United Imaging Intelligence
一家有国家资本背书、已有真实部署规模、私募估值约 $1.4B 的中国医学影像 AI 龙头,但缺少可支撑该价格的独立经济数据。
United Imaging Intelligence 具备真实的国资背书部署规模和可信平台,但没有任何独立经济性披露,公开证据很难支撑约 $1.4B 标记。
封面要素
公司概况
United Imaging Intelligence(UII)是一家上海私营医疗 AI 公司,2017 年 12 月成立,是上海上市公司 United Imaging Healthcare(UIH,688271.SH)旗下专门的 AI 子公司。其 uAI 平台覆盖 12 个产品平台、100 多个 AI 应用,横跨 CT、MR、PET/CT、X-ray 和超声;2025 年开始转向多模态医疗基础模型「Yuanzhi」以及 uAI NEXUS 套件。公开证据支持其监管组合很深(13 张 NMPA 三类证、15 项 FDA、31 项 CE)、覆盖中国 4,000 多家机构,并在 2025 年 6 月以接近 $1.4B 估值完成 Series A,但公司未披露独立收入、利润率、员工人数或股权条款。
- 成立时间
- 2017-12-01
- 创始人
- Xiang "Sean" Zhou, Dijia Wu
- 创立地点
- Shanghai, China
- 总部
- Shanghai, China
- 产品
- uAI 软件套件覆盖疾病检测、量化、自动报告和临床工作流应用,经由 uAI Clinical Portal 交付;同时包括 Yuanzhi 多模态基础模型,以及 uAI NEXUS 图像、语言、语音和视频模型,既可嵌入 UIH 扫描仪,也可运行在第三方硬件上。
- 客户
- 医院和医疗机构,尤其是中国三甲医院,借助母公司 UIH 已安装的影像设备基础,并在欧洲和美国逐步增加部署。
- 商业模式
- 向医疗机构销售 AI 软件,以影像工作流附加软件和母公司设备渠道变现;独立定价和收入模式未公开披露。
- 阶段
- Series A private
- 融资情况
- 2025 年 6 月完成约 RMB 1B($139M)Series A,由 E Fund Management 和 Shanghai Guotou 共同领投,投后估值约 $1.4B,母公司 UIH 参与;此前轮次和累计融资总额未公开披露。
执行摘要
主要优势
- E Fund、上海国投等国资关联锚定方,加上上市母公司 UIH,带来少见的融资耐久性和政策一致性。
- 已在中国 4,000 多家机构、全球 3,000 多家机构落地,显示真实临床拉力。
- 监管组合很深:13 项 NMPA Class III、15 项 FDA、31 项 CE 许可,相比获证更薄的竞争者形成门槛。
- uAI 平台覆盖广(12 个平台、100+ 应用),2025 年又转向多模态基础模型,支撑长期平台叙事。
- 可优先使用母公司 UIH 的影像数据、硬件分销渠道和品牌信任,独立创业公司很难匹配。
主要风险
- 未披露独立收入、利润率或留存,因此约 $1.4B 标记隐含的倍数不可知,也可能偏高。
- 中国医院 AI 变现弱,Airdoc 等上市同行仍在亏损,说明付费意愿是真风险。
- 对母公司 UIH 的结构性依赖很重,集中关键股东和治理风险,也让独立估值更复杂。
- 地缘政治暴露(BIOSECURE 式措施、FDA 对中国器械的审查、数据治理张力)威胁出口乐观情景。
- NMPA Class III 变更控制和不成熟的 AI 支付政策,可能让获证变成成本中心,而不是收入驱动。
未决问题
- United Imaging Intelligence 的独立收入、ARR、增长和母公司相关收入拆分完全未披露。
- uAI 软件的毛利率、实际定价和捆绑 / 付费组合未公开。
- 4,000+ 部署足迹中的付费机构数量、使用深度和留存未知。
- 与母公司 UIH 围绕分销、IP、算力和定价的关联交易条款未披露。
- Series A 后股权结构表、董事会构成和独立员工数未公开。
目录
01公司概览
1.1 身份、产品与商业模式
United Imaging Intelligence 将自己定位为一家国际化医疗 AI 公司,构建覆盖完整临床路径的人工智能软件,从筛查、诊断延伸到治疗和随访。其中国法律主体为上海联影智能科技股份有限公司,总部位于上海市徐汇区云锦路 701 号。公司是 United Imaging Healthcare Technology Group(UIH)旗下专门的 AI 子公司;UIH 是上海证券交易所科创板上市的医学影像硬件制造商(688271.SH)。UII 以「uAI」品牌销售软件,覆盖 CT、MR、PET/CT、X-ray 和超声等模态。 对一家私营医疗 AI 公司而言,UII 的产品面异常宽。到 2025 年中,管理层称公司已有 12 个产品平台、100 多个 AI 应用,并围绕 uAI Clinical Portal(uCP)组织,把疾病检测、量化和报告应用打包进医院工作流。2025 年,UII 将叙事重心转向名为「Yuanzhi」(元智)的多模态医疗基础模型,以及由图像、语言、语音和视频模型组成的 uAI NEXUS 套件,从单病种检测器转向多病种「agent」产品,例如一次胸部 CT 扫描可筛查 73 项胸部发现的 agent。 商业模式是向医疗机构销售软件,并高度借力母公司已安装的影像设备基础。UII 称其 AI 产品已进入中国 4,000 多家医疗机构,软件既作为嵌入式智能运行在 UIH 扫描仪上,也可运行在第三方硬件上。公司称约 80% 员工为研发人员,更像一家软件和模型公司,而非设备制造商。公开信息仍缺少独立收入、利润率和定价,投资人无法脱离 UIH 单独承销这门生意。[CO001, CO002, CO003, CO004, CO005, CO006]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 |
|---|---|---|---|---|
| 成立 | 2017 年 12 月(上海) | 2017-12-22 | 高 | |
| 总部 | 上海市徐汇区云锦路 701 号 | 2026-06-23 | 高 | |
| 母公司 | United Imaging Healthcare(688271.SH) | 2026-06-23 | 高 | |
| 最新融资 | RMB 1B(约 $139M)Series A | 2025-06-20 | 高 | |
| 投后估值 | 约 $1.4B(投前约 RMB 9B) | 2025-06-20 | 中 | 确认准确的投前 / 投后拆分及任何优先权条款 |
| AI 应用 | 12 个平台上 100+ 个 | 2025-07 | 中 | 将应用数量与产生收入、由医院付费的应用对齐 |
| 监管许可 | 13 项 NMPA Class III;15 项 FDA;31 项 CE | 2025-07 | 中 | 核对厂商口径与 FDA 38 个获批设备清单 |
| 已部署机构 | 中国 4,000+;全球 3,000+ | 2025-07 | 中 | 拆分已安装与已付费;定义活跃部署 |
| 收入 / ARR | 2026-06-23 | 低 | UII 未披露独立收入;索取财务数据 | |
| 员工人数 | null(>80% 研发) | 2026-06-23 | 低 | 索取当前员工人数和研发占比 |
公开指标混合了 NMPA/FDA/CE 许可,以及中国与全球部署数量;null 表示所审阅材料未公开披露。
[CO001, CO007, CO013, CO016, CO018, CO020]展示母公司连接、数据、基础模型和装机基础如何在 UII 经营模型中衔接。
[CO003, CO005, CO006, CO007, CO022, CO031]1.2 创始人、管理层与治理
UII 的领导层以 Xiang "Sean" Zhou 为核心;公司和第三方资料均将其列为联合创始人、(联席)CEO 及公司法定代表人。Zhou 长期研究医学影像和机器学习,也是公司在行业论坛上的公开面孔,包括 2025 年 World Artificial Intelligence Conference(WAIC),他在会上阐述了多模态 agent 战略。Dijia Wu 被列为首席技术官,资料数据库还显示公司有一支深厚的研究副总裁团队,包括 Feng Shi、Yaozong Gao 和 Shanhui Sun。公司强调其全球化招募的 AI 研究团队分布在上海、北京、武汉,以及位于马萨诸塞州 Burlington 的波士顿地区子公司 UII America, Inc.。 治理由母公司关系主导。United Imaging Healthcare 既是公司母公司,也是 Series A 的持续投资方,意味着 UIH 创始人 Xue Min 所在集团拥有实质控制权,AI 子公司的战略与上市母公司紧密绑定。这种绑定是一把双刃剑:UII 因此获得影像数据、硬件分销渠道和品牌信任的优先入口,但也集中了承担关键人物和关键股东风险,并让独立治理叙事更复杂。 公开信息没有披露董事会构成、Series A 后股权表,以及 UII 与 UIH 之间经济利益的精确划分。36Kr 在 2025 年报道称 UII 正在考虑独立上市;若推进,治理和财务披露会被迫更充分。在此之前,对董事席位、投资人权利和与母公司的关联交易条款做尽调,主要仍依赖私有材料。研究团队深度是真实优势,但治理不透明和母公司控制,是身份图景中最清楚的开放问题。[CO008, CO009, CO010, CO011, CO012, CO034]
| 人员 | 职务 | 背景 | 创始人市场匹配或职能覆盖 | 关键人物依赖 |
|---|---|---|---|---|
| Xiang "Sean" Zhou(Zhou Xiang,联合创始人) | 联合创始人、(联席)CEO、法定代表人 | 长期从事医学影像和机器学习研究;曾在 WAIC 2025 和行业论坛公开发声。 | 直接匹配医学影像 AI 领域;负责战略、融资和多模态智能体路线图。 | 高 |
| Dijia Wu | 首席技术官 | 资深医疗 AI 研究者,负责模型和平台工程。 | 掌握核心技术和基础模型开发。 | 高 |
| Feng Shi;Yaozong Gao;Shanhui Sun(研发负责人) | 副总裁,研发 | 资料库点名的研究负责人,支撑深厚 AI 梯队。 | 证明创始人之外的研究深度。 | 中 |
| United Imaging Healthcare(Xue Min group,母公司) | 母公司及控股股东 | UIH 是上海上市母公司,剥离出 UII 并继续提供支持。 | 提供数据、渠道、资本和品牌;同时集中控制权。 | 高 |
本表只列出审阅公开来源中可见的领导者和控股母公司;不是完整组织架构或董事会名单。
[CO008, CO009, CO010, CO011, CO034, CO035]1.3 融资历史、投资人和规模
UII 融资历史中最硬的数据点,是 2025 年 6 月的 Series A。36Kr、DealStreetAsia 和 Insight 等多家媒体报道称,UII 完成约 RMB 1B(约 $139M)融资,由 E Fund Management(易方达)和 Shanghai Guotou Investment(Shangguo,上国投资)共同领投,Shanghai Alliance/United Investment、Sun Rock Capital、Suzhou Capital Group、Tsing Song(Qingsong)Capital、Junwei Venture Capital、Zhangjiagang Venture Capital、Wuhan Optics Valley industrial investment,以及母公司 United Imaging Healthcare 参与。不同来源对共同领投方和参投方命名略有差异;部分转载将 SIG Property Management 与 E Fund 并列为领投方,这一点值得与股权表核对。 估值口径把 UII 明确推入独角兽区间。报道显示其投前估值接近 RMB 9B(约 $1.25B),融资后投后估值约 $1.4B,使 UII 在成立约七年后成为中国估值最高的独立医疗 AI 公司之一。公司称融资将用于多模态大模型研发、医院数智化系统和商业化铺开,而非战略转向。 规模声明很强,但对分母敏感。UII 引用 12 个产品平台、100 多个 AI 应用、13 张 NMPA 三类医疗器械证、15 项美国 FDA 清除和 31 项 CE 认证,产品已部署到中国 4,000 多家医疗机构,并在全球完成 3,000 多家医院部署。母公司 UIH 报告 2025 年收入约 RMB 13.8B,同比增长约 34%;这能说明 UII 所依托的渠道,但不是 UII 自身收入。公开记录支持其监管和部署规模已相当可观,但 UII 的独立收入、ARR、员工人数和烧钱速度仍未披露,必须视为缺口。[CO013, CO014, CO015, CO016, CO017, CO018]
| 利益相关方 | 角色 | 控制权或经济重要性 | 尽调问题 |
|---|---|---|---|
| E Fund Management(易方达) | Series A 联席领投方 | 联席领投 2025 年融资;锚定本轮的大型中国资产管理机构。 | 索取领投条款、董事 / 观察员权利和持股比例。 |
| Shanghai Guotou / Shangguo(上国投资) | Series A 联席领投方 | 国资相关上海投资管理人,联席领投本轮。 | 澄清国资背书影响和治理权利。 |
| United Imaging Healthcare(688271.SH) | 母公司及参与投资方 | 控股母公司;提供数据、硬件渠道和品牌。 | 梳理关联交易协议、IP 授权和收入分成条款。 |
| Shanghai Alliance / United Investment | 参与投资方 | 本轮点名的上海相关投资方。 | 确认持股和任何战略任务。 |
| 投资方包括 Sun Rock Capital;Suzhou Capital Group | 参与投资方 | 财务基金和区域国资相关基金共同进入投资团。 | 确认参与金额和 pro-rata 权利。 |
| 投资方包括 Tsing Song(Qingsong)Capital;Junwei VC;Zhangjiagang VC | 参与投资方 | 创投和区域基金补足投资团。 | 将完整投资团名单与 cap table 核对。 |
| Wuhan Optics Valley industrial investment(产业投资方) | 参与投资方 | 与 UII 武汉布局相关的区域产业投资平台。 | 确认任何本地部署或设施承诺。 |
投资者名单由新闻报道重建;cap table 和精确持股未公开,联席领投名称在不同来源中略有差异。
[CO013, CO014, CO015, CO017, CO038]截至 2026-06-23,按证据权重评估监管规模、资本可得性、母公司杠杆和披露质量。
[CO013, CO020, CO022, CO024, CO038, CO039]1.4 里程碑、监管进展和负面信号
UII 的里程碑从 2017 年 12 月成立开始,随后快速累积监管清除并发布基础模型。公司在 2020 年凭 COVID-19 胸部 CT 分诊软件获得全国关注,之后取得多张 NMPA 三类证;自 2024 年起,又获得美国 FDA 510(k) 清除,包括用于颅内出血分诊的 uAI Easy Triage ICH,以及 uAI Portal 后处理平台。按 FDA 2025 年底统计,「United Imaging」持有 38 项已清除放射 AI 设备,在所有厂商中排名第五,仅次于 GE HealthCare、Siemens Healthineers、Philips 和 Canon,并领先 Aidoc。 产品发布动能延续到 2025–2026 年。RSNA 2025 上,UII 发布 uAI Insight 图像到报告 agent 和具身 AI 机械臂超声系统;ECR 2026 上,公司展示 30 多个 CE 认证应用和 uCP 门户的 60 多个应用;2026 年 4 月,公司开源 uAI NEXUS MedVLM,这是一款被 CVPR 2026 接收的医学视频语言模型。2025 年 6 月 Series A,以及报道称公司考虑独立 IPO,共同构成一段推进很快的近期历史。 负面和怀疑信号主要是结构性问题,而非丑闻驱动。UII 所处的中国医疗 AI 市场拥挤且部分商品化,AI 报销机制仍不成熟,许多已获批产品难以把清除转化为医院付费收入;公司对母公司 UIH 在数据、渠道和资本上的高度依赖,本身就是集中度风险;中美技术和数据治理摩擦,也会给一家寻求西方市场的中国医疗 AI 厂商带来出口和采购不确定性。市场、财务、风险和估值章节会展开这些主题。[CO024, CO025, CO026, CO027, CO028, CO029]
| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2017-12 | UII 在上海成立,作为 United Imaging Healthcare 的 AI 子公司。 | 创立 | 公司成立 | United Imaging Healthcare;Xiang Zhou(相关方) | 确立成立日期和母公司关联。 |
| 2020 | UII 在疫情期间部署 COVID-19 胸部 CT 分诊 AI。 | 产品 | 疫情 AI 部署 | UII;中国医院 | 早期全国知名度和临床 AI 可信度。 |
| 2024-09 | UII 获得美国 FDA 510(k) 许可,包括 uAI Easy Triage ICH。 | 监管 | FDA 许可 | 美国 FDA;UII America | 打开进入美国市场的受监管路径。 |
| 2025-06-20 | UII 完成人民币 10 亿元 Series A,由易方达和上海国投联席领投。 | 融资 | 约 $139M,投后约 $1.4B | E Fund;Shanghai Guotou;UIH;其他 | 将 UII 定价为医疗 AI 独角兽,并为模型建设提供资金。 |
| 2025-07 | WAIC 2025 上,UII 发布多模态「元智」模型和智能体战略。 | 产品 | 基础模型发布 | UII;Xiang Zhou | 将 UII 从单病种工具重新定位为多模态智能体。 |
| 2025-07 | 36Kr 报道 UII 正在考虑独立上市。 | 治理 | IPO 考虑中 | UII;36Kr | 预示未来可能出现披露和流动性事件。 |
| 2025-12 | RSNA 2025 上,UII 首发 uAI Insight 图像到报告智能体和机械臂超声。 | 产品 | 智能体产品发布 | UII;RSNA | 推向端到端「智能体放射科医生」工作流。 |
| 2026-03 | FDA 2025 年底清单显示 United Imaging 有 38 个获批放射 AI 设备,总排名第 5。 | 规模 | 38 个 FDA 获批设备 | 美国 FDA;The Imaging Wire | 独立确认相对全球同行的监管规模。 |
| 2026-04 | UII 开源 uAI NEXUS MedVLM 医学视频模型,并被 CVPR 2026 接收。 | 产品 | 开源模型发布 | UII;CVPR 2026 | 展示前沿研究产出和开发者触达。 |
| 2026-05 | ECR 2026 上,UII 展示 30+ 个 CE 认证应用和 uCP 60 应用门户。 | 产品 | 欧洲展示 | UII;ECR | 释放欧洲商业和监管扩张信号。 |
时间线反映审阅公开材料中明确标注日期的里程碑;仅在该公开来源集合内穷尽。
[CO001, CO013, CO024, CO025, CO026, CO027]追踪 UII 从 2017 年成立,到 FDA 许可、2025 年 Series A 轮,以及 2026 年基础模型和欧洲里程碑。
[CO001, CO013, CO024, CO025, CO026, CO028]1.5 图表
02市场分析
2.1 市场边界与现状替代方案
United Imaging Intelligence 竞争的是医学影像 AI 市场:这是一类把机器学习用于 CT、MRI、X-ray、超声、PET 和核医学扫描的软件,服务检测、量化、分诊和报告。这个边界最直接相关:纳入的支出,是医院和影像中心用于诊断 AI 软件、工作流工具及其背后基础模型的预算;排除的支出,是影像硬件本身(母公司 UIH 的业务)、通用医院 IT 和非影像临床软件。几乎每个 UII 应用的现状替代方案,都是没有 AI 辅助的人类放射科医生阅片;这个替代方案在决策点上不额外收费,才是 AI 必须替代或增强的真正竞争者。 公开机构对这一品类的测算跨度很大且口径不一。Grand View Research 估计全球 AI 医学影像市场 2025 年约 USD 1.8B,到 2033 年升至 USD 20.2B,CAGR 约 35%;The Business Research Company 估计 2026 年为 USD 5.29B,到 2030 年达到 USD 20.28B,CAGR 约 39.9%;Coherent Market Insights 则仅将 2026 年估为 USD 2.20B。因此,2026 年全球数字大致落在 USD 2.2–5.3B,取决于使用哪家的品类定义;任何单一 TAM 幻灯片都应被打上问号。 在该品类中,需求按应用和模态集中。神经学是 2025 年收入占比最高的单一应用领域,约 37%;计算机断层扫描是领先模态,约占 34–42%;肿瘤学则因癌症发病率上升,属于增长最快的应用之一。医院是主导终端用户,占支出超过 60%;全球放射科医生短缺,是把 AI 拉入工作流的最清晰结构性驱动因素。UII 宽广的 uAI 组合——胸部 CT、神经、肿瘤、心血管等——直接映射到这些最大细分市场。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 相关性 |
|---|---|---|---|---|
| AI 医学影像软件(核心) | 面向 CT、MRI、X-ray、超声、PET 的检测、分诊、量化和报告 AI | 影像硬件、通用医院 IT、非影像临床软件 | 医院放射科及医院资本 / IT 预算 | UII uAI 产品组合的直接相关核心市场 |
| 中国医学图像分析软件 | 面向中国医院的云端和本地部署图像分析平台 | 纯 PACS 存储和设备维护 | 三级医院;医院或省级预算 | UII 最窄且直接相关的公开中国视角 |
| 临床决策支持与报告 AI | 图像到报告智能体、放射报告 NLP、量化 | 通用医院 ERP 和计费 | 临床医生为使用者;医院 IT 为付款方 | UII 正通过 uAI NEXUS 和智能体扩展进入的邻近市场 |
| 更广义医疗 AI | 药物发现、远程医疗、医院运营、基因组分析 | AI 范围内但影像之外的项目不纳入 | 混合企业和政府医疗预算 | 仅作上行视角;会夸大 UII 当前可服务收入 |
| 现状替代方案(无辅助阅片) | 既有人员编制吸收放射科医生劳动力成本 | 净新增软件许可 | 医院人员预算 | AI 必须替代或增强的决策时免费基准 |
纳入支出反映 UII 公开销售的影像 AI 任务;排除支出标出只与其部分重叠的邻近硬件和软件预算。
[CM001, CM007, CM010, CM016, CM020]把最窄、最直接相关的中国影像软件视角,与 UII 叙事触达的更广中国和全球资金池放在一起。
这些是语境视角,不是可相加层级;底部节点明确是公开数据缺口,而非数值估计。
[CM002, CM005, CM010, CM011, CM018]2.2 规模测算视角与公开 SAM 逻辑的局限
