At the very start of the Year of the Horse, the largest embodied intelligence financing round has already emerged.
It is learned that Galbot (银河通用机器人) has announced the completion of a new funding round totaling RMB 2.5 billion. The investor lineup is exceptionally strong, including the National AI Industry Investment Fund (Phase III), Sinopec, CITIC Group Investment Holdings, Bank of China, SAIC Motor Financial Holdings, SMIC Juyuan, Beijing E-Town Capital, Future Industry Investment, Kunpeng Fund, Wuxi Venture Capital, Fujian Industrial Investment, as well as multiple existing shareholders who continued to increase their stakes.
Notably, this round comes just two months after the company’s previous financing. The pace is striking. “The entire process took less than two months and was effectively an oversubscribed round,” said a source close to Galbot. Investor enthusiasm in the primary market has exceeded expectations, with multiple institutions reportedly lining up following the company’s appearance at the Spring Festival Gala.
With this latest round, Galbot now ranks first in China’s embodied intelligence sector by total capital raised, and its valuation continues to lead the humanoid robotics industry.
A battle among giants has officially begun.
Just in: the year’s biggest embodied AI deal, valuing the company at over RMB 20B.
In just three years, Galbot has become a phenomenon in China’s venture capital circle. As the market has observed, nearly every category of top-tier investment institution now stands behind the company.
By the end of 2025, Galbot completed another record-breaking round of over USD 300 million. The round was led by China Mobile (中国移动), with participation from CICC Capital (中金资本), Suzhou Ventures Group (苏创投), and the CCTV Media Convergence Fund (央视融媒体基金). International investors from Singapore and the Middle East also joined, alongside existing shareholders. At the time, sources indicated that Galbot’s post-money valuation had reached USD 3 billion (exceeding RMB 20 billion).
Now, the latest RMB 2.5 billion round has surfaced, bringing together another group of prominent industrial and strategic investors. The expanded shareholder base significantly broadens Galbot’s application landscape across energy, advanced manufacturing, and commercial real estate.
For example, Sinopec (中国石化)’s nationwide gas station network, petrochemical high-risk environments, and thousands of Easy Joy convenience stores could provide extensive deployment scenarios in smart energy and intelligent retail. CITIC Group (中信集团)’s diversified portfolio, spanning advanced manufacturing bases to commercial properties, offers substantial real-world applications in intelligent manufacturing and smart services. Meanwhile, SAIC Motor’s automotive production lines open further opportunities for Galbot in flexible industrial automation.
Within just two months, Galbot has once again set new domestic records in both single-round and cumulative financing within the embodied intelligence sector. Its valuation continues to rise accordingly, consolidating its leadership position in China’s humanoid robotics industry.
Led by a post-90s professor, the Spring Festival Gala’s first autonomous working robot went viral overnight.
Why has Galbot become the publicly highest-valued humanoid robotics company in China?
As the saying goes, investing is ultimately about investing in people. At the helm of Galbot is Wang He (王鹤), who has left a strong impression on the venture capital community. Born in 1992, he earned his bachelor’s degree from the Department of Electronic Engineering at Tsinghua University and later pursued a Ph.D. at Stanford University. He is regarded as one of the earliest scholars globally to conduct research on end-to-end embodied foundation models.
After returning to China, Wang He (王鹤) joined Peking University (北京大学) as a faculty member, where he founded and led the Embodied Perception and Interaction Lab (EPIC Lab). In 2022, he also established the Embodied Intelligence Research Center at the Beijing Academy of Artificial Intelligence (北京智源人工智能研究院) and served as its director.
In May 2023, he went on to found Galbot . The team he assembled is widely regarded as a “golden lineup” in China’s embodied intelligence field, bringing together leading AI and robotics experts from Peking University (北京大学), Tsinghua University (清华大学), Stanford University (斯坦福大学), and Beihang University (北京航空航天大学), among others.
In Wang’s vision, “The future of general-purpose robots lies in integrating the ‘cerebellum’—responsible for bodily control and dexterous physical interaction—with the ‘brain,’ which handles cognition, understanding, and planning.”
It was under this framework that Galbot independently developed AstraBrain (银河星脑). Across the industry, most players still develop the “brain” (task planning), “cerebellum” (motion control), and “hands” (dexterous manipulation) as separate modules, lacking a unified end-to-end foundation model. This fragmented architecture often results in information loss between modules, leading to delayed responses and weak generalization.
