虎嗅

What do the medals from the Robot Olympics really signify?

原文:机器人运动会的奖牌,到底说明什么?

Summary of Key Points

The 2026 World Humanoid Robot Games have concluded, with Zhiyuan leading the way with 46 medals (first in the corporate merger category) and Tiangong with 45 medals (first in the team category), followed by Yushu with 10 medals and Yinhe General with 5 medals. This competition served as a comprehensive test for the industry, addressing the issue of previous inability to compare demonstration videos across different companies. It evaluated technical strengths, product coverage, and organizational capabilities through unified rules. The scale of the competition doubled, and there was a significant improvement in performance (for example, the 400-meter dash time was reduced from 1 minute and 28 seconds to 38 seconds). The focus has shifted from pure athletic performance to practical application scenarios, with platform ecosystems and full autonomy becoming new priorities. The leading manufacturers adopted distinct strategies: Zhiyuan aimed for a high total number of medals by participating in multiple products and events; Tiangong focused on high-dynamic movements; Yinhe General delved into long-duration, fully autonomous tasks; and Yushu, despite reducing the number of entries due to production commitments, still maintained its long-term advantages. The entry of large companies (such as Honor) and the rise of niche brands (like Xiaobiga) also reflect new industry trends.

Why is this competition important? — The industry finally has a "unified benchmark"

In the past, robot demonstration videos were either shot in internal laboratories or heavily edited, making it impossible for viewers to compare the robots' capabilities. Questions like "What venue was used for the 100-meter run?" or "Were the objects for grasping pre-placed?" remained unanswered. This competition provided a unified standard with the same venue and rules, making it clear to compare aspects such as running speed, grasping accuracy, and task completion. More importantly, it distinguished between autonomous and remotely controlled robots, with fully autonomous tasks earning double the score of remotely controlled ones. This not only showcased technical prowess but also reflected companies' resource investment (e.g., the number of teams sent) and organizational skills (e.g., on-site debugging capabilities), effectively acting as a stress test.

The rivalry between Zhiyuan and Tiangong — Different metrics, different strengths

The debate over which company was the best (Zhiyuan or Tiangong) actually stems from the difference between the two ranking systems:

  • Team Category: Tiangong's team won first place with 5 gold and 4 bronze medals.
  • Corporate Merger Category: Zhiyuan, by combining the scores of its own teams in Shanghai, Chengdu, and other locations with those of third-party teams using its platform, secured first place with a total of 46 medals.

Zhiyuan's strengths: It adopted a broad strategy, participating with multiple products like the精灵G2 (for hotel services) and OmniHand (for tasks like bean picking and wiring) in various cities, resulting in a higher total number of medals. The OmniHand won 7 out of 8 gold medals, demonstrating its precision in fine-grained operations. Zhiyuan also emphasized that it used commercially available hardware, which is a strong selling point for a company planning to list on the Hong Kong stock market.

Tiangong's strengths: Its expertise lies in high-dynamic movements, with impressive performance in the 400-meter dash (38.15 seconds, more than twice as fast as last year) and in relay events. This success is attributed to continuous investment in the Beijing Humanoid Robot Innovation Center. The presence of third-party teams in its medal tally indicates a robust platform ecosystem.

The differentiated approaches of the leading manufacturers — Each path to success

The strategies of the top four manufacturers reflect their development directions:

  • Yinhe General: Focusing on challenging tasks (such as household, catering, and supermarket applications), with 3 out of 5 medals in fully autonomous categories. This reflects the complexity of these tasks, which require the robot to perceive its environment, make decisions, and complete tasks independently. The next step will be to see if it can operate smoothly in real-world stores with minimal errors.
  • Yushu: Although it won fewer medals (10 this year) due to reduced entries, its long-term strengths, including motion control in quadruped robots, cost-effective productization, and supply chain capabilities, remain significant. However, the industry increasingly values practical application skills, and Yushu may fall behind if it fails to improve in this area.

Hidden signals beyond the competition — The arrival of large companies and the power of marketing

Besides the leading manufacturers, several other factors are worth noting:

  • Entry of large companies: Honor entered with its "Lightning" robot, which performed nearly as well as Tiangong in the 400-meter dash, joining the top tier. Consumer electronics companies have advantages in chips, batteries, and supply chains, which can accelerate hardware innovation and drive down costs.
  • The power of marketing: The niche brand Xiaobiga, with its 75-cm robot, gained popularity through its distinctive design and adorable behaviors. This shows that marketing is crucial for brand recognition, but the brand must be closely associated with its products to be remembered by consumers.
  • Platform ecosystems are becoming a trend: The success of platforms like the football platform, which was used by third-party teams for soccer matches, indicates that they are easy to use and have active ecosystems. These platforms can test robots' coordination and decision-making abilities, which are transferable to other tasks.

What does this competition tell us? — The industry's shift from "showcasing technology" to "practical application"

The outcomes of this competition highlight several key points:

1. Medals do not equate to general intelligence: Fast running and precise grasping do not guarantee the robot's ability to handle various tasks flexibly like humans; general intelligence is still a distant goal.

2. Shift from athletics to practical applications: The focus has shifted from pure athletic performance to real-world tasks like hotel services and supermarket management, indicating that the industry is evaluating robots' practical usefulness.

3. Platforms and ecosystems are crucial: Teams with open platforms and active ecosystems perform better, as they can attract more developers and rapidly enhance overall capabilities.

4. Practicality is key: Even with excellent competition results, it's essential to see if robots can be mass-produced and operate reliably in real-world scenarios. Whether these capabilities can be translated into practical applications will be more interesting in future competitions.

In summary, these games reflect the current state of the humanoid robot industry: technology is advancing rapidly, but there is still a way to go before robots become a common part of our lives. While each manufacturer has its unique approach, they all face the same question—can their robots truly help people with practical tasks?