编者荐语： 日经中文网与36氪开展内容交换合作。将精选36氪的精彩独家财经、科技、企业资讯，与读者分享。 以下文章来源于36氪Pro ，作者海若镜 36氪Pro . 36氪旗下官方账号。深度、前瞻，为1%的人捕捉商业先机。 在薛翀的认知里：产品创新，是商业世界里 最大的杠杆。 文 ｜ 海若镜 封面来源 ｜ 日经中文网 2025年连获创新医疗等投资的3轮融资；医疗SaaS外的AI新业务，签约ARR （年度经常性收入） 增长1 2倍，达到六七千万元，2026年签约合 同额有望达1.5亿元。 一位医疗AI从业者讲道，以前他并不知道这家公司，直到2025年全诊拿下了广安门医院、常州市第一人民医院的标。广安门医院很重视数字化； 且这两个医院大模型的标，单价不低，说明他家基础不错，有现成的东西。 在薛翀看来，"有现成产品可以演示，而不是存在于PPT里"，是全诊的竞争力之一。访谈时，他顺手打开手机里"全诊通"APP，说了一段患者自 述，软件即时识别语音后，形成了结构化病历。 美国做AI智能病历的公司Abridge，2025年5月实际ARR突破1亿美元，估值拉升到53亿美元。除了在2B业务上对标Abridge，在薛翀的蓝图 ...

Core Viewpoint - Citigroup maintains a "Buy" rating for Yidu Tech (02158) and raises its target price to HKD 11, indicating a potential upside of over 96% for the stock [1] Group 1: Company Positioning - Yidu Tech is positioned as a leading player in the domestic AI healthcare sector, leveraging favorable policies, deep industry experience, and a comprehensive product layout to secure a core position in the wave of healthcare AI infrastructure development [1] - The company has embedded itself in national AI healthcare application pilot projects in regions like Beijing and Henan, enhancing operational efficiency in medical institutions through products like Doctor Copilot [1] Group 2: Policy and Market Opportunities - The macro policy push for "AI + Healthcare" in China provides a clear growth path for Yidu Tech, with plans to establish multiple national pilot bases by 2027 and achieve widespread AI technology coverage in secondary hospitals by 2030 [1] - The company's WeChat mini-program "Yidu Qianyu" demonstrates strong technical capabilities, ranking alongside top products in accuracy and traceability, thus supporting both To-G (government and hospitals) and To-C (consumer) business models [1] Group 3: Financial Projections - Citigroup recognizes Yidu Tech's substantial accumulation in medical data governance and knowledge construction, attributing unique advantages to the quality and authority of its data sourced from top hospitals [2] - The company's To-G business segment shows more predictable profitability compared to consumer applications, benefiting directly from increased hospital AI budgets and national project tenders [2] - Financial forecasts have been adjusted, with expectations for the big data platform revenue to grow by 17% and 30% year-on-year in fiscal years 2026 and 2027, respectively, with a potential break-even point in fiscal year 2026 [2]

Core Viewpoint - Citigroup maintains a "Buy" rating for Yidu Tech and raises its target price to HKD 11, anticipating a potential upside of over 96% for the stock [1] Group 1: Company Positioning - Yidu Tech is positioned as a leading player in the domestic AI healthcare sector, leveraging favorable policies, deep industry experience, and a comprehensive product portfolio to secure a core position in the wave of healthcare AI infrastructure development [1] - The company has embedded itself in national AI healthcare application pilot projects in regions like Beijing and Henan, enhancing operational efficiency in medical institutions and alleviating the burden on healthcare personnel through products like Doctor Copilot [1] Group 2: Policy and Market Opportunities - The macro policy push for "AI + Healthcare" in China provides a clear growth path for Yidu Tech, with plans to establish multiple national pilot bases by 2027 and achieve widespread AI technology coverage in secondary hospitals by 2030 [1] - The company's WeChat mini-program "Yidu Qianyu" demonstrates solid technical capabilities, ranking alongside top products in accuracy and traceability, thus supporting both To-G (government and hospitals) and To-C (consumer) business models [1] Group 3: Financial Projections - Citigroup recognizes Yidu Tech's strong foundation in medical data governance and knowledge construction, attributing its unique advantage to the large scale and quality of data sourced from top hospitals [2] - The company's To-G business segment shows more predictable profitability compared to consumer applications, benefiting directly from increased hospital AI budgets and national project tenders [2] - Citigroup has adjusted its financial forecasts for Yidu Tech, projecting a 17% and 30% year-on-year growth in big data platform revenue for fiscal years 2026 and 2027, respectively, with a potential break-even point in fiscal year 2026 [2]