UII 的核心地域是中国;最干净的窄口径公开视角,是中国医学图像分析软件。Grand View Research 将该市场 2022 年规模估为 USD 422.8M,并预计到 2030 年 CAGR 约 24.9%。这是最直接相关的分析师品类,但相对于 UII 自称覆盖 4,000 多家机构而言,该口径既小又旧,意味着 2022 年后市场可能快速增长,或 UII 捕获的份额已超出这个窄定义。 更宽口径的市场大得多,但也更松。中国整体医疗 AI 市场 2023 年约 USD 1.59B,按一家机构数字,预计 2028 年达到 USD 7.33B(CAGR 约 42%),2030 年达到 USD 18.88B;另一家机构预计中国医疗 AI 2025 年为 USD 2.67B,到 2035 年升至 USD 84.11B。这些数字圈出了 UII 的相邻上行空间——临床决策支持、报告、医院运营——但它们把药物发现、远程医疗和非影像分析也纳入其中,因此高估了 UII 今天可触达的收入。专门的中国放射 AI 平台市场如今被作为 2026 年快速增长细分市场单独跟踪,是最接近 UII 实际服务市场的代理指标。 诚实的尽调结论是:已审阅的公开来源没有给出干净的 UII 专属 SAM 或 SOM,无法把其付费医院收入机会与相邻池子拆开;中国估算本身在不同出版方之间也相差一个数量级。软件和解决方案,而非硬件,是中国医疗 AI 支出的主导且最快增长板块;政府投资也支撑该品类,例如 2018 年宣布的 USD 2.1B 北京 AI 产业园。北美在 2025 年仍持有最大区域份额(约 44–45%),亚太则是增长最快区域,因此 UII 位于最高增长地域,但不是最大收入地域。[CM010, CM011, CM012, CM013, CM014, CM015]
| 发布方 | 年份 | 地域 | 数值 | CAGR | 方法 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|
| Grand View Research | 2025 | 全球 | USD 1.8B | 到 2033 年 35.1% | AI 医学影像市场快照 | 中 | 类别定义不同于同行;影像 AI 范围较宽 |
| The Business Research Company | 2026 | 全球 | USD 5.29B | 到 2030 年 39.9% | 医学影像 AI 市场报告 | 中 | 2026 年数值最高;类别边界可能最宽 |
| Coherent Market Insights 预测 | 2026 | 全球 | USD 2.20B | 到 2033 年 34.8% | 医学影像 AI 市场报告 | 中 | 2026 年数值最低;范围较窄,显示估计分散 |
| Grand View Research | 2022 | 中国 | USD 422.8M | 到 2030 年 24.9% | 中国医学图像分析软件市场 | 中 | 范围窄但数据偏旧;是最接近 UII 的直接中国视角 |
| Made-in-China Insights | 2023 | 中国 | USD 1.59B | ~42% 至 2028 年 | 中国医疗 AI 市场估算 | 低 | 包含非影像 AI;高估了 UII 可服务市场 |
| Towards Healthcare | 2025 | 中国 | USD 2.67B | 41.2% 至 2035 年 | 中国医疗 AI 前景 | 低 | 长期预测区间很宽;医疗 AI 口径过宽 |
这些口径只能框出机会边界,无法定义干净的 SAM;仅 2026 年全球数字就横跨 USD 2.2-5.3B, 中国估算相差一个数量级。
[CM002, CM003, CM004, CM005, CM010, CM011]保留 2026 年全球 AI 医学影像市场低、中、高公开估计之间的差距。
所有数值均为十亿美元。区间刻意展示估计不确定性,而不是制造一个虚假的单一 TAM;中位只是简单中点,不是第四个来源。
[CM002, CM003, CM004, CM005, CM010]2.3 买方、用户和付款方分层
UII 的主要买方是中国三级医院,尤其是最先进的三甲医院;这些机构影像量最高,拥有资本预算和临床研究任务,也最先采用 AI。到 2026 年,已有 1,200 多家三甲医院采用 AI 辅助诊断系统,用于放射、病理或两者兼有,说明顶层医院已进入主流渗透阶段。第三方影像中心是医院之外正在出现的次级渠道。 医院内部,用户和付款方并不相同。用户是放射或病理科室及其临床医生;付款方是医院资本和 IT 预算,有时由省级卫健部门采购补足,因为全国没有专门为这类软件直接付费的 AI 报销项目。因此,预算归属分散在医院行政、科室和省级部门之间,AI 软件要与其他资本优先事项竞争,而不是动用一个封闭资金池。典型采用路径从科室试点开始,到科室部署,再到全院使用,偶尔扩展为省级铺开;每一步都被采购规则和临床验证要求卡住。 UII 享有一种由母公司身份带来的结构性买方优势:United Imaging Healthcare 的影像设备销售,把 UII 软件带入同一批医院账户,因此 AI 会与买方已在采购的扫描仪一起销售或嵌入其中。更广义上,国内厂商受益于中国集中化医疗数据和极大患者量,相比面临数据本地化壁垒的西方进入者,拥有训练数据优势。中国医院采用 AI,主要是为了缓解放射科医生短缺、标准化诊断,并满足国家数字健康任务;这恰好与 UII 领先的应用方向对齐。[CM021, CM022, CM023, CM024, CM025, CM026]
| 细分市场 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 三甲医院 | 医院管理层与科室主任 | 放射科医生与病理科医生 | 医院资本开支 / IT 预算 | 大量 CT/MRI/X 光阅片与报告 | 医院管理层 | 放射科医生短缺与数字医疗硬性要求 |
| 省级 / 集中采购 | 省级卫健部门 | 区域内多家医院 | 省级医疗预算 | 区域内标准化诊断 | 省级卫健部门 | 国家 AI 战略与医疗公平目标 |
| UIH 设备客户(母公司渠道) | 采购 United Imaging 扫描仪的医院 | 影像技师与放射科医生 | 设备 + 软件合并预算 | 嵌入扫描仪或随扫描仪销售的 AI | 医院采购部门 | 硬件加 AI 捆绑采购决策 |
| 第三方影像中心 | 独立影像中心运营方 | 中心放射科医生 | 中心运营预算 | 外包阅片与分诊 | 中心所有者 | 吞吐量与成本压力 |
| 低线 / 县级医院 | 县级医院管理层 | 全科临床医生 | 受限的医院预算 | 专科医生稀缺场景下的筛查与分诊 | 医院管理层 | 专科医生短缺;远程诊断推进 |
各细分市场里,用户(临床医生)与付款方(医院或省级预算)并不相同;AI 没有单独报销项目, 支出要和其他资本开支优先级竞争。
[CM021, CM022, CM023, CM024, CM025, CM026]UII 可以通过多类医院和采购买方赢单,但每类买方的付款方和采用触发点都不同。
[CM021, CM022, CM025, CM026, CM027, CM031]2.4 增长驱动、约束与估值相关性
最强增长驱动来自政策和劳动力。中国国家 AI 战略和数字健康任务推动医院采用 AI,放射科医生短缺则让自动分诊和量化在经济上说得通。监管方面,NMPA 根据临床风险,将 AI 诊断软件作为二类或三类医疗器械监管;其创新医疗器械特别审查程序为开创性 AI 提供加速审批路径,UII 已借此累积三类证。2026 年,NMPA 发布更新后的 AI 分类指导,并提出覆盖 AI 和影像的 80 多项行业标准计划;第 52 号公告(2026)则规范了产品如何重新分类,例如从二类转为三类。这些举措增加清晰度,也抬高了新进入者需要提交的证据和成本门槛。 约束同样具体,并会压在估值上。第一,报销:中国 DRG 医院支付体系通常不单独报销 AI 影像软件,成本由医院预算吸收;报销不成熟叠加临床验证门槛高,使许多已获批产品卡在试点,难以放大为付费部署。第二,数据治理:根据中国 PIPL,医疗健康数据属于敏感个人信息,需要单独同意和数据本地化;2026 年 1 月 1 日生效的跨境数据传输认证措施,也给任何把数据移出境外的厂商增加合规负担。第三,市场估算宽且互相矛盾,投资人无法用单一可信 TAM 锚定 UII 估值。 从估值看,市场背景是双刃剑。UII 位于最高增长地域和最大需求应用领域,拥有特权母公司渠道和可信的创新器械监管路径——这些都支撑其约 USD 1.4B 估值。但缺少报销项目、规模来源相差一个数量级、三类审批负担和数据治理成本,正是把巨大理论市场与真实付费收入隔开的摩擦。市场真实且快速增长;尽调要证明的是,UII 能把部署转化为可报销或已纳入预算的经常性收入。[CM031, CM032, CM033, CM034, CM035, CM036]
| 驱动因素 / 约束 | 方向 | 时间 | 含义 | 尽调问题 |
|---|---|---|---|---|
| 国家 AI 战略与数字医疗硬性要求 | 驱动 | 当前 | 自上而下的政策推动医院采用 AI 诊断。 | UII 需求有多少来自政策拉动,多少来自临床拉动? |
| 放射科医生短缺与影像量上升 | 驱动 | 当前 | 自动分诊和定量分析在经济账上变得说得通。 | UII 每个部署点带来多少可衡量的生产率提升? |
| NMPA 创新医疗器械通道 | 驱动 | 当前 | 加速审评奖励先发 AI,更快拿到注册批准。 | UII 有多少项批准走了创新器械通道? |
| AI 软件没有单独 DRG 报销 | 约束 | 当前 | 成本由医院预算吸收,付费规模化放慢。 | UII 部署中,付费、捆绑或免费各占多少? |
| PIPL 数据本地化与跨境规则 | 约束 | 当前 | 敏感健康数据规则抬高合规成本,也限制出口。 | UII 出海时如何处理数据治理? |
| 三类审批与临床验证负担 | 约束 | 近期 | 高证据门槛拉长从获批到收入的周期。 | UII 从 NMPA 获批到首笔付费销售,平均需要多久? |
| 市场估算宽且互相矛盾 | 约束 | 当前 | 没有一个可信的 TAM 能托住估值。 | 管理层能否提供分细分市场收入,用来落出更干净的 SAM? |
同一组政策力量既创造需求,也抬高证明门槛;UII 必须证明,部署能转成可报销或已纳入预算的经常性收入。
[CM031, CM032, CM033, CM034, CM035, CM036]展示中国医院通常如何从试点走向付费、规模化 AI 部署,以及商业化在哪里卡住。
漏斗数值是顺序等级,不是实测转化率;它们展示公开材料所描述的部署顺序和商业化掉队点。
[CM028, CM031, CM033, CM038, CM039]2.5 图表
03竞争对手
3.1 竞争格局与对手画像
United Imaging Intelligence 面向一个高度拥挤的市场销售产品:公开追踪器显示,截至 2026 年,中国活跃医学影像 AI 厂商有数百家,获得充足融资的竞争者也有数十家。竞争集合横跨六层——中国本土直接纯软件公司、生态巨头、平台型设备厂商、全球纯软件公司、现状替代方案和医院内部自建——UII 与每一层都有重叠。 直接中国同行是最显眼的竞争者。Infervision(Tuixiang)2016 年成立于北京,八轮融资约 USD 213M,聚焦 X-ray 和 CT 诊断。DeepWise(Shenrui)成立于 2017 年,获 Legend Capital 和 CICC 支持,融资约 USD 115M,专注癌症和卒中放射 AI,并在 2026 年中国医学装备展发布新的多模态影像平台。Airdoc(Eaglevision,02251.HK)2021 年 11 月在港交所成为中国第一家 AI 医学影像 IPO,估值约 RMB 5.8B,专攻视网膜筛查,并报告 2025 年收入约 USD 24.1M。Shukun Technology 同样成立于 2017 年,领先于心血管和肿瘤 AI,2026 年正准备由 Goldman Sachs、Sequoia 和 Qiming 支持的香港 IPO;Yitu Healthcare 则在 600 多家顶级医院运行癌症筛查和 NLP 诊断。 纯软件公司之外,是生态巨头和既有厂商。Ping An Smart Healthcare 以一体化金融医疗 AI 套件覆盖 170 多个城市、40,000 多家机构;Tencent Miying(AIMIS)提供云端影像 AI,拥有已获三类证的产品,并接入 Tencent 生态。平台型既有厂商——GE HealthCare、Siemens Healthineers、Philips 和 Canon——把 AI 直接嵌入扫描仪和 PACS,并主导 FDA 授权数量。UII 的位置很独特:像既有厂商一样,它通过母公司 UIH 掌握硬件渠道;又像纯软件公司一样,它销售宽广的独立 AI 组合。[CP001, CP002, CP003, CP004, CP005, CP006]
| 竞争对手 | 类别 | 规模 / 融资 | 目标细分市场 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Infervision (Tuixiang) | 中国直接专营厂商 | 8 轮约 USD 213M;Series D | X 光与 CT 诊断、肺部 | 早期进入者,注册批准组合大 | 2026 年没有巨额融资;估值未披露 |
| DeepWise (Shenrui) | 中国直接专营厂商 | 约 USD 115M;Series D(Legend、CICC) | 肿瘤与卒中放射 AI | 新多模态平台仍在证明规模 | 肿瘤细分赛道拥挤;定价不透明 |
| Airdoc (Eaglevision, 02251.HK) | 中国直接专营厂商(上市) | 2021 年 HKEX IPO;约 RMB 5.8B;2025 年收入约 USD 24.1M | 视网膜 / 糖尿病视网膜病变筛查 | 首家上市 AI 影像厂商;全球筛查触达 | 眼科聚焦较窄;收入底盘薄 |
| Shukun Technology | 中国直接专营厂商 | VC 支持;筹备 2026 年香港 IPO(Goldman、Sequoia、Qiming) | 心血管与肿瘤 AI | 心血管细分领导力强 | IPO 时点有风险;估值未披露 |
| Yitu Healthcare | 中国直接专营厂商 | VC 支持;600+ 顶级医院 | 肿瘤筛查与临床 NLP | 医院关系深,聚焦筛查 | 母公司重组历史;透明度低 |
| Ping An Smart Healthcare | 生态巨头 | 隶属 Ping An 集团;40,000+ 家机构 | 金融医疗一体化 AI,覆盖 3,000+ 种疾病 | 保险 + 医疗生态交叉销售 | 影像只是多条业务线之一;不是影像专科厂商 |
| Tencent Miying (AIMIS) | 生态巨头 | Tencent 支持;已获三类批准 | 云端影像 AI(肺炎、青光眼、结肠) | 云规模与 Tencent 生态整合 | 影像是战略副线,不是核心收入 |
| GE / Siemens / Philips / Canon | 平台在位者(硬件 + AI) | FDA 放射 AI:130 / 95 / 58 / 48 | 扫描仪与 PACS 内嵌 AI | 硬件渠道与最多注册批准数量 | 在细分应用上不如专营厂商灵活 |
| Aidoc / Viz.ai / Annalise.ai 竞品 | 全球专营厂商 | Aidoc >USD 500M;Viz.ai 约 USD 252M(估值 $1.2B) | 企业临床 AI、卒中、广覆盖 | 基础模型、CMS 报销、规模 | 中国准入有限;监管与数据壁垒 |
规模数字取自最新公开披露;中国专营厂商估值大多未披露,美国 / 全球同业则运行在不同报销制度下。
[CP002, CP003, CP005, CP006, CP007, CP008]按中国医院分销强度和企业软件广度与开放性绘制竞争对手;UII 分销强,但在开放企业广度上处于中游。
坐标轴是基于已引用融资、审批和部署数据得出的有证据支撑的序数评分,并非实测市场坐标。
[CP011, CP012, CP014, CP019, CP030, CP035]3.2 能力、定价和分销对比
从监管和清除规模看,平台型既有厂商大幅领先。FDA 2026 年 Q1 名单显示,GE HealthCare 以 130 项放射 AI 授权排名第一,其后是 Siemens Healthineers(95)、Philips(58)和 Canon(48);United Imaging 以 40 项排名第五,领先 Aidoc(33)。全球纯软件公司中,Aidoc 2026 年 4 月完成由 Goldman Sachs 领投的 USD 150M Series E,累计融资超过 USD 500M,并运营 aiOS 企业平台和 CARE 临床基础模型;后者获得首个综合基础模型分诊系统 FDA 清除,已部署到近 2,000 家医院,每年分析 6,000 多万例病例。Viz.ai 已融资约 USD 252M,估值 USD 1.2B,覆盖约 1,600–2,000 家医院,并在 CMS 报销的大血管闭塞卒中检测上领先。Annalise.ai 现为 Harrison.AI 子公司,以宽度差异化,单个工具覆盖 124 多项胸部 X-ray 发现和 130 多项 CT 脑部发现。 资本和变现图景暴露了 UII 的竞争缺口。分析师评论指出,资本正转向具备规模、监管深度和工作流所有权的企业基础设施 AI 厂商,而不是单一用途工具;仅 Aidoc 一家的累计融资就已超过 UII 约 USD 139M Series A 的三倍。在美国,CMS 报销——包括 2026 年用于冠状动脉和主动脉钙化分析的新计费代码——是主要商业驱动,Viz.ai 已获 CMS 报销且盈利;中国没有同等全国 AI 报销,因此美国纯软件公司能跑 SaaS 订阅模式,中国厂商大多不能。 中国定价不透明:多数厂商通过医院采购销售,或把 AI 与设备打包,而不是公布每次检查价格。UII 的差异化在于垂直整合——AI 嵌入 UIH 扫描仪和 PACS,再叠加宽广的 uAI 组合;但相对 Aidoc 和 Viz.ai,其限制在于离开自家硬件之后,企业级全院软件集成叙事不够开放,也不够灵活。[CP011, CP012, CP013, CP014, CP015, CP016]
| 采购标准 | United Imaging Intelligence | 中国专营同业 | 全球专营厂商(Aidoc/Viz) | 平台在位者(GE/Siemens) |
|---|---|---|---|---|
| 硬件渠道 / 内嵌 AI | 强(借助母公司 UIH) | 弱(仅软件) | 弱(仅软件) | 强(自有扫描仪) |
| 监管批准广度 | 高(NMPA 三类、40 项 FDA、CE) | 中(三类持证者) | 高(33+ 项 FDA、CMS) | 很高(130/95 项 FDA) |
| 基础模型 / 多模态 | 是(uAI NEXUS、Yuanzhi) | 起步中(DeepWise 大模型) | 是(Aidoc CARE) | 部分具备(内嵌套件) |
| 企业软件开放度 | 中(自有硬件上最佳) | 中 | 高(厂商无关 aiOS) | 中(自有生态) |
| 报销支撑的变现 | 弱(中国没有 AI 项目) | 弱(中国没有 AI 项目) | 强(CMS 代码) | 混合 |
| 中国医院分销 | 很强(三甲触达) | 强(国内) | 弱(准入有限) | 强(装机基础) |
单元格是基于已引用公开来源的序位判断;未标出的强项代表缺少公开确认,并不代表已确认同等水平。
[CP011, CP012, CP014, CP016, CP019, CP025]| 厂商组 | 定价 / 合同模式 | 包含能力 | 折扣或未知项 | 含义 |
|---|---|---|---|---|
| United Imaging Intelligence | 医院采购;常与 UIH 设备捆绑 | uAI 产品组合应用、内嵌扫描仪 AI | 独立软件定价未披露 | 渠道捆绑有助采用,但模糊了软件 ARR |
| 中国专营同业 | 医院采购;按项目授权 | 单病种或多模态模块 | 按次检查与折扣条款不公开 | 竞争密集压低价格和利润率 |
| 全球专营厂商(Aidoc/Viz.ai) | SaaS 订阅 / ARR,美国可获 CMS 报销 | 企业平台、多病种覆盖 | 合同金额随医疗系统规模变化 | 报销支撑耐久的经常性收入 |
| 平台在位者 | 嵌入硬件资本采购或服务合同 | 扫描仪与 PACS 内的 AI 功能 | AI 价值很少单独列项 | AI 作为硬件差异化出售,而非独立业务线 |
| 内部自建 / 开源 | 内部成本;无许可费 | 研究级模型,可定制 | 维护与验证负担 | 免费替代品压住厂商定价权 |
已审阅来源没有披露中国厂商透明的按次检查定价;最清晰的定价反差,是美国有报销支撑的 SaaS, 对比中国的捆绑采购。
[CP018, CP021, CP022, CP031, CP033]对比决定中国影像 AI 采购的各项标准下,不同玩家的能力强弱。
[CP011, CP016, CP018, CP022, CP025, CP026]3.3 护城河、切换成本和被替代风险
UII 最强护城河,是母公司 UIH 的硬件渠道:购买 United Imaging 扫描仪的医院账户,也会被带入 UII AI;国内数据规模和中国本土监管位置进一步强化这一点。AI 一旦嵌入扫描仪工作流和医院 PACS,切换成本就会上升,有利于 UII 和平台型既有厂商这类整合供应商;分销权也偏向硬件销售商(GE、Siemens、Philips、UIH),纯软件初创公司难以匹配。 但每条护城河都有可信威胁。渠道优势会被部分抵消,因为中国医院会多宿主,并行运行多家厂商的 AI;市场密度高——数百家厂商——压低定价,也加速单病种检测器商品化。监管宽度真实但可复制:Infervision、DeepWise 和 Tencent 也持有三类证。基础模型规模正在成为新前沿,资本更厚的全球纯软件公司投入超过 UII——Aidoc 的 CARE 对比 UII 的 uAI NEXUS;Annalise 和 Aidoc 的宽度战略也与 UII 的多病种 agent 路线收敛,因此单靠宽度不太可能长期差异化。医院内部自建和开源医疗模型,正在研究型医院成为增长中的现状替代方案;Airdoc 和 Shukun 的 IPO 路径也表明同行在融资进入公开市场,追逐同一批医院预算。 从估值相关定位看,UII 约 USD 1.4B 的估值高于上市 Airdoc(IPO 时约 RMB 5.8B),但低于全球龙头的融资基础,意味着它背负一个需要由竞争力证明的溢价。没有公开来源显示 UII 在与具名同行的正面替代中获胜;Viz.ai 的盈利能力和 CMS 报销也给美国对手带来 UII 尚不能匹配的商业化优势。耐久结论是,UII 在渠道和数据上确有护城河,但护城河可被挑战;只有把这些优势转化为可报销、抗多宿主的经常性收入,竞争边才守得住。[CP023, CP024, CP025, CP026, CP027, CP028]
| 护城河主张 | 威胁 | 严重程度 | 缓释 / 尽调问题 |
|---|---|---|---|
| 母公司 UIH 硬件渠道把 AI 带进医院 | 医院同时使用多家厂商 AI | 高 | 量化 UIH 客户中只运行 UII AI、而非竞品 AI 的占比 |
| 监管批准广度(NMPA/FDA/CE) | 同业也持有三类和 FDA 批准 | 中 | 比较付费部署转化,而不是批准数量 |
| 国内数据规模优势 | PIPL 限制存在,竞争对手也能获取中国数据 | 中 | 确认自有且权利清晰的训练数据量 |
| 基础模型 / 多模态领先 | Aidoc 与在位者在模型研发上投入高于 UII | 高 | 用独立数据集对标 uAI NEXUS 与 CARE |
| 垂直整合 / 切换成本 | 开放企业平台削弱锁定 | 中 | 测试 UII AI 移植到第三方扫描仪的可行性 |
| 中国本土市场地位 | 没有报销项目,变现能力低于美国同业 | 高 | 跟踪 UII 参与的省级 AI 捆绑付费试点 |
| 相对上市 Airdoc 的溢价估值 | 竞争与准备 IPO 的同业压缩倍数 | 中 | 用同业收入倍数压力测试估值 |
严重程度取决于每项威胁对护城河的直接削弱程度;反复出现的尽调主题,是能否把渠道和注册批准优势转成可报销、可防守的收入。
[CP023, CP024, CP025, CP027, CP028, CP029]截至 2026-06-23,按证据加权评估 UII 相对行业的竞争耐久度。
[CP023, CP024, CP027, CP033, CP040]3.4 图表
04财务
4.1 收入模式和变现架构
United Imaging Intelligence 是上市母公司 United Imaging Healthcare(UIH,688271.SH)旗下独立融资的医疗 AI 业务,变现对象是软件和临床 AI 应用,而不是影像硬件。作为私营子公司,它不披露独立收入、ARR 或净利润,因此只能从产品披露和母公司报表重建收入架构。UII 通过两个可见渠道变现 AI:嵌入或打包进 UIH 扫描仪和 PACS 的应用,以及销售给医院的独立 uAI 软件许可。公司报告截至 2025 年中,已有 12 个产品平台、100 多个 AI 应用,部署于 4,000 多家医疗机构,并由 13 张 NMPA 三类证、15 项 FDA 清除和 31 项 CE 认证支撑。 未公开的信息与已公开信息同样关键。UII 不公布按检查或按席位的标价;与多数中国影像 AI 厂商一样,它通过医院采购或设备打包销售,这遮住了实际定价和独立收入归因。公司在 FDA 2026 年 Q1 放射 AI 授权名单中排名第五(约 40 项已清除设备),说明持续投入产品和研发,但清除宽度只是投入代理指标,不是收入数字。因此,收入模型方向上看得清——向庞大装机基础销售打包加独立软件——但收入质量和规模仍藏在母公司合并报表里。[CI001, CI002, CI003, CI004, CI005, CI006]
