赛力斯机器人曝光 车企为何自研机器人？

This article features Seres' first public unveiling of its self-developed embodied intelligent robot, and outlines the current industry trend of automakers rushing into the humanoid robot track. Key takeaways are as follows:

1. Seres' new "Xiaosai" robot series adopts a differentiated R&D path: the robot hardware is developed in-house, while the AI "brain" is built in partnership with Volcano Engine. It took less than one year from the signing of the partnership to the deployment of the robot on the production line, a go-to-market speed far outpacing the industry average.

2. Instead of following the crowd to develop general-purpose full humanoid robots, Seres defined robot forms based on specific industrial tasks, and built a heterogeneous robot cluster to fill the gap in flexible tasks that traditional industrial robots cannot complete. The robots have already been deployed at Seres' super factory.

3. Chinese regulators have launched a special initiative for on-site training of humanoid robots and embodied intelligence, explicitly naming industrial manufacturing as the priority application scenario. The industry broadly views 2026 as the first year of mass production for humanoid robots, and the track has now entered a critical phase of use case validation.

This article outlines the current industry trend of automakers entering the embodied intelligent robot space, and provides multiple insights for brand owners on product R&D and new track expansion. Key takeaways are as follows:

1. For R&D strategy: Seres adopts a division-of-labor development model, focusing on its core strength of robot hardware development and outsourcing AI brain development to a specialized tech provider. This model replicates the successful partnership that built the AITO brand with Huawei, drastically shortening the R&D cycle and reducing in-house development risk — a model well worth learning for brands entering new tracks.

2. For growth outlook: Against the backdrop of intense competition in both hardware and smart vehicle features, embodied intelligent robots have become a core new growth track for automakers. Leading players including Xiaomi, Xpeng and BYD have already entered the space, and policy clearly supports industrial adoption. Beyond B2B manufacturing, C-end retail, service and companion robots represent large untapped consumer markets with broad growth potential.

3. For production value: Embodied intelligent robots can fill the gap in flexible assembly work, adapt to the faster product iteration cycle of smart vehicles, improve production efficiency, and help brands keep up with the industry's fast iteration trend.

This article discloses the current development status of the embodied intelligent robot track, and provides opportunity and risk insights for industry practitioners. Key takeaways are as follows:

1. Market opportunities: Policy has explicitly designated industrial manufacturing as the priority landing scenario for embodied intelligence, with 2026 marked as the first year of mass production for humanoid robots. The track is entering a phase of accelerated commercial deployment, and embodied data has become a core competitive resource. Multiple embodied data startups have secured hundreds of millions of yuan in Series A funding, and large tech firms are all building out their own data platforms, creating significant growth opportunities for related component and service suppliers.

2. Model reference: Seres has pioneered a third path alternative to both full-stack in-house development and the traditional robot-vendor pilot model: in-house hardware development + third-party AI brain + task-based robot definition. This approach delivers far faster deployment, achieving production line integration in less than a year, with lower costs and smaller trial-and-error risks than full-stack development — a strong model for new entrants to learn from.

3. Risk warning: The industry is now increasingly focused on return on investment. It remains to be seen whether launching multiple robot categories in a short timeframe can deliver profitability. Practitioners should avoid blind follow-the-crowd expansion, and prioritize real-world deployment performance and ROI.

This article shares Seres Super Factory's practical experience in deploying embodied intelligent robots, which offers strong reference value for factories pursuing digital and intelligent upgrades. Key takeaways are as follows:

1. Clear upgrade needs: For most auto plants, the stamping, welding and painting processes already achieve over 80% automation, with some leading plants reaching 100% automation. However, assembly line automation only hits 20% to 30%, as a large number of non-standard flexible tasks cannot be completed by traditional programmed equipment. This is exactly the core entry point for embodied intelligent robots, which can fill the gap in existing automation systems.

2. replicable deployment strategy: Instead of blindly chasing general-purpose humanoid robots, Seres reverse-engineered robot forms based on the plant's existing automation workflow and specific tasks. It matched different task requirements with different robot forms including wheeled robots, humanoid robots with inspection guns, and fixed robotic arms, to build a heterogeneous embodied intelligence cluster. This approach seamlessly integrates with existing factory systems and directly improves production line efficiency, making it far more practical than the general-purpose robot-first approach.

