36Kr Review of Financing for 20 Embodied Intelligence Companies: Tsinghua-Backed Teams and Young Founders Attract Concentrated Investment

In 2026, the pace of fundraising for embodied intelligence startups continued to accelerate. ManuPrax Robotics completed two rounds of financing totaling nearly ten million US dollars about two months after its registration; Luobo Party, founded by Huang Yi, who was born in 2004, counted Xiaomi and CATL among its investors; and Zhi Pingfang disclosed in June that its new series of financing totaled close to 5 billion yuan.

The team backgrounds that capital focuses on are also fairly concentrated. A statistic released by IT Juzi in August reviewed 1,155 embodied intelligence entrepreneurs, of whom 96 have a Tsinghua background, spread across 75 companies; in that statistic, the median financing for companies with a Tsinghua background was 210 million yuan. This is a statistical result for a specific sample and cannot directly explain the fundraising performance of any single company.

However, although they likewise carry the labels of prestigious universities and young teams, these companies' product directions are not consistent: some make complete humanoid robots, some develop robot "brains," some focus on dexterous hands, and some provide data and simulation environments for robots.

The following selects 20 companies that are representative in terms of founding time, young founding teams, or financing scale, and reviews their founders, technology products, and financing progress one by one. The information is current as of November 3, 2026; the numbering follows the order of the text; ages are estimated based on publicly available birth years; and financing is marked respectively as a single round, a series of rounds, or a cumulative amount.

ManuPrax Robotics completed its registration in September 2026. CEO Li Puhao was born in 2001, about 25 years old, studied as an undergraduate in the Department of Automation at Tsinghua University, and is currently pursuing a doctorate under the supervision of Zhu Songchun. CTO Chen Zeyuan was born in 2000, about 26 years old, holds a master's degree from Peking University, and has participated in dexterous manipulation research such as ClutterDexGrasp and CADGrasp.

The company develops a software-hardware integrated dexterous hand solution, with a focus on fine motion control: after the robot understands "pick up the cup," how the fingers approach, apply force, and adjust when the object slips is the problem it is trying to solve. The team plans to train motion models with reinforcement learning.

On November 2, the company disclosed that it had completed two consecutive financing rounds totaling nearly 10 million US dollars, led respectively by Alphaist Partners and Yaotu Capital, with follow-on investment from Jieli Angel and the actual controller of a certain component industry party.

02 | LatentVerse: founded by a Tsinghua doctoral student, raised hundreds of millions of yuan in a seed round within months of establishment

LatentVerse was founded in May 2026. Founder and CEO Hu Yucheng was born in 2001, about 25 years old, is a doctoral student at Tsinghua University's Institute for Interdisciplinary Information Sciences, studied under Chen Jianyu, founder of Xingdong Jiyuan, and previously participated in embodied foundation model research at ByteDance Seed.

The company proposes the Unified Tactile Action Model (UTAM), which treats touch as a native input and places environment understanding, prediction, and action generation in a unified architecture. Its goal is to enable robots to continuously correct actions based on feedback after contacting objects while performing tasks, with key research directions including dexterous manipulation, model pretraining, and on-device deployment.

In August 2026, the company announced the completion of a seed round of hundreds of millions of yuan, with investors including Gaorong Ventures, Qingliu Capital, Inno Fund, Zhiyuan, and Xingdong Jiyuan.

Kunlun Xing Robotics was registered on March 16, 2026. Founder and CEO Ren Geng was born in 1982, about 44 years old, and previously served as general manager of Huawei's Myanmar branch, president of Alibaba Cloud China, and president of ENN Group. Co-founder Lang Xianpeng previously served as senior vice president of Li Auto and was long responsible for intelligent driving R&D.

The company simultaneously develops general-purpose humanoid robots and embodied models, with the Kunlun World Model (KWM) as the core, attempting to introduce physical causality into robot decision-making. Hardware R&D covers drive motors, reducers, and integrated joint modules, adopting a route of co-developing the body and the model.

On June 23, the company announced the completion of three series of financing rounds, with a cumulative scale of several billion yuan. Investors include Gaorong Ventures, Gaorong Ventures, CAS Star, Zhongding Capital, Huaye Capital, Innovation Works, Xin Capital, and Jianfa Capital.

04 | Physis: 22-year-old Peking University co-founder, world model secures over 100 million US dollars in financing

Physis was founded in 2026, co-founded by Chen Boyuan, Ji Jiaming, and others, and incubated by the Beijing Academy of Artificial Intelligence. Chen Boyuan was born in 2004, about 22 years old, graduated from Peking University Yuanpei College in the artificial intelligence direction, and serves as co-founder and CRO; the company's CEO is Ji Jiaming.