| 收入流 | 机制 | 单位 | 当前价值 / 状态 | 收入质量 | 尽调问题 |
|---|---|---|---|---|---|
| 内嵌 / 捆绑 AI | AI 应用随母公司 UIH 扫描仪和 PACS 内置或搭售 | 按系统 / 设备交易 | 已覆盖 UIH 装机基础;价值未单独报告 | 战略价值高,但归因藏在硬件交易里 | 要求披露 AI 附加率,以及从捆绑设备销售中拆出的 AI 收入 |
| 独立 uAI 软件许可 | uAI / uAI NEXUS 应用通过采购卖给医院 | 按应用 / 医院 / 席位 | 100+ 款应用已在 4,000+ 家机构上线 | 有经常性潜力,但定价不透明且无法报销 | 要求披露独立软件 ARR、续约率与实际成交价格 |
| 基础模型 / 智能体平台 | 多模态医疗智能体与大模型升级(Series A 融资用途) | 平台 / 订阅(起步中) | 商业化早期;由 2025 年 Series A 资助 | 变现尚未验证;研发投入重 | 要求披露从路线图到收入的计划,以及试点合同金额 |
| 科研 / 科学协作 | 医院科研、病历书写与科学应用 | 项目 / 许可 | 已上线,但规模小且未披露 | 直接收入可能较低 | 要求披露科研收入与临床部署收入占比 |
| 母公司 / 关联方支持 | 来自 UIH (688271.SH) 的共享研发、渠道与品牌 | 公司间 | 规模重要,但未单独披露 | 交叉补贴遮住独立经济性 | 要求提供关联方交易和成本分摊明细 |
UII 不公布独立收入;收入流只能靠产品披露和母公司文件重构, 当前数值是状态描述,不是已报告数据。
| 产品 | 定价基础 | 标价 vs 实收 | 折扣 / 未知项 | 来源依据 |
|---|---|---|---|---|
| uAI 临床应用 | 医院采购 / 按应用授权 | 无公开标价;实收价格未披露 | 与设备打包遮住实收价格 | 公司产品页和交易报道 |
| 设备打包 AI | 纳入 UIH 扫描仪 / PACS 采购 | 边际标价实际为零;由交叉补贴支撑 | 无法从硬件价格中拆出 AI 价值 | 母公司披露和产品页 |
| 多模态 agent 平台 | 新兴订阅 / 平台(Series A 之后) | 尚未公开定价 | 商业模式仍在成形 | 融资轮次和管理层评论 |
| 报销挂钩变现 | 中国全国层面没有 AI 软件报销项目 | 不适用;仅靠医院直接付费 | 缺少 DRG / 医保代码压住定价权 | 监管和市场准入来源 |
中国影像 AI 定价不透明;各行描述变现基础和具体未知项, 而不是已公布价格。
UII 如何把母公司锚定的装机基础和软件组合转成收入;独立收入和利润率层仍明确未知。
这座桥展示产品披露和母公司文件中可见的变现逻辑;报告刻意不编造 UII 未公布的收入数字。
[CI001, CI003, CI005, CI011, CI020]4.2 公开牵引力和销售效率代理指标
从牵引力看,UII 更像一个真实部署的平台,而不是试点阶段工具:4,000 多家机构运行其应用,清除组合在 FDA 名单中全球排名第五。但这种牵引力的效率主要靠推断。UII 的核心 GTM 动作是母公司 UIH 的硬件渠道,把 AI 带入医院设备账户,并提供软件型对手难以匹配的分销经济性。反面是,打包销售模糊了 AI 与硬件各自贡献多少收入;UII 也不披露销售周期、合同条款、CAC、回本期或收入集中度。 同行证据把收入质量问题衬得很尖锐。资本深度是一个竞争变量:UII 约 USD 139M 的 Series A,相比 Aidoc 等全球纯软件公司(累计融资超过 USD 500M)偏小,限制了基础模型研发火力。中国 2026 年创投反弹——亚洲一季度融资约 USD 27.4B,其中多数流向中国且由 AI 领跑——支持后续资本可得性。但上市中国可比公司令人警醒:Airdoc 报告 2024 年净亏损 CN¥255M,尽管收入增长,仍已连续五年净亏损。这些代理指标说明 UII 有规模和渠道,但其独立变现效率无法从公开数据证明。[CI008, CI009, CI010, CI011, CI029, CI030]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| 独立收入 / ARR | 未披露 | 低 | 衡量规模的核心输入;并入母公司口径 | 从数据室获取经审计独立损益表 |
| 毛利率 | 未披露(母公司研发约占收入 19%) | 低 | 决定软件经济性假设是否成立 | 要求提供独立 COGS 和毛利率桥表 |
| CAC / 回本期 | 无公开指标 | 低 | 检验渠道优势是否高效 | 要求按医院层级提供混合 CAC 和回本期 |
| 净收入留存 | 未披露 | 低 | 检验软件收入韧性 | 要求按 cohort 提供总留存和净留存 |
| 部署覆盖 | 4,000+ 机构;100+ 应用 | 中 | 能确认触达范围,但不能确认变现深度 | 对齐已部署机构与付费机构 |
所有达到承保级别的单位经济性指标都未公开;公开证据只支持部署覆盖。
公开证据能支撑的财务区间很少:Series A 轮规模、隐含估值,以及界定 UII 资源背景的母公司收入规模;UII 独立区间仍无法推导。
中点只是为便于阅读做的算术处理;母公司数字是经审计的点值,公开资料无法推导 UII 独立收入或烧钱区间。
[CI012, CI014, CI022, CI024, CI029]4.3 成本结构、利润率和资本强度
UII 业绩并入母公司 UIH,未单独报告,因此公开数据无法判断其独立毛利率——这是结构性限制,不是临时缺口。母公司的审计数字仍能勾勒 UII 所贡献的经济背景。UIH 报告 2025 年收入 RMB 13.80B,同比增长 33.98%;归母净利润 RMB 1.87B,同比增长 48.14%。2025 年半年报显示收入 RMB 6.02B(增长 12.79%),归母净利润 RMB 998M,海外收入 RMB 1.14B,服务收入 RMB 0.82B。 研发强度是资本强度最清晰的信号:UIH 2025 年投入 RMB 2.62B 研发(增长 15.95%),约占收入 19%;上半年单独投入 RMB 1.14B,其中相当部分投向 UII 参与构建的 AI 和平台技术。利润率面临国内逆风:中国集中带量采购(VBP)对影像设备形成降价压力,母公司通过高端产品和快速增长的海外组合来对冲。2026 年一季度业绩——收入 RMB 2.91B(增长 17.34%),净利润 RMB 0.40B(增长 7.78%)——显示利润增速慢于 2025 全年。关键在于,AI 诊断软件通常不在中国医保或 DRG 体系下报销,因此厂商以较低费用直接与医院签约;Ping An 等盈利同行,主要靠生态交叉销售变现,而不是独立影像 AI 定价。[CI012, CI013, CI014, CI015, CI016, CI017]
公开证据更清楚地揭示 UII 的渠道和部署输入;要搭完整单位经济模型,利润率、成本和留存层仍缺失。
缺失的 CAC、利润率和留存数据被明确标成未知节点,而不是藏在虚假的精确度后面。
[CI010, CI019, CI020, CI021, CI035]4.4 资本充足性和融资依赖
资本获取能力是 UII 公开财务图景中最强的一块。公司在 2025 年 6 月完成约 RMB 1B(约 USD 139M)Series A,由 E Fund Management 和 Shanghai Guotou(SIG)共同领投,Shanghai United Investment 等参与。该轮隐含投前估值约 RMB 9B(约 USD 1.3B),投后约 USD 1.4B。管理层称资金将用于大型医疗 AI 模型研发、覆盖影像、文本和语音的多模态医疗 agent,以及医院数字化和智能化业务系统;UII 也据称正在考虑独立上市。 资产负债表图景在承销真正需要开始的地方戛然而止。UII 不披露现金余额、月度烧钱或 runway,也没有记录到公开债务或项目融资义务。融资时间线已在公司概览章节覆盖;出于财务分析目的,相关事实在此作为有独立来源的本地论断重述。相对全球同行,这轮融资规模不大——仅 Aidoc 的累计融资就超过 UII Series A 的三倍——但 2026 年中国资本环境支持融资,母公司也提供深厚资产负债表和渠道。对 UIH 的财务依赖(共享研发、合并报表和硬件销售渠道)既是 UII 最大资产,也是集中度风险,因为独立投资人无法把 UII 的偿付能力从母公司中剥离出来。[CI022, CI023, CI024, CI025, CI026, CI027]
| 项目 | 公开信号 | 财务解读 | 盲点 |
|---|---|---|---|
| Series A 融资 | ~RMB 1B(约 USD 139M),2025 年 6 月 | 增长资本获取能力强 | 募资用途拆分和稀释细节 |
| 估值 | ~RMB 9B 投前估值(~USD 1.3B);投后 ~USD 1.4B | 成立 7 年子公司的私募估值偏高 | 没有独立收入,收入倍数无法验证 |
| 账上现金 / 烧钱 / runway | 未披露 | 无法估算偿付能力或 runway | 月度烧钱和现金余额完全未公开 |
| 债务 / 项目融资 | 未披露 | 看不到杠杆 | 与母公司的集团内融资条款未知 |
| 下一轮 / IPO 触发点 | 据称正在考虑独立上市 | 保留未来资本选择权 | 时间、地点和前置条件未披露 |
资本获取能力有充分证据;资产负债表和 runway 输入完全未披露, 这里复述的是融资报道,而不是财务报表。
融资可得性信号很强,但公开记录没有给出独立承保所需的资产负债表和利润率细节。
该图把公开证据能支撑的内容,与仍需尽调室财务材料回答的问题拆开。
[CI013, CI022, CI027, CI031, CI038]4.5 财务结论和尽调阻断项
最可支撑的承销观点是:UII 资本获取能力强、母公司有力,但作为独立实体还无法干净承销。积极面真实存在:以约 USD 1.4B 估值完成 USD 139M Series A,部署基础超过 4,000 家机构,FDA 清除数位居全球前五,母公司 2025 年经审计收入 RMB 13.80B、利润 RMB 1.87B。公开财务信号中最强的是 Series A 与母公司经审计增长的组合,而不是 UII 自身财报——后者并未公开存在。 风险同样具体。主要尽调阻断项,是未披露的独立收入、毛利率、烧钱速度、runway 和留存。同行证据具有警示意义:Airdoc 2025 年年报警告,即便获得 NMPA 三类批准,商业化仍存在风险;分析师预计其要到 2026 年左右、在约 102% 收入增长基础上才能盈亏平衡——这提醒投资人,中国医疗 AI 变现很难,且缺少报销支持。净判断是:对资本获取和母公司实力信心高,对部署规模信心中等,对独立单位经济和偿付能力信心低。任何投资论证都应把独立收入质量、利润率路径和母公司依赖结构列为一阶尽调事项,而不是收尾问题。[CI031, CI032, CI033, CI036, CI037, CI038]
| 缺失指标 | 对承保判断的影响 | 具体尽调路径 |
|---|---|---|
| 独立收入和 ARR | 无法估算业务规模,也无法检验估值 | 在 NDA 下获取经审计独立财务报表 |
| 毛利率和成本结构 | 软件经济性假设尚未验证 | 要求提供 COGS、服务交付成本和利润率桥表 |
| 烧钱、现金和 runway | 偿付能力和融资依赖未知 | 要求管理层提供现金流模型和银行余额 |
| 留存和收入集中度 | 收入韧性和客户风险不清楚 | 要求提供 cohort 留存和头部客户集中度 |
| 集团内 / 关联方条款 | 母公司交叉补贴扭曲独立经济性 | 要求提供转让定价和成本分摊明细 |
这些缺口是任何 UII 承保判断的闸门项;它们都属于标准数据室披露, 公开来源无法替代。
4.6 图表
05产品与技术
5.1 产品定义与模块地图
United Imaging Intelligence 交付的是统一医疗 AI 平台,品牌为 uAI,运行在临床工作流内,而不是作为独立分析工具存在。从客户视角看,平台接收影像检查,运行 AI 分析,生成结构化诊断报告,支持治疗计划,并把数据送入医院运营仪表盘。UII 称其产品组合包含 12 个产品平台、100 多个 AI 应用,覆盖 CT、MR、X-ray、超声和病理,并通过 uAI Clinical Portal(uCP)/ uAI Discover 部署;该门户与医院 PACS 集成,并支持跨模态的 60 多个应用。 最重要的产品转向,是从窄任务自动化转向能在诊疗路径中理解、推理和行动的协作式「AI agents」。RSNA 2025 上,UII 展示 uAI Insight,一个执行多模态推理和临床关联的影像到报告 agent;uAI Agent for Chest CT Reporting 可从一次扫描中检测 73 种胸部疾病,AUC 约 95%;Brain MRI agent 可识别 47 种神经系统疾病;uAI Agent for Ultrasound 以具身机械臂形态交付;另有 uAI Avatar 助手和医院管理 agent。这些 agent 定义了平台现在瞄准的用户任务:更快报告、自主采集、患者预诊互动和运营质量监测。[CE001, CE002, CE003, CE004, CE005, CE006]
| 模块 / 产品线 | 用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| uAI Clinical Portal(uCP)/ uAI Discover 产品 | 放射科、医院 IT | 已上线;60+ 应用;多区域部署 | 跨模态、集成 PACS 的单一平台 | 正常运行时间、可靠性和单应用使用率未披露 |
| uAI Insight(影像转报告 agent) | 放射科医生 | RSNA 2025 展示;正在商业化 | 多模态推理和临床关联 | 日常使用中的独立准确性验证 |
| uAI Agent for Ultrasound(具身机械臂) | 超声技师、诊所 | RSNA 2025 首发;早期阶段 | 借助视觉 / 触觉线索自主扫描 | 成熟度、安全记录和监管状态 |
| uAI NEXUS MedVLM(医疗视频 LLM) | 研究人员、开发者、外科团队 | 2026 年 4 月开源;CVPR 2026 | 首个开源医疗视频 LLM,4B/7B 参数 | 从研究模型走向获批产品的路径 |
| DeepRecon / 图像重建 | 影像技师 | 已获批,并随 uMR 系统出货 | 与母公司 UIH 做软硬件协同设计 | 独立价值 vs 硬件打包价值的拆分 |
成熟度基于公开披露;与已获批的门户应用相比,agent 产品更新, 独立验证也更少。
| 用户任务 | 当前工作流 | UII 方案 | 可衡量收益 | 限制 |
|---|---|---|---|---|
| 阅读并出具胸部 CT 报告 | 人工阅片和口述录入 | uAI Agent for Chest CT Reporting 产品 | 一次扫描识别 73 种胸部疾病,AUC 约 95% | 需要放射科医生确认;日常使用准确性未验证 |
| 分诊疑似卒中 / 出血 | 按工作列表顺序阅片 | uAI Easy Triage ICH(FDA 已批准) | 在非增强头颅 CT 上标记疑似 ICH,加快通知 | 仅用于分诊;不是诊断设备 |
| 分析血管 CT 以支持规划 | 人工描绘血管 | uAI Portal(FDA 已批准分割) | AI 分割主动脉、冠状动脉和肺血管 | 范围仅限列出的解剖结构 |
| 采集超声检查 | 依赖操作者的手工扫描 | uAI Agent for Ultrasound(机器人) | 自主、一致的采集 | 早期阶段;临床和安全验证仍待完成 |
| 管理科室质量 / 运营 | 人工审计和电子表格 | uAI 医院管理 agent | 监控影像 / 报告质量和设备 | 运营 ROI 未经独立量化 |
收益来自公司或会议披露;限制反映的是已展示能力与独立验证日常表现之间的差距。
医院里,一次影像检查如何从采集走到 AI 辅助报告和运营洞察。
该流程基于公开产品描述,展示设计中的临床路径;每一步自动化程度会随应用和医院而变。
[CE003, CE004, CE005, CE006, CE019]5.2 架构与基础模型运营模式
UII 产品由多模态医疗大模型驱动,整合临床语言、视觉和语音。架构可追溯到 RSNA 2024 发布的 uAI PLUS(Pretrained Large Models for Unified Solutions),它结合医学影像基础模型、医疗 LLM 和多模态 LLM;更新的则是 uAI NEXUS 套件。2026 年 4 月,UII 开源 uAI NEXUS MedVLM,被描述为首个开源医学视频大语言模型,并被 CVPR 2026 接收,显示其研究深度获得全球计算机视觉社区认可。 技术核心记录在一篇公开研究论文中(arXiv:2512.06581)。MedVLM 建立在 MedVidBench 之上;MedVidBench 是一个包含 531,850 个视频指令对的基准,横跨八类临床来源。训练流程先做监督微调,再接 MedGRPO——一种强化学习框架,引入跨数据集奖励归一化和医疗 LLM 裁判,以防训练崩塌。尽管只使用 4B 和 7B 参数,该模型优于 GPT-5.4 和 Gemini 3.1 等更大的通用模型;在手术安全评估上准确率达到 89.4%,而 GPT-5.4 为 1.8%。UII 在 GitHub(UII-AI/MedGRPO-Code)发布开源代码、模型权重、MedVidBench 数据集和公开排行榜。生产环境中,同一模型谱系支撑的应用会整合 DICOM 摄取、AI 分析、结构化报告和 PACS/RIS 连接,而不是作为孤立点工具运行。[CE007, CE008, CE009, CE010, CE011, CE012]
| 层级 / 组件 | 角色 | 依赖 | 风险 |
|---|---|---|---|
| 多模态基础模型(uAI PLUS / uAI NEXUS) | 横跨视觉、语言、语音的核心感知和推理 | 大规模临床数据集;GPU 算力 | 泛化到未见过的真实临床场景 |
| MedVLM + MedGRPO 训练流水线 | SFT 加 RL,并做跨数据集奖励归一化 | MedVidBench 数据集;开源工具链 | 可复现性和随时间漂移 |
| 应用层(100+ uAI 应用 / agents) | 面向具体任务的检测、报告、分诊 | 基础模型谱系;监管批准 | 各应用验证广度不一 |
| uAI Clinical Portal 集成层 | DICOM 接入、结构化报告、PACS/RIS | 医院 IT 系统;互操作标准 | 多厂商集成复杂度 |
| 硬件协同设计(uMR 上的 DeepRecon) | 针对母公司 UIH 扫描仪调校的 AI | 母公司 UIH 设备装机基础 | 差异化绑定母公司硬件覆盖 |
架构基于研究论文、产品页和会议披露重构;内部 serving、MLOps 和安全层未公开详述。
分层展示 UII 如何把临床数据转成 AI 智能体,从基础模型一路到接入 PACS 的应用。
各层来自研究论文、产品页面和会议披露的推断;内部服务和安全层没有公开文件。
[CE007, CE009, CE011, CE015, CE030]5.3 部署、集成、可靠性和路线图
UII 的 AI 已广泛部署:它通过 uAI Clinical Portal 运行在中国 4,000 多家医疗机构,并安装于欧洲、亚洲、北美和南美;该门户集中处理 DICOM 摄取、AI 分析、结构化报告和医院 IT 集成。这是基础设施式部署——PACS/RIS 集成和结构化报告——而不是研究试点,符合医院要求的可靠性门槛。 产品路线图可从 UII 的主要会议节奏中看见。RSNA 2025 上,公司以「Agent Radiologist: Are We There Yet?」为主题,展示覆盖报告、超声采集和医院管理的下一代 agent。ECR 2026 上,公司强调放射 AI 的验证和规模化,引用 60 多个应用和 30 多项 CE 认证,并突出与 Massachusetts General Hospital 等领先中心进行的多中心、多厂商验证合作。最新里程碑是 2026 年 4 月开源 MedVLM;这与 ECR 2026 的规模化叙事一起,标志着 UII 正从清除驱动的应用厂商,转向基础模型和 agent 平台。集成宽度和区域支持是优势;不够可见的是最新自主 agent 的独立可靠性和 uptime 数据。[CE006, CE016, CE017, CE018, CE019, CE020]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源依据 |
|---|---|---|---|---|
| RSNA 2024 | uAI PLUS 基础模型平台首发 | 已发布 | 确立基础模型战略 | 官方新闻稿 |
| RSNA 2025 | AI agents(Insight、Ultrasound、Avatar、管理) | 已展示 / 正在商业化 | 从任务型 AI 转向 agentic AI | 官方新闻稿和报道 |
| April 2026 | uAI NEXUS MedVLM 开源;CVPR 2026 | 已开源发布 | 主张开放基础模型领导地位 | 官方新闻稿、GitHub、arXiv |
| ECR 2026 | 验证和规模化;60+ 应用、30+ CE | 已发布 / 正在放量 | 强调验证和覆盖广度 | 官方新闻稿和报道 |
| 2026 年以后 | 覆盖诊疗路径的多模态 agents | 开发中 | 商业化路径仍在成形 | 会议评论 |
路线图条目锚定在有日期的会议发布和开源发布;前瞻条目是管理层方向性表述, 不是承诺交付日期。
影响 UII 产品表现、部署和差异化的外部依赖。
这些依赖来自产品、监管和母公司披露;内部云、安全和供应依赖没有公开细节。
[CE015, CE016, CE030, CE031, CE032, CE034]5.4 差异化——技术、数据和监管护城河
UII 最强技术差异化,是在医学视频理解上拥有开源基础模型领导力,并由 CVPR 2026 接收以及公开基准、代码、权重和排行榜验证;这些资产邀请全球开发者社区参与,而这种透明姿态在通常闭源的中国医疗 AI 厂商中很少见。开放既是可信度和招聘信号,也是在为行业设定评估标准。 还有三条护城河强化技术领先。第一,与母公司 United Imaging Healthcare 的垂直整合,让 UII 能与扫描仪共同设计 AI——例如 uMR 系统上的 DeepRecon 图像重建——形成纯软件对手难以复制的软硬件循环。第二,大规模、多样化、来自临床的训练语料,经由医院合作不断精炼,是会随部署复利的数据护城河。第三,横跨 FDA、NMPA 和 CE 的监管宽度本身就是差异化资产;UII 在 FDA 2026 年 Q1 放射 AI 授权名单上以约 40 项已清除设备排名第五,优于更窄的竞争者。开放问题在于耐久性:开源旗舰模型会吸引快速跟随,因此护城河更多取决于数据、硬件共同设计和清除宽度,而不是模型保密。[CE026, CE028, CE029, CE030, CE031, CE032]
截至 2026-06-23,UII 主要能力领域的相对成熟度和证据强度。
成熟度和证据评分是基于已引用披露做出的证据加权序数判断,并非实测基准。
[CE022, CE028, CE036, CE037, CE040]5.5 信任、安全、安保与合规
