In the rapidly evolving landscape of artificial intelligence, embodied intelligence has emerged as a transformative force, capturing the attention of venture capitalists, tech giants, and innovators alike. Li Zhu, founding partner of InnoAngel Fund and president of the Beijing Frontier International Artificial Intelligence Research Institute, shares his insights on why embodied intelligence represents a fertile ground for AI entrepreneurs. According to Li, the embodied intelligence sector offers a more level playing field compared to other AI domains, with immense market potential that could reshape future lifestyles and industries. As an early-stage investment firm, InnoAngel Fund has strategically positioned itself in this space, backing companies like Yun Shen Chu, Qian Jue Technology, Lu Ming Robotics, MegaRobo, and Accelerated Evolution, among others. This shift to technology-focused investments since 2019 underscores a broader trend toward embodied intelligence and humanoid robots, which Li believes are still in their infancy, requiring time for technological convergence and practical application exploration.
The concept of embodied intelligence revolves around robots that interact with the physical world, enabling perception, decision-making, and execution in real-time environments. Li Zhu categorizes the new generation of AI into three key paradigms: information intelligence, embodied intelligence, and brain-like intelligence. Information intelligence includes products such as ChatGPT and DeepSeek, which process and generate data. In contrast, embodied intelligence elevates robots by allowing them to engage dynamically with their surroundings, while brain-like intelligence draws from neuroscience to simulate human cognitive processes for autonomous learning and reasoning. This framework highlights the unique position of embodied intelligence as a bridge between digital and physical realms, making it a critical area for innovation and investment.
-
Embodied Intelligence: A Level Playing Field for AI Entrepreneurs
Li Zhu emphasizes that the embodied intelligence sector is particularly favorable for AI entrepreneurs due to its early developmental stage. Unlike established AI domains where tech behemoths dominate, embodied intelligence places startups and industry giants on a similar footing. Both must undertake fundamental tasks such as developing embodied large models, researching cerebellar motion control, and collecting extensive real-world data. This lack of entrenched advantage fosters a collaborative and competitive environment where innovation can thrive. Li observes that diverse players, including top academic researchers and executives from major tech firms, are entering the embodied intelligence arena, driven by the potential for groundbreaking discoveries. The absence of a clear frontrunner means that startups have a genuine opportunity to lead in embodied intelligence advancements, making it a hotspot for entrepreneurial success.
Moreover, the embodied intelligence market’s nascent state allows for experimentation and adaptation. Entrepreneurs can define their niches without facing immediate pressure from monopolistic practices. Li notes that this dynamic is rare in the tech world, where scalability and resources often favor incumbents. However, in embodied intelligence, the focus on real-world interaction and physical embodiment demands novel approaches that reward creativity and technical prowess. As a result, embodied intelligence startups can leverage their agility to pivot and refine their strategies based on technological progress and market feedback. This adaptability is crucial in a field where standards are still evolving, and embodied robot applications are being tested across various sectors.
-
Investment Philosophy: Evaluating Embodied Intelligence for Long-Term Impact
InnoAngel Fund’s investment strategy in embodied intelligence is guided by two core principles: assessing the industry’s potential to drive monumental change and identifying viable application scenarios. Li Zhu explains that embodied intelligence is poised to trigger significant shifts in global operations, with projections indicating a multi-trillion-yuan market in the coming years. Analysts predict that as embodied intelligence permeates households, its growth could accelerate exponentially, spawning super-unicorn companies valued in the hundreds of billions. This transformative power stems from the泛化性 and versatility of embodied intelligence, which enables robots to operate in non-standardized environments—a departure from traditional, rhythm-based industrial robots.
The fund also prioritizes sectors where embodied intelligence can achieve tangible落地. Industries such as manufacturing, agriculture, and energy present ample opportunities for embodied robot deployment. Li highlights that no single company can monopolize the entire embodied intelligence landscape; instead, the market will likely fragment into specialized providers offering tailored solutions. For instance, an embodied intelligence firm focusing on agricultural robotics might develop machines for precision farming, while another targets logistics with autonomous embodied robots. By深耕 specific domains, companies can build sustainable revenue streams. InnoAngel Fund seeks out teams with core technological expertise, product definition capabilities, and the foresight to navigate the evolving embodied intelligence ecosystem. Li stresses the importance of dynamic adjustment, as success in embodied intelligence hinges on aligning innovations with emerging needs.
-
The Current State of Humanoid Robots: Challenges and Opportunities in Embodied Intelligence
Humanoid robots, a prominent subset of embodied intelligence, have sparked intense debate within the investment community. While capital floods into embodied intelligence startups—with financing in robotics and embodied intelligence exceeding 200 billion yuan in the first half of 2025, surpassing the previous year’s total—skeptics like Zhu Xiaohu of GSR Ventures and Lee Kai-fu, former chairman of Innovation Works, question their practicality. Li Zhu acknowledges these concerns but argues that the industry is in a phase of exploration and divergence. Technological paradigms for humanoid robots have not yet converged, meaning that researchers and companies are still experimenting with various approaches to achieve general artificial intelligence through embodied intelligence.