The company develops a general world foundation model and released "Wujie · Physis-v0.1" in June. Its research focuses on predicting the physical state after an action occurs, providing model capabilities for robot manipulation, physical simulation, and industrial scenarios. In July, the company established a joint world model laboratory with Peking University.

In March, the company completed a first round of financing exceeding 10 million US dollars, and in June completed a seed++ round exceeding 100 million US dollars. The latter was participated in by Matrix Partners China, Guanghe Ventures, Morningside Venture Capital, BAI Capital, Zhongding Capital, and others, with strategic investment from Ant Group and additional investment from Gaorong Ventures and Yanyuan Ventures.

05 | LiberAI: founded by a post-00s Tsinghua doctoral student, cumulative financing exceeding 1 billion yuan in less than a year since establishment

LiberAI's corporate entity is Beijing Jiangxian Technology Co., Ltd., founded in December 2025. Founder Liu Songming was born in 2000, about 26 years old, is a doctoral student in Tsinghua University's Department of Computer Science and Technology, a recipient of the Tsinghua University Special Scholarship, and studied under Zhu Jun. Before founding the company, he led research on the RDT series of embodied foundation models.

The company focuses on physical world models and dexterous manipulation, emphasizes large-scale pretraining, and attempts to align video data with data containing robot actions and physical interaction information, in order to alleviate the problem of insufficient robot training data. In 2026, the company has publicly released a high-degree-of-freedom dexterous manipulation foundation model.

As of September 2026, the company has raised more than 1 billion yuan in cumulative financing. Investors include ZhenFund, Sequoia China, Shunwei Capital, Meituan Dragon Ball, Innovation Works, and others; the several-hundred-million-yuan Pre-A round disclosed in June was led by Shunwei Capital.

06 | OriginFlow: a 25-year-old Tsinghua doctoral student turns electromyographic signals into robot training data

OriginFlow was founded in 2025. Founder and CEO Qin Shentao was born in 2001, about 25 years old, graduated from Harbin Institute of Technology for his undergraduate degree, and later pursued a doctorate at Tsinghua University.

Unlike companies that directly manufacture complete machines, OriginFlow mainly addresses the problem of robot data collection. Its NeuroScale solution uses a non-invasive motor neural interface as its entry point, collecting electromyographic and other signals during human operation, attempting to supplement action intention, force exertion, and contact information that are difficult to record with video alone, thereby providing data for embodied model training.

In May 2026, the company disclosed that it had successively completed angel, strategic, and Pre-A1 financing rounds, totaling more than 500 million yuan. Investors include BlueRun Ventures, Oasis Capital, 58 Strategic Investment, Puhua Capital, Shuimu Tsinghua Seed Alumni Fund, Lisi Capital, and Yuanhe-affiliated institutions, among others.

07 | RoboParty: a founder born in 2004, moving from a dorm-room open-source project to a complete-machine platform

RoboParty was founded by Huang Yi, born in 2004. He is about 22 this year and graduated from Harbin Institute of Technology. During his undergraduate years, Huang Yi developed the bipedal robot AlexBot in his dormitory and opened up the related design and code.

Founded in 2025, the company develops open-source bipedal robots for developers, research, and the education market, with business covering joint modules, complete machines, motion control, and development tools. In January 2026, the company open-sourced the ROBOTO Origin prototype, releasing hardware drawings, control code, and motion control algorithms.

In July 2026, the company disclosed that it had successively completed an Angel++ round and a Pre-A round, totaling nearly 500 million yuan across the two rounds. The former received additional investment from existing shareholders such as Shunwei Capital, Matrix Partners China, Xiaomi Strategic Investment, and SenseTime Guoxiang; the latter was exclusively invested by CATL.

08 | Zerith: a post-00s Tsinghua team, first having robots do cleaning and retail

Zerith Robotics was established in Hefei in January 2025. Founder and CEO Min Yuheng was born in 2000, about 26 years old, studied in the Mingyue Innovation Experimental Class at Chongqing University for his undergraduate degree, and later pursued a master's degree in artificial intelligence at Tsinghua University. The company's core team comes from the Tsinghua AI&Robot Laboratory.

The company develops the Zerith V series embodied action models, the H1 wheeled-arm robot, and a supporting data platform, with application directions including cleaning, product grasping, and unmanned retail. Its H1 robot has been used in robot shops in Hefei and other places to identify, grasp, and deliver products.

