What happens when a robot damages itself on stage during a corporate event, or suddenly malfunctions and kicks a bystander during a mall interaction—how can losses from such unexpected risks be mitigated when robots "go to work"? Clearly, as embodied AI robots rapidly integrate into countless industries, their insurance protection systems are undergoing a critical transformation—shifting focus from the manufacturing end to the application end. Although insurers had previously begun exploring embodied AI robot insurance, industry insiders note that coverage largely focused on production and manufacturing processes, making it difficult to apply across numerous real-world scenarios. However, with recent market developments—such as Ping An Property & Casualty Insurance collaborating with Shanghai Electric to launch the nation’s first insurance policy for an embodied AI robot financial leasing project in Shanghai on January 4, and Qintianzu partnering with PICC Property and Casualty Insurance in December last year to introduce customized protection plans—the path for insurers to cover application-scenario risks through "insurance + leasing" models is becoming clearer. The implementation of the "insurance + leasing" model is helping to safeguard the commercial application of robots. As cutting-edge carriers of artificial intelligence, embodied AI robots are already being deployed in diverse scenarios such as mall services, event performances, and restaurant interactions—yet they urgently require insurance protection tailored to application-level risks. Recent market observations indicate that the "insurance + leasing" model is beginning to support the large-scale commercial application of embodied AI robots. On January 4, Ping An Property & Casualty Insurance formally signed a cooperative insurance agreement with Shanghai Electric Financial Leasing Co., Ltd. and Shanghai Electric Insurance Broker Co., Ltd., marking the country’s first "insurance + financial leasing" policy for embodied AI robots in a real commercial application scenario in Shanghai. According to He Ying, General Manager of Ping An Property & Casualty Insurance Shanghai Branch, the insurance plan not only covers conventional robot hardware protection but also includes comprehensive safeguards such as third-party liability, product quality liability, and information leakage liability. Additionally, in December last year, Qintianzu and PICC Property and Casualty Insurance Hangzhou Branch officially announced the launch of customized risk protection for robots, covering risks such as equipment damage and malfunctions during leasing, as well as compensation for third-party personal injury or property damage caused by accidents during robot operation. "Every robot on our platform is required to have insurance. As an emerging technological product, various unexpected incidents or risks are inevitable for robots. Having insurance protection is a highly effective safeguard for both robot users and owners," said Li Yiyan, CEO of Qintianzu, in an interview. In fact, insurers had already made some inroads into the embodied AI robot sector, with several companies previously launching related insurance products. However, in He Ying’s view, insurance protection needs in this field still face three major challenges: first, core technical data is highly confidential, making it difficult for insurers to accurately assess risks; second, most robot manufacturers are in their startup phase, and traditional insurance service models struggle to meet their flexible needs and cost considerations; third, moral hazard prevention is more challenging under single-device ownership models. These factors contribute to a mismatch between supply and demand for insurance services in the embodied AI robot industry, creating bottlenecks in product innovation and market implementation. "Previously, insurance products in this field were mostly applied to risk protection during the robot manufacturing process but lacked coverage for specific application scenarios. On the other hand, the high purchase price of individual robots and issues of information asymmetry made it difficult to cover single-unit risks with insurance, which is a core reason why such insurance hasn’t been widely adopted," He Ying explained. Li Yiyan also believes that insurance can deliver greater value across the diverse application scenarios of robots compared to the factory environment. Industry experts suggest that the "insurance + leasing" model, with leasing platforms acting as central intermediaries, treats batches of robots as collective insured objects, covering both robot damage and third-party liability risks in application scenarios. By collaborating to build localized data analysis platforms, partners can effectively address challenges such as difficulties in insuring single devices and assessing risks. The insurance sector is eyeing new opportunities in a market projected to be worth billions. According to Li Yiyan’s estimates, 2025 is being called the "first year of mass robot production" by the industry, and the robot leasing market is expected to exceed a scale of 10 billion yuan by 2026. Although the "insurance + leasing" model has taken initial steps, there is clearly vast room for expansion in the embodied AI robot insurance sector, with future upgrades possible in coverage scope, service boundaries, and technological empowerment. In terms of coverage scope, Li Yiyan stated that current insurance products mainly cover embodied AI robot damage and basic third-party liability, but could be extended in the future to include risks from natural disasters, fires, earthquakes, and other force majeure events, as well as damage during robot transportation, to meet diversified needs as scenarios expand. Regarding service boundaries, He Ying suggested using the current "insurance + financial leasing" model as a blueprint to extend services to the entire lifecycle of embodied AI robots—including R&D, production, testing, operation, maintenance, product liability, and cybersecurity—while offering bundled financial protection solutions such as corporate property insurance, employer liability insurance, and credit guarantee insurance to comprehensively support the high-quality development of the robot industry. On data sharing, to overcome challenges posed by missing historical data and difficulties in data acquisition for embodied AI robot insurance, He Ying recommended that insurers collaborate with industry players and academic institutions to build a shared risk control database. By jointly researching, accumulating, and sharing risk data from embodied AI robots across different scenarios, they can proactively explore and lead the establishment of risk assessment standards, technical specifications, and industry service guidelines to promote a transparent, standardized, and healthy insurance service ecosystem.