UII 的质量与合规底座来自监管准入。uAI Easy Triage ICH(K242292)在 2024 年 9 月获得 FDA 510(k) Class II 许可,可在非增强头部 CT 上分诊疑似颅内出血;uAI Portal(K240411)也已获 FDA 许可,用于 AI 驱动的血管 CT 分割;uAI EasyTriage-Rib 和 DeepRecon MR 重建工具继续扩充已获批产品组合。到 2025 年中,UII 称约有 15 项 FDA 许可、31 项 CE 认证、13 项 NMPA Class III 批准,其中 NMPA Class III 是中国最高风险器械类别,需要严格临床评价。FDA 申报由专业监管伙伴 MCRA 支持,说明公司进入美国市场所需的质量体系已经有一定成熟度。 透明度上,UII 比多数同行走得更远:公开发布开放基准 MedVidBench,并提供第三方评估排行榜。但公开记录里仍有三个信任缺口。最新自主智能体的独立第三方临床验证证据仍有限。公司很少披露平台的数据隐私架构、安全控制,以及符合 PIPL 的数据处理细节。从开源基础模型走向付费、受治理的医院产品,商业化路径也尚未完整说明。综合看,监管许可和基准透明度很强,但独立验证、隐私 / 安全披露、开源到收入的桥梁更薄;这些是基础模型泛化到真实临床场景之外的主要产品技术风险。[CE022, CE023, CE024, CE025, CE027, CE033]
| 控制 / 认证 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| FDA 510(k) 批准 | 约 15 项获批(例如 K242292 ICH、K240411 Portal) | 美国市场;CT/MR 分诊和分割 | 日常使用中的上市后表现未公开 |
| NMPA III 类批准 | 持有 13 项(2025 年中) | 中国;最高风险等级诊断 AI | 单应用临床评价细节未公开 |
| CE 认证 | 持有 31 项;ECR 2026 提到 30+ 项 | MDR/MDD 下的 EU/EEA 市场 | 证书与具体应用的映射不清楚 |
| 开放基准 / 透明度 | MedVidBench + 公开排行榜已上线 | 医疗视频模型评估 | 全产品组合的独立基准不清楚 |
| 数据隐私 / 安全控制 | 未公开详述 | 医院数据处理;对齐 PIPL | 安全架构和 DPA 条款未披露 |
获批数量来自公司和注册库披露;视频模型透明度强, 但隐私、安全和独立临床验证披露很薄。
5.6 图表
06客户
6.1 客户基础与分层
United Imaging Intelligence 服务医院和医疗系统,主要买方是放射科和医院 IT,使用者是放射科医生和临床医生,付款方则是医院本身,因为中国没有面向独立诊断 AI 软件的全国医保支付项目。客户基础沿三条可见轴线分层。按地域,核心市场是中国公立医院体系,尤其是大型三甲和三级教学医院;海外版图则借母公司 United Imaging Healthcare(UIH)渠道进入欧洲、北美、亚洲和南美。按用例,客户采用 uAI 做分诊、病灶检测、分割、结构化放射报告,并越来越多用于胸部 CT、脑 MRI、超声等 AI 智能体。按渠道,UII 用两种方式触达客户:AI 原生嵌入 UIH 扫描仪和 PACS;以及向可能运行多厂商设备的医院授权独立 uAI Clinical Portal(uCP)软件。 规模信号是真实的。UII 称其 12 个平台、100 多个应用已部署到 4,000 多家医疗机构;母公司称装机基础覆盖 100 多个国家和地区,到 2025 年底美国累计系统超过 640 套。中国国家医疗体系在 2025–2026 年把 AI 辅助诊断推入 1,200 多家三甲医院,正是 UII 瞄准的池子。但公开披露把机构数、部署应用数和硬件账户混在一起,因此外部无法拆出付费 UII 软件客户数量,也无法区分这些医院只是拥有带嵌入式 AI 的 UIH 扫描仪,还是单独购买了软件。[CU001, CU002, CU003, CU004, CU005, CU006]
| 分群 | 买方 / 用户 / 付款方 | 主要使用场景 | 规模信号 | 战略价值 | 尽调缺口 |
|---|---|---|---|---|---|
| 中国三级 / 三甲医院 | 放射科采购;放射科医生使用;医院付款 | 分诊、检测、分割、结构化报告 | 4,000+ 机构的核心;全国 1,200+ 三甲已部署 AI | 量最大、具参考价值的学术账户 | 未拆分付费 uAI 软件站点与硬件账户 |
| UIH 硬件打包账户(全球) | 影像科采购扫描仪;技师 / 临床医生使用;医院付款 | uMR / uCT 采集和重建中的原生 AI | 100+ 国家;到 2025 年底美国累计 640+ 套系统 | 借母公司装机基础形成分销护城河 | AI 价值无法从硬件收入中拆出 |
| 独立 uAI Clinical Portal 授权客户 | 医院 IT / 放射科采购;放射科医生使用;医院付款 | 单一 PACS 集成工作流界面承载 60+ 应用 | 60+ 应用统一;多区域部署 | 不依赖硬件更新的经常性软件上行 | 未披露授权数量、定价或续约 |
| 国际学术 / 私立中心 | 影像科 / 放射科采购;临床医生使用;医院付款 | MRI/CT 影像加 CE 认证 AI 应用 | 美国、意大利、法国、立陶宛具名站点 | 可信度和西方市场验证 | 很少有站点专门拆出 uAI 软件采用情况 |
| 研究 / 验证合作者 | 学术 PI 采购 / 合作;研究人员使用;经费 / 医院付款 | agent 和模型的多中心临床验证 | ECR 2026 覆盖四大洲验证 | 证据管线反哺临床可信度 | 合作很少等同于付费生产部署 |
分群基于截至 2026-06-23 运行日的公司、客户侧和行业披露重构; 尽调缺口列标出各分群公开记录尚未证明的内容。
截至 2026-06-23,医院通常如何从采购 United Imaging 硬件,扩展到 uAI 软件和 AI 智能体。
[CU004, CU011, CU035, CU036]6.2 采用轨迹与部署深度
UII 的采用故事更像真实生产部署,而不是试点阶段实验;但支撑指标大多描述平台广度或母公司硬件足迹,并非离散的 UII 软件账户。uAI Clinical Portal 在单一工作界面整合 60 多个应用;到 2025 年中,UII 称约有 13 项 NMPA Class III 批准、15 项 FDA 许可、31 项 CE 认证,并在 ECR 2026 展示 30 个 CE 认证应用。这种监管广度说明产品已获准在多个司法辖区用于常规临床。第三方对中国医院 AI 的报道提到,放射报告工作流改善约 30%;UII 也指向覆盖数千病例的多中心验证。 但部署漏斗混用了无法拼成清晰转化图的单位。最上层是至少运行一个 uAI 应用的数千家机构;下面是 UIH 在 100 多个国家的硬件账户;再下面是更少的、具名且可公开引用的医院部署;底部则是 RSNA 2025 和 ECR 2026 展示、仍在成形的 AI 智能体,其独立生产级验证仍有限。UII 没有披露单应用使用量、活跃使用率、复购或续约数量,也没有给出能把「4,000 家机构」换算成席位、阅片量或经常性收入的分母。轨迹向上且覆盖面广,但深度和耐久性只能推断,尚未被计量。[CU009, CU010, CU011, CU012, CU013, CU014]
| 指标 | 数值 | 截至日期 | 来源依据 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 运行 uAI 应用的机构 | 4,000+ | 2025 年中 | 公司 / 官方披露 | 中 | 生产落地覆盖面广 | 未披露席位数、阅片量或单机构收入 |
| uAI Clinical Portal 上的应用 | 60+ | 2025-2026 | 公司 / 活动披露 | 中 | 多应用工作流触点很深 | 未披露单应用使用量或活跃使用率 |
| 通过 CE 认证的 AI 应用(欧洲) | 30 | ECR 2026(2026 年 3 月) | 官方新闻稿 / 新闻 | 高 | 可在欧盟常规临床中使用 | 未披露欧盟客户数或收入 |
| 母公司累计美国影像系统 | 640+ | 2025 年底 | 母公司官方披露 | 高 | 渠道触达美国客户 | 搭载付费 uAI 软件的占比未知 |
数值混合了公司口径与母公司口径,单位也不一致;分母为 "null" 的指标,表示截至 2026-06-23 运行日公司尚未公开披露。
从广泛机构覆盖到具名生产级案例的示意性推进;单位不同,不能按字面理解为转化率。
漏斗混合了不同公开单位(机构、国家数、硬件系统、认证应用和具名案例),用来可视化证明质量收窄,而不是单一经审计的转化分母。
[CU002, CU009, CU013, CU019]6.3 具名客户证明与引用质量
具名且可外部核验的客户,是检验采用情况的最强证据;在这一点上,UII 的证据并不均匀。最清晰的公开引用是 UC Davis Health,其放射设备清单显示一套 United Imaging 1.5T uMR 680 MRI 系统已在生产中使用——这是客户侧独立披露,不是厂商案例。行业报道还点名 UIH 在意大利(包括一台 uCT 960+)、法国、德国、立陶宛的装机;旗舰 uMR Omega 3T 超宽孔系统也已进入 X-Ray Associates of New Mexico 等美国影像服务机构。在中国,复旦大学附属中山医院等领先学术中心被提及为 uAI 用户。ECR 2026 上,UII 展示了其 30 个 CE 认证应用在欧洲、亚洲、北美、南美的多中心验证。 尽调中有两个限定条件很关键。第一,大多数具名证明挂在 UIH 硬件账户上;披露很少单独说明哪些站点把 uAI 软件作为独立项目付费,因此硬件引用不能自动等同于软件收入引用。第二,最新自主 AI 智能体——RSNA 2025 和 ECR 2026 展示的报告撰写、超声机器人、图像到报告智能体——已在研究环境中演示和验证,但缺少具名站点的独立生产级结果证据。因此,证明矩阵应把已获批影像应用评为成熟度高、佐证扎实;智能体层的雄心很高,但独立留存和结果可见度偏低。[CU019, CU020, CU021, CU022, CU023, CU024]
| 客户 | 细分 | 部署 / 使用场景 | 生产使用还是试点 | 已披露结果 / 信号 | 证据限制 |
|---|---|---|---|---|---|
| UC Davis Health | 美国学术医学中心 | United Imaging 1.5T uMR 680 MRI 已进入临床服务 | 生产使用 | 列入科室自有 MRI 设备清单 | 硬件引用;未单独说明 uAI 软件采用情况 |
| X-Ray Associates of New Mexico 客户 | 美国门诊影像服务商 | 安装 uMR Omega 3T 超宽孔径 MRI | 生产使用 | 旗舰系统落地,用于扩大社区 MR 可及性 | 来自厂商 / 行业来源;未披露结果指标 |
| 欧洲医院(意大利、法国、立陶宛) | 国际学术 / 公立医院 | uCT 960+ 与 uMR 680 影像设备,叠加 CE 认证 AI 应用 | 生产使用 | 西欧多个具名国家有装机 | 未逐项列出具体站点结果和 uAI 使用情况 |
| 复旦大学附属中山医院(上海) | 中国三甲教学医院 | uAI 诊断与工作流工具 | 具名引用 | 作为中国头部中心的 uAI 用户被引用 | 引用来自二级来源;无合同或规模细节 |
| ECR 2026 多中心验证队列 | 跨洲研究 / 临床站点 | 验证 30 个 CE 认证应用和 AI 智能体 | 验证 / 展示 | 覆盖四大洲的多中心证据 | 属于验证合作,未确认付费部署 |
公开具名客户集合只是不完全列举:截至 2026-06-23 运行日,可在已抓取的公司、客户侧和行业页面看到的部署。它不是完整客户名单,大多数引用绑定的是 United Imaging Healthcare 硬件账户,而非独立 uAI 软件合同。
[CU019, CU020, CU021, CU022, CU023, CU025]按结果具体度、独立佐证、留存可见度和生产成熟度加权,比较 UII 各能力层的证明质量。
评分是截至 2026-06-23、基于已引用披露做出的证据加权序数判断,并非实测分数。
[CU021, CU024, CU026, CU030, CU033]6.4 留存、耐久性与切换成本
耐久性上,UII 有很强的结构性论据,却几乎没有披露数据。结构性逻辑在于,uAI 常常原生嵌入 UIH 扫描仪,并集成进医院 PACS/RIS,从而抬高切换成本:替换 AI 通常意味着打断影像工作流或底层硬件关系;uAI Clinical Portal 又整合 60 多个应用,医院一旦标准化使用,就会积累培训和集成锁定。监管广度也强化粘性,因为医院在临床路径中验证过已获批应用后,通常不会无故重新验证竞品。这些都是深度集成临床基础设施常见的真实留存顺风。 数据层面基本是空的。UII 不发布净收入留存、毛留存、流失率、续约率、合同期限或客户满意度分数,也没有公开队列展示随时间的重复使用。多数采用与硬件捆绑,连母公司文件也暴露不出 AI 软件续约行为。诚实的判断是,耐久性在机制上可信,但指标上未被证明;每一项留存主张都应转化为带明确尽调问题的空值:在 NDA 下索取按医院层级拆分的 NRR/GRR、应用级活跃使用和续约队列、合同条款、流失原因。Airdoc 等亏损上市同行说明,广泛临床采用并不保证收入耐久且有利润。[CU029, CU030, CU031, CU032, CU033, CU034]
| 指标 | 数值 | 细分 | 置信度 | 尽调问题 |
|---|---|---|---|---|
| 净收入留存(NRR) | null | 所有细分 | 未披露 | 要求按医院等级、软件单售 vs 捆绑拆分 NRR |
| 总留存 / 流失率 | null | 独立 uAI 授权客户 | 未披露 | 获取续约和流失队列,并说明流失原因 |
| 合同期限 / 续约条款 | null | 医院账户 | 未披露 | 确认期限、自动续约和价格上调条款 |
| 客户满意度 / NPS | null | 放射科医生用户 | 未披露 | 要求提供调研结果并开放客户访谈 |
| 切换成本机制(定性) | 高(嵌入式 + PACS 集成) | 捆绑账户 | 中 | 在真实续约中验证锁定效应与重新验证成本 |
所有量化留存指标均为 null,因为截至 2026-06-23 运行日 UII 没有公开披露;只有定性的切换成本机制能被证据支撑。
6.5 扩张路径与集中度风险
UII 的扩张引擎,是锚定母公司渠道的落地再扩张循环:医院先购买带嵌入式 AI 的 UIH 扫描仪,再为更多已获批应用采用 uAI Clinical Portal,随后在新智能体和基础模型能力商业化后继续扩展;海外增长则跟随 UIH 快速国际化铺开(母公司 2025 年海外收入增长超过 50%)。这种打法有力,因为硬件关系先筛选了账户,也让 AI 获客成本低于纯软件对手。 同一结构在三条线上制造集中度风险。第一,渠道和所有权集中:UII 的分销、品牌和很大一部分装机基础依赖母公司 UIH,因此客户关系只有一部分属于 UII 自己。第二,地域和客群集中在中国三级医院市场;该市场受带量采购和预算压力挤压价格,AI 软件又没有全国医保支付项目。第三,增长前沿的地缘政治集中风险:美国和欧洲扩张面对已有记录的信任、供应链安全、数据隐私和中美贸易紧张阻力,这些阻力可能拖慢或阻断采购,恰好发生在 UII 最想拿下的学术账户。公开披露没有量化前几大客户收入集中度,因此单账户或单渠道依赖的幅度仍是重大开放尽调项,而不是已测量数字。[CU035, CU036, CU037, CU038, CU039, CU040]
| 扩张驱动因素 | 集中度风险 | 风险兑现后的影响 | 尽调路径 |
|---|---|---|---|
| 借 UIH 硬件捆绑先落地、再扩张 | 渠道和所有权依赖母公司 UIH | 客户关系和续约只由 UII 部分掌控 | 梳理集团内渠道条款和账户归属 |
| uAI Clinical Portal 交叉销售 60+ 个应用 | 单一账户内依赖单一平台 | 关键应用表现不佳时,平台可能被替换 | 复核单应用使用量和被替换历史 |
| 借 UIH 国际业务 50%+ 爬坡推动海外增长 | 暴露于中美信任、安全和贸易摩擦 | 西方学术站点采购可能放慢或受阻 | 用出口和数据本地化风险压力测试管线 |
| 新 AI 智能体与基础模型 | 最新层级的生产经济性尚未验证 | 若智能体仍处于商业化前,展示与收入会脱节 | 跟踪具名付费智能体部署和结果证据 |
| 中国三甲市场深度 | VBP 定价下的地域 / 细分集中 | 价格受压,且没有 AI 软件单独报销项目 | 量化中国收入占比和实际定价趋势 |
集中度风险根据截至 2026-06-23 运行日披露的渠道和地域结构推断;UII 未公布收入集中度拆分,因此没有任何一项能量化。
硬件关系如何扩展成经常性 uAI 软件预算和更新的 AI 智能体预算,以及它制造出的集中度风险。
[CU035, CU036, CU037, CU038]6.6 图表
07风险
7.1 风险概览与严重性排序
United Imaging Intelligence 的风险面,不太像一家初创公司的脆弱性,更像一家依附上市影像母公司的中国医疗 AI 公司所处的政策与结构环境。按剩余严重性排序,最高层是外部风险,且大多不受管理层控制:美国针对中国生物技术和影像采购的国家安全限制、中国 NMPA 对高风险 AI 软件的审批复杂性,以及 PIPL 和《数据安全法》下的健康数据隐私与本地化规则。第二层是对母公司 United Imaging Healthcare(UIH)的结构性依赖,分销、资本、品牌和装机基础都集中在单一交易对手身上。第三层是运营和技术风险——临床 AI 安全、模型泛化、召回、医院部署系统的网络安全。第四层是财务模型和执行风险,母公司实力会压低这一风险,但独立经济数据未披露又把风险抬高。 这里的严重性综合了发生概率、影响和缓释成熟度。监管和地缘政治风险影响高、发生概率上升、公司侧缓释能力低,因为 UII 不能绕开 BIOSECURE Act 或改写 NMPA 证据要求,只能适应。依赖风险影响高,但母公司的财务健康和一致激励能部分缓释。运营风险为中等,监管许可和质量体系是部分缓释因素。克制的读法是,UII 的主要威胁是外生且缓慢移动,而非急性风险,因此可以监测:下方每项顶层风险都带有可度量的论点击穿触发器;诉讼敞口、安全事件、收入集中等未确认事项,会以明确尽调路径标出,而不是默认良性。[CR001, CR002, CR003, CR004, CR005, CR006]
截至 2026-06-23,UII 主要风险的发生可能性、影响、缓释成熟度和剩余严重性(序数、证据加权)。
评级是基于截至 2026-06-23 所引用披露、按证据权重给出的序数判断,并非实测概率。
[CR004, CR009, CR012, CR027, CR035]7.2 监管与法律风险
监管敞口是 UII 的定义性风险,而且双向延伸。在中国,NMPA 将高风险诊断和决策支持 AI 归为 Class III——最高风险级别——要求用本地代表性数据做强制临床评价、证明统计稳健性,为会演进或自我修改的算法预先设定变更控制,并在任何用于训练或验证的患者数据上遵守 PIPL、《数据安全法》和《网络安全法》。NMPA 的 2025–2026 年标准项目和高端器械指导,继续把门槛推向国际接轨,同时保留中国特有的临床和网络安全要求。因此,每个新 uAI 应用、每次模型更新,都是一个新的监管项目,而不是一次性许可。健康数据本地化和跨境传输安全评估,也进一步限制 UII 如何用多站点数据训练模型,或从中国基础设施服务海外客户。 在美国和盟友市场前沿,法律风险正在升级。BIOSECURE Act 已于 2025 年 12 月 18 日作为 FY2026 NDAA 第 851 节成为法律,限制联邦机构以及与联邦相关的承包商和受资助方,从被列为「受关注生物技术公司」的主体采购「生物技术设备或服务」;OMB 名单应在 2026 年 12 月前发布,而国防部第 1260H 条「中国军事公司」指定已在 2026 年扩展。即便 UII 或 UIH 尚未被具体指定,寒蝉效应叠加 FDA 对中国来源器械的供应链和网络安全审查,也可能拖慢联邦资助美国医院的采购。公开记录中没有发现直接点名 UII 的 IP 诉讼,但 medtech 本就诉讼密集,放射 AI 的专利 / NPE 活动正在上升;因此没有披露案件是尽调缺口,不是健康证明。[CR009, CR010, CR011, CR012, CR013, CR014]
| 规则 / 制度 / 事项 | 司法辖区 | 状态 | 可能性 | 严重性 | 缓释措施 | 剩余暴露 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| BIOSECURE Act(FY2026 NDAA Sec. 851)采购禁令 | 美国 | 2025 年 12 月 18 日生效;OMB 名单应于 2026 年 12 月发布 | 中高 | 极高 | 调整美国商业化路径;争取非联邦渠道 | 若被列名,涉联邦美国采购可能被禁止 | 跟踪 OMB BCC 名单和 UII/UIH 的 1260H 列名 |
| NMPA 三类审批与变更控制 | 中国 | 制度仍在执行;2025–2026 年标准更新 | 高 | 高 | 本地临床证据;质量体系;已持有 13 个三类证 | 每个应用和模型更新都是新的监管项目 | 获取获批管线、时间表和变更控制日志 |
| PIPL / 数据安全法 / 健康数据本地化 | 中国 | 已生效;数据出境需 CAC 安全评估 | 高 | 高 | 境内数据托管;同意与 DPIA 流程 | 跨境训练 / 服务受限;披露很薄 | 核验数据架构、DPIA 和出境审批 |
| FDA 对中国来源设备的审查 | 美国 | 供应链 / 网络安全审查趋严 | 中 | 高 | 15 项 FDA 准入许可;MCRA 监管支持 | 审批可能放慢,采购也可能降温 | 复核 FDA 申报历史及任何机构问询 |
| IP / 专利诉讼暴露 | 全球 | 未披露点名 UII 的案件 | 中低 | 中 | 专利组合;FTO 分析(未见) | 放射 AI 领域 NPE 活动上升;没有案件并不等于确认无暴露 | 开展 FTO 和诉讼检索;要求 IP 赔偿承诺 |
各行按剩余严重性排序;“未披露案件”仅表示截至 2026-06-23 运行日未见公开记录,并不确认没有暴露。
[CR009, CR010, CR011, CR013, CR016, CR017]7.3 运营、质量与技术风险
UII 的运营和技术风险集中在临床安全、模型可靠性,以及嵌入医院影像工作流的软件安全。uAI 应用会影响分诊、检测、分割和报告;一旦出现假阴性,或模型无法从训练数据泛化到新医院的扫描仪、协议或患者组合,就可能造成临床伤害并引发监管行动。NMPA 变更控制规则意味着,修复或更新已部署模型本身也是受治理、证据负担重的流程,而不是静默补丁。RSNA 2025 和 ECR 2026 展示的最新自主智能体把风险抬得更高,因为自主性越高,基准表现与可追责真实世界临床行为之间的距离越大;具名站点的独立生产级验证仍有限。 网络安全和数据处理是第二条运营轴线。系统一旦集成进医院 PACS/RIS,某些情况下还原生嵌入 UIH 扫描仪,就会形成攻击面,也带来 PIPL 和《数据安全法》下的患者数据保护义务;UII 对安全架构、控制措施或事件历史公开披露很少,因此韧性必须验证,不能假设。供应链和制造风险主要落在母公司 UIH 硬件业务上,但会通过捆绑部署,以及任何组件或出口限制传导到 UII。诚实判断是,UII 的质量和许可广度是真实缓释因素,但使用层面的可靠性数据、安全披露和上市后监测证据并不公开,必须取得;一旦确认临床事件或召回,就是清晰的论点击穿触发器。[CR019, CR020, CR021, CR022, CR023, CR024]
| 失效模式 | 可能性 | 严重性 | 缓释成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|
| 模型无法泛化到新站点 / 扫描仪 / 人群 | 中 | 高 | 中(准入许可、验证研究) | 本地验证薄弱时存在临床错误风险 | 未披露站点级使用量或错误率 |
| 临床安全事件或设备召回 | 中低 | 极高 | 中(质量体系、上市后监测) | 一旦发生,会伤害声誉并引发监管后果 | 未公开披露事件或召回历史 |
| 自主智能体实践表现偏离基准 | 中 | 高 | 低(独立生产验证有限) | 自主性越高,问责缺口越大 | 没有独立具名站点的智能体结果证据 |
| PACS 集成系统遭遇网络安全入侵 | 中 | 高 | 中低(控制措施未披露) | 患者数据暴露并触发 PIPL 责任 | 安全架构和事件历史未公开 |
| 供应链 / 组件或出口中断(经母公司传导) | 中 | 中高 | 中(母公司规模和服务网络) | 捆绑部署暴露于硬件约束 | 未披露组件依赖或出口暴露 |
可能性和严重性是截至 2026-06-23 运行日、按证据加权后的序数判断;“未解决缺口”标记 UII 未公开披露的运营数据。
7.4 伙伴、依赖与集中度风险