In April 2026, the company announced the completion of a new round of more than 100 million yuan in financing, led by Runze Group, with Ningbo Dongli, Chuangye Jieli, and Pinghu Zexin participating. In September, the company disclosed another financing round, with investors including Wuxi Guolian Xizhou Fund, Qingke Chantou, Cowin Capital, and Hefei Chantou, among others.

09 | Lingchu Intelligence: co-founded with a post-00s reinforcement learning researcher, angel and Pre-A rounds totaling 2 billion yuan

Lingchu Intelligence was founded in 2024, co-founded by Wang Qibin, Chen Yuanpei, and others. CEO Wang Qibin previously worked in R&D at Yunji Technology and JD Robotics; co-founder and reinforcement learning lead Chen Yuanpei was born in 2001, about 25 years old, and previously studied and conducted research in Yang Yaodong's team at Peking University and Fei-Fei Li's laboratory at Stanford University.

The company focuses on dexterous manipulation, developing the PSI series models and the Psi-SynEngine data collection engine, attempting to use human operation data to train robots. Its technology products target scenarios such as logistics warehousing and manufacturing, and the company says it has completed small-scale validation in real logistics customer warehouses.

In March 2026, the company announced the completion of angel and Pre-A financing rounds, totaling 2 billion yuan across the two rounds. Subsequently, the company completed a Series A round, with participation from Tuopu Group, Ruicheng Fund under Chery Holding, Lens Technology, 37 Interactive Entertainment, Wuhu Municipal Investment Holding Group, and Fosun Chuangfu, among others.

10 | Linker Hand: a post-00s co-founder in charge of design, laying out three dexterous hand routes

Linker Hand was established in 2023, co-founded by Zhou Yong, Jia Xiaoyou, and others. Jia Xiaoyou was born in 2003, about 23 years old, is a graduate student in the interdisciplinary information art and design program at Tsinghua University, and serves as co-founder and design director, responsible for the appearance and structural design of the Linker Hand series of dexterous hands.

The company's products cover three technical routes: linkage, tendon-driven, and direct-drive, including the L20, L30, and O30/R30 series, and also provide the Open TeleDex teleoperation system and the LinkerSkillNet skill library. Customers target humanoid robot complete-machine enterprises, research institutions, and industrial automation scenarios.

In February 2026, the company completed a Series B round of nearly 1.5 billion yuan; in April it announced the completion of a B+ round, with investors including the Zhongguancun Science City Fund, Bank of China Asset Management, and Fosun Chuangfu, among others. The specific amount of the B+ round was not disclosed.

11 | UniX AI: a post-00s Yale doctoral student returns to China, with robots starting from hotel scenarios

UniX AI was founded in 2024. Founder and CEO Yang Fengyu was born in 2000 and is about 26 years old. He completed his undergraduate studies in computer science at the University of Michigan, then went to Yale University for a doctorate, and in 2024 paused his studies to return to China and start a business. Chief scientist Wang Hesheng is a professor at Shanghai Jiao Tong University.

The company develops robot products such as Wanda and Panther. Wanda 2.0 adopts a wheeled dual-arm form factor and targets scenarios such as hotels, property management, and retail. Its technology system includes the UniFlex imitation learning framework, the UniTouch tactile model, and task planning, with a focus on operations such as cleaning, tidying, grasping, and carrying.

In March 2026, the company announced the completion of a new round of equity financing of nearly 300 million yuan; in August it received a strategic investment of tens of millions of yuan from Keli Sensing. According to media reports in October, the company had completed five rounds of financing, with cumulative proceeds approaching 1 billion yuan.

12 | Songyan Power: a 28-year-old Tsinghua entrepreneur who brings consumer-grade humanoid robots to under 10,000 yuan

Songyan Power was founded in September 2023. Founder Jiang Zheyuan was born in 1998 and is about 28 years old. He was admitted to the Department of Electronic Engineering at Tsinghua University for his undergraduate studies and dropped out during his doctoral studies in 2023 to start a business.

The company is developing both bipedal humanoid robots and biomimetic robots, with products including N2 and "Xiaobumi." Among them, "Xiaobumi" was unveiled in 2025 at a price of 9,998 yuan, targeting the consumer, education, and developer markets, and integrating functions such as walking, dancing, voice interaction, and graphical programming.

In March 2026, the company announced the completion of a Series B round led by Chendao Capital, the industrial investment platform affiliated with CATL, with participation from Guoke Investment, Jingguosheng Fund, Jiuhe Venture Capital, and others. The Series B round totaled nearly 1 billion yuan, and as of that disclosure the company had completed nine rounds of financing.

