Zhongyi Embodied Raises Nearly ¥100M Angel Round to Build Robot Operating System
Industrial capital is betting on the "robot operating system" track. On September 21, Zhongyi (Shenzhen) Embodied Technology announced it had completed an angel round of nearly 100 million yuan, led by the controlling shareholder of Anda Intelligent, with Shenzhen-Hong Kong Hengtian and Beijing Junrui participating.
Just two months earlier, the company had closed a seed round in the tens of millions of yuan, pushing its post-money valuation past several hundred million yuan. Two rounds in three months is rare in the embodied intelligence sector. Its Pre-A round is already underway, with several top institutions having expressed interest and started due diligence.
The proceeds will be directed entirely to three core areas: advancing hardcore underlying technologies for embodied intelligence, iterating a complete product matrix, and recruiting top technical talent. The company will also participate in building industry standards for physical robot interaction and promote coordinated ecosystem development.
Zhongyi Embodied was only registered in April 2026 in Shenzhen's Nanshan district. Its parent, Zhongyi Group, is an AIoT veteran founded in 2016 with over 400 employees, more than 50% of whom are R&D engineers; it closed a nearly 100 million yuan A round in late 2023.
Unlike most players focused on robot bodies, Zhongyi Embodied targets a more underlying position: building an "operating-system-level" foundation and becoming an intelligent base provider for the robot industry.
Technically, it has developed a three-tier architecture: a cloud brain for global scheduling and model training, an edge midbrain for regional coordination and real-time decisions, and a device-side cerebellum for single-unit control. Standardized interfaces connect the layers and are compatible with mainstream chips.
Through its physical-world OS and device-side foundation, the company says it can unify access, management, and scheduling of all heterogeneous devices and agents within a space.
Beyond the tech stack, Zhongyi Embodied's "Tiangong Ji" full-lifecycle service platform, launched in 2026, is even more noteworthy. It offers an asset-light, integrated model for robot leasing, secondary development, maintenance, and supply chain services, and has aggregated over 40 leading hardware manufacturers.
Public information indicates the company already holds nearly 100 million yuan in maintenance and leasing orders.
Placing Zhongyi Embodied's funding in the industry context makes the signal clearer. In 2026, the embodied intelligence sector is visibly stratifying: some companies build bodies (the Figure route), some build foundation models (the Generalist AI route), and others build operating systems and base platforms.
A 21st Century Business Herald analysis noted that "the real competition in embodied intelligence today is not simply a model contest, but a contest over one's position in the industrial division of labor."
Zhongyi Embodied has chosen the base layer. That choice has a practical basis: 76% of domestically produced robots still rely on ROS, and the self-sufficiency rate of underlying systems is below 40%.
Huawei-affiliated Deep OpenHarmony has led the launch of the M-Robots open-source community, building a distributed heterogeneous robot operating system based on OpenHarmony, with the same goal of "moving robots from单体 intelligence to swarm intelligence."
Industry analysts argue that as embodied intelligence moves from proof-of-concept to large-scale deployment, the value of underlying infrastructure is rapidly emerging. By choosing the "operating system" path, Zhongyi Embodied avoids the red ocean of robot bodies and hits a critical need on the eve of an industry explosion.