Summary of Key Points
Industrial humanoid robots are rapidly entering automobile manufacturing and battery production facilities, but they are currently in the “training phase.” The main challenges include high costs (which need to be reduced by 60% to become cost-effective) and low efficiency (they can only replace about half of a worker’s workload). Automobile manufacturers and battery companies are building their own ecosystems through a combination of in-house development and external purchases. There may be small-scale tenders by the end of the year, but widespread adoption will depend on further improvements in technology and cost reduction.
1. Why Are Automobile/Battery Factories the “First Testing Ground” for Humanoid Robots?
Automobile and battery production are ideal areas for humanoid robots because traditional robotic arms are unsuitable for modern production lines:
- The rapid iteration of new energy vehicle models and mixed-line production (for example, producing different models on the same line) make traditional robotic arms inflexible and costly to modify.
- Battery manufacturing involves many high-voltage and hazardous processes (such as battery pack insertion and removal), which pose significant risks for humans.
- Humanoid robots have the advantages of being able to adapt to various tasks (moving around like humans, picking up objects), learning from experience, and handling dangerous work.
Companies like CATL are using wheeled humanoid robots like “Xiaomo” for high-voltage tasks, while BYD and Geely have introduced UbiBot robots for training purposes, targeting these specific challenges.
2. What Are the Challenges That Prevent Humanoid Robots from Being Widely Used?
There are two major hurdles before humanoid robots can replace workers on a large scale:
1. High costs: A robot currently costs six to seven hundred thousand yuan, and automobile manufacturers believe the cost must be reduced to around two hundred fifty thousand yuan (a 60% reduction) to make it more economical compared to hiring workers.
2. Low efficiency: A robot can only replace about half of a worker’s workload. Wheeled robots are more efficient than bipedal ones, but overall, they are still slower.
3. Technological maturity: The stability and fault resolution capabilities of humanoid robots have not yet reached industrial standards. Questions such as long-term durability and rapid repair in case of malfunctions remain unresolved.
Therefore, the relationship between automobile manufacturers and robot companies is one of collaboration and training, with engineers from robot companies working directly in the factories to debug and collect data, starting with simpler tasks like material handling and visual inspection.
3. Are Automobile Manufacturers Simply Buying Robots or Also Developing Them Themselves?
Leading automobile companies are adopting a “dual-track approach” of both purchasing and developing robots to build their own ecosystems:
- External purchases: Companies like BYD and Geely are introducing robots from UbiBot to gather data on different applications.
- In-house development: Companies like Xpeng and BYD are transferring automotive technologies (such as battery technology and autonomous driving algorithms) to develop their own robots for use in their factories. The feedback from these robots can then be used to improve the robots.
CATL has taken an even more proactive approach: the robots developed by its invested company, Galaxy General, and its ecosystem partner, Qianxun Intelligence, all use CATL’s batteries, creating a closed loop where costs and technology are better controlled.
4. Will There Be a Wave of Robot Orders by the End of the Year?
Industry analysts predict that there may be small-scale tenders in the fourth quarter of 2026, but large-scale purchases are still ahead:
- Reasons for optimism:
- The Ministry of Industry and Information Technology’s special initiative will be evaluated by the end of the year.
- Automobile manufacturers typically plan their technological upgrades for the following year in the fourth quarter, which is the time to purchase equipment.
- Pilot projects from the first half of the year will be assessed by the end of the third quarter, laying the groundwork for tenders.
- The promotion of the “RaaS (Robot as a Service)” model (renting robots and paying based on usage) reduces initial investment for manufacturers.
- Reasons for caution:
- Technology is not yet mature enough in terms of precision and stability.
- Costs remain high, so manufacturers are cautious about the return on investment, leading to likely small-scale trials rather than widespread adoption.
5. Can Humanoid Robots Change Automobile Production in the Future?
In the long run, if costs decrease and efficiency improves, humanoid robots could transform production processes:
- They could move from replacing repetitive tasks (such as material handling and inspection) to more complex jobs (assembling and testing).
- Automobile manufacturers can integrate their supply chains with robots, further reducing costs.
- Companies like BMW’s leasing model and Tesla’s self-production and sales strategies are exploring viable commercial approaches.
However, in the short term, the focus is on solving the practical issues of whether robots are feasible and cost-effective. The initial results will become clear after the evaluation of the special initiative at the end of the year.
In summary, the rollout of industrial humanoid robots is proceeding in a “small steps, fast pace” approach, with trials in automobile and battery factories to gather data and experience. Only when costs and technology meet industry standards will they be widely adopted in production facilities. The current training phase is essential for laying the foundation for their future use.