13 | ScanArk: a 1998-born Tongji PhD entrepreneur building an interactive digital world for robots

ScanArk was founded in November 2024 by a doctoral team from the RAIL laboratory at Tongji University. Founder and CEO Peng Yun was born in 1998 and is about 28 years old; he previously participated in core projects in the 3D grid algorithm team at Huawei's 2012 Laboratory. Co-founders Su Shuai and Liu Chuangwei were both his fellow students and his teammates on the RoboCup field.

The company works on robot training infrastructure, using 3D reconstruction technologies such as 3D Gaussian Splatting (3DGS) to transform real environments into digital scenes containing object attributes, part relationships, and interaction states, allowing robots to try operations in a simulated environment, observe the results, and then verify them in the real environment.

In October 2026, the company disclosed that within four months it had completed three financing rounds—seed, seed+, and pre-angel—totaling tens of millions of yuan, with investors including Xiaomiao Langcheng, Guohai Innovation Capital, leading embodied robot companies, and industrial investment institutions.

14 | X Square Robot: a Tsinghua undergraduate and master's founder developing complete machines around embodied foundation models

X Square Robot was founded in December 2023. Founder and CEO Wang Qian completed his undergraduate and master's studies at Tsinghua University, then went to the University of Southern California for a doctorate, with research directions covering robot learning and human-robot interaction. In an early 2026 media interview, he was 37 years old.

The company adopts an end-to-end model approach, self-developing WALL-A, WALL-B, and the open-source model WALL-OSS, and developing wheeled dual-arm robots such as "Quantum 2." Its technical focus is to combine vision, language, action, and world model capabilities so that robots can complete multi-step operations in open environments.

In January 2026, the company announced the completion of a 1 billion yuan A++ round, with investors including ByteDance and HongShan. According to the company's official website in June, it had consecutively completed Series B, B+, B++, and C rounds, all four of which had closed, with the specific amounts not fully disclosed.

15 | AI² Robotics: a post-80s big-tech technical leader starts a business, with a new series of financing approaching 5 billion yuan

AI² Robotics was founded in 2023. Founder and CEO Guo Yandong is a post-80s entrepreneur who completed his undergraduate and master's studies at Beijing University of Posts and Telecommunications and his doctorate at Purdue University in the United States. He previously served as a researcher at Microsoft, chief scientist at XPeng Motors, and chief scientist of intelligent perception at OPPO.

The company self-develops embodied models such as GOVLA and the AlphaBot (Aibao) series of general-purpose robots, with application directions including automobile manufacturing, semiconductor displays, public services, and retail. In 2026, the company further proposed the NeuroVLA brain-inspired architecture, exploring ways to reduce model training costs, speed up response, and enable on-device learning.

In February 2026, the company disclosed the completion of a Series B series of financing exceeding 1 billion yuan; in June, Shenzhen News Network reported that the company had completed another series of new financing totaling nearly 5 billion yuan.

16 | Spirit AI: an industrial robotics veteran partners with a Tsinghua scholar, raising nearly 5 billion yuan in a few months

Spirit AI was founded in early 2024. Founder and CEO Han Fengtao is a post-80s entrepreneur who graduated from Zhejiang University and previously co-founded Rokae Robotics and served as its CTO. Co-founder and chief scientist Gao Yang is a post-90s researcher with research backgrounds at Tsinghua University and the University of California, Berkeley, and is currently an assistant professor at the Institute for Interdisciplinary Information Sciences at Tsinghua University.

The company develops both embodied foundation models and humanoid robots, with products including the Spirit series of models and the Moz1 robot. In 2026, the company released Spirit v1.6, continuing to advance the training of general manipulation capabilities using real-world interaction data.

In February 2026, the company completed two consecutive financing rounds totaling nearly RMB 2 billion, and in April completed another RMB 1 billion round, co-led by Shunwei Capital and Yunfeng Fund; it subsequently announced the completion of a RMB 1.5 billion Series A+ round. According to the company's combined disclosures for this period, the four rounds within a few months totaled nearly RMB 5 billion.

17 | Tashi Zhiheng: A startup founded by technology and management teams from Huawei and Baidu, with a Pre-A round exceeding USD 450 million

Tashi Zhiheng officially marks February 2025 as its formal founding date, while its R&D entity, Shanghai Tashi Zhiheng Technology Co., Ltd., was registered in July 2024. Founder and CEO Chen Yilun was born in 1984 and is about 42 years old. He earned his bachelor's and master's degrees from Tsinghua University and his doctorate from the University of Michigan, and previously served as CTO and chief scientist of Huawei's autonomous driving system. Co-founders include former Baidu executive Li Zhenyu and former Huawei "Genius Youth" Ding Wenchao, among others.