Application scenarios for humanoid robots remain in flux, limiting their immediate value proposition. Currently, many embodied robots demonstrate basic functions like walking or jumping, but they lack the泛化性 to handle diverse tasks without retraining. For example, a humanoid robot trained to grasp a cup may struggle with an apple or banana, highlighting the gaps in embodied intelligence capabilities. Li points out that while expectations are high for embodied intelligence to deliver universal adaptability, current limitations necessitate focused trials. Some ventures, such as those involving humanoid robots in exhibition displays, generate modest revenue while refining their embodied intelligence systems. These efforts, though preliminary, contribute to the gradual maturation of embodied robot technology.
-
Path to Mainstream Adoption: The Future of Embodied Intelligence in Daily Life
Looking ahead, Li Zhu envisions a future where humanoid robots integrated with embodied intelligence become commonplace in households, assisting with chores, eldercare, and other domestic duties. However, he cautions that this vision requires patience and incremental progress. Today’s humanoid robots are not yet ready for home environments, but companies are using creative applications to enhance their performance. For instance, Accelerated Evolution, an InnoAngel portfolio company, explores soccer-playing humanoid robots in entertainment settings. While these robots are currently controlled via remote devices, advancements in embodied intelligence could enable autonomous decision-making, such as tracking ball movement and scoring goals.
This evolutionary process focuses on improving robustness, lightweight design, and battery life—key factors for practical embodied robot deployment. Li believes that activities like marathons, boxing matches, and soccer games serve as testing grounds for embodied intelligence, driving innovations that will eventually support broader adoption. As embodied intelligence algorithms become more sophisticated, humanoid robots could transition from novelty acts to reliable assistants, embodying the promise of intelligent, interactive machines. The journey involves not only technical refinements but also societal acceptance, as embodied robots must seamlessly integrate into human-centric spaces.
-
Broader Implications: How Embodied Intelligence Could Reshape Industries and Society
The ripple effects of embodied intelligence extend beyond personal convenience to economic and social transformation. Li Zhu elaborates that embodied robots could revolutionize sectors like healthcare, where they might assist in surgeries or patient monitoring, and education, by providing personalized tutoring. In industrial settings, embodied intelligence enables flexible automation, reducing reliance on rigid assembly lines and adapting to variable production demands. This adaptability is central to the embodied intelligence paradigm, as it allows machines to learn from and respond to dynamic environments.
Furthermore, the growth of the embodied intelligence market could spur job creation in research, development, and maintenance of embodied robots. While concerns about automation displacing workers exist, Li argues that embodied intelligence may generate new roles that require human-robot collaboration. For example, technicians skilled in embodied robot programming could see increased demand. The ethical dimensions of embodied intelligence also warrant attention, as developers must ensure that embodied robots operate safely and equitably. InnoAngel Fund monitors these trends, recognizing that responsible innovation in embodied intelligence will be crucial for sustainable growth.
-
Investment Trends and Data in the Embodied Intelligence Sector
The surge in funding for embodied intelligence reflects growing confidence in its potential. In the first half of 2025, robotics and embodied intelligence attracted over 200 billion yuan in investments, indicating robust investor interest. This capital influx supports R&D in embodied robot hardware and software, from sensor technologies to AI models that enhance perceptual abilities. Li Zhu notes that early-stage investments like those from InnoAngel Fund target companies with disruptive ideas in embodied intelligence, even if commercial returns are not immediate. The focus is on long-term value creation, as embodied intelligence applications mature.
Data from market analyses suggest that the embodied intelligence sector could reach trillion-yuan scales within years, driven by adoption in logistics, services, and smart homes. For instance, embodied robots in warehouses are already improving efficiency, while prototypes for domestic helpers are in testing. Li emphasizes that while numbers are promising, the true measure of success lies in how well embodied intelligence solves real-world problems. InnoAngel’s portfolio includes firms that have achieved early sales through niche applications, demonstrating the viability of embodied robot solutions in specific contexts. As the industry evolves, Li expects more concrete metrics to emerge, guiding future investments in embodied intelligence.
-
Conclusion: Embracing the Embodied Intelligence Revolution with Strategic Patience
In summary, embodied intelligence represents a pivotal shift in AI, offering unique opportunities for entrepreneurs and investors alike. Li Zhu’s perspective underscores the importance of patience and strategic focus in navigating this emerging field. While humanoid robots and other embodied intelligence applications are still developing, their potential to transform industries and daily life is undeniable. InnoAngel Fund remains committed to supporting innovations in embodied intelligence, betting on teams that combine technical excellence with market insight. As embodied robot technologies advance, they could unlock new levels of automation and interaction, ultimately bringing intelligent machines into homes and workplaces worldwide. The journey may be long, but the rewards of embodied intelligence—in terms of economic growth and societal benefit—make it a pursuit worth championing.
For now, the embodied intelligence community continues to experiment, learn, and iterate. From soccer fields to factory floors, each application of embodied robots brings us closer to a future where machines not only think but also act intelligently in the physical world. Li Zhu’s optimism serves as a reminder that in the realm of embodied intelligence, the greatest achievements may still lie ahead, waiting for the convergence of technology and imagination.