Summary of Key Highlights
The 2026 World Robot Conference revealed a crucial shift in the robotics industry: from focusing on “showcasing technological prowess” to competing to “see who can make money first.” While there was a surge in new products, they were highly homogenized. Companies have shifted from performing acrobatic demonstrations and dancing to showcasing robots performing actual tasks, but the progress of these implementations varies greatly. Central state-owned enterprises (SOEs), as the largest consumers, have moved from being mere spectators to becoming key players in the industry, with the “Procurement Day” becoming a formal event for the first time, facilitating the transition from merely displaying products to making actual transactions. The data collection segment has become particularly promising due to the massive amount of data required for embodied intelligence, creating business opportunities for companies that provide the necessary tools and technologies.
Overall, the industry has entered a new phase characterized by a focus on practical profitability, but overcoming the gap between demonstration in exhibition halls and real-world applications remains a significant challenge.
1. Surge in New Products, Yet “First Releases” Lose Their Value
This year’s conference featured 150 new products in a single session, which is half of the total number released last year (330) and less than 80 products throughout 2024—indicating an incredibly fast growth rate. However, upon closer inspection, many of these new products are not complete robots; components such as dexterous hands, sensors, and chips make up a large portion, suggesting that the supply chain is indeed maturing. Nevertheless, “first releases at the exhibition” do not necessarily mean the products are ready for production.
Why do manufacturers release so many new products? Simply put, it’s a way to maintain their presence in the market—after all, releasing something shows they are still active in the industry. The problem is that when almost everyone can release new products, the value of such releases diminishes. It’s similar to the Nokia era of feature phones: dozens of models were released each year, but 99% of them were quickly forgotten by consumers. The same situation applies to humanoid robots today—there are many models, but few are actually viable for practical use.
2. From “Showcasing Skills” to “Performing Tasks,” but Still a Long Way to Go
Last year, the most popular demonstrations involved robots performing acrobatics; this year, the focus shifted to more practical tasks such as sorting, handling, and palletizing. Companies like Zhishen Technology showcased robots capable of these tasks, as well as logistics systems with adaptive capabilities, and Hikvision’s cleaning robots—all indicating a move towards more useful applications.
However, there is still a significant gap between theory and practice: some robots merely repeat simple actions in simulated environments, while manufacturers claim they are still in the testing phase. Others provide real data, such as stating that their robots have undergone months of training in a factory and completed a certain number of tasks successfully. For example, some parents describe robots folding towels as “unsteady,” indicating that their movements have been slowed down for stability reasons.
In essence, the gap between demonstrating robots capable of performing tasks and actually doing so efficiently lies in the need for adaptation to real-world working conditions. No matter how impressive the demonstrations are, robots may fail when faced with dust, noise, or irregular objects in a factory setting—these issues can only be resolved through extensive on-site testing.
3. SOEs Enter the Scene + Procurement Day: The Industry Is Finally Ready to Make Money
The most significant change this year is the participation of SOEs, which have traditionally been buyers rather than exhibitors. Forty-nine SOEs, including State Grid and PetroChina, participated in the conference. Their display boards featured practical requirements such as “power inspection needs” and “pain points in oil and gas pipeline maintenance,” along with actual order lists waiting to be fulfilled by companies.
Even more crucial is the fact that the “Procurement Day” has been officially incorporated into the conference schedule for the first time. This marks an official recognition that the robotics industry must move beyond just showcasing technology towards actual business transactions. The presence of SOEs, as major buyers, brings much-needed financial momentum to the industry.
4. Data Collection Becomes a Hotcomer: Companies Selling Tools to Make Money
Another noteworthy change is the rise of data collection companies at the conference. Companies like Luming Robotics and Guanglun Intelligence have showcased their capabilities in collecting first-person perspective data, while others have provided open-source datasets of human behavior (hundreds of thousands of hours). Remote control devices and motion capture systems, which were previously used in laboratories, now have their own dedicated exhibition booths and sales teams.
The reason for this trend is that embodied intelligence requires a vast amount of real-world physical data. While language models are trained with billions of hours of data, high-quality physical interaction data globally amounts to only a few hundred thousand hours—there is a huge gap that needs to be filled. As a result, companies providing data collection tools have seen rapid growth. Guanglun Intelligence raised 1 billion yuan in funding in March and secured 550 million yuan in orders in the first quarter; Yiren Technology expects to turn a profit in 2025 and already received more orders in the first quarter of 2026 than it did throughout 2025. These companies, which previously worked in the field of intelligent driving data, have now found new opportunities in the robotics industry.
Conclusion
The core message from this year’s conference is clear: everyone has access to similar technologies. The competition now revolves around who can generate revenue first. The surge in new products, the involvement of SOEs, and the profitability of data collection all indicate a shift from a technology-driven to a business-driven industry. However, the real test lies ahead: will the robots capable of performing practical tasks truly gain a foothold in factories? And will the demands of SOEs translate into actual sales? These aspects are far more critical than simply adding more new products to exhibition displays.