en.Wedoany.com Reported - June 5 news, Yuanli Lingji recently completed a merger with logistics robotics company Atomix through equity acquisition, while simultaneously closing a new round of strategic financing. Investors in this round include Zhipu, StepFun, SenseTime, Alibaba, as well as industrial capital Huaqin and SAIC Hengxu. Post-merger, the company will form a more complete full-stack layout centered on embodied intelligence large models, robotics infrastructure, hardware platforms, and logistics application scenarios.
The core value of this merger lies in bringing "embodied model capabilities" and "real-world logistics scenario data" under one roof. The biggest difference between embodied intelligence and pure software large models is that the model must complete perception, planning, control, and execution in the physical world. Training data cannot come solely from text, images, or videos; it also requires data from real devices, real tasks, and real operational workflows. Logistics scenarios feature high frequency, standardization, quantifiability, and scalable replication. Warehouse sorting, transportation, picking, path planning, shelf scheduling, and multi-robot collaboration continuously generate data related to robot actions, environmental changes, task completion rates, and anomaly handling. Atomix has previously accumulated a flexible warehousing network covering over 20 countries and nearly 100 brands in the logistics robotics field, with clients including Uniqlo, Coca-Cola, and Mixue Bingcheng, handling over 600,000 daily shipments. For Yuanli Lingji, this real-world operational data and customer scenarios provide key fuel for transitioning embodied large models from laboratory training to an industrial closed loop.
Tang Wenbin, founder and CEO of Yuanli Lingji, stated that the scaling of embodied large models must move from manual collection to real industrial scenarios, and the massive logistics data accumulated by Atomix is the key fuel driving the flywheel.
The main bottleneck currently facing the embodied intelligence industry has shifted from "whether there is a robot body" to "whether robots can reliably execute tasks in complex scenarios." The hardware supply chain is gradually maturing, with increasing availability of robotic arms, mobile chassis, sensors, controllers, and end effectors. However, what truly enables robots to achieve generality and transferability is how the model understands scenarios, decomposes tasks, adapts to different devices, and continuously learns in real environments. Yuanli Lingji previously released the embodied native large model DM0, the general embodied open-source framework Dexbotic, and co-launched the real-robot evaluation platform RoboChallenge with HuggingFace. Atomix brings logistics robots, flexible warehousing automation, and a commercial customer network. After integration, Yuanli Lingji can connect model training, robot software and hardware, evaluation platforms, and actual warehousing operations into a single chain, allowing the model to move beyond demonstrations and research tasks to directly undergo validation in high-frequency logistics operations. Public information shows that Yuanli Lingji plans to release a warehouse logistics three-level sorting system centered on DM0, supporting mixed operations of multiple robot types, on June 15, and plans to launch the next-generation large model DM0.5, the first general-purpose robot, and new application infrastructure in July. If these products can be implemented, Yuanli Lingji will have the opportunity to further transform from an embodied intelligence model company into a robot platform company with scenario delivery capabilities.
The structure of investors in this financing round is also noteworthy. The simultaneous entry of large model and AI companies such as Zhipu, StepFun, SenseTime, and Alibaba indicates that embodied intelligence is becoming an important direction for model companies to move from Token competition to Action competition. Software large models address information understanding, content generation, and task planning, while embodied models need to further connect mechanical execution, physical environments, and industrial workflows. Due to high data density, clear commercial value, and relatively easy ROI evaluation, logistics scenarios are becoming a key track for the first large-scale validation of embodied intelligence. The participation of industrial capital Huaqin and SAIC Hengxu also means that embodied intelligence is not only attracting model companies but also forming closer ties with smart manufacturing, supply chains, the automotive industry, and hardware engineering capabilities. For investors, the "model + hardware + scenario + data" structure formed by the merger of Yuanli Lingji and Atomix is closer to a sustainable commercial closed loop than simply betting on a single robot body or a single algorithm model.
Going forward, the key challenges for Yuanli Lingji will focus on merger integration efficiency, large-scale replication of logistics scenarios, robot system stability, and the generalization capability of embodied models. The warehousing network and customer resources brought by Atomix can provide real business entry points, but ensuring stable operation of robot systems across different warehouses, goods, operational rhythms, and human-robot collaboration environments still requires long-term engineering validation. If this merger can drive a positive cycle between model capabilities and logistics data, Yuanli Lingji has the potential to become an important example of embodied intelligence transitioning from technical demonstrations to industrial implementation in China.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