Core Insights - Wang Xiaochuan has returned to the public eye after nearly a year of silence, focusing on the medical AI sector with Baichuan's new products, Baichuan-M3 and Baichuan-M3 Plus, which claim to surpass GPT-5.2 and have significantly low hallucination rates [2][4] - The medical industry presents unique challenges compared to the internet sector, where consumer sovereignty is limited, and communication barriers exist between IT and medical professionals [3][5] - Despite skepticism from the medical community, Wang Xiaochuan remains optimistic about the potential of AI in healthcare, viewing it as a long-term commitment [4][19] Group 1 - Baichuan has launched Baichuan-M3 and Baichuan-M3 Plus, with hallucination rates of 3.5% and 2.6% respectively, claiming superior diagnostic capabilities [2] - Wang Xiaochuan believes that the lack of investment from major players in the medical field presents an opportunity for Baichuan [4] - The company is focusing on developing AI products for both consumers and businesses, with plans to expand internationally [11][29] Group 2 - Communication issues between the IT and medical sectors are highlighted, with experts noting that both fields operate under different logics and terminologies [3][5] - Wang Xiaochuan's approach to medical AI has been criticized for lacking understanding of the medical industry's complexities, with experts suggesting that his views may be overly simplistic [5][6] - The medical AI landscape is competitive, with many companies struggling to find a successful business model, indicating that Baichuan faces significant challenges ahead [6][29] Group 3 - Baichuan's strategy includes leveraging partnerships with hospitals and healthcare professionals to develop AI solutions, although skepticism remains regarding the effectiveness of these collaborations [10][36] - The company aims to address the challenges of integrating AI into healthcare by focusing on localized deployment and ensuring that their products meet the specific needs of medical professionals [28][29] - Wang Xiaochuan's vision for AI in healthcare emphasizes the importance of maintaining trust between doctors and patients while enhancing medical decision-making [17][19]

Core Viewpoint - The article discusses the challenges and opportunities faced by Baichuan, a company led by Wang Xiaochuan, as it transitions into the medical AI sector, emphasizing the complexities of integrating AI into healthcare and the need for effective communication between tech and medical professionals [3][4][5]. Group 1: Company Overview and Leadership - Wang Xiaochuan has returned to the public eye after a year, making significant donations and launching new AI products, including Baichuan-M3 and Baichuan-M3 Plus, which claim to outperform existing models with lower hallucination rates [3][4]. - Baichuan's focus on healthcare is seen as a personal passion for Wang, who believes that the lack of investment from major tech companies in healthcare presents an opportunity for his company [7][8]. Group 2: Industry Challenges - The article highlights a fundamental disconnect between the IT and healthcare sectors, where the former operates on market-driven principles while the latter is characterized by a more authoritarian service model [6][8]. - Experts express skepticism about Baichuan's ability to succeed in healthcare, citing the need for deep understanding of the medical field and the challenges of communication between tech and healthcare professionals [9][10]. Group 3: Product Development and Market Strategy - Baichuan is developing AI products aimed at both consumers and businesses, with plans to collaborate with major telecom and logistics companies for market promotion [15][16]. - The company is focusing on creating AI solutions that can assist in medical decision-making without compromising the trust between doctors and patients [27][28]. Group 4: Future Outlook and Industry Dynamics - The article suggests that while the medical AI field is fraught with challenges, there is potential for growth as the industry evolves and adapts to new technologies [42][43]. - Experts believe that AI could significantly enhance the capabilities of healthcare professionals, but the integration of AI into clinical practice will require careful consideration of ethical and legal implications [46][47].

