Leading Embodied AI Brain Company Secures Hundreds of Millions in RMB in Financing, with Multiple Overseas Funds Increasing Investments; Previously Backed by Alibaba | Exclusive Report from Yingke

Author | Huang Nan

Editor | Yuan Silai

Yingke has learned that Qiongche Intelligence, an embodied intelligence enterprise, has recently completed hundreds of millions of yuan in Series A financing. This round was led by C Capital, with multiple overseas industrial players such as Sea Limited and Puhua Capital, as well as leading domestic financial investors following suit. The old shareholder, Prosperity7 Ventures, also made an additional over - subscribed investment.

This financing came just a few months after the company received an investment from Alibaba Group in October 2025.

Among them, C Capital has rich industrial and channel resources in Asia, Europe, and Australia. As a leading Internet platform in Southeast Asia, Sea Limited's e - commerce, logistics, and digital payment ecosystems are highly compatible with the implementation scenarios of embodied intelligence. The participation of these two investors will help Qiongche Intelligence expand its global market by providing strategic access and local collaborative capabilities, and accelerate the implementation of the "data - model - scenario" closed - loop in diverse overseas environments.

Founded in November 2023, Qiongche Intelligence is one of the earliest domestic startups focusing on the field of embodied intelligence.

Lu Cewu, the co - founder of the company, is currently the deputy dean of the School of Artificial Intelligence at Shanghai Jiao Tong University, a specially - appointed professor of the Changjiang Scholars Program, the winner of the best papers at IROS and ICRA, and has been nominated for the best paper at RSS twice. He has published more than 200 papers in top international conferences/journals such as Nature, T - RO, IJRR, and RSS. Co - founder Wang Shiquan is a doctoral graduate from Stanford University, who has long been engaged in research on biomimetic dexterous manipulation and artificial intelligence. He founded the general robotics company FeiXi Technology in 2016 and has complete experience from technology R & D to industrialization. Many other core members come from fields such as robot systems and large - model algorithms, forming a composite team with both profound R & D experience and industrial implementation capabilities.

The real - world demands for commercial implementation often occur in dynamic, non - standardized, and uncertain scenarios, requiring robots to understand ambiguous instructions, autonomously plan multi - step tasks, and respond to environmental changes in real - time. This long - chain closed - loop ability has become the key threshold to distinguish between laboratory demonstrations and industrial applicability.

The core product of Qiongche Intelligence, Noematrix Brain (Qiongche Embodied Brain), provides robots with a complete decision - making closed - loop ability from instruction understanding, task planning to environmental perception, and has currently achieved large - scale deployment across different bodies and scenarios. In terms of robot forms, it has been adapted to various bodies such as wheeled single - arm, wheeled double - arm, and humanoid double - arm robots; in terms of implementation scenarios, it has covered real - world operating environments such as retail pharmacies and hotel laundries.

On the one hand, through a unified model architecture, this product can efficiently adapt to multiple bodies, which helps to lower the hardware threshold and deployment costs for robot intelligence; on the other hand, the well - verified scenario generalization ability in differentiated scenarios also reflects the system scalability of the model from single - task to multi - task and cross - scenario operations.

Specifically in scenario applications, the evolution of the capabilities of Noematrix Brain is also prominently demonstrated in complex environments such as intelligent pharmacies. When the robot brain receives an order, it does not mechanically execute a series of pre - programmed actions but needs to dynamically complete the task chain of "receiving the order - navigating to the warehouse - locating the medicine - precisely grasping it - checking and packing".

Qiongche Smart Pharmacy (Image source/Enterprise)

This requires it to have three key capabilities: efficiently planning the optimal operation sequence when multiple tasks are concurrent, precisely navigating to quickly reach the storage location, completing end - to - end medicine location, grasping, and putting it into the packing bag, and being able to assist in the learning and coverage of complex tasks and extreme cases through human - robot collaboration, achieving an evolution from "able to execute" to "high - efficiency and high - success - rate", and stably completing the entire process from order to shipment.

Next, Qiongche Intelligence will further integrate edge - cloud collaboration, post - model training, and a new generation of end - to - end embodied large models to provide stronger scenario generalization capabilities, enabling robots to continuously optimize decision - making strategies during actual operation.

Behind the stability and reliability of decision - making lies a more fundamental support, namely high - quality and large - scale real - world scenario data. The evolution of the model highly depends on learning from interaction data in the physical world. However, the scarcity of high - quality real - machine data and the high cost of data collection have always been the core bottlenecks in the industry.

In terms of data collection methods, Qiongche Intelligence was the first to propose the "accompanying data collection" solution. Through self - developed data collection devices such as the exoskeleton CoMiner and the portable RoboPocket, data collection is extended to diverse real - world environments such as homes, offices, and industries in a lightweight and low - cost manner, building a physical - world database for embodied digital genes. Currently, Qiongche Intelligence has hundreds of thousands of hours of high - quality real - machine data.

Qiongche's Accompanying Data Collection Devices (Image source/Enterprise)

In the medium - to - long - term competition in the development of the embodied intelligence industry, the core lies in "data acquisition efficiency" and "scenario definition ability". Qiongche Intelligence is trying to achieve a closed - loop iterative path of "data - model - scenario", that is, data is generated from real - world scenarios, the model's capabilities are improved through data training, and models with stronger generalization capabilities can be quickly deployed and implemented in new scenarios, which in turn feed back more available data.

Yingke has learned that in addition to the full - process implementation in the intelligent pharmacy scenario, Qiongche Intelligence has started cooperation with enterprises in fields such as hotel services and logistics sorting to explore the intelligent application in scenarios such as laundry services and item organization. The overseas market layout is also advancing simultaneously. The company is responding to the needs of overseas customers and adjusting technical solutions according to the scenario characteristics of different regions.

In addition, Qiongche Intelligence has reached strategic cooperation with leading humanoid robot enterprises such as Ubtech and domestic and foreign data collection centers. On the one hand, it expands its data reserves; on the other hand, it builds a data collaboration channel for overseas markets.

After this round of financing, Qiongche will continue to promote the R & D and iteration of large models, focusing on the training of the base model. With a model featuring strong generalization ability and autonomous decision - making, it will be adapted to a variety of standardized robot hardware. The company plans to release a smart pharmacy solution after the Spring Festival. Its embodied brain can, with high robustness, handle medicines in different packages and switch different end - effectors for grasping, providing core technical support for the implementation of this solution.