Rushi Robotics Closes 100 Million Yuan Pre-A Round to Build Embodied Intelligence Platform for Real-World Elder Care Scenarios | Tongcheng Family

Background

Recently, Dalton Venture portfolio company Shanghai Rushen Robotics Technology Co., Ltd. (hereinafter referred to as "Rushen Robotics") announced the completion of a 100-million-yuan Pre-A funding round, with investment from Qingsong Capital, leading listed computing power company RangeRay Technology, and Pinghu Zexin. This round of financing will strongly support Rushen Robotics' continuous operation in real elderly care scenarios, data generation, and model iteration. This marks another key milestone in the company's capital development following Dalton Venture's sole investment in Rushen Robotics' angel+ round in September 2025.

Founded in June 2023, Rushen Robotics is an embodied intelligence robotics company focused on smart health and elderly care scenarios, dedicated to building intelligent health and elderly care robots that can truly enter homes and accompany the elderly.

Rushen Robotics chooses to start from real elderly care scenarios, acquiring task data through continuous robot operation in real environments, and using data to drive model capability iteration, gradually forming a closed-loop evolutionary system of "scenario—data—model—scenario." In this system, real scenarios are used for continuous capability generation and validation, operational data becomes the core training asset, and models continuously evolve through real tasks and feed back into scenario capabilities.

Centered on this technical approach, the company has launched the world's first dual-modal embodied elderly care robot combining human-carrying and service capabilities, achieving the integration of mobility assistance and proactive care.

Swipe left or right to see more

At the core capability level, the product is equipped with a seven-degree-of-freedom 20kg high-load force-controlled compliant robotic arm, enabling stable operation and safe interaction in complex institutional and home environments. It has built an autonomous task capability system around elderly care scenarios, covering high-frequency service actions such as feeding, delivery, transport, and care assistance. Through a bilateral force-feedback embodied data collection system, it achieves force-feedback data accuracy better than ±1N in real interactions, providing long-term high-quality data support for model training. At the safety interaction level, the system achieves fine adjustment of contact force through force control and compliant control, allowing the robot to maintain stable and safe human-robot interaction capabilities under high-degree-of-freedom operations, meeting the core requirements of "low impact, high reliability, and predictable behavior" in elderly care scenarios, ensuring reliable operation in complex real-world scenarios such as elderly care.

Swipe left or right to see more

To date, Rushen Robotics has accumulated over 1,000 hours of real elderly care scenario operational data collection, completing the critical leap from simulation validation to real-world closed-loop. Based on multi-modal data from real scenarios, Rushen Robotics is gradually building an embodied intelligence system for elderly care scenarios from embodiment-data-model, continuously improving robot autonomy through real home data closed-loop: QiCare OS. The system includes a three-layer collaborative model architecture of VLM brain + QiCare SkillFormer cerebellum + dedicated Safety Supervisor, where the VLM handles cognitive planning, the Skill Policy cerebellum is driven by QiCare SkillFormer, a proprietary policy model independently developed by Qijia, forming elderly care scenario-specific policy model assets and atomic skills. At the same time, it innovatively proposes a Safety Supervisor safety strategy, responsible for determining safety boundaries in real scenarios, autonomy level judgment, and human takeover warnings, ensuring robot safety assurance in real scenarios. Rushen has thus become the first domestic team to build an autonomous model for elderly care services.

In relevant deployments for "intelligent elderly care service robots," the National Development and Reform Commission and other departments have explicitly proposed promoting a real-scenario-data-collection and application validation system centered on real elderly care scenarios, emphasizing field operation and task validation in institutional and home environments to form standardized data systems and accelerate the transition of intelligent elderly care robots from experimental validation to scaled application. Rushen Robotics, through continuous operation in real elderly care institutions and home scenarios, forms structured data and drives model iteration.

On the institutional side, the company has carried out continuous operation validation in elderly care and health care institutions in multiple regions including Shanghai and the Beijing-Tianjin-Hebei area, with cumulative operation exceeding 45 days, and is simultaneously advancing 5–10 institutional scenario POC testing projects.

Swipe left or right to see more

On the home side, the company's first real home entry scenario has been implemented, and the "Thousand Households Plan" has been launched to promote robots entering broader family service systems. At the same time, as a participant in the Ministry of Industry and Information Technology's "Intelligent Elderly Care Service Robot Paired Tackling and Scenario Application Pilot," the company is deeply involved in the national-level elderly care robot scenario validation system construction.

On the commercialization path, Rushen Robotics is transitioning from a single product delivery model to a long-term operation system of "robot + service + data," forming data assets through continuous operation in real scenarios, and using this to drive capability iteration and service capability upgrades, promoting the shift from equipment attributes to long-term service attributes.

Supporting this system is a composite team with continuous successful entrepreneurship experience, international embodied intelligence research background, and medical robotics industrialization experience. Core members come from universities and research institutions including University of Hamburg, Tsinghua University, Shanghai Jiao Tong University, and Oxford University, covering key technical directions such as robot embodied intelligence, control planning, and human-robot interaction, and possessing medical robotics and health and elderly care scenario industrialization experience, forming a composite team capability system of "embodied intelligence + medical health and elderly care."

Dalton Venture Partner Lincheng Lin stated: "We have long been optimistic about technological breakthroughs and service upgrades in the elderly care and rehabilitation robotics track. Rushen Robotics insists on starting from real elderly care scenarios, acquiring task data through continuous operation, and using data to drive model capability iteration, forming a closed-loop evolutionary system of 'scenario—data—model—scenario.' Qijia Q1 is not a functional demonstration product, but a robot that can truly serve the elderly and caregivers. We look forward to witnessing its widespread application in real usage environments, and will long-term accompany the team in jointly promoting the construction of a smart health and elderly care ecosystem."

In the future, Rushen Robotics will continue to expand real scenario deployment scale, accelerate embodied intelligence capability system construction, and promote robots from elderly care services to broader family service scenarios.

---

ID: daltonventure

Long press to follow

Recommended Reading