On the morning of October 18, the 2025 Sustainability Global Leaders Conference was held at the Expo Park in Huangpu District, Shanghai. Co-hosted by the World Green Design Organization and Sina Group, SEER Robotics, as an innovative enterprise in the robotics field, was invited to attend.

During the "Exploring the Trends of Scenario-based Robotics Applications" roundtable forum, SEER Robotics Founder and CEO Zhao Yue engaged in an in-depth dialogue with multiple industry guests. Centering on the path and practice of robotics technology moving from proof-of-concept to large-scale implementation, he shared the company's cutting-edge insights and strategic layout in promoting the intelligent transformation of industries.

Zhao Yue stated during the dialogue that the future of the robotics industry lies in "openness and diversity." The ultimate form of robots should not be limited to "humanoid" but should be the "optimal solution" defined by specific scenarios. The key to achieving widespread accessibility of robots lies in continuously lowering the development barrier through technological innovation and building an ecosystem where global participants can co-innovate.

The following content is compiled from the live dialogue transcript:

Question 1.

From "dancing and showing off skills" to "real work," what are the bottlenecks hindering the large-scale implementation of humanoid robots?

Zhao Yue:

SEER Robotics is an intelligent robotics company that provides underlying control systems for global automation companies and machinery companies. So far, we have helped over 1500 clients worldwide create various types of robots. The robot mentioned in the host's question refers more narrowly to bipedal humanoid robots. In our view, there are two core bottlenecks for their large-scale implementation.

The first bottleneck comes from the sensor level. The current development of humanoid robots is severely constrained by sensor technology, especially in the realm of touch. Unlike autonomous driving, which primarily relies on visual data that can be processed systematically, human operation relies on millions of fine-grained tactile feedback points distributed across the skin. To achieve truly intelligent humanoid robots capable of handling complex physical interactions, the key lies in obtaining pixel-level tactile sensors through breakthroughs in materials science or biology, thereby providing a valuable data foundation for model training.

The second bottleneck lies in robot collaboration. "No single company can develop all robots, and no factory will use only one type of robot." Future robot application scenarios will be highly diverse, inevitably giving rise to robots with different forms and functions. No single company can produce all robots to meet all needs. Therefore, the second key challenge is to establish open, universal collaboration standards and an ecosystem, enabling different categories and functions of robots to collaborate efficiently within the same scenario. This is essential to maximize the value of the entire robot system and requires the collaboration and effort of the entire industry.

Question 2.

What will the form of future humanoid robots be like? Must they resemble humans? Are there other "optimal solution" forms for certain scenarios?

Zhao Yue:

The form of future robots should not be confined to the humanoid form. Pursuing a "global optimal solution" applicable to all scenarios is neither realistic nor necessary. Specific scenarios will inevitably have "optimal solutions" with specific forms.

For example, future autonomous vehicles will involve the cars themselves becoming intelligent, rather than humanoid robots driving traditional cars. The ultimate direction of technology is to make the environment and tools themselves intelligent, rather than mechanically imitating the human form to operate old tools.

The value of humanoid robots lies in their flexibility, making them suitable for scenarios requiring a high degree of imitation of human actions, such as climbing stairs in factories or performing complex equipment maintenance. However, in most production links, such as handling and sorting, specialized industry robots like robotic arms and mobile robots can perform these tasks more efficiently.

In the concept of the "unmanned factory" proposed by Musk, a high-level role is envisioned for humanoid robots: in a truly unmanned factory, humanoid robots would act as "commanders," responsible for maintaining and managing other specialized robots, thereby achieving truly unmanned operations.

SEER Robotics' mission is to "Accelerate an open and diverse intelligent civilization, Build your own robot fleet within days!". We firmly believe that the forms of future robots will be diverse and flourishing. We will continuously lower the barrier to intelligent robots, empowering all industries to define and develop the robots they need, in various forms.

Question 3.

With the development of AI, robots are beginning to undertake more complex tasks. In the next 3-5 years, what fundamental changes will occur in the collaboration model between frontline employees and their robot colleagues? What new requirements does this pose for corporate organizational structures and talent skills? How can human-robot collaboration maximize the release of social value?

Zhao Yue:

In the next 3-5 years, the fundamental changes in human-robot collaboration will be reflected in both actual production and organizational coordination. Within organizational coordination, we are applying a large number of AI Agents to reshape the way organizations collaborate.

