At the 2026 World Robot Conference held in Beijing from August 19 to 23, Liu Yulong, co-founder and head of product research at Zhishen Technology, sat down to address some of the sector's most pressing questions. His insights covered Unitree Technology's recent stock market debut, whether the embodied intelligence track is experiencing a bubble, and the complex relationship between public listings and the widespread losses seen across the industry.
When asked about a Unitree investor's prediction that "only three companies in embodied intelligence will thrive, ten will barely survive, and the rest will perish," Liu offered a distinctly different perspective. He suggested this reflects short-term market sentiment rather than a long-term reality, arguing that embodied intelligence represents a genuine industrial revolution. Drawing a parallel to the information age, he noted that the number of listed companies generated during that era far exceeded just a few, because the technology penetrated countless sectors. If one only pursues the pinnacle of technology in a single domain, perhaps only a few leaders would emerge, but a productivity revolution that transforms production relations will reshape opportunities across the entire industrial chain.
The internet era produced numerous listed companies and countless more private ones, and embodied intelligence will follow a similar trajectory. Beyond the head companies racing to go public, Liu believes there will be abundant opportunities for firms that genuinely apply embodied intelligence across vertical industries. "This track is not a single-head pattern; it must develop as a more complete pyramid or matrix structure," he explained.
Turning to the recent IPO of Unitree Technology and the reality that most embodied intelligence companies are still loss-making, Liu addressed the relationship between listing and corporate development. He emphasized that going public is fundamentally a stage in a company's growth, not the finish line. The embodied intelligence industry is massive in scale, requiring heavy investment in core technology research and development, which demands continuous capital support. Moreover, the industry is inherently data-driven, with data originating from real-world scenarios, and the development and implementation of these scenarios require substantial time and resources. Before technology matures, it must undergo extensive validation and iterative refinement, making sustained early-stage investment an objective law of industrial development and an inevitable path.
Regarding commercial pacing, Liu revealed that while pursuing long-term technological breakthroughs, the company is simultaneously identifying application scenarios that can be deployed immediately. In certain vertical scenarios, they have already partnered with collaborators to achieve a positive cycle of commercial operations. This "invest while producing" rhythm allows the company to balance its long-term vision with short-term survival needs.
Discussing the decades-long development of humanoid robots and why the sector is only now experiencing a new wave of enthusiasm, Liu attributed this not to any single technological breakthrough but to the completion of the entire technological landscape. He recalled that Honda began developing humanoid robots in the 1980s when electronic technology was underdeveloped, and mechanical engineers had to piece things together to make robots walk unsteadily—that was the technological paradigm of that era. In his view, humanoid robots represent the long-term direction for two key reasons: first, they can do everything humans can do, serving as direct replacements without requiring modifications to human living environments; and second, learning from humans is the most natural approach—"just like teaching a child"—which is far faster than deriving formulas or programming from scratch.
"The term 'embodied intelligence' has only emerged in the past two or three years, but its underlying logic has been accumulating for a long time," Liu pointed out. The current moment is unique because the five major modules—brain, cerebellum, body, energy, and perception—have finally come together. AI large models, built on thirty years of internet development, provide the "brain"; batteries, motors, and sensors give robots a "body"; and advances in tactile perception enable dexterous manipulation. "When the landscape is complete enough, it enters the fast lane," Liu predicted, forecasting exponential growth for the industry. Once that threshold is crossed—when the ceiling of technology surpasses the floor of demand—the inflection point for application arrives.
When asked which scenarios are most likely to achieve commercial viability first, Liu responded confidently: "Many scenarios can survive, whether B2B or B2C." He revealed that Zhishen Technology is practicing a strategy of "laying eggs along the way," finding applicable solutions in specific vertical scenarios before the full embodied intelligence technology chain is completely mature, working with partners to get operations moving first. Regarding the technical roadmap, Liu disclosed that the company is not limited to quadruped robots' mobility capabilities but is making "embodied manipulation" its primary focus for the second half of the year. "Humanity evolved from crawling on all fours to freeing up our hands—this remains one of our paths for laying eggs along the way," he said. For quadruped robots, adding arms on top of patrol and transport functions creates more integrated operational capabilities, which he identified as a core competitive advantage. In the coming months, Zhishen Technology will prioritize advancing embodied manipulation data and model implementation.
Among the many distinguished guests at the conference, two cross-disciplinary figures stood out: Xue Qikun, an academician of the Chinese Academy of Sciences and winner of the nation's highest science and technology award, representing the pinnacle of China's basic scientific research; and Lin Yuan, a legendary retail investor who grew from 8,000 yuan to billions in assets, embodying the ultimate practical experience of navigating bull and bear markets.