The company develops the AWE embodied large model and supporting robots, adopts a data collection approach centered on human manipulation, and prioritizes precision processes such as wire harness assembly in automotive and 3C manufacturing. In March 2026, the company publicly introduced the technical architecture and industrial applications of AWE 3.0.

In April 2026, the company announced the completion of a Pre-A round exceeding USD 450 million, led by Hillhouse Capital and Sequoia China. Previously, its two angel-stage financing rounds totaled approximately RMB 1.65 billion.

18 | Galaxea: Led by a post-90s Tsinghua alumnus, expanding from robot hardware to VLA models

Galaxea was founded in September 2023. Founder and CEO Gao Jiyang was born in 1992 and is about 34 years old. He earned his bachelor's degree from Tsinghua University and his doctorate from the University of Southern California, and previously worked on autonomous driving R&D at Waymo and Momenta. The co-founding team also includes Tsinghua scholar Zhao Xing, among others.

The company's main approach is the joint development of complete robots and models. Its products include the R1 series wheeled dual-arm robots, the R1 Pro, robotic arms, and the G0 and G0 Plus models. It provides hardware to research institutions and robot developers, as well as models and development tools.

In February 2026, the company announced the completion of a RMB 1 billion Series B round. Investors included Jinding Capital, BAIC Industrial Investment, Bihong Investment, Zhengxin Valley Capital, Qianhai Ark, and Yifeng Capital, among others, with several existing shareholders increasing their investments. The company's official website currently discloses cumulative financing of nearly RMB 5 billion.

19 | Galbot: Tsinghua undergraduate, Stanford PhD, a single round of RMB 2.5 billion

Galbot was founded in May 2023 by Wang He and Yao Tengzhou. Founder and CTO Wang He was born in 1992 and is about 34 years old. He earned his bachelor's degree from the Department of Electronic Engineering at Tsinghua University and his doctorate from Stanford University, and is currently a researcher at Peking University. Yao Tengzhou graduated from the Robotics Institute of Beihang University and previously worked on robot R&D at ABB.

The company develops the "Galaxy Brain" AstraBrain embodied model and general-purpose robots. Its products include the Galbot G1 and the industrial handling robot Galbot S1, with applications covering retail, healthcare, logistics, and factory handling. Its training approach combines human demonstrations, simulated synthetic data, and reinforcement learning, seeking to reduce reliance on real-machine data collection.

In March 2026, the company announced the completion of a new RMB 2.5 billion financing round. Investors included the National Artificial Intelligence Industry Fund, Sinopec, Bank of China, SAIC Group Financial Holdings, SMIC Capital, and E-Town Capital, among others, with several existing shareholders increasing their investments.

20 | Robot Era: Founded by a post-90s Tsinghua doctoral advisor, securing another RMB 1 billion in July

Robot Era was founded in August 2023 and was incubated by the Institute for Interdisciplinary Information Sciences at Tsinghua University, among others. Founder and CEO Chen Jianyu was born in 1992 and is about 34 years old. He earned his bachelor's degree from Tsinghua University and his doctorate from the University of California, Berkeley. After returning to China, he joined the faculty at Tsinghua and has long conducted research on reinforcement learning and robot control.

The company conducts R&D spanning data, embodied models, motion control, dexterous hands, and robot hardware. Its products cover full-size bipedal humanoid robots, wheeled service robots, and fully direct-drive five-finger dexterous hands, with commercial applications including logistics sorting and handling.

In April 2026, the company completed a financing round exceeding USD 200 million; in July, it announced another RMB 1 billion round led by Chengtong Fund, with participation from Jiangxi Guokong, Yufu Zhongxin Fund, and Hangzhou Capital, and follow-on investment from CICC Renault, Kunlun Venture Capital, and Hony Capital, among others. The two rounds together totaled close to RMB 2.5 billion.

When these companies obtained financing, their products were at markedly different stages: some were still developing models or prototypes, while others had begun scenario validation and small-batch delivery. For this group of new companies, the indicators more worth tracking next will be whether customers make repeat purchases, whether the robots can work stably, and whether delivery and after-sales costs can support sustained business operations.

This article is from the WeChat public account "Xinzhineng Toutiao," author: He Tongxue, published by 36Kr with authorization.