Heart Capital Invests in Xi Xuan Chuang Zhi's Angel Round to Advance Embodied AI Edge-Side General-Purpose High-Performance Computing Platform | Heart Star PORTFOLIO

Heart Capital has completed a 500 million RMB angel round investment in Xixuan Chuangzhi Technology (Wuxi) Co., Ltd. ("Xixuan Chuangzhi").

Xixuan Chuangzhi was jointly founded by MetaX and UBTECH, focusing on embodied AI edge chips and computing platforms. The company's first in-house chip is positioned as a "brain-and-cerebellum integrated" SoC for robots, benchmarked against NVIDIA Thor. It is being developed around the needs of embodied AI scenarios — multimodal perception, cognitive decision-making, and motion coordination — with parallel work on chips, compute modules, and supporting software.

Yan Han, founding partner of Heart Capital, said: "We have long been bullish on the enormous value and market potential of domestic computing power. Five years ago, we invested in the computing GPU company MetaX; last year, betting on the inference market, we made an early investment in Xi Wang, which focuses on inference GPU chips. This angel round in Xixuan Chuangzhi marks the beginning of our布局 in edge AI chips and another show of support for the MetaX and UBTECH teams. Xixuan Chuangzhi combines MetaX's GPU technology积累 with UBTECH's experience in robot application scenarios. We look forward to accompanying the team as they translate these assets into competitive products and drive domestic computing power into more real-world scenarios."

The funds will be primarily used for core chip and computing platform R&D, software toolchain development, team expansion, and joint validation with robot manufacturers and industry customers to advance mass production and delivery. The company is currently developing its first in-house robot main-control SoC and accompanying hardware-software system, with plans to provide a development and validation platform to industry partners in 2027 for adaptation and testing, and to launch its first SoC product and complete embodied AI solution in 2028.

Chip design driven by real robot tasks

Robots performing tasks in real environments must handle perception, reasoning, and motion decision-making under tight constraints on power, heat dissipation, and space. Model deployment, sensor data processing, and the coordination of computing and control systems all affect how well a robot responds to environmental changes and completes tasks.

Based on this understanding, Xixuan Chuangzhi has redefined the domain-control solution for the robot "brain-and-cerebellum integrated" architecture. Combining MetaX's积累 in GPU architecture, chip engineering, and software ecosystems with UBTECH's experience in robot hardware, system integration, and application scenarios, the company brings real robot task requirements into chip design. From the product definition stage, it considers model runtime requirements alongside energy efficiency, interface adaptation, and system deployment, driving coordinated validation of chip design and robot systems.

In-house chip architecture design

Xixuan Chuangzhi is currently advancing the architecture design of its in-house main-control chip. In its early phase, the company is primarily targeting robot applications in semi-structured environments, developing a general-purpose edge computing platform around robots' perception, understanding, and autonomous decision-making needs across different tasks.

On the AI hardware acceleration technology roadmap, the company uses general-purpose GPU (GPGPU) as the core computing unit and, around a CUDA-like development ecosystem, is working to adapt mainstream AI frameworks, operators, and models to the edge platform. It plans to build, together with ecosystem partners, a complete toolchain covering model development, simulation evaluation, deployment debugging, and production operation — supporting smooth migration of robot and model makers across different computing platforms, reusing existing models and applications, and reducing the software rework, adaptation validation, and deployment costs caused by changing underlying computing hardware.

Bringing together chip R&D and industry application experience

For the R&D and delivery of its embodied AI computing platform, Xixuan Chuangzhi has assembled talent and experience across SoC architecture, GPU architecture, chip engineering, software ecosystems, robot systems, and industry applications, closely linking chip definition, software development, scenario validation, and commercialization needs.

CEO Yang Xianbo previously served as Senior Vice President and President of the Industry BG at H3C, with extensive experience in industry markets, solutions, and commercial operations. He will drive deeper collaboration with robot makers, model makers, and industry customers to accelerate product validation and application rollout.

Building an embodied AI computing foundation with industry partners

At a critical window for the large-scale adoption of embodied AI, Xixuan Chuangzhi will continue to deepen collaboration with shareholders, investors, and industry partners to jointly advance the R&D, validation, and rollout of edge computing platforms.

Chen Weiliang, Chairman of Xixuan Chuangzhi and founder of MetaX, said:

"Device-cloud collaboration is an important path toward large-scale embodied AI adoption. Xixuan Chuangzhi closely integrates general-purpose high-performance GPUs and a complete software stack with robot scenarios, driving the coordinated evolution of cloud training, simulation, and edge applications. We will uphold independent innovation and open compatibility, so that developers' existing work carries forward and model innovations translate into real applications faster, working with industry partners to build a solid domestic computing foundation for embodied AI."

James Zhou, Vice Chairman of Xixuan Chuangzhi and founder of UBTECH, said:

"For robots to enter real-world scenarios, computing platforms must work closely with the hardware body and algorithms. We want to bring requirements accumulated from robot R&D and application into chip design earlier, and validate products through real tasks, so that R&D stays closer to what robot companies actually need."

Yang Xianbo, CEO of Xixuan Chuangzhi, said:

"The value of embodied AI ultimately comes down to whether robots can solve real problems. We will focus on the development and delivery needs of robot makers and model makers, improving chips, software toolchains, and deployment support to lower the barrier from model validation to product application — letting AI better serve production and daily life and driving embodied AI toward large-scale adoption."

Founded in 2022, Heart Capital is a venture fund focused on early-stage Chinese technology startups.

Heart Capital's team consists mainly of the founding partners and core investors of Lightspeed China Partners, along with seasoned investors from industry. The team's past investments include MetaX (688802.SH), Xpeng Motors (NYSE:XPEV, 09868.HK), Full Truck Alliance (NYSE:YMM), SUNMI (06810.HK), RoboSense (02498.HK), Ambiq Micro (NYSE: AMBQ), Hanshow Technology (301275.SZ), FinVolution (NYSE:FINV), HERE (NASDAQ:HERE), as well as LandSpace, MicroNano Space, Baichuan, Yunmanman Cold Chain Logistics, World Logistics, FanDeng Reading, and Lanhu.

Rooted in China with a global outlook, Heart Capital is dedicated to accompanying and supporting early-stage founders with the potential to become future world-class companies in China's tech sector. Heart Capital champions the value of "heart," believing technology can be a bridge that connects people. Heart Capital looks forward to accompanying more young Chinese entrepreneurs as they step onto the world stage.