Orient: Accelerating the commercialization of AI image, improving data foundation, and opening up platform space.

Orient released a research report stating that as of the end of June 2026, the national medical insurance cloud imaging index has reached nearly 440 million records, with 5 provinces completing the provincial-level cloud imaging procurement. The report believes that the medical insurance cloud imaging will drive the industry's business model from single-hospital software projects to provincial platform construction, continuous operation and maintenance, and value-added application service upgrades. The AI imaging revenue model is also expected to shift from one-time projects to software authorization, subscription, and payment based on the number of examinations. As the scarcity of registration certificates decreases, industry competition is shifting focus to landing capabilities, and enterprises with equipment coordination, regional platform construction, and cross-selling capabilities across multiple disease types are expected to drive revenue realization at lower marginal customer acquisition costs. Key points from Orient are as follows: Data infrastructure is gradually improving, opening up space for AI imaging applications. As of the end of June 2026, a total of 31 provinces and the Xinjiang Production and Construction Corps have uploaded nearly 440 million records to the medical insurance cloud imaging index. 5 provinces and regions, including Guizhou, Xinjiang, and Guangxi, have completed the provincial-level cloud imaging procurement, while Tianjin, Inner Mongolia, Hainan, and other regions are progressing. With the gradual standardization of image data storage, retrieval, and interface standards, the data sources for AI imaging are expected to expand from single-hospital data to regional imaging platforms. The application scenarios will also extend from assisting in the diagnosis of single diseases such as pulmonary nodules and fractures to cross-hospital retrieval, image quality control, identification of repeated examinations, and comprehensive analysis of multiple diseases. The report believes that medical insurance cloud imaging will drive the industry's business model from single-hospital software projects to provincial platform construction, continuous operation and maintenance, and value-added application services upgrades. Industrial value is expected to shift from single algorithm authorization to data governance, PACS transformation, cloud resources, and AI application platform diffusion. The connection between hospital-end applications and payments is gradually improving, and the AI imaging revenue model is expected to upgrade. For AI imaging to be commercialized in volume, it needs to pass through the three stages of imaging data access, clinical use in hospitals, and medical service charges. On the application side, as of the end of March 2026, 85 designated medical institutions in Jinan have been connected to the medical insurance cloud imaging, with a cumulative collection and upload of 8.07 million imaging examination data, forming a business cycle of image data collection, indexing upload, patient retrieval, and direct settlement of medical insurance. This provides a scalable application foundation for embedding AI algorithms in clinical workflows. On the payment side, the pricing policy for radiological examinations has included AI-assisted diagnostic technologies such as fast medical imaging diagnosis in the price extension items, indicating that AI imaging is shifting from the cost of hospital informatization to measurable medical service value. The report believes that with the expansion of hospital access and the implementation of local fee policies, the AI imaging revenue model is expected to shift from free trials, scientific cooperation, and one-time project construction to software licensing, annual subscriptions, and payment based on the number of examinations. Core business indicators will also shift from the number of trial hospitals to the number of paid examinations, single-service prices, renewal rates, and revenue cycles. Product supply is rapidly expanding, and industry competition is shifting focus to landing capabilities. The newly approved scenes for AI imaging have expanded from mature areas such as CT and DR to nuclear medicine PET-CT, reproductive genetic pathology, digestive endoscopy, cervical pathology, and grassroots tuberculosis screening. In June, Union Imaging's intelligent chest nuclear medicine image assisted triage software obtained the first category three certificate in the domestic nuclear medicine AI field, bringing the company's total number of AI applications that have obtained NMPA category three certificates to 20; Numerical Technology's number of category three certificates has reached 19, and the product matrix of leading companies is gradually improving. The report believes that as the scarcity of registration certificates decreases, industry competition will increasingly depend on hospital access, clinical workflow compatibility, fee closure, and scalable delivery capabilities, and investment evaluation should shift from the number of certificates obtained to the number of paid hospitals, single-hospital examination volume, renewal rates. Enterprises with equipment coordination, regional platform construction, and cross-selling capabilities across multiple disease types are expected to drive revenue realization at lower marginal customer acquisition costs. Risk warning The procurement and implementation of the medical insurance cloud imaging may not meet expectations, and the implementation of local fee policies may not meet expectations; AI product hospital access, clinical use, and fee conversion may be lower than expected; Medical institutions' investment in informatization and pressure on returns may pose risks.