Summary of Key Points
Zhi Ping Fang is a humanoid robot company that has achieved a valuation of over 20 billion yuan in just three years since its establishment. It has raised funds 12 times in the past year, relying on its approach of "full-stack self-development" (developing both the AI model for the robot's "brain" and the hardware for its "body"), earning it the nickname "the Chinese version of Tesla." However, it now faces three challenges:
1. Its competitors have also started to adopt full-stack development, diluting its initial advantage;
2. Technology is evolving rapidly, and its core model is reaching its limits, necessitating a shift towards new directions;
3. Although it aimed to create industrial productivity tools, the difficulty of implementing these in real industrial settings has led it to explore consumer markets such as coffee shops and tourist attractions. While capital is betting on its potential, whether it will ultimately succeed depends on the actual implementation of its products.
I. Valuation Doubled in Two Years: Why Is Zhi Ping Fang So Highly Pursued by Capital?
Zhi Ping Fang's fundraising pace is impressive: its valuation just exceeded 10 billion yuan four months ago and has now surpassed 20 billion yuan, with 12 rounds of financing in one year, backed by significant investors including state-owned entities, Maotai, and CICC. The key reason for capital's enthusiasm lies in its commitment to full-stack self-development. While others were focusing on purchasing off-the-shelf components to build robot bodies in 2023, Zhi Ping Fang invested in developing the VLA large model, which enables robots to integrate what they see and hear with their physical actions. Only Tesla and Zhi Ping Fang dared to take this approach at that time, and their forward-thinking strategy coincided with the explosive growth of the humanoid intelligence sector, attracting substantial investment.
However, challenges have emerged: competitors like Yin He Tong and Zi Bian Xing have also started full-stack development, diminishing Zhi Ping Fang's lead. Additionally, Zhi Ping Fang itself acknowledges that the purpose of its funding is to widen the gap between its AI model and those of its rivals, as technology development is extremely costly. For example, Yu Shu Technology spent only tens of millions on research and development each year when it did not focus on the AI model.
II. Full-Stack Self-Development as a Ace, but Can It Keep Up with Rapid Technological Progress?
Full-stack self-development means developing everything from the robot's AI model to its physical components from scratch. This was once Zhi Ping Fang's core competitive advantage, but technology is evolving too quickly:
The VLA model has been criticized for reaching its limits, and more advanced models are becoming the new standard (NVIDIA predicts that world-class models will become the foundation for robots by 2026). Zhi Ping Fang plans to release a "neurological architecture" model called NeuroVLA in 2026 to give robots instinctive responses. However, technological advancements are relentless; as soon as one issue is resolved, another emerges, requiring continuous investment in upgrades.
III. From Industrial Powerhouse to Coffee Shop Employee: Why Has Zhi Ping Fang Changed Its Focus?
The company's founder initially stated that it would focus on industrial productivity tools rather than entertainment robots. Yet, it has now expanded into coffee and ice cream robot services, entering commercial and tourist areas. The reason for this shift is twofold:
- Industrial orders have been lukewarm, with deals mainly being strategic partnerships (e.g., with Hikvision, a semiconductor company), without specific financial details. The only concrete contract is for 1,000 robots to be deployed over three years, worth approximately 500 million yuan.
- Industrial implementation is extremely challenging: factory owners are cautious about investing in robots that could cause production disruptions, and the high cost of these robots (several hundred thousand yuan each) makes them less cost-effective compared to skilled workers. Competitors like Zi Bian Xing have already ventured into household and retail applications, forcing Zhi Ping Fang to seek opportunities in consumer markets.
IV. Why Does It Still Take 3-5 Years for Robots to Be Adopted in Factories?
The adoption of humanoid robots in factories faces significant hurdles:
1. High Risk: A robot failure can halt production lines, posing a major risk for manufacturers.
2. Low Cost-Effectiveness: The cost of these robots is much higher than that of traditional industrial robots, making them less viable compared to skilled labor.
3. Long Implementation Period: The process involves 3-6 months of concept validation, followed by 12 months of small-scale testing and pilot deployments, and finally, large-scale deployment takes about two years.
Statistics support this: in 2025, only 13,000 humanoid robots were shipped globally, with most of these going to industrial applications. For example, UbiSoft delivered just over 500 industrial robots, and Yu Shu's revenue from industrial sales accounted for only 2.6%. Experts predict that it will still take at least 3-5 years before industrial robots become widely adopted.
Conclusion: The Challenges Behind the Capital Boom
Zhi Ping Fang's story resembles that of a technology star—high valuation, extensive funding, and advanced technology. However, it faces practical challenges related to technological innovation, market transformation, and time-consuming implementation. While capital is optimistic about its potential, success depends on its ability to turn its technology into viable products that are both useful in factories and popular in consumer contexts. After all, no amount of financing can replace actual customer orders as the ultimate proof of a business's viability.