UII 最独特的结构性风险,是对母公司 United Imaging Healthcare 的依赖。UIH 提供把嵌入式 uAI 带进医院的硬件渠道、很大一部分品牌和全球服务足迹、很大一块装机基础,并通过更大集团提供资本和战略方向。这种集中是一把双刃的缓释工具:母公司 2025 年人民币 138.0 亿元收入和强资产负债表,降低了融资和分销风险;但也意味着 UII 的客户关系、定价权,甚至监管命运,都只有一部分掌握在自己手里。UIH 一旦遭遇冲击——1260H 列名、出口限制、带量采购下需求下滑——会直接传导到子公司。国资相关所有权和投资者又增加了双向政治对齐风险:在中国有本土偏好,在美国则引发更高疑虑。 母公司之外,依赖风险还包括:少数基础模型和 uAI Clinical Portal 作为价值交付的单一平台;中国带量采购压低医院采购预算;下一轮融资对资本提供方的依赖。UII 没有披露前几大账户收入拆分,客户集中度幅度未知,因此单账户或单渠道依赖是重大开放项。下方依赖地图和登记表把每个关键交易对手——母公司、监管机构、重点医院、资本提供方、平台组件——对应到失败场景、严重性评级和尽调路径;失去母公司支持或被美国正式指定,是最清晰的依赖论点击穿点。[CR027, CR028, CR029, CR030, CR031, CR032]
| 依赖 | 交易对手 | 角色 | 集中度 | 失效场景 | 严重性 | 缓释措施 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| 分销与装机基础 | 母公司 UIH(688271.SH) | 硬件渠道、品牌、服务、资本 | 很高 | 母公司需求下滑、被列名或遭美国指定 | 极高 | 母公司财务实力;激励一致 | 客户关系只由 UII 部分掌控 |
| 监管审批 | NMPA / FDA / 欧盟公告机构 | 市场准入守门人 | 高 | 规则收紧、执法加强或准入许可被撤回 | 高 | 既有准入许可;监管合作伙伴 | 每个新应用都依赖重新审批 |
| 平台与模型 | uAI Clinical Portal + 基础模型 | 价值交付的单一界面 | 高 | 关键应用表现不佳或模型失效 | 中高 | 多应用组合;持续研发 | 账户内存在平台替换风险 |
| 资本提供方 | Series A 投资者 / 下一轮市场 | 为 AI 智能体建设融资 | 中 | 降估值融资或下一轮停滞 | 中 | 母公司强;2026 年中国 AI 融资反弹 | 独立现金消耗和现金续航未披露 |
| 关键客户 / 渠道 | 三甲医院;UIH 设备账户 | 收入和背书 | 未知(未披露) | 失去头部账户或渠道转向 | 中高 | 公司声称机构基础广 | 未披露头部客户集中度 |
集中度评级根据截至 2026-06-23 运行日披露的结构推断;“未知”表示没有公开量化该依赖。
UII 依赖的关键交易对手,以及各自失效会传导出的影响。
[CR027, CR029, CR031, CR034]7.5 财务、执行风险、缓释措施与退出标准
财务模型风险是 UII 主要敞口里最可控的一类,但并非可以忽略。2025 年 6 月 Series A 融资约 1.39 亿美元、估值约 14 亿美元,加上资金雄厚的上市母公司,使近期急性偿付风险不高。风险在质量和耐久性,而不是生存:UII 没有披露独立收入、毛利率、烧钱速度或 runway;中国没有诊断 AI 软件的全国医保支付项目;带量采购压缩价格;Airdoc 等亏损上市同行——连续五年净亏损——说明广泛临床采用未必能转化为有利润、可持续的收入。执行和人员风险集中在留住稀缺 AI 与临床监管人才、扩大尚未形成收入的 AI 智能体商业化,以及管理从开源到付费产品的桥梁;这些都没有公开量化。 缓释纪律让这些风险变得可投资。每项顶层风险都对应一个可监测指标和论点击穿触发器:UII 或 UIH 被 BIOSECURE/1260H 正式指定,NMPA 执法行动或撤回许可,已报道临床安全事件或召回,母公司 UIH 财务显著恶化,下一轮融资 down-round 或停滞,或出现单一客户收入集中超过审慎阈值的证据。下方传导图展示每项风险如何流向收入、客户、利润率、融资和估值;退出标准表把每项风险配上可度量触发器和行动含义。凡证据缺失之处——安全事件、诉讼、合同集中、独立经济数据——本章记录具体尽调问题,而不是凭直觉给出安全感。[CR035, CR036, CR037, CR038, CR039, CR040]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| AI / ML 研究人才 | 留住稀缺的基础模型和 CV 人才 | 中 | 高 | 开源可见度;母公司资源 | 复核流失率、股权和关键人覆盖 |
| 临床监管事务 | 能否并行推进 NMPA/FDA/CE 项目 | 中 | 高 | 已证明能拿下多司法辖区准入许可 | 确认监管团队人数和外部合作伙伴 |
| 商业 / 市场进入领导力 | 在硬件捆绑之外扩张付费软件 | 中 | 中高 | 母公司渠道降低部分 CAC | 要求提供销售配额达成情况和销售组织结构 |
| 创始人 / 关键人集中 | 依赖 CEO 和 CTO 领导力 | 中低 | 中 | 领导团队已成形 | 确认接班安排和关键人风险覆盖 |
人员风险评级根据截至 2026-06-23 运行日的公开披露推断;UII 未公布流失率、员工人数或关键人覆盖数据。
| 风险 | 可监测触发器 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 美国国家安全限制 | OMB BCC 名单 / DoD 1260H 指定 | UII 或 UIH 被列为受关注公司 | 下调美国增长假设;重新评估估值 |
| NMPA / 中国监管 | 执法行动或准入许可被撤回 | 任何 Class III 暂停或关键应用被驳回 | 暂停中国产品扩张假设 |
| 临床安全 | 披露事件或召回 | 确认与 uAI 产品相关的伤害事件 | 重新评估质量与责任敞口 |
| 母公司依赖 | 母公司 UIH 财务 / 战略冲击 | 收入显著下滑或战略性剥离 | 重新评估渠道、资本与品牌支持 |
| 融资 / 模型 | 下一轮条款与烧钱披露 | 下轮估值下调,或现金 runway 不足 12 个月 | 收紧偿付能力与稀释假设 |
| 客户集中度 | 头部客户收入披露 | 单一客户或渠道超过审慎占比 | 对收入质量施加集中度折价 |
触发条件是截至 2026-06-23 本轮研究日期可监控的尽调阈值,不是预测;每一项都把可量化事件与行动含义配对。
UII 的主要风险如何传导到客户、收入、利润率、融资和估值。
[CR012, CR014, CR028, CR036]7.6 图表
08估值
8.1 建议与价格纪律
基于今天的公开证据,United Imaging Intelligence 的正确姿态是继续研究(research-more),不是回避,也不是买入。战略和临床故事很强:2025 年 6 月 Series A 融资约 1.39 亿美元,投后估值接近 14 亿美元,由易方达、上海国资投资方和上市母公司 United Imaging Healthcare 锚定。部署真实存在,覆盖 4,000 多家机构,监管堆栈也很深,拥有 NMPA Class III、CE 和 FDA 许可。问题在价格纪律。United Imaging Intelligence 没有披露独立收入、毛利率或留存,因此 14 亿美元估值隐含的倍数,从公开证据看根本无法得知。这迫使承销判断依赖战略叙事,而不是分母。建议谨慎,是因为核心未解问题是估值支撑,而不是临床相关性;举证责任现在完全落在尽调上。[CV001, CV002, CV003, CV005, CV010, CV015]
| 维度 | 取值 | 理由 |
|---|---|---|
| 建议 | 继续研究 | 战略背书和真实落地都成立,但估值押注的是尚未披露的独立经济性。 |
| 置信度 | 中 | 融资头条已有充分交叉验证,但核心经营分母在公开记录中完全缺位。 |
| 风险评级 | 高 | 在任何经审计单位经济性出现前,偏高的私募估值先撞上变现、地缘政治和母公司依赖风险。 |
| 估值立场 | 偏高 | 在未披露收入基数上打出 $1.4B 标记,已经把中国同业仍难证明的 AI 变现成功定价进去。 |
| 决策含义 | 为战略可选性保持跟进,但在承销价格前,必须拿到独立收入、毛利率和留存。 | 闸门不是战略相关性,而是分母披露和变现证明。 |
该建议对价格和证据都敏感;更清晰的独立披露,或更低的入场价,都可能推动立场变化。
[CV001, CV003, CV005, CV015, CV031, CV038]| 论点 | 正方 | 什么会改变判断 |
|---|---|---|
| 战略背书 | 国家相关锚定方易方达和上海国投,加上上市母公司 UIH,给了少见的融资耐久度和政策同向性。 | 如果母公司支持或国家资本胃口降温,融资与需求顺风都会减弱。 |
| 落地证明 | AI 工具覆盖 4,000 多家机构,并拥有约十三项 NMPA Class III 许可,证明临床拉动真实存在。 | 如果已安装部署很浅、靠捆绑免费铺开,或变现很低,平台叙事会变薄。 |
| 监管护城河 | 大量 NMPA Class III 加上 CE 和 FDA 许可,抬高了相对许可更薄竞争者的门槛。 | 如果审批无法转化为付费、可报销使用,护城河就是监管成本,而不是收入。 |
| 估值支撑 | 如果独立收入已在 $100M-$200M 区间且具备软件毛利率,$1.4B 标记可以说得通。 | 如果收入低得多,或毛利率很薄,相比上市影像 AI 可比公司,价格会显得很贵。 |
| 变现现实 | 母公司 UIH 的设备带动效应,可能让 UII 把 AI 作为附加软件层变现。 | 如果中国医院对 AI 的支付意愿像同业亏损所暗示的那样持续偏弱,变现就会滑坡。 |
反方核心是分母和变现风险,并不是说 United Imaging Intelligence 缺少临床牵引。
[CV003, CV005, CV008, CV015, CV022, CV031]结论仍偏谨慎:战略背书和部署都是真实的,但价格锚已经跑在任何已披露的独立经济性之前。
[CV001, CV003, CV005, CV008, CV015, CV031]United Imaging Intelligence 在背书和部署上得分较高,但独立经济性可见度和当前价格支撑偏弱。
分数是锚定公开引用证据的 0-10 序数判断,并非公司披露的管理层 KPI。
[CV005, CV008, CV015, CV022, CV031, CV034]8.2 可比公司纪律与母公司锚点
私募估值缺少分母时,上市可比公司是必要的纪律工具。最接近的规模化参照是母公司本身:United Imaging Healthcare 在 2026 年中市值接近人民币 850 亿元,约为收入的六倍;2025 年收入约人民币 138 亿元,净利润接近人民币 18.7 亿元。这是硬件主导的倍数,不是纯软件 AI 倍数,但能给预期设边界。影像 AI 同行中,中国上市公司 Airdoc 的企业价值 / 收入约为 1.8 倍,且仍在亏损;更广泛的中国影像 AI 同行聚集在低个位数收入倍数。对照美国上市 Tempus 约 7.1 倍、Viz.ai 2022 年 12 亿美元估值等私募参照,United Imaging Intelligence 的 14 亿美元位于同业高端。信号不是 UII 已被高估,而是其估值已经嵌入相对公开影像 AI 经济模型更高的战略和国资背书溢价。[CV003, CV010, CV016, CV017, CV018, CV019]
| 可比对象 | 指标 | 倍数 / 估值 / 状态 | 相关性 | 局限 |
|---|---|---|---|---|
| United Imaging Intelligence(标的) | 2025 年 6 月 Series A 轮投后 | 未披露独立收入,对应约 $1.4B 估值 | 正在检验的估值标记;现有最佳私募锚点。 | 未披露独立收入,因此隐含倍数不可知。 |
| United Imaging Healthcare(母公司,688271.SH) | 2026 年 6 月市值 / 市销率 | 约 CNY 85B 市值,约 6x P/S | 最接近的规模化可比公司,也是 UII 的直接股东。 | 硬件主导的母公司,不是纯软件 AI 倍数。 |
| Airdoc(HKG 2251) | 2026 EV / 收入(上市,亏损) | 约 1.8x EV/收入;约 5.4x 市值/收入 | 直接上市的中国影像 AI 同业,可校验变现现实。 | 规模更小、亏损,且受公开市场折价。 |
| Tempus AI(NASDAQ) | 2026 EV / 收入(上市) | 约 7.1x EV/收入,对应约 $10B EV | 高端健康 AI 的规模化诊断加数据基准。 | 美国上市、基因组学主导,监管和支付体系不同。 |
| Viz.ai | 2022 年 Series D 私募估值 | 约 $1.2B 估值;据称 2025 年已盈利 | 接近 UII 标记的影像分诊 AI 私募同业。 | 美国护理协调模型,且私募轮次较久。 |
| Infervision | 据报 2026 年估值 / 收入估计 | 约 $215M 估值,对约 $12M 收入(约 17x,异常值) | 直接中国影像 AI 同业,可观察私募规模。 | 第三方估计;高倍数很可能有数据噪声。 |
| Yitu Technology | 最新参考私募估值 | 横跨广义 AI,约 $2.4B 估值 | 上海 AI 同业,显示中国平台 AI 估值标记。 | 业务多元,超出医学影像;IPO 已搁置。 |
这组可比对象混合了一个申报级母公司可比、上市同业和私募参考,因为没有单一公司能完全匹配 UII 这种依附母公司的影像 AI 位置。
[CV003, CV016, CV017, CV018, CV019, CV020]14 亿美元估值对应的收入倍数,取决于未披露的分母,差异会很大;上市同业给低端区间提供锚点。
条形图把假设收入水平下的隐含所需倍数与上市同业锚点放在一起;由于 UII 未披露收入,这些只是估值纪律工具,不是预测。
[CV003, CV016, CV018, CV019, CV031]8.3 情景、上行与下行
United Imaging Intelligence 的牛市情景是连贯的。如果 AI 能作为高毛利软件附着在母公司扩张中的装机基础上变现,如果出口能在欧洲和盟友市场放量,如果到 2028 年独立收入已在 2 亿至 3.5 亿美元区间,8 至 12 倍收入倍数可以支撑 20 亿至 40 亿美元价值。基准情景更窄:UII 随母公司增长,把许可转化为适度付费使用,使当前估值只算勉强可辩护,约为 6 至 9 倍收入。熊市情景也可信,因为这里的估值对缺失分母高度敏感。如果中国医院 AI 付费意愿仍弱,正如上市同行亏损所暗示;如果地缘政治通过 BIOSECURE 式限制封顶出口;或如果收入远低于估计,那么上市影像 AI 约 2 至 5 倍收入区间就会成为相关锚点,下行会迅速扩大。在 14 亿美元估值上,投资者已经提前折现了相当多好消息,给变现失望留下的空间很小。[CV015, CV022, CV024, CV026, CV031, CV034]
| 情景 | 概率信号 | 2028 收入假设 | 估值逻辑 | 隐含股权价值 |
|---|---|---|---|---|
| 乐观 | AI 作为高毛利附加层,随 UIH 增长中的装机基础变现,出口也在国际市场放量。 | $250M-$350M | 对耐久增长、战略稀缺性和国家背书给 8x-12x 收入 | $2.0B-$4.2B |
| 基准 | UII 随母公司增长,并把许可转化为适度付费使用,但估值纪律仍需约束。 | $130M-$200M | 以高端但不过激的中国软件倍数给 6x-9x 收入 | $0.9B-$1.8B |
| 悲观 | 医院对 AI 的支付意愿持续偏弱,地缘政治限制出口,变现像上市同业一样滞后。 | $60M-$110M | 更接近中国上市影像 AI 区间,给 2x-5x 收入 | $0.15B-$0.55B |
| 决策锚点 | 在 $1.4B 标记下,投资人已经在承销基准情景的上半段。 | 当前已披露的私募锚点 | 价格几乎不给变现失望留空间 | $1.4B |
情景是公开证据约束工具,不是管理层指引;区间很宽,因为 UII 没有披露独立收入分母。
[CV003, CV015, CV022, CV024, CV026, CV031]| 触发条件 | 阈值 | 对论点的传导 | 行动含义 |
|---|---|---|---|
| 披露的独立收入远低于估计 | 经验证 UII 收入明显低于约 $100M,且没有异常高毛利率 | $1.4B 标记会立刻失去分母支撑。 | 不按 Series A 头条价格承销。 |
| 医院 AI 变现持续偏弱 | 经营亏损持续,类似 Airdoc 等上市同业 | 高端倍数逻辑会塌向公开影像 AI 区间。 | 按 2x-5x 收入可比重新切估值。 |
| 地缘政治出口天花板变硬 | BIOSECURE 式限制或 FDA 行动阻断美国及盟友销售 | 国际上行空间和出口驱动的乐观情景都会消失。 | 下调乐观概率,并扩大下行情景。 |
| 母公司依赖转为不利 | UIH 重新分配资本、重组 UII,或关联方条款变化 | 头条估值不再代表可独立投资的经济性。 | 暂停,直到独立治理和经济性清晰。 |
| 许可无法转化为付费使用 | NMPA Class III 审批停留为无报销的成本中心 | 监管护城河会变成费用,而不是耐久收入。 | 将护城河主张视为未证实,并相应折价。 |
这些触发条件刻意绑定估值传导,而不是宽泛经营风险。
[CV015, CV022, CV031, CV034, CV035, CV038]基于公开证据,区间横跨当前估值;只有变现和出口规模同时跑通,结果才会落在当前估值之上。
区间是基于公开证据的简单情景输出;收入分母缺失,因此刻意避免给出虚假的 DCF 精度。
[CV003, CV015, CV031, CV034, CV038]8.4 估值基准及其含义
外部倍数基准有助于判断 United Imaging Intelligence 的估值有多激进。进入 2026 年,HealthTech 并购估值大多在中个位数收入倍数、低双位数 EBITDA 附近成交;优质 AI 和数据平台则能高于这一带。AI 初创顾问工作提到的分布更宽,最强平台型公司可达到未来收入的 10 至 50 倍。香港市场上,Chapter 18A 的未盈利 biotech 和医疗 AI 上市公司,在短期收入可预测时估值接近未来收入 15 至 30 倍,差异化平台则更高。如果 UII 的独立收入规模不大但增长快、且具备软件利润率,United Imaging Intelligence 也许能证明双位数收入倍数合理,从而让 14 亿美元估值落在高端私募区间之内,而非之外。决定性未知仍是分母:同一个标价是否合理或偏贵,完全取决于公司尚未披露的收入数字。[CV007, CV011, CV012, CV013, CV014, CV030]
8.5 下一步必须证明什么
决定性尽调路径很直接。投资者需要 United Imaging Intelligence 的独立收入,至少需要管理层认证的收入桥;需要经常性与项目型收入拆分、扣除算力和支持后的毛利率、4,000 多家机构足迹中的留存和使用深度,以及与母公司的关联交易条款。公开证据足以说明 UII 在临床和战略上重要;但不足以说明新投资者应在没有这些材料时接受 14 亿美元估值。地缘政治敞口还增加了第二道门槛:必须量化依赖美国和盟友市场的管线占比,并对照 BIOSECURE 式和 FDA 风险,因为出口限制可能拿掉牛市情景的一条核心支柱。在这些输入到位之前,本章信息很简单:保持兴趣,把公司留在积极尽调中,但让它自己挣来溢价,而不是把战略欣赏自动转化成价格接受。[CV031, CV032, CV034, CV035, CV036, CV037]
| 主题 | 缺失证据 | 为什么重要 | 责任方或尽调路径 |
|---|---|---|---|
| 独立收入与 ARR 桥接 | UII 单独口径确认收入、经常性占比、增长,以及母公司相关收入拆分。 | 价格对目前未披露的分母极其敏感。 | 财务尽调、管理层认证收入桥接,以及母公司剥离明细。 |
| 毛利率与变现模型 | 单应用定价、捆绑与付费组合,以及扣除算力和支持后的毛利率。 | 只有 AI 软件经济性真正有吸引力,高端倍数才站得住。 | 财务和产品尽调,叠加定价兑现分析。 |
| 留存与部署深度 | 付费机构数量、使用深度,以及 4,000 多家覆盖中的续约或流失。 | 只有客户名单广度、没有深度,撑不起耐久溢价。 | 客户访谈和部署遥测审查。 |
| 母公司依赖与关联方条款 | UII 与 UIH 之间的分销、IP、算力和定价条款。 | 经济现实可能与头条私募标记大幅偏离。 | 法律顾问审查关联方协议和股权结构表。 |
| 地缘政治出口敞口 | 依赖美国及盟友市场的项目管线占比,并暴露于 BIOSECURE 和 FDA 风险。 | 出口限制可能拿掉乐观情景的一条核心支柱。 | 监管与贸易法律顾问做情景分析。 |
如果这些问题不能清楚回答,正确姿态是耐心等待,而不是基于战略叙事激进接价。
[CV015, CV031, CV032, CV034, CV035, CV038]8.6 图表
免责声明
本报告基于截至运行日期的公开信息;在 United Imaging Intelligence 未提供经审计或独立运营披露的部分,使用了明确标注的第三方估计和上市同业可比数据。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | United Imaging Intelligence is a Shanghai-based medical-AI company and the AI subsidiary of United Imaging Healthcare. | 高 | SO003, SO005, SO007 |
| CO002 | UII's Chinese legal entity is Shanghai United Imaging Intelligence Co., Ltd. (上海联影智能科技股份有限公司). | 高 | SO002, SO003, SO021 |
| CO003 | UII is headquartered at 701 Yunjin Road, Xuhui District, Shanghai. | 高 | SO002, SO021 |
| CO004 | UII develops AI software across the full clinical pathway covering screening, diagnosis, treatment, and follow-up. | 中 | SO002, SO003 |
| CO005 | UII markets its software under the uAI brand spanning CT, MR, PET/CT, X-ray, and ultrasound modalities. | 中 | SO001, SO002 |
| CO006 | In 2025 UII reframed around a multimodal medical foundation model called Yuanzhi and the uAI NEXUS model suite. | 中 | SO008, SO015 |
| CO007 | UII operates 12 product platforms and more than 100 AI applications. | 中 | SO008 |
| CO008 | Xiang "Sean" Zhou (Zhou Xiang) is co-founder and (co-)CEO of UII and its legal representative. | 中 | SO008, SO006 |
| CO009 | Dijia Wu is identified as UII's chief technology officer. | 低 | SO006 |
| CO010 | UII's research bench includes named vice-presidents such as Feng Shi, Yaozong Gao, and Shanhui Sun. | 低 | SO006 |
| CO011 | UII's public board composition, post-Series-A cap table, and division of economics with parent UIH are not disclosed. | 中 | SO005, SO008 |
