Chuangwei believes that highly humanoid robots are an entirely new category, and there will be space for new high-end brands to survive.

Written by丨Li Anqi

Edited by丨Gong Fangyi

We have learned exclusively that recently, the embodied intelligence company Chuangwei Power (Kinetix AI) completed a series of angel + round financing exceeding 500 million yuan, with investments from Singapore's Temasek Holdings under VC institutions such as Xiangfeng Capital, Fangguo Capital, and Wanshi Capital. It is understood that this financing will be used to accelerate the iteration of "data-model-ontology" and promote the landing of embodied intelligent products.

Chuangwei Power was established in September 2025, focusing on highly humanoid robots. Before this round of financing, Chuangwei Power had completed three rounds of financing. The company currently has nearly 200 employees and launched its first full-size humanoid robot product, KAIBot, in April this year.

Chuangwei has three founders, CEO Yu Jie, who was the former general manager of commercial vehicle autonomous driving at Huawei, was at Huawei's level 22, and was responsible for the commercialization of Huawei's L4 mining truck; co-founder Luo Ping is the deputy dean of research at the School of Computer Science and Data Science at the University of Hong Kong, and head of HKU MMLAB (Hong Kong University Multimedia Laboratory), with his PhD under Tang Xiaowu. The other co-founder is Zheng Cunyuan, who was the head of the entire machine in the robot department of a leading new force automotive company before starting his own business.

Talent and funding are entering the embodied intelligence industry. Different companies have chosen different paths: some develop low-cost ontologies, some use wheeled chassis, some enter fixed scenarios first, and some focus on models and data. Most companies cannot fulfill the technical capabilities and landing progress advertised previously, so when seeking financing today, they must better articulate how the entrepreneurial path ultimately leads to product formation and revenue narrative.

One of the co-founders, Zheng Cunyuan, previously communicated with "Wandian Auto", believing that the current lower limit of embodied intelligence capabilities is still too low, "low enough that directly making products now seems a bit foolish." Chuangwei plans to first develop a highly humanoid ontology and dexterous hand while also engaging in the development of first-person data (egocentric data) and world models, ultimately shaping commercially viable products. This also means higher costs, more complex supply chains, and longer verification cycles.

Most infrastructure in real-world environments is designed according to human body shapes and activities, and robots also need to adopt forms that are closer to human. The closer the robot's form is to that of a person, the easier it is to utilize human motion data and adapt to existing environments. The closer the ontology is to the planned product form, the less the algorithm's results are affected by subsequent hardware updates.

Zheng Cunyuan has a background in motion control and machine learning algorithms. During his graduate studies, he observed that doctoral candidates in the robotics industry often needed to spend several years building hardware before they could begin algorithm research. According to his description, to publish an algorithm paper, someone spent the first three years on hardware, the fourth year on algorithms, and the fifth year writing the paper. Hardware, ontology form, and models are updated every few years, making it difficult for previously accumulated results to be directly applied to the new generation of systems.

Chuangwei's first-generation highly humanoid robot ontology KAIBot stands 1.73 meters tall and weighs about 65 to 70 kilograms, with a total of 115 degrees of freedom. The team referenced over 200 degrees of freedom of the human body during the design.

Header image source: Chuangwei Power

- FIN -

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。