3. Tangible efficiency gains: Embodied intelligent robots adapt to new vehicle models via visual recognition and large model inference, eliminating the months-long programming debugging required by traditional automation. This adapts to the twice-as-fast product iteration cycle of smart vehicles compared to traditional fuel vehicles, and helps factories speed up new product launches. Seres' robots have already completed initial deployment validation, proving the commercial path is viable.

This article sorts out the current deployment trends and customer pain points in the embodied intelligence industry, and points out clear business directions for related service providers. Key takeaways are as follows:

1. Industry development trend: Embodied intelligence has now entered a critical phase of industrial deployment, with 2026 as the first year of mass production for humanoid robots. China's Ministry of Industry and Information Technology and State-owned Assets Supervision and Administration Commission have jointly launched a special initiative that names industrial manufacturing as the priority application scenario. Driven by both policy and market demand, demand for embodied intelligence-related services in the industrial sector will grow substantially, creating large market opportunities for service providers.

2. Core customer pain points: Manufacturing factories do not actually need single-function general-purpose humanoid robots; they need system solutions that directly improve overall production line efficiency. Currently, many robot vendors blindly pursue general-purpose designs that cannot integrate with factories' existing digital ecosystems and fail to solve real production problems — this is exactly the core opportunity for service providers.

3. Partnership and business directions: Industrial specialization is becoming increasingly clear: manufacturing players excel at hardware development and on-site deployment, and need professional service providers to deliver core technologies and services including large AI models, cloud-edge collaboration, and embodied data. The Seres-Volcano Engine partnership is a typical example. In addition, embodied data has become an industry-wide刚需, creating large market demand for data operation and data supply services.

This article outlines the changing demand for platforms driven by the development of the embodied intelligence industry, and provides reference for platforms to layout related businesses and avoid industry risks. Key takeaways are as follows:

1. Demand insight: The commercialization of embodied intelligence is now following a clear trend of specialization, with a new deployment path that separates hardware development from AI brain development. Manufacturing players with production capacity and on-site scenarios take charge of hardware development, and need professional platforms to provide core technical support including large AI models and multi-modal cloud-edge collaboration. Seres' partnership with Volcano Engine is a typical example, representing a new growth opportunity for technology platforms.

2. Core business direction: Embodied data has become a core resource competed by the entire industry. Large tech firms including Baidu and JD.com have already moved first to layout embodied data platforms to capture market share in data operation. Platforms can capitalize on this industry trend, and build targeted embodied data services and industrial matching platforms to meet industry development needs.

3. Reference for recruitment and operation: A growing number of manufacturing players with production capacity are entering the robot space, and these companies already have in-house scenarios and deployment capabilities. Platforms can carry out targeted recruitment, build a matching ecosystem aligned with the division-of-labor industry trend, and connect players across hardware, AI technology, data and other segments to reduce deployment costs for the entire industry. This also helps avoid the industry pitfall of blindly chasing general-purpose humanoid robots while ignoring real-world deployment performance.

This article discloses the latest developments in the commercialization of the embodied intelligence industry, and provides new cases and research directions for industry research. Key takeaways are as follows:

1. New industry development trend: A large number of automakers are now entering the embodied intelligent robot space, and have forged a path different from the traditional approach of existing robot vendors, which expand from C-end to B-end or vice versa. They have created a new "factory ecosystem + robot" paradigm, where robots are natively deployed in factories and reverse-defined based on the factory's existing digital ecosystem. This is an entirely new commercialization paradigm worthy of in-depth research.

2. Business model innovation: Seres has developed a third R&D path alternative to full-stack in-house development and traditional robot vendor-factory partnership, with a division-of-labor business model of "in-house hardware + outsourced AI brain". This replicates the successful cooperation experience of the AITO brand, delivers far faster deployment than traditional models, and provides a new sample for industrial division of labor and collaboration.

3. New research topics: The industry has shifted from the technology validation phase to a focus on return on investment. When automakers enter this track, they aim not only to fill the gap in flexible production work and adapt to the fast iteration industry trend, but also to capture the core resource of embodied data. This entirely new industry competition logic creates new research topics for studying modern industrial competition. In addition, the government's launch of a special initiative to drive industry deployment also provides new context for industrial policy research.