来源：子弹财经 正是在这一行业风口下，2026年1月，上海万怡医学科技股份有限公司（以下简称"万怡医学"）向港交 所主板递交招股书，由光大证券国际担任独家保荐人。 万怡医学在招股书中将自己定义为医学学术、教育及研究综合AI解决方案的头部企业，推出面向医疗 从业者的AI循证平台MedEvidence。 但翻开万怡医学近二十年的发展史，这家公司更底层的身份却是一家地道的会展公司。 "我是拒绝把AI引入我们医生的日常生活。"近期，著名医生张文宏拒绝AI进入病历系统的发言，在医疗 与科技圈激起了不小反应。 张文宏担忧年轻医生若习惯了AI生成的答案，将无法得到临床思维的训练和能力提升。 对此，AI医疗公司百川智能创始人王小川持反对意见，认为张文宏为了保护医生训练体系，而不是从 患者利益出发。 这场关于医生如何应用AI的辩论，揭示了AI医疗行业的另一面：相比于辅助患者端的问诊挂号等功 能，AI在医生端的知识赋能与临床循证支持，正成为一个更吸引人、也更具争议的市场。 随着全球医学文献与临床指南呈爆炸式增长，中国医师数量已从2019年的390万人增长到2024年的510万 人。在信息过载与诊疗压力下，如何利用AI从海量数据中 ...

2022年，创业第七年，薛翀开始跑步，用汗水来对抗郁闷和失意。 以下文章来源于36氪Pro ，作者海若镜 36氪Pro . 36氪旗下官方账号。深度、前瞻，为1%的人捕捉商业先机。 在薛翀的认知里： 产品创新，是商业世界里最大的杠杆。 文 ｜ 海若镜 来源｜ 36氪Pro（ID：krkrpro） 封面来源 ｜ Unsplash 资本寒冬、市场骤冷，谈好的5000万融资无法到账，他创立的医疗SaaS公司全诊医学只好收缩战线，退守浙江。彼时，医疗行业普遍面临大裁员的苦 楚，薛翀也是。濒死挣扎后，他决定还是挤出资源，保留一支10人小队，探索并不清晰的AI创新业务。 没想到，正是这个反常识的决策，让全诊医学翻了盘，接住了大模型时代的"馈赠"： 2025年连获创新医疗等投资的3轮融资；医疗SaaS外的AI新业务，签约ARR （年度经常性收入） 增长1 2倍，达到六七千万元，2026年签约合同额有 望达1.5亿元。 一位医疗AI从业者讲道，以前他并不知道这家公司，直到2025年全诊拿下了广安门医院、常州市第一人民医院的标。广安门医院很重视数字化；且这两 个医院大模型的标，单价不低，说明他家基础不错，有现成的东西。 在薛翀 ...

Group 1: Economic Developments - Shenzhen's Nanshan District has become the first district in China to surpass a GDP of 1 trillion yuan, achieving an average annual growth rate of over 5.8% from 2016 to 2025, with a total economic output of 652.7 billion yuan at the end of the 13th Five-Year Plan [1] - The global electric vehicle market is projected to exceed 12.1 million units sold by 2025, driven by the decreasing costs of batteries and improved charging infrastructure, with a significant contribution from the entry-level market [27] - China's sports industry is expected to surpass 5 trillion yuan in total scale by 2025, with the sports goods market reaching 2.49 trillion yuan, reflecting a surge in domestic sports consumption [27] Group 2: Corporate Actions and Strategies - Anta Sports has agreed to acquire approximately 29% of Puma's shares for 1.5 billion euros, becoming the largest shareholder of the German sports brand [3] - Ideal Auto plans to close a small number of inefficient retail stores this year, clarifying that this is a normal operational adjustment and not indicative of a significant change in business operations [11] - BYD has abandoned its plan to invest 290 million dollars in a lithium project in Chile due to slow government responses and declining lithium prices [11] Group 3: Market Trends and Performance - Tesla's vehicle registrations in Europe dropped by 20% year-on-year in December, with a total annual decline of 27%, while BYD's registrations surged over twofold in the same period [11][13] - Nike plans to lay off 775 employees to enhance profitability and accelerate automation, amid ongoing challenges with sales growth and profit margins [11] - OpenAI's advertising prices for ChatGPT are three times higher than those on Meta platforms, with projected advertising revenue exceeding 10 billion dollars by 2027 [11]