| CO012 | 36Kr reported in 2025 that UII is weighing an independent public listing. | 中 | SO008 |
| CO013 | UII completed a roughly RMB 1 billion (about $139 million) Series A in June 2025. | 高 | SO007, SO009, SO010 |
| CO014 | The Series A was co-led by E Fund Management and Shanghai Guotou Investment (Shangguo). | 高 | SO007, SO008, SO009 |
| CO015 | Series A participants included Shanghai Alliance/United Investment, Sun Rock Capital, Suzhou Capital Group, Qingsong Capital, and United Imaging Healthcare. | 中 | SO007, SO010, SO012 |
| CO016 | UII's Series A implied a pre-money valuation near RMB 9 billion and a post-money valuation around $1.4 billion. | 中 | SO009 |
| CO017 | Some reposts credit SIG Property Management as a co-lead alongside E Fund, differing from the E Fund plus Shanghai Guotou framing. | 低 | SO010 |
| CO018 | UII holds 13 Class III NMPA medical-device certificates. | 中 | SO008 |
| CO019 | UII reports 15 US FDA clearances and 31 CE certifications for its AI applications. | 中 | SO008 |
| CO020 | The FDA's end-2025 list shows United Imaging with 38 cleared radiology AI devices, ranking fifth among all vendors. | 中 | SO020 |
| CO021 | UII says its AI products have entered more than 4,000 medical institutions across China. | 中 | SO008 |
| CO022 | UII reports more than 3,000 hospital deployments worldwide. | 中 | SO002 |
| CO023 | UII says proceeds from the Series A will fund multimodal large-model development and hospital digital-intelligence systems. | 中 | SO008 |
| CO024 | At RSNA 2025 UII debuted the uAI Insight image-to-report agent and an embodied-AI robotic-arm ultrasound system. | 高 | SO015, SO018, SO019 |
| CO025 | A UII chest-CT agent can identify 73 thoracic findings from a single scan. | 中 | SO008, SO015 |
| CO026 | UII unveiled its multimodal foundation-model and agent strategy at the 2025 World AI Conference (WAIC). | 中 | SO008 |
| CO027 | UII gained early prominence in 2020 for deploying COVID-19 chest-CT triage software. | 中 | SO003, SO021 |
| CO028 | UII received US FDA 510(k) clearances including uAI Easy Triage ICH and the uAI Portal platform from 2024. | 高 | SO024, SO020 |
| CO029 | In April 2026 UII open-sourced uAI NEXUS MedVLM, a medical-video language model accepted to CVPR 2026. | 中 | SO025 |
| CO030 | At ECR 2026 UII presented more than 30 CE-certified applications and the uCP portal's 60-plus apps. | 中 | SO016 |
| CO031 | UII does not publicly disclose its standalone revenue, ARR, gross margin, or pricing. | 中 | SO005, SO008 |
| CO032 | UII says roughly 80% of its staff are research-and-development personnel. | 低 | SO006 |
| CO033 | UII's software runs both as embedded intelligence on UIH scanners and on third-party hardware. | 中 | SO002, SO008 |
| CO034 | UII maintains offices in Shanghai, Beijing, Wuhan, and a Boston-area subsidiary, UII America, Inc. | 中 | SO022, SO006 |
| CO035 | UII America, Inc. is located in Burlington, Massachusetts, anchoring UII's North American presence. | 中 | SO022 |
| CO036 | UII is structurally dependent on parent UIH for imaging data, hardware distribution channel, and capital. | 中 | SO003, SO007, SO008 |
| CO037 | United Imaging Healthcare participated as an investor in UII's Series A, reinforcing parent control. | 高 | SO007, SO009, SO010 |
| CO038 | UII's Series A investors include state-linked vehicles such as Shanghai Guotou and Suzhou Capital Group. | 中 | SO007, SO012 |
| CO039 | UII's standalone headcount, board composition, and cap-table terms remain undisclosed. | 中 | SO005, SO008 |
| CO040 | Parent United Imaging Healthcare reported 2025 revenue of about RMB 13.80 billion, up roughly 34% year on year. | 高 | SO017, SO023 |
| CO041 | UII operates in a crowded, partly commoditized Chinese medical-AI market where AI reimbursement remains immature. | 低 | SO003, SO008, SO026 |
| CO042 | As a Chinese medical-AI vendor, UII faces US-China technology and data-governance tensions affecting Western expansion. | 低 | SO020, SO024 |
| CM001 | United Imaging Intelligence competes in the AI-in-medical-imaging software market, which applies machine learning to CT, MRI, X-ray, ultrasound, PET and nuclear scans. | 中 | SM001, SM005 |
| CM002 | The global AI medical imaging market was estimated at about USD 1.8 billion in 2025 and is projected to reach USD 20.2 billion by 2033 at roughly 35% CAGR. | 中 | SM001 |
| CM003 | An alternative analyst estimate sizes the 2026 global AI medical imaging market at USD 5.29 billion, reaching USD 20.28 billion by 2030 at about 39.9% CAGR. | 中 | SM005 |
| CM004 | A third analyst values the 2026 global AI medical imaging market at only USD 2.20 billion, reaching USD 17.77 billion by 2033 at 34.8% CAGR. | 中 | SM006 |
| CM005 | Public 2026 global AI-imaging estimates span roughly USD 2.2-5.3 billion, a wide spread reflecting inconsistent category definitions across publishers. | 中 | SM001, SM005, SM006 |
| CM006 | Neurology was the largest application segment at around 37% of 2025 revenue and CT the leading modality at roughly 34-42%. | 中 | SM001, SM006 |
| CM007 | Hospitals are the dominant end-user of AI medical imaging, accounting for over 60% of demand. | 中 | SM006, SM001 |
| CM008 | Oncology is among the fastest-growing application segments for AI medical imaging, supported by rising cancer incidence. | 中 | SM008 |
| CM009 | The shortage of radiologists is a primary structural driver of AI medical imaging adoption worldwide. | 中 | SM001, SM013 |
| CM010 | China medical image-analysis software was valued at USD 422.8 million in 2022, growing at roughly 24.9% CAGR to 2030. | 中 | SM002 |
| CM011 | China's broader AI-in-healthcare market reached about USD 1.59 billion in 2023 and is projected to USD 7.33 billion by 2028 (~42% CAGR) and USD 18.88 billion by 2030. | 中 | SM015 |
| CM012 | A different analyst projects China's AI-in-healthcare market at about USD 2.67 billion in 2025 rising to USD 84.11 billion by 2035 at roughly 41% CAGR. | 低 | SM014 |
| CM013 | China's radiology-AI-platforms market is now tracked as a distinct, fast-growing 2026 segment, the closest public proxy to UII's served market. | 中 | SM013 |
| CM014 | China AI-imaging and healthcare-AI estimates diverge by an order of magnitude across publishers, so any single China TAM figure is unreliable without management segmentation. | 中 | SM002, SM014, SM015 |
| CM015 | North America held the largest regional share (about 44-45% in 2025) while Asia-Pacific is the fastest-growing region for medical-imaging AI. | 中 | SM001, SM006 |
| CM016 | Software and solutions, not hardware, is the dominant and fastest-growing segment of China's healthcare-AI spend. | 中 | SM003 |
| CM017 | China's government announced large AI investments such as a USD 2.1 billion Beijing AI industrial park in 2018, underpinning the healthcare-AI sector. | 中 | SM003 |
| CM018 | No reviewed public source isolates a clean UII-specific SAM or SOM separating its paid hospital revenue opportunity from adjacent pools. | 中 | SM002, SM013 |
| CM019 | The narrowest directly relevant public sizing lens for UII is China medical image-analysis software (USD 422.8 million, 2022). | 中 | SM002 |
| CM020 | Broader China healthcare-AI lenses (USD 1.59-7.33 billion) bracket UII's adjacency upside but overstate its current addressable revenue. | 中 | SM015, SM014 |
| CM021 | China's tertiary hospitals, especially Tier-3A institutions, are the primary buyers and deployment sites for AI imaging. | 中 | SM016, SM013 |
| CM022 | By 2026 more than 1,200 Tier-3A hospitals in China had adopted AI-assisted diagnostic systems for radiology and/or pathology. | 中 | SM016 |
| CM023 | The user of hospital AI imaging is the radiology or pathology department, while the payer is the hospital's capital and IT budget rather than a separate AI reimbursement line. | 中 | SM016, SM026 |
| CM024 | Budget ownership for AI imaging in China is fragmented across hospital administration, departments and provincial health bureaus driving centralized procurement. | 中 | SM016, SM013 |
| CM025 | Parent United Imaging Healthcare's imaging-equipment sales pull UII software into the same hospital accounts, giving UII a privileged distribution channel. | 中 | SM013 |
| CM026 | Chinese hospitals adopt AI imaging primarily to relieve the radiologist shortage, standardize diagnosis and meet national digital-health mandates. | 中 | SM016, SM001 |
| CM027 | China's centralized health data and large patient volumes give domestic vendors a training-data advantage over Western entrants. | 中 | SM015 |
| CM028 | The hospital AI adoption path runs from departmental pilot to departmental deployment to hospital-wide use and occasionally provincial roll-out, gated by procurement and validation. | 中 | SM013, SM016 |
| CM029 | Third-party imaging centres are a secondary channel for AI imaging emerging alongside hospital deployments in China. | 中 | SM013 |
| CM030 | Buyer budget ownership for AI software is contested because there is no dedicated reimbursement line, so spend competes with other hospital capital priorities. | 中 | SM026 |
| CM031 | China's national AI strategy and digital-health mandates are a primary policy driver accelerating hospital AI adoption. | 中 | SM014, SM003 |
| CM032 | The NMPA regulates AI diagnostic software as Class II or Class III medical devices depending on clinical risk. | 高 | SM017, SM020 |
| CM033 | NMPA's Innovative Medical Device special-review programme offers an accelerated approval pathway for pioneering AI devices. | 中 | SM022, SM018 |
| CM034 | In 2026 the NMPA issued updated AI classification guidance and an 80-plus item industry-standards plan covering AI and imaging. | 中 | SM018, SM021 |
| CM035 | NMPA Announcement No. 52 (2026) standardizes how medical devices are reclassified, for example from Class II to Class III, affecting AI device makers. | 中 | SM022 |
| CM036 | Medical and health data is classified as sensitive personal information under China's PIPL, requiring separate consent and data localization. | 高 | SM023, SM024 |
| CM037 | China's cross-border data-transfer certification measures took effect on 1 January 2026, adding compliance burden for medical-data export. | 高 | SM025, SM024 |
| CM038 | AI imaging software is generally not separately reimbursed under China's DRG hospital-payment system; its cost is absorbed into hospital budgets. | 中 | SM026 |
| CM039 | Immature AI reimbursement and high clinical-validation bars leave many cleared products stuck in pilots rather than scaled paid deployment. | 中 | SM026, SM016 |
| CM040 | Class III approval burden, procurement complexity and data-governance costs are the main constraints on China medical-AI market growth. | 中 | SM017, SM023, SM026 |