Traditional management models view processes as "optimal practice templates," aiming to fix employees as "cogs" in the execution chain to reduce human uncertainty. But I strongly oppose this logic: a company becoming stronger should not come at the cost of "employees becoming cogs." Instead, it should be that "the company gets better because its employees get stronger." Employees must never be cogs.

The future logic of organizational coordination is shifting towards being "customer-centric." Processes should be chains that "create value for the customer end-to-end," starting from customer needs and ending with customer satisfaction and success. That is, "from the customer, to the customer." To create value for customers, we cannot rely on fixed process templates. Customer needs are diverse, and processes should also be diverse accordingly.

For example, our customer management system allows regional teams to customize differentiated processes based on the characteristics of local customers (e.g., state-owned enterprises in North China, foreign enterprises in East China, private enterprises in South China). This is supported by a unified data platform, with AI agents autonomously designing processes to deliver appropriate value to different customers.

When we were defining our strategy and values, we had a key consideration: our mission mentions "openness and diversity," so the company itself must first achieve diversity. To this end, we abolished the traditional values system to truly practice the concept of openness and diversity, avoiding limiting employees' creativity with a unified set of values.

Question 4.

How can we avoid the "digital divide"? While promoting intelligence, how can we ensure that workers of different ages and skill levels can integrate into the new ecosystem instead of being marginalized?

Zhao Yue:

Our core internal action is focusing on empowering employees, helping them use tools to optimize internal processes. Externally, our core is helping various enterprises independently develop robots. We always believe that robots should not be the patent of any particular industry giant, but should be an open, diverse undertaking of ecosystem co-construction.

Over the past decade, we have helped many automation companies achieve intelligent transformation. We have observed that many companies, after purchasing robots, find that most robot companies are unwilling to cooperate when they want to adjust the robot's functions due to scenario needs.

For instance, a client once needed a special robotic arm with a "five-meter reach but a payload of only one kilogram." While this was a niche demand, it held real value. However, mainstream suppliers could not meet this demand due to its small volume. This creates a "digital divide". Valuable demands cannot be realized due to technical barriers.

The ideal way to bridge this divide is to enable the demand side to develop independently. But this is undoubtedly challenging, as they may not possess the professional expertise in robot kinematics or supply chain management. To address this, SEER Robotics has focused on two core tasks over the past decade:

First, comprehensive compatibility with core components. We have adapted to almost all mainstream core robot components on the market. Taking LiDAR as an example, we have integrated single-line and multi-line, domestic and imported, solid-state and non-solid-state, repetitive scanning and non-repetitive scanning types. Clients can freely choose based on their specific requirements for cost, performance, safety level, etc., without needing to invest resources in underlying hardware adaptation.

Second, lowering the barrier to robot development. We are committed to making robot development simple by encapsulating complex algorithms and models. Our vision is that in the future, even an ordinary person who only knows how to use a smartphone can create the robot they need using our open platform, without needing to understand the underlying technical principles.

In the future of human-robot coexistence, it's not about us solving problems for everyone, but rather, through technological inclusivity, enabling everyone to possess the tools to "break the mold" and independently create robots that solve their own pain points.

Question 5.

China's robotics industry chain has obvious advantages. Is global expansion an inevitable choice? Under the general trend of going global, such as the "Belt and Road" initiative, what strategies and challenges do your companies have?

Zhao Yue:

Global expansion is an inevitable choice, and SEER Robotics' overseas business has also achieved rapid growth in recent years. Facing the complex international environment, the core of our overseas strategy can be summarized in two points.

First, we do not view overseas as a single market, but break it down into multiple differentiated markets. Based on the industrial characteristics, customer needs, and development stages of different regions, we carry out refined layouts and formulate targeted market strategies.

Second, we are actively exploring a new cooperation paradigm, similar to a reverse operation of the early joint venture car model. That is, we export core technology and solutions, and establish joint ventures or co-brands with local enterprises. We fully understand and respect various countries' desires to establish their own independent industry chains. Our goal is to empower overseas partners, leveraging their local supply chains and market resources, to produce robot products that better meet regional needs. For example, China's forklift industry already holds an important position globally. When overseas forklift enterprises wish to transform and upgrade to produce intelligent forklifts, we provide the key technology to support their local manufacturing of intelligent forklifts.

We always firmly believe that robots are not the privilege of any single company or country, but are tools to serve all humanity. Therefore, SEER Robotics' globalization is not simply about exporting technology or products. With "localized collaboration" as the core concept, we aim to deeply integrate with overseas markets, help local areas build robotics industry chains through cooperation, and ultimately achieve win-win development.