Xixuan Chuangzhi Technology (Wuxi) Co., a Chinese startup building chips and software for "embodied intelligence" — the industry term for AI that controls physical robots — has raised roughly 500 million yuan, or about $70.5 million, in an angel round. Jinyuan Capital and Morpheus Ventures co-led the financing, which was backed by humanoid robot maker UBTech and domestic GPU designer Moxin, according to a report from 10jqka dated November 11, 2026. No valuation was disclosed.
What is an "embodied intelligence" chip, and what does Xixuan build?
Robots that walk, balance or grasp objects need to process camera feeds, inertial sensors and motor commands in real time, directly on the machine, because sending that data to a distant cloud server and waiting for a response is far too slow for anything that has to stay upright or catch a falling object. Xixuan is building the chips, computing platforms and software toolchains designed to sit inside a robot and handle exactly that workload.
The company's stated goal is to become a leading domestic maker of "embodied intelligence edge-side computing platforms" — essentially the onboard brain hardware for humanoid and industrial robots — rather than a general-purpose chipmaker. Xixuan is registered in Wuxi, with research and development teams spread across Shanghai, Shenzhen and Beijing, and its near-term plan is to validate its chips jointly with robot manufacturers before moving into mass-production delivery.
Why would a humanoid robot maker and a GPU designer back a chip startup?
UBTech, one of China's largest humanoid robot makers and known for its Walker line of bipedal robots, has an obvious interest in securing chips tuned specifically for its own machines rather than depending entirely on general-purpose processors bought off the shelf. Moxin, a domestic GPU chip designer, contributes silicon design expertise that a pure software or robotics company would otherwise lack.
Together, the two backers give Xixuan both a flagship customer and hardware credibility from day one — a vertical-integration strategy similar to how other hardware companies design their own chips rather than rely on third-party suppliers, aiming for tighter performance and cost control across the whole robot.
How does this fit into the broader robotics chip race?
Nvidia's Jetson and Thor families currently dominate robotics edge computing worldwide, powering a large share of humanoid and mobile robot prototypes outside China. Chinese robot makers have growing incentive to find domestic alternatives as export controls on advanced chips tighten, and Xixuan is one of several Chinese startups now chasing that opening, competing for the same pool of robot manufacturers that need onboard compute for navigation, vision and motor control.
What happens next?
Xixuan's immediate priority, per the funding report, is joint validation of its chips and software toolchains with robot manufacturers and industry customers, working toward mass-production delivery. Given UBTech's involvement, its first real-world proving ground is likely to be UBTech's own humanoid lineup, with sales to other robot makers expected to follow if the chips perform well in that deployment.