AstraBrain fundamentally reconstructs this paradigm. It is the world’s first—and currently the only—end-to-end embodied foundation model that fully integrates “brain–cerebellum–neural control.” The system connects high-level multimodal perception with low-level real-time feedback control across the entire chain, enabling deep integration of full-body coordination and precise hand manipulation. As a result, the robot’s movements are fluid and accurate, resembling human motor performance.
As is well known, large models depend heavily on data, yet large-scale real-world data collection is neither efficient nor always feasible. To address this, Galbot built what it describes as the world’s largest embodied intelligence dataset at the hundred-billion scale and introduced a data infrastructure platform called AstraSynth (银河星坊). The company pioneered a hybrid training paradigm that relies primarily on synthetic simulation data supplemented by a limited amount of real-robot data.
Within high-fidelity physical simulation environments, the system can generate massive and diverse scenarios, allowing robots to traverse extreme edge cases in virtual space before fine-tuning with a small volume of real-world data. According to disclosed figures, this training approach is said to be 1,000 times more efficient than Tesla’s, with trained models achieving a reported 99% task success rate.
Wang He has previously noted that from humanoid robots performing dances on national television to participating in robot marathons, the industry is currently in what he calls the “era of motion.” However, he argues that the sector is transitioning from a “motion era” to a “productivity era.” The true breakthrough, in his view, lies in the coordinated integration of hands, eyes, and brain—an essential capability for enabling humanoid robots to rapidly empower a wide range of industries.
From science fiction to reality
At the outset of 2026, China’s embodied intelligence sector has entered an intense phase of competition.
In just the past two months, multiple large-scale financing rounds have been announced across the domestic embodied intelligence landscape. Beneath this wave of capital, however, a quiet elimination round is unfolding. Third-party statistics show that over the past year, more than 20 robotics companies globally have faced bankruptcy, large-scale layoffs, or business divestitures.
Investor sentiment is also shifting. One VC partner remarked bluntly, “In 2026, we will not look at embodied intelligence startups that have neither mass production nor revenue.” Wang Tianmiao (王田苗), a professor at Beihang University has similarly cautioned that beneath the sector’s prosperity lies hidden risk, with capital accelerating toward leading players and upstream core components with higher certainty. As investors put it, “Out of more than 100 companies, perhaps only three to five will survive. Over 90% of capital and resources will be sacrificed. That’s the reality.”
All of this signals the emergence of a pronounced Matthew Effect: capital, resources, and talent are converging on top-tier players. The ability to close a commercial loop has become the dividing line within the industry. Financing alone can no longer secure a seat at the table; self-sustaining revenue generation is now the defining test of survival.
As many observers have noted, Chinese robotics is transitioning from “running on the competition field” to “working in factories” and “serving in households.” To date, Galbot has taken an early lead in building a full industrialization loop of “high-quality datasets + end-to-end foundation model + scaled deployment.” Its technologies have been deeply implemented across industrial, retail, and healthcare scenarios, translating technical advantages into productivity gains. As Wang He once stated, “We don’t aim for robots that can do 100 things. We aim for robots that can do 10 things to industrial-grade standards.”
In industrial manufacturing, Galbot has secured deep cooperation with global leaders including CATL (宁德时代), Bosch (博世), Toyota (丰田汽车), BAIC Group (北汽集团), and SAIC Motor (上汽集团), with cumulative orders reaching several thousand units.
Notably, the company introduced the industry’s first heavy-duty industrial embodied robot, Galbot S1, designed specifically to address high-load challenges in manufacturing environments. Its dual arms support a maximum payload of 50 kilograms, meeting industrial-grade standards and operating reliably under complex conditions such as dust, vibration, and temperature fluctuations. Equipped with an embodied logistics model and autonomous battery-swapping capability, the S1 fills a key gap in industrial-grade embodied intelligence applications.
In instant retail and smart pharmacy operations, Galbot has set a global industry record by operating a 24/7 unmanned instant retail warehouse for over one year. The company has deployed smart retail pharmacies in 24 cities, with each store managing over 5,000 SKUs, transforming the concept of “round-the-clock retail” into a practical reality.
Healthcare and eldercare have also become important application domains. Galbot has established deep partnerships with leading Grade-A tertiary hospitals, including Xuanwu Hospital (宣武医院) and West China Hospital (华西医院), deploying robots in wards, pharmacies, and guidance services to enhance medical operational efficiency.
From the Spring Festival Gala stage to industrial production lines, from smart pharmacies to urban landmarks, Galbot is undergoing a transition from technological pioneer to provider of general-purpose productivity. This evolution challenges previous skepticism regarding the practical viability of embodied intelligence in China.
The momentum is unmistakable. Many investors are increasingly convinced that Chinese players are poised to take a leading position in the global race for embodied intelligence.