西南证券发布研报称，中国医疗人工智能行业空间较大，政策推动行业快速发展；讯飞医疗科技 (02506)具备自研基座大模型，性能上限持续抬升。星火医疗大模型X1能够显著降低幻觉现象，提升推 理的正确性、专业性与可解释性；基于大模型和大数据技术平台，构建多元化GBC客户产品矩阵。首 次覆盖，建议关注。 西南证券主要观点如下： 基于大模型和大数据技术平台，构建多元化GBC客户产品矩阵 截至2024年底，公司已向超过7万家基层医疗机构提供产品和服务，业务覆盖全国30多个省份、670多个 区县，以及500多家等级医院，其中包含了中国百强医院中的40多家和十强医院中的7家。1)To G基层医 疗服务业务25H1收入增速52.3%；区域管理平台解决方案25H1收入增速178.1%：基层医疗服务+区域管 理平台解决方案是基本盘，基层医疗服务业务线主要由智医助理和慢病管理组成，区域管理平台解决方 案主要是由智慧卫生和智慧医保组成；2)To B业务25H1收入下滑10%：主要由于大模型持续更新。业务 包括智慧医院+AI诊疗助理，提升医疗服务能力与效率；3)To C业务25H1收入增速10%：包括影像云平 台+诊后管理+讯飞晓医+智 ...

Core Insights - The company reported a revenue of 299 million yuan for the first half of 2025, representing a year-on-year growth of 30.26% [1] - Gross profit reached 154 million yuan, with a year-on-year increase of 26.96% [1] - Despite ongoing losses, the net loss attributable to the parent company decreased to 74.09 million yuan, a reduction of 42.86% year-on-year [1] Business Segments - Government Segment: Revenue from grassroots solutions grew by 52.3% to 83.81 million yuan, with the core product "Smart Medical Assistant" covering 697 districts and counties and over 75,000 grassroots institutions, completing over 1.01 billion AI-assisted diagnostic suggestions [2] - Regional Solutions: Revenue surged by 178.1% to 57.63 million yuan, with AI infectious disease prevention solutions applied in multiple provinces [2] - Hospital Segment: Revenue from hospital solutions reached 52.86 million yuan, a decrease of 10% year-on-year, with the deployment of the company's large medical model in over 20 leading hospitals [2] - Consumer Segment: Patient management service revenue was 104 million yuan, up 10.1%, with the AI health assistant "Xunfei Xiaoyi" achieving over 24 million downloads and completing 140 million AI consultations [2]

Core Insights - Keda Xunfei reported a data resource of 61.74 million yuan in its 2024 annual report, making it the highest listed company in the software and information technology services sector in terms of reported amounts [2] - The company has been focusing on the intelligent voice and artificial intelligence sectors while also expanding into the smart healthcare sector, with R&D investment showing a consistent upward trend over the past three years [2][11] - In 2024, Keda Xunfei's R&D investment reached 1.996 billion yuan, a year-on-year increase of 24.62% from 1.602 billion yuan in 2023 [2] Smart Healthcare Development - Keda Xunfei's AI diagnostic assistant is expected to experience rapid growth, with significant room for penetration in the 2,843 counties across mainland China [7] - As of December 31, 2024, the AI diagnostic assistant has covered over 670 counties in more than 30 provinces, with applications in over 70,000 grassroots medical institutions, providing approximately 910 million AI-assisted diagnostic suggestions [7] - The accuracy of assisted diagnoses has improved from about 90% in 2021 to approximately 95% currently, while prescription review accuracy has also increased from about 93% to 95% [7] Competitive Landscape - Keda Xunfei has previously challenged the market share claims of its competitor, Yunzhisheng, which is attempting to go public after multiple unsuccessful attempts [8][10] - In 2023, Keda Xunfei's medical AI revenue was the largest in China's medical AI industry, with a market share of 61.5% in the CDSS market for grassroots medical institutions [11] - Other companies in the software services sector, such as Donghua Software and Shenzhou Digital, are also actively entering the smart healthcare field, with various AI-based products and solutions being developed [12][13] R&D Investment Comparison - Among the listed companies, Daily Interactive has the highest R&D investment as a percentage of revenue at 36.13%, followed by Keda Xunfei at 19.62%, while Shenzhou Digital has the lowest at only 0.32% [15]

Core Viewpoint - The article emphasizes that AI will not replace traditional Chinese medicine (TCM) but rather enhance the capabilities of TCM practitioners who can effectively utilize AI tools [6][21]. AI in Traditional Chinese Medicine - AI has the potential to assist in diagnosing and treating patients by providing reference prescriptions based on symptoms, as demonstrated by tools like DeepSeek [4][5]. - The consensus among industry professionals is that the future of TCM lies in practitioners who can leverage AI rather than AI replacing practitioners [6][21]. Current Applications of AI in TCM - Young TCM doctors reported limited use of AI during face-to-face consultations, primarily utilizing it for administrative tasks like creating teaching materials and work summaries [9]. - AI is seen as a tool for analyzing complex cases and expanding diagnostic thought processes, particularly in Western medicine contexts [10]. Challenges in AI Implementation - The complexity of TCM's theoretical framework poses challenges for AI, as it requires structured data for effective training, which is often lacking [11]. - AI struggles with the "four diagnostic methods" (observation, listening, inquiry, and pulse-taking) essential in TCM, leading to potential misdiagnoses [12]. Future Directions for AI in TCM - AI can improve patient-doctor matching by assessing TCM practitioners' capabilities and guiding patients to suitable doctors based on their symptoms [17]. - In the diagnostic and treatment phases, AI can assist in analyzing patient information and drafting medical documents, allowing doctors to focus more on patient interaction [18]. AI's Role in Enhancing TCM Practice - The development of AI tools aims to elevate the diagnostic capabilities of TCM, with products like AI-assisted consultation and digital representations of renowned TCM practitioners being explored [19]. - AI can also facilitate patient follow-ups and health management, enhancing the overall patient experience [20]. Conclusion - The article concludes that the integration of AI into TCM is not about replacement but about collaboration, where AI acts as a supportive tool for practitioners to improve their practice and patient outcomes [21].