IDC 报告显示：上半年智元出货规模与营收断层领先

Overall: The most important takeaway for general readers is that humanoid robots have moved from concept showcases to real jobs. To assess industry maturity, look at three indicators: shipment volumes, deployment scenarios, and actual daily working hours.

1. Shipment data: In the first half of 2026, global humanoid robot shipments reached approximately 25,000 units, up 432.1% year-over-year; China shipped more than 19,000 units, accounting for 77.9% of the global total. Zhiyuan shipped over 8,600 units, representing 35% of the global market and 45% of China's market, making it No.1 in both shipments and revenue.

2. Shifting priorities: The competitive focus has moved from robot appearance, hardware specs, and motion capabilities to understanding the physical world, task planning, manipulation execution, and continuous learning. Whether a robot is "smart and useful" now matters more than whether it can walk or jump.

3. Real-world deployments: Zhiyuan deployed more than 300 robots at Chimelong, covering seven scenarios including entertainment and commercial performances, science education tours, guided navigation and shopping, retail services, intelligent companionship, hotels, and sports, operating eight hours per day. At Aishida store, 100 robots were deployed as shopping guides, each on duty for over 10 hours per day.

4. Outlook: IDC forecasts global shipments will exceed 750,000 units by 2030, as humanoid robots enter more industrial, commercial-service, and consumer markets.

Overall: For humanoid robot brands, the real competition is no longer at the spec-sheet launch event; it lies in whether you can build a full-stack self-developed technology foundation and achieve large-scale shipment and real-world deployment the way Zhiyuan has.

1. Brand building: Zhiyuan uses its "Three Intelligences in One" architecture as a technology brand, validating its products in landmark projects such as Chimelong and Aishida. The brand story is supported by deployed cases, not concepts.

2. Channels and in-store experience: Deploying 100 Lingxi X2 robots as "silicon-based shopping guides" across shelves, checkout counters, and experience zones shows that half-sized humanoid robots are better suited for physical retail space. Brands can use them to proactively greet customers, explain products, and guide them through the store, forming a human-machine collaborative service loop.

3. Product R&D: Zhiyuan offers four product lines---Yuanzheng, Lingxi, Jingling, and Kuto---which suggests that brands should define product forms by scenario: full-sized, half-sized, wheeled, and quadruped robots each serve different roles.

4. Consumer trends: Over three consecutive days of livestreaming, the robots proactively engaged with customers and participated in selling, reflecting rising consumer acceptance of tech-enabled interactive service. The emergence of a retail toolkit means brands can standardize service workflows and replicate them across more stores.

Overall: The most useful insight for sellers is that humanoid robots are becoming a new productivity tool in stores and livestreams. The retail scenario already has replicable toolkits, creating fresh service and monetization opportunities worth watching.

1. Market opportunity: In the first half of 2026, China shipped more than 19,000 humanoid robots, accounting for 77.9% of global shipments; IDC predicts global shipments will exceed 750,000 by 2030. If your business touches robot rental, experiential retail, or livestream content, the growth upside is clear.

2. Demand shift: Consumers are already accepting proactive greeting, product explanation, and wayfinding from robots in stores. Robots can also become a content anchor in livestreams to attract viewers and drive interaction, proving that "human-machine collaborative selling" is a workable retail expression.

3. Practices to adopt: Aishida deployed 100 robots in store as shopping guides, each on duty for over 10 hours daily, and ran livestreams for three consecutive days. Zhiyuan has distilled a retail-scenario toolkit, which sellers can adapt to connect greeting, guidance, and livestreaming into a standardized process.

4. Risks and entry points: Large-scale robot deployment requires significant upfront investment, so it suits branded chains looking to pilot a model first. Smaller sellers can start with a single store or a few robots, testing traffic-driving effects via event marketing and livestream content before deciding whether to scale up.

Overall: For factories, this article offers two takeaways: first, humanoid robot manufacturing itself creates new production orders; second, when factories introduce robots, they should start from real job requirements rather than buying machines as showpieces.

1. Manufacturing and supply-chain opportunity: China shipped more than 19,000 humanoid robots in the first half of the year, with Zhiyuan shipping over 8,600 and emphasizing mass-production delivery capability. This indicates growing demand for structural parts, motors, sensors, and complete-machine OEM services upstream.

2. Product design needs: Zhiyuan uses four product types---full-size, half-size, wheeled, and quadruped---to cover different scenarios. This suggests factories involved in robot-related manufacturing should design and configure around the customer's actual space, task content, and operating hours, rather than competing on appearance alone.

