan Pre-A+ Round to Advance End-to-End VLA Embodied Robot Commercialization | Yunqi Capital
Today, embodied intelligence startup **"Zhi Square" (智平方)** announced the completion of a Pre-A+ round worth hundreds of millions of RMB, with **Yunqi Capital participating in the investment.** This round of funding will primarily accelerate the continued iteration of its end-to-end VLA model and the scaled commercial deployment of embodied intelligence robots.

Today, embodied intelligence startup "Zhipingfang" (智平方) announced the completion of a Pre-A+ funding round worth hundreds of millions of RMB, with Yunqi Capital participating in the investment. The proceeds will primarily accelerate the continuous iteration of its end-to-end VLA model and the scaled commercial deployment of embodied intelligent robots.
Embodied intelligence is a sector Yunqi has long cultivated, and VLA (Vision-Language-Action) is a technical route we are particularly focused on. We look forward to its advances enabling embodied intelligence with stronger perception, more natural interaction, and more efficient decision-making.
Zhipingfang is among the earliest domestic companies to commit to the "end-to-end" technical paradigm and achieve breakthroughs in critical VLA domains. Its RoboMamba end-to-end embodied large model for manipulation robots, launched in June 2024, led the industry in both application timing and generalization capability. This edition of "Yunqi Partners" takes you deeper into Zhipingfang and the details of this funding round.
Chengyu Mao Founding Partner, Yunqi Capital
Embodied intelligence is becoming one of the key variables reshaping industrial form after the mobile internet era. As an early-stage investor deeply rooted in frontier technology sectors, Yunqi has always focused on the intersection of technological innovation and industrial demand. The Zhipingfang team combines internationally top-tier embodied intelligence algorithm R&D capabilities with a deep understanding of the underlying logic of Chinese manufacturing — this "double helix DNA" enables it to adapt to complex scenario deployment. Yunqi continues to double down on the convergence of embodied intelligence and the real economy, having already mapped out the complete industrial chain from core components to system integration. We look forward to helping Zhipingfang achieve technical resonance in flexible manufacturing and other fields through ecosystem collaboration, jointly pushing embodied intelligence from single-point breakthroughs toward systematic evolution.
The following content is excerpted from "QbitAI"
On March 6, 2025, Zhipingfang announced the rapid completion of a new Pre-A+ funding round exceeding 100 million RMB, with investors including Yunqi Capital, Dunhong Asset Management, and SDIC Chuangying, among other financial and industrial funds. Zhipingfang is the earliest domestic enterprise to explore the end-to-end VLA technical route and has consistently maintained industry-leading status, achieving the commercialization of intelligent robots equipped with end-to-end VLA embodied large models in 2024. Since the beginning of 2025, Zhipingfang has swiftly completed two funding rounds worth hundreds of millions within just two months. This round of financing will be mainly used to accelerate the continuous iteration of the end-to-end VLA model and to promote the deployment of embodied intelligent robots in scaled commercial scenarios.
Zhipingfang's self-developed AI2R Brain embodied large model has been successfully deployed in the Alpha Bot series of robots. The latest product in the Alpha Bot series, Alpha Bot 1S, combines wheeled design, a folding lift structure, and high-DOF robotic arm features, enabling it to accurately perceive the world in highly flexible, adaptable, and diverse scenarios, achieving "operation of any object, completion of diverse tasks, in any scenario."
01 The Positive Flywheel of Product, Data, and Commercialization Is Zhipingfang's First-Principles Thinking
The company is the earliest domestic general-purpose intelligent robot company to achieve actual commercial deployment. To date in its commercial progress, Zhipingfang has signed contracts with multiple domestic and international top-tier automakers, high-end manufacturers, and internet giants, having passed rigorous POC processes with numerous clients, and achieved tens of millions in recognized revenue in 2024.
From its founding, Zhipingfang established a clear commercialization path: from high-end manufacturing to public services, then to personal wellness. Its experience covering the three major smart terminals — mobile phones, automobiles, and robots — along with mass production experience in AI hardware-software integration, helped the company rapidly streamline commercial operations. 2025 will be the mass production inaugural year for embodied intelligence, and Zhipingfang will continue to leverage its distinct advantages in rapid technical deployment and clear product commercialization pathways to further accelerate its commercialization process.
02 The Entrepreneurial Journey Runs Through a Main Thread of Technological Foresight and Strategic Execution
Recently, humanoid robot company Figure AI launched the Helix system, a Vision-Language-Action (VLA) model for humanoid robots that can help Figure AI achieve generalization for scenario-based tasks, while Zhipingfang is the earliest domestic company to establish the "end-to-end" technical paradigm and achieve breakthroughs in critical end-to-end VLA domains.
Last June, Zhipingfang collaborated with Peking University to launch RoboMamba, an end-to-end embodied large model for manipulation robots. This model was not only made publicly available earlier than Physical Intelligence's OpenVLA, but also significantly surpassed Google's RT series models in generalization capability on unseen tasks. The release of this breakthrough established Zhipingfang's technical leadership and earned it a spot at NeurIPS 2024, a top-tier AI conference — the first time a Chinese embodied startup has made its international voice heard in the end-to-end VLA direction. Additionally, Zhipingfang's domestic embodied large model AI2R Brain has already substantially surpassed leading Silicon Valley models in general spatial perception and generalized manipulation success rates.

03 A Team DNA of Both "Having Built Cars" and "Having Built Phones" Lays a Deep Foundation
Founder & CEO Dr. Yandong Guo is a nationally recognized leading innovation expert. He previously served as a researcher at Microsoft headquarters' core AI team in the United States. In 2018, he joined Xpeng Motors as Chief Scientist, responsible for deep learning-based intelligent system R&D, and drove the mass production deployment of deep learning technology in automotive-form hardware terminals, creating significant demonstration effects in the industry. In 2020, he joined OPPO as Chief Scientist, leading the development of world-class intelligent perception and interaction platforms and multimodal large models, and driving the intelligentization of various smart terminals (mobile phones, IoT, quadruped robots, etc.). Years of traversing the industry-academia-research nexus across intelligent hardware and autonomous driving have given Guo deep cross-Pacific industrial experience. Beyond him, the company has assembled top talent in AI, robotics, and smart terminals, with core members from internationally leading enterprises including Microsoft, Xpeng, OPPO, and Momenta, as well as renowned institutions including Tsinghua University, Peking University, Chinese Academy of Sciences, UC Berkeley, and Columbia University.
Zhipingfang believes that today's intelligent robots can possess greater generality and generalization, capable of performing various jobs and serving diverse industries, transcending the predicament of traditional robots confined to single positions doing single tasks. The core to realizing this vision lies in AI and engineering scale-up capabilities. Breaking this down specifically: AI model structural innovation, infrastructure, training architecture, supply chain capabilities, and the comprehensive optimization between underlying capabilities and compute power, along with endgame-oriented system design — all require layered accumulation and refinement, placing extremely high demands on a team's technical leadership. Currently, including among tech giants, only a very small number of robotics companies can achieve this.
As the only domestic embodied intelligence startup with full-stack AI system capabilities, Zhipingfang's vision is to make general-purpose intelligent robots as accessible as smart cars and smartphones, becoming smart terminals available to everyone. Today, as one of Shenzhen Nanshan District's "Eight Robot Giants," Zhipingfang will in 2025 accelerate the commercial deployment of world-class end-to-end VLA applications on embodied robots.





