Summary of Key Highlights
The 2026 World Robot Conference revealed a crucial shift in the humanoid robot industry: from being a showcase of technological prowess to becoming a practical tool for real work. However, commercialization still faces challenges such as high costs, a lack of actual orders, and various design considerations. 97% of all humanoid robot shipments globally come from China, and interest in financing continues to rise (93.5 billion yuan in the first half of the year, a fivefold increase). However, the investment approach has shifted from focusing on the overall robot design to emphasizing orders and core technologies. Wheeled, dual-arm robots have been adopted in factories due to their safety and cost advantages, while bipedal robots hold unique value in complex scenarios, with the two types complementing each other in the future. The industry's large-scale adoption will require three steps, currently hindered by insufficient data and technical bottlenecks.
Detailed Analysis
1. The Industry is Evolving: From Showoff to Practical Use
At previous robot conferences, performance robots that could dance and do acrobatics were common. This year, such demonstrations have decreased, replaced by practical applications such as loading and unloading in factories, pharmacy sorting, and unmanned retail. For example, the robots displayed by Zhi Ping Fang can accurately pick up different types of PCB boards in factories, and those from JD Retail have already been tested in real-world settings.
Why the change? The industry has realized that just having impressive capabilities is not enough; robots must be practical and profitable. The focus has shifted from whether they can be manufactured to whether they can be effectively used—whether they can perform tasks consistently, whether customers will repurchase them, and whether the investment will be profitable (ROI). However, being usable does not mean being efficient; for instance, robots still only perform about 30% as well as humans in long-duration tasks (over 1 minute) and need further optimization.
2. Commercialization Hurdles: Few Real Orders, High Costs, and Poor Stability
Despite China exporting over 40,000 humanoid robots in the first half of the year, the number of actual commercial orders is limited. For instance, 73% of Yushu Technology’s revenue comes from research and education, with only 9% from industrial applications. Many robots are used internally for technology iteration and algorithm training, not for selling to customers.
Costs are also a major issue: the total lifecycle cost of humanoid robots (hardware, debugging, maintenance) is much higher than that of human labor, and the gap can only be narrowed in scenarios where they operate 24/7. The localization rate of key components (such as micro-screws) is less than 20%, leading to high import prices.
Stability is another concern: in factories, power outages or malfunctions can cause problems; bipedal robots may fall and damage valuable materials, while wheeled robots, though safer, have lower task success rates. For example, Xiaomi’s “Tie Da” robot has a 98% success rate in automotive assembly, but its efficiency still lags behind that of humans in long tasks.
3. The Debate over Robot Forms: Bipedal vs. Wheeled – Which Will Take the Lead?
The industry is divided into two camps: wheeled, dual-arm robots are being adopted first, with bipedal robots following.
Wheeled robots are favored because they are safer (do not fall in accidents), cheaper (up to 10 times less expensive than bipedal robots), and more efficient (faster movement, lower energy consumption). In flat workshops, wheeled robots offer better value for money.
Bipedal robots are not useless, though. They can handle uneven surfaces or outdoor inspections where wheeled robots cannot. Xiaomi continues to develop bipedal robots because it aims for robots to adapt to human environments without needing factory modifications and to be versatile across multiple workstations. In the future, the two types will complement each other: wheeled robots for flat areas and bipedal robots for complex terrains.
4. A Major Shift in Investment Logic: From Robots to Core Technologies
In the first half of 2026, 93.5 billion yuan was invested in the industry, a fivefold increase from the same period last year, but the focus of funding has changed. Previously, investments were made in the complete robots (even prototype models that could walk), but now they are focused on core components and underlying technologies.
Examples of areas of investment include:
- Localization of Critical Components: Reducing reliance on imported micro-screws and hollow cup motors.
- **Robot “Brains”: Developing AI models that enable autonomous decision-making and training data (requiring millions of hours of real-world data).
- Vertical Industry Solutions: Companies specializing in factory inspections or logistics sorting will receive funding if they can demonstrate a viable ROI.
Leading robot companies may still attract investment, but smaller firms without practical application capabilities will find it harder to secure funding as the industry becomes more selective.
5. Three Steps to Large-Scale Adoption
Experts suggest three stages for the industry’s adoption:
- Short-term (within 3 years): Focus on simple tasks such as inspections, loading/unloading, and transportation (which require low error tolerance).
- Medium-term: Move on to logistics sorting (which demands handling non-standard materials but still allows for some error tolerance).
- Long-term: Aim for core industrial processes (such as automotive welding, which require high precision and speed, with proven ROI).
The main barriers are data and technology: robot models need extensive real-world training data, which is currently scarce, leading to unstable models. Many companies are investing in data collection equipment to address this issue.
Conclusion
The humanoid robot industry is moving from a phase of hype to practical application. To become widely adopted, challenges such as cost, stability, and data must be overcome. Wheeled robots are leading the way, while bipedal robots will follow. Core technologies and components will be the key to future competitiveness. For ordinary consumers, there’s no need to rush into buying household robots; it’s better to wait and see if robots can effectively perform tasks in industrial settings first.