Frontier News | Zhipingfang Unveils New-Generation Intelligent Robot AlphaBot 2 and Announces Construction of Its Own Production Line

Text by | Zhang Ziyi

Edited by | Yuan Silai

On April 17th, Zhipingfang, an enterprise focusing on embodied intelligent robots, launched its new-generation general intelligent robot AlphaBot 2 in Shenzhen and comprehensively upgraded its "AGI Terminal" strategy. This strategic upgrade covers the upgrade of core embodied large model technology, product iteration, the construction of self-owned production lines, and commercialization goals.

As a new terminal defined by general intelligence, general intelligent robots rely on their multi-modal perception, rapid learning, autonomous decision-making, and execution capabilities, and are expected to become the fourth-generation revolutionary intelligent terminals. Moreover, a key difference between general intelligent robots and previous robots lies in the driving force of the general embodied large model.

At the press conference, Zhipingfang announced that its original embodied large model brand AI2R Brain was officially upgraded to Alpha Brain. Guo Yandong, the founder of Zhipingfang, introduced that the most crucial feature of Alpha Brain is that it is equipped with the globally and omni-body Vision-Language-Action (GOVLA: Global&Omni-body Vision-Language-Action) large model independently developed by Zhipingfang in a full-stack manner. That is, it is a vision-language-action large model with the ability to handle tasks from the desktop to the open environment, from single-arm to full-body coordination, and from simple tasks to long-range reasoning. This is a comprehensive upgrade of VLA technology, aiming to achieve efficient interaction and autonomous action of intelligent terminals (such as robots) in full-body coordinated control and full-scenario task coverage.

The GOVLA large model consists of three parts: the spatial interaction basic model, the slow system, and the fast system. By parsing user instructions (such as voice instructions), real-time environmental information, and the robot's status, the two systems work in a division of labor: the slow system System2 is responsible for complex logical reasoning, task decomposition, and outputting language interaction content; the fast system System1 outputs the robot's full-body control actions and movement trajectories, taking into account both real-time response and complex decision-making capabilities. It is worth mentioning that the conventional VLA large model only outputs the actions of the robotic arm, while the GOVLA large model proposes to output full-body control and movement trajectories for the first time.

According to Zhipingfang, it has integrated DeepSeek's technology into the construction process of the VLA large model for the first time, improving the reasoning ability of GOVLA, endowing the robot with the ability to understand and analyze long-range complex tasks, and creating the strongest globally and omni-body VLA model that is domestically controllable.

In terms of the design and optimization of the whole-machine hardware, Zhipingfang has adopted a series of technologies such as more reasonable structural design, optimized module selection, and optimized motion control underlying layer to enable the whole-machine hardware to fully exert the upper limit of Alpha Brain's capabilities and adapt to diverse tasks in automobile manufacturing, semiconductor manufacturing, public services, biotechnology, etc.

In addition, at this press conference, Zhipingfang and Bloomage Biotechnology officially announced the establishment of a strategic cooperation.

In the future, Zhipingfang's robots will perform operations such as material collaborative transfer, intelligent unpacking and disinfection, intelligent visual inspection, and multi-material collaborative intelligent feeding in Bloomage Biotechnology's factories, covering the entire process of automated operations. Moreover, in high-risk scenarios such as aseptic product filling and microbial culture monitoring, robots can replace humans to avoid cross-contamination and reduce the risk of microbial contamination; their end-to-end learning ability can also quickly adapt to the needs of process changes. The two parties will also rely on Bloomage Biotechnology's mature supply chain system across the country to promote the implementation of robots in industrial scenarios and improve productivity.