| CP001 | The China medical-imaging-AI market is densely populated, with hundreds of active vendors and dozens of well-funded contenders as of 2026. | 中 | SP013, SP021 |
| CP002 | Infervision (Tuixiang), founded 2016 in Beijing, has raised about USD 213 million over eight rounds and focuses on X-ray and CT diagnosis. | 中 | SP002 |
| CP003 | DeepWise (Shenrui), founded 2017 in Beijing, has raised about USD 115 million backed by Legend Capital and CICC, focused on cancer and stroke radiology AI. | 中 | SP003, SP004 |
| CP004 | DeepWise unveiled new multimodal AI imaging platforms at the 2026 China Medical Equipment exhibition. | 中 | SP004 |
| CP005 | Airdoc (Eaglevision, 02251.HK) became China's first AI-medical-imaging IPO on the Hong Kong exchange in November 2021, valued around RMB 5.8 billion. | 高 | SP005, SP010 |
| CP006 | Airdoc specializes in retinal AI for diabetic-retinopathy screening and reported about USD 24.1 million (RMB 173.3 million) of revenue for 2025. | 中 | SP009, SP010 |
| CP007 | Shukun Technology, founded 2017, specializes in cardiovascular and tumor AI and was preparing a Hong Kong IPO in 2026 backed by Goldman Sachs, Sequoia and Qiming. | 中 | SP022 |
| CP008 | Yitu Healthcare deploys cancer-screening and NLP diagnostic AI across more than 600 top Chinese hospitals. | 中 | SP017, SP021 |
| CP009 | Ping An Smart Healthcare reaches more than 40,000 medical institutions across 170-plus cities with an integrated finance-healthcare AI ecosystem covering 3,000-plus diseases. | 中 | SP018, SP023 |
| CP010 | Tencent Miying (AIMIS) offers cloud-based imaging AI with Class III-cleared products for pneumonia, glaucoma and colon conditions, integrated into the Tencent ecosystem. | 中 | SP016 |
| CP011 | Platform incumbents lead FDA radiology-AI authorizations - GE HealthCare 130, Siemens Healthineers 95, Philips 58, Canon 48 - embedding AI into hardware and PACS. | 中 | SP001, SP011 |
| CP012 | United Imaging ranks fifth on the FDA's Q1-2026 radiology-AI authorization list with 40 cleared devices, ahead of Aidoc's 33. | 中 | SP001 |
| CP013 | Aidoc (Israel) raised a USD 150 million Series E in April 2026 led by Goldman Sachs, bringing total funding to over USD 500 million. | 高 | SP007, SP006 |
| CP014 | Aidoc runs the aiOS enterprise platform and CARE clinical foundation model - which gained a first FDA clearance for a comprehensive foundation-model triage system - deployed across nearly 2,000 hospitals analyzing 60 million-plus cases annually. | 中 | SP007, SP015 |
| CP015 | Viz.ai (US) has raised about USD 252 million, reached a USD 1.2 billion valuation, deployed across roughly 1,600-2,000 hospitals, and leads CMS-reimbursed large-vessel-occlusion stroke detection. | 中 | SP019, SP024 |
| CP016 | Annalise.ai (a Harrison.AI subsidiary) differentiates on breadth, covering 124-plus chest X-ray findings and 130-plus CT-brain findings in single tools. | 中 | SP020, SP025 |
| CP017 | Capital is shifting toward enterprise-infrastructure AI vendors with scale, regulatory depth and workflow ownership rather than single-use modular tools. | 中 | SP006 |
| CP018 | In the US, CMS reimbursement - including new 2026 billing codes for coronary and aortic calcium analysis - is a major commercial driver absent in China. | 中 | SP006 |
| CP019 | UII's differentiation is vertical integration - AI embedded in UIH scanners and PACS plus a broad uAI portfolio - versus pure-play cloud platforms. | 中 | SP001, SP021 |
| CP020 | UII's relative limitation versus Aidoc and Viz.ai is a less open, less flexible enterprise-wide software-integration story outside its own hardware. | 中 | SP012, SP021 |
| CP021 | Pricing in China medical-AI is opaque; most vendors sell via hospital procurement or bundle AI with equipment rather than transparent per-study pricing. | 中 | SP013, SP021 |
| CP022 | US pure-play vendors increasingly use SaaS subscription and ARR models supported by reimbursement, a monetization path largely unavailable in China. | 中 | SP019, SP006 |
| CP023 | UII's strongest moat is the parent UIH hardware channel that pulls its AI into the same hospital accounts that buy United Imaging scanners. | 中 | SP001, SP021 |
| CP024 | UII's channel moat is partly neutralized because Chinese hospitals multi-home, running several vendors' AI side by side. | 中 | SP021, SP013 |
| CP025 | Regulatory clearance breadth (NMPA Class III, FDA, CE) is a real but replicable moat, as peers like Infervision, DeepWise and Tencent also hold Class III clearances. | 中 | SP001, SP016 |
| CP026 | Foundation-model scale is becoming the new competitive frontier, with Aidoc's CARE and UII's uAI NEXUS both pursuing multi-condition models. | 中 | SP007, SP015 |
| CP027 | Better-capitalized global pure-plays out-fund UII on foundation-model R&D, with Aidoc alone raising over USD 500 million versus UII's roughly USD 139 million Series A. | 中 | SP006, SP007 |
| CP028 | The Chinese market's density of hundreds of vendors pressures pricing and accelerates commoditization of single-disease detectors. | 中 | SP013, SP021 |
| CP029 | Switching costs rise once AI is embedded in scanner workflows and hospital PACS, favoring integrated vendors like UII and the platform incumbents. | 中 | SP011, SP001 |
| CP030 | Distribution power favors incumbents (GE, Siemens, Philips, UIH) that sell hardware, giving them a structural channel pure-play startups lack. | 中 | SP001, SP011 |
| CP031 | Internal hospital builds and open-source medical models are a growing status-quo substitute, especially at research hospitals. | 中 | SP021 |
| CP032 | Airdoc's and Shukun's IPO paths show Chinese peers seeking public capital, intensifying competition for the same hospital budgets. | 中 | SP005, SP022 |
| CP033 | Viz.ai's profitability and CMS reimbursement give US peers a commercialization edge UII cannot yet match in China. | 中 | SP019, SP006 |
| CP034 | Annalise's and Aidoc's breadth strategies converge with UII's multi-disease agent roadmap, so differentiation on breadth alone is unlikely to be durable. | 中 | SP016, SP007 |
| CP035 | UII's clearest durable advantage is the combination of domestic data scale, parent channel and a China-home-market regulatory position. | 中 | SP019, SP021 |
| CP036 | The competitive set spans direct Chinese peers, global pure-plays, platform incumbents, ecosystem giants (Ping An, Tencent), substitutes and internal builds. | 中 | SP013, SP021 |
| CP037 | Likely future entrants and escalating threats include well-capitalized global pure-plays expanding into Asia and Chinese cloud giants deepening imaging AI. | 中 | SP007, SP016 |
| CP038 | No public source shows UII winning competitive head-to-head displacement data against named peers. | 中 | SP012, SP021 |
| CP039 | UII's roughly USD 1.4 billion valuation sits above listed Airdoc (about RMB 5.8 billion at IPO) but below the global leaders' funding base, implying a premium competition must justify. | 中 | SP005, SP010 |
| CP040 | Differentiation durability ultimately depends on UII converting its hardware-channel and data moats into reimbursed, multi-homing-resistant recurring revenue. | 中 | SP021, SP006 |
| CI001 | United Imaging Intelligence is the independently financed medical-AI subsidiary of listed United Imaging Healthcare (UIH, 688271.SH), monetizing AI software rather than imaging hardware. | 中 | SI005, SI015 |
| CI002 | As a private subsidiary, UII discloses no standalone revenue, ARR, or net income. | 中 | SI005, SI017, SI020 |
| CI003 | UII monetizes its AI through two channels - applications embedded in or bundled with parent UIH scanners and PACS, and standalone uAI software licenses sold to hospitals. | 中 | SI015, SI005 |
| CI004 | UII publishes no per-study or per-seat list pricing; Chinese imaging-AI is sold via hospital procurement or equipment bundling. | 中 | SI015, SI021 |
| CI005 | UII reports 12 product platforms and more than 100 AI applications deployed across over 4,000 medical institutions. | 中 | SI005, SI015, SI016 |
| CI006 | UII holds 13 NMPA Class III certificates, 15 FDA clearances, and 31 CE certifications as of mid-2025. | 中 | SI005, SI016 |
| CI007 | UII ranks fifth on the FDA Q1-2026 radiology-AI authorization list with about 40 cleared devices, signaling sustained product and R&D investment. | 中 | SI023 |
| CI008 | UII's primary go-to-market motion is the parent UIH hardware channel, which pulls AI into hospital equipment accounts. | 中 | SI015, SI005 |
| CI009 | UII does not publicly disclose its hospital sales cycle, contract terms, or revenue concentration. | 低 | SI005, SI020 |
| CI010 | No public CAC, payback, or customer-acquisition-efficiency metric exists for UII. | 低 | SI020, SI017 |
| CI011 | The UIH channel gives UII distribution economics few software-only rivals match, but bundling obscures standalone AI revenue attribution. | 中 | SI015, SI005 |
| CI012 | Parent UIH reported 2025 revenue of RMB 13.80 billion, up 33.98% year on year. | 高 | SI002, SI004, SI022 |
| CI013 | Parent UIH's 2025 net profit attributable to shareholders was RMB 1.87 billion, up 48.14% year on year. | 高 | SI002, SI022 |
| CI014 | Parent UIH 2025 R&D investment reached RMB 2.62 billion, up 15.95% year on year. | 中 | SI002, SI004 |
| CI015 | Parent UIH H1 2025 revenue was RMB 6.02 billion (up 12.79%) with net profit attributable to the parent of RMB 998 million. | 高 | SI001, SI003 |
| CI016 | UIH H1 2025 overseas revenue reached RMB 1.14 billion (18.99% of revenue) and service revenue RMB 0.82 billion (13.56%). | 中 | SI003, SI001 |
| CI017 | UIH H1 2025 R&D investment was RMB 1.14 billion, up 12.09% year on year. | 中 | SI003 |
| CI018 | UIH Q1 2026 revenue was RMB 2.91 billion (up 17.34%) with net profit of RMB 0.40 billion (up 7.78%), a slower profit trajectory than full-year 2025. | 中 | SI002, SI004 |
| CI019 | China's volume-based procurement (VBP) imposes downward price pressure on domestic imaging equipment, compressing parent margins absent a premium and overseas mix. | 中 | SI004, SI022 |
| CI020 | UII's standalone gross margin is not knowable from public data because results are consolidated within parent UIH and not separately reported. | 中 | SI001, SI020 |
| CI021 | UIH's high R&D intensity of roughly 19% of revenue reflects the capital intensity of the AI and imaging platform UII contributes to. | 中 | SI002 |
| CI022 | UII closed a Series A of about RMB 1 billion (approximately USD 139 million) in June 2025. | 高 | SI011, SI012, SI014 |
| CI023 | The UII Series A was co-led by E Fund Management and Shanghai Guotou (SIG), with Shanghai United Investment and others participating. | 中 | SI011, SI005 |
| CI024 | The round implied a pre-money valuation of about RMB 9 billion (~USD 1.3 billion), roughly a USD 1.4 billion post-money mark. | 中 | SI011, SI017 |