文 | 海若镜 2022年，创业第七年，薛翀开始跑步，用汗水来对抗郁闷和失意。 资本寒冬、市场骤冷，谈好的5000万融资无法到账，他创立的医疗SaaS公司全诊医学只好收缩战线，退 守浙江。彼时，医疗行业普遍面临大裁员的苦楚，薛翀也是。濒死挣扎后，他决定还是挤出资源，保留 一支10人小队，探索并不清晰的AI创新业务。 没想到，正是这个反常识的决策，让全诊医学翻了盘，接住了大模型时代的"馈赠"： 2025年连获创新医疗等投资的3轮融资；医疗SaaS外的AI新业务，签约ARR（年度经常性收入）增长12 倍，达到六七千万元，2026年签约合同额有望达1.5亿元。 一位医疗AI从业者讲道，以前他并不知道这家公司，直到2025年全诊拿下了广安门医院、常州市第一 人民医院的标。广安门医院很重视数字化；且这两个医院大模型的标，单价不低，说明他家基础不错， 有现成的东西。 在薛翀看来，"有现成产品可以演示，而不是存在于PPT里"，是全诊的竞争力之一。访谈时，他顺手打 开手机里"全诊通"APP，说了一段患者自述，软件即时识别语音后，形成了结构化病历。 美国做AI智能病历的公司Abridge，2025年5月实际ARR突破1亿美元， ...

Core Insights - OpenEvidence has rapidly gained traction in the medical field, achieving a valuation of $12 billion and annual revenue exceeding $150 million within just four years of its establishment [1] - The company addresses a critical gap in the medical industry by providing a free tool for doctors that significantly reduces the time needed to access reliable medical information [4][5] - OpenEvidence's business model revolves around monetizing the attention of healthcare professionals and providing targeted advertising for pharmaceutical companies [7][9][10] Group 1: OpenEvidence's Rise - OpenEvidence has become the primary entry point for doctors by effectively addressing the overwhelming volume of medical knowledge and the limitations of traditional databases [2][3] - The platform utilizes a retrieval-augmented generation (RAG) approach, allowing doctors to obtain accurate information in just three seconds, thus enhancing decision-making efficiency [4] - The company has achieved viral growth, with monthly active users reaching 400,000 and covering approximately 34% of practicing physicians in the U.S. [5] Group 2: Revenue Generation - OpenEvidence generates revenue by providing targeted advertising to pharmaceutical companies during critical decision-making moments for doctors [8][9] - The platform's ability to deliver compliant and relevant advertising content has made it an attractive option for drug companies looking to reach physicians effectively [10][12] - Additionally, OpenEvidence sells its core capabilities as APIs to hospitals and medical schools, further diversifying its revenue streams [11] Group 3: Challenges for Chinese Competitors - Chinese companies face significant challenges in replicating OpenEvidence's success due to data integration difficulties and the lack of open access to authoritative medical databases [15][16] - Trust issues arise in China regarding pharmaceutical advertising alongside clinical decision tools, making it difficult for companies to monetize similar models [17][18] - The high workload of Chinese doctors limits their ability to engage with tools like OpenEvidence, necessitating a more practical approach tailored to local conditions [19][20] Group 4: Competitive Landscape - JD Health focuses on a model that combines tools, supply chain, and services, but faces trust issues due to potential biases in its recommendations [23][24] - Alibaba Health aims to develop a comprehensive medical operating system but struggles with the transactional aspect of its services [25][26] - Ant Group's approach with its AI tool "Afu" seeks to integrate deeply into the medical workflow, potentially offering a more complex but rewarding business model [27][28] Group 5: Future Outlook - The medical AI market in China is expected to diversify, with different players targeting various segments, such as serious medical scenarios and primary care [29] - The key lesson from OpenEvidence for Chinese companies is to effectively use free tools to capture high-value users and monetize their needs [29]