3. Digital and e-commerce implications: Zhiyuan kept robots on duty for over 10 hours per day at Aishida and participated in three days of livestreaming. Factories can apply similar approaches to factory showrooms and production-line livestreams, turning robot employees into digital marketing touchpoints.

4. Deployment evaluation: The article emphasizes moving from production line to real job positions. When evaluating a robot, factories should assess whether it can enter the actual workflow, operate reliably every day, and create value, rather than remaining a demonstration.

Overall: For service providers, the core message is that customers need "deployment-ready" robots that can stay on the job continuously, not just a piece of hardware. This generates extensive service demand around long-duration operation, scenario adaptation, and outcome-based operation.

1. Technology trend: Humanoid robots are integrating large models, world models, VLA, and data loops. The competitive focus has shifted to understanding the physical world, task planning, manipulation execution, and continuous learning, giving service providers opportunities to offer expertise in algorithm tuning, data collection, and scenario training.

2. Customer pain points: Chimelong's cultural tourism scenarios involve high footfall and strong interference, requiring eight hours of continuous daily operation; Aishida's retail scenario demands more than ten hours of on-duty time per day. The core pain point is ensuring stable, error-free, maintainable robot performance.

3. Solution approach: Zhiyuan uses its "Three Intelligences in One" technology foundation and has distilled a toolkit from its retail projects. Service providers can turn specific scenario workflows into standardized toolkits and provide one-stop services spanning deployment, training, livestream operation, and remote maintenance.

4. Business opportunity: From hundreds of interactive points in theme parks to storefront shopping guides, the service scope has moved beyond selling equipment to scenario planning, content updates, human-robot collaboration workflow design, and data analytics. Service providers can pivot to a long-term operational service model.

Overall: For marketplace operators, this article shows that humanoid robots are forming a new commercial supply: they can serve as in-store service carriers and as livestream content carriers. Platforms should build their infrastructure around scenario toolkits, data loops, and operating rules.

1. Platform demand: Merchants need to put robots into use quickly, without figuring out deployment on their own. Marketplace operators can integrate robot hardware, scenario solutions, livestream operations, and data monitoring, offering a "Robot-as-a-Service" matching and management system.

2. Latest practice: Zhiyuan and Aishida built a store shopping-guide team with 100 robots and distilled a retail-scenario toolkit. Platforms can adopt this combination of "toolkit + store + livestream" and export standard templates to chain merchants.

3. Recruitment and operations: A large customer like Chimelong deployed over 300 robots and hundreds of interactive points in one go. Platforms can recruit robot solution providers around theme parks, commercial streets, and chain retail, and match them through a project-based model.

4. Risk mitigation: The industry is moving from a hardware-specs race to a competition in embodied-AI comprehensive capabilities. Platforms should not only look at motion performance when recruiting vendors but also assess task planning, scene understanding, continuous learning, and on-site service ability, and accept projects based on criteria such as long-duration operation and high-footfall interference.

Overall: This article provides empirical evidence that the humanoid robot industry has entered a critical phase of commercialization: rapid market growth, a paradigm shift in competition, and the emergence of business models. Researchers should continue to observe from technological, industrial, and governance perspectives.

1. Industry data: Global shipments in the first half of 2026 were about 25,000 units, up 432.1% year-over-year, with China accounting for 77.9%. Zhiyuan shipped over 8,600 units, representing 35% globally and 45% in China, making it the first company to reach large-scale shipment and commercial deployment.

2. Technology paradigm: The competitive focus has shifted from hardware structure and motion performance to comprehensive embodied-AI capabilities. Large models, world models, VLA, and data loops together shape new research directions. A robot's ability to "learn and adapt to its environment" is now more critical than the physical body itself.

3. Business model: Zhiyuan uses its "Three Intelligences in One" foundation to support deployment-oriented projects in cultural tourism and retail, and forms replicable solutions through retail-scenario toolkits. This indicates that the embodied-AI business model is moving from selling devices to selling job-ready services and scenario operations.

4. Policy and governance issues: Large-scale robots operating long-term in open spaces with high footfall raise new questions around human-robot safety, privacy protection, service standards, data governance, and liability allocation. Researchers can use the Chimelong and Aishida cases to propose impact assessments for job replacement, standards for robot deployment, and governance recommendations for data loops. Given IDC's forecast of over 750,000 shipments by 2030, governance rules need to be prepared in advance.