| CI025 | Planned use of UII Series A proceeds covers large medical-AI model R&D, multimodal medical agents spanning imaging, text and voice, and hospital digital business systems. | 中 | SI005, SI013 |
| CI026 | UII is reportedly considering an independent listing or IPO. | 中 | SI005 |
| CI027 | UII's cash on hand, monthly burn, and runway are not disclosed. | 低 | SI005, SI020 |
| CI028 | No public debt or project-finance obligations are disclosed for UII. | 低 | SI017, SI020 |
| CI029 | UII's roughly USD 139 million Series A is small relative to global pure-play peers such as Aidoc (more than USD 500 million raised), constraining foundation-model R&D firepower. | 中 | SI011, SI023 |
| CI030 | China's 2026 venture funding rebound - first-quarter Asia funding of about USD 27.4 billion, most flowing to China and led by AI - supports follow-on capital availability for UII. | 中 | SI010 |
| CI031 | Listed peer Airdoc remained loss-making, reporting a CN¥255 million net loss for FY2024. | 中 | SI007, SI008 |
| CI032 | Analysts project Airdoc to breakeven only around 2026, requiring roughly 102% annual revenue growth - illustrating Chinese medical-AI monetization difficulty. | 中 | SI007 |
| CI033 | Airdoc's 2025 annual report warned that ultimate commercialization remains at risk despite NMPA Class III approval, due to slow hospital adoption and uncertain budgets. | 中 | SI006 |
| CI034 | Airdoc has posted net losses for five consecutive years despite rising revenue, a cautionary comparable for UII's revenue quality. | 中 | SI008, SI009 |
| CI035 | AI diagnostic software is generally not reimbursed under China's health-insurance or DRG system, forcing direct hospital deals at modest fees. | 中 | SI021, SI025 |
| CI036 | Ping An's integrated finance-healthcare model shows profitability is achievable mainly via ecosystem cross-sell rather than standalone imaging-AI fees. | 中 | SI024, SI026 |
| CI037 | UII's strongest public financial signal is the USD 139 million Series A at a roughly USD 1.4 billion valuation plus the parent's audited growth, not any standalone UII statements. | 中 | SI011, SI002 |
| CI038 | The primary financial diligence blockers are undisclosed standalone revenue, gross margin, burn, runway, and retention. | 中 | SI020, SI005 |
| CI039 | Financial dependency on parent UIH - channel, shared R&D, and consolidated reporting - is both UII's biggest asset and a key concentration risk. | 中 | SI015, SI005 |
| CI040 | Public evidence supports high confidence in capital access and parent strength, but low confidence in standalone unit economics and solvency. | 中 | SI001, SI017 |
| CE001 | UII delivers a unified medical-AI platform (uAI) spanning imaging analysis, diagnosis, structured reporting, treatment planning, and hospital operations, embedded in the clinical workflow. | 中 | SE014, SE013 |
| CE002 | The uAI portfolio comprises 12 product platforms and more than 100 AI applications across CT, MR, X-ray, ultrasound, and pathology. | 中 | SE014, SE016 |
| CE003 | UII has shifted from narrow task-AI to collaborative AI agents spanning the care pathway, including uAI Insight, the uAI Agent for Ultrasound, uAI Avatar, and a hospital-management agent. | 中 | SE009, SE018 |
| CE004 | uAI Insight is an imaging-to-report AI agent performing multimodal reasoning and clinical correlation. | 中 | SE009, SE011 |
| CE005 | The uAI Agent for Chest CT Reporting detects 73 thoracic diseases from a single scan at about 95% AUC, and the Brain MRI agent identifies 47 neurological conditions. | 中 | SE018, SE011 |
| CE006 | The uAI Clinical Portal (uCP) / uAI Discover integrates with hospital PACS and supports more than 60 AI applications across imaging modalities. | 中 | SE013 |
| CE007 | UII's products are powered by multimodal medical large models integrating clinical language, vision, and speech (uAI PLUS and uAI NEXUS). | 中 | SE010, SE001 |
| CE008 | uAI PLUS, debuted at RSNA 2024, is a foundation platform combining imaging foundation models, medical LLMs, and multimodal LLMs. | 中 | SE010 |
| CE009 | UII open-sourced uAI NEXUS MedVLM, described as the first open-source medical video large language model, accepted to CVPR 2026. | 高 | SE001, SE003 |
| CE010 | MedVLM is built on MedVidBench, a benchmark of 531,850 video-instruction pairs across eight clinical sources. | 高 | SE001, SE003 |
| CE011 | MedVLM is trained via supervised fine-tuning plus MedGRPO, a reinforcement-learning framework using cross-dataset reward normalization and a medical-LLM judge. | 中 | SE003, SE002 |
| CE012 | Despite using only 4B and 7B parameters, MedVLM outperforms GPT-5.4 and Gemini 3.1 on medical-video tasks, reaching 89.4% surgical-safety accuracy versus 1.8% for GPT-5.4. | 高 | SE001, SE003 |
| CE013 | UII released open code on GitHub (UII-AI/MedGRPO-Code), model weights, the MedVidBench dataset, and a public leaderboard. | 高 | SE002, SE003, SE004, SE005 |
| CE014 | The MedVLM research was produced by a United Imaging research team (Terrence Chen, Ziyan Wu, and colleagues) and published on arXiv. | 中 | SE003 |
| CE015 | uAI products integrate DICOM ingestion, AI analysis, structured reporting, and hospital-IT/PACS/RIS connection rather than operating as isolated point tools. | 中 | SE013, SE007 |
| CE016 | UII's AI runs across more than 4,000 medical institutions in China and is deployed in Europe, Asia, North America, and South America. | 中 | SE013, SE014 |
| CE017 | UII demonstrated next-generation AI agents at RSNA 2025 under the theme "Agent Radiologist - Are We There Yet?" | 中 | SE009, SE011 |
| CE018 | At ECR 2026 UII presented validation and scaling of radiology AI, citing more than 60 applications and over 30 CE certifications. | 中 | SE013 |
| CE019 | A uAI Agent for Ultrasound debuted as an embodied robotic arm performing autonomous scanning with integrated visual and tactile cues. | 中 | SE009, SE011 |
| CE020 | UII collaborates with leading centers such as Massachusetts General Hospital for multi-center, multi-vendor validation. | 中 | SE013 |
| CE021 | UII's latest 2026 product milestone is the April 2026 open-source MedVLM release together with the ECR 2026 scaling narrative. | 中 | SE001, SE012 |
| CE022 | uAI Easy Triage ICH (K242292) received FDA 510(k) Class II clearance on September 24, 2024 for triage of suspected intracranial hemorrhage on non-contrast head CT. | 高 | SE006, SE007 |
| CE023 | uAI Portal (K240411) is FDA Class II cleared for AI-based segmentation of vascular CT including aorta, coronary, pulmonary, head/neck, and lower-extremity vessels. | 中 | SE008, SE007 |
| CE024 | uAI EasyTriage-Rib (2021) is FDA-cleared for acute rib-fracture triage on chest CT, and DeepRecon provides deep-learning MR denoising on uMR systems. | 中 | SE007 |
| CE025 | UII holds 13 NMPA Class III certificates for high-risk diagnostic AI across CT, MR, and ultrasound. | 中 | SE016, SE020 |
| CE026 | UII ranks fifth on the FDA Q1-2026 radiology-AI authorization list with about 40 cleared devices. | 中 | SE019 |
| CE027 | Across regimes UII reports about 15 FDA clearances, 31 CE certifications, and 13 NMPA Class III approvals as of mid-2025. | 中 | SE014, SE016 |
| CE028 | UII's strongest technical differentiation is open-source foundation-model leadership in medical video understanding, validated by CVPR 2026 acceptance. | 中 | SE001, SE003 |
| CE029 | UII's open benchmark, dataset, and leaderboard invite the global developer community, a transparency posture rare among Chinese medical-AI vendors. | 中 | SE002, SE012 |
| CE030 | Vertical integration with parent UIH lets UII co-design AI with scanners (e.g., DeepRecon on uMR), a hardware-software loop software-only rivals lack. | 中 | SE007, SE014 |
| CE031 | UII's large, diverse, clinically sourced training data, refined through hospital collaboration, is a core data moat. | 中 | SE010, SE013 |
| CE032 | Regulatory breadth across FDA, NMPA, and CE is itself a differentiation asset versus narrower competitors. | 中 | SE014, SE019 |
| CE033 | FDA 510(k) clearances, supported by regulatory partner MCRA, evidence quality-system maturity for US market entry. | 中 | SE015, SE006 |
| CE034 | NMPA Class III is China's highest-risk device class, requiring rigorous clinical evaluation. | 中 | SE020, SE021 |
| CE035 | UII provides open benchmarking via MedVidBench and a public leaderboard for transparent third-party evaluation. | 中 | SE005, SE003 |
| CE036 | Independent third-party clinical-validation evidence for UII's newest autonomous agents is still limited in the public record. | 中 | SE011, SE013 |
| CE037 | UII discloses little public detail on data-privacy architecture, security controls, and PIPL-aligned data handling for its platform. | 低 | SE013, SE007 |
| CE038 | The commercialization path from open-source foundation models to paid, governed hospital products is not yet fully detailed publicly. | 中 | SE001, SE013 |
| CE039 | Open benchmarking currently centers on the video model; the breadth of independent benchmarking across the full clinical portfolio is unclear. | 中 | SE003, SE005 |
| CE040 | Key product and technical risks include foundation-model generalization to real clinical settings, multi-vendor integration complexity, and differentiation tied to parent hardware. | 中 | SE011, SE007 |
| CU001 | UII's primary buyers are hospital radiology departments and IT, its users are radiologists and clinicians, and the payer is the hospital because China lacks a national reimbursement line for standalone diagnostic-AI software. | 中 | SU019, SU014 |
| CU002 | UII reports 12 platforms and more than 100 AI applications deployed across over 4,000 medical institutions. | 高 | SU019, SU017 |
| CU003 | UII's core market is China's public hospital system, particularly large Tier-3 and Tier-3A teaching hospitals. | 中 | SU002, SU014 |
| CU004 | UII reaches customers through two channels - AI embedded natively in UIH scanners and PACS, and standalone uAI Clinical Portal software licensed to hospitals. | 中 | SU020, SU012 |
| CU005 | Customers adopt uAI for triage, lesion detection, segmentation, structured radiology reporting, and AI agents for chest CT, brain MRI, and ultrasound. | 中 | SU012, SU013 |
| CU006 | China's national health system pushed AI-assisted diagnostics into more than 1,200 Tier-3A hospitals during 2025-2026, the pool UII targets. | 中 | SU014, SU002 |
| CU007 | The count of paying uAI software customers cannot be separated from hospitals that own a UIH scanner with embedded AI from public disclosures. | 中 | SU019, SU015 |
| CU008 | UII's overseas customer base rides the parent UIH channel into Europe, North America, Asia, and South America. | 中 | SU005, SU011 |
| CU009 | Parent UIH's installed base spans more than 100 countries and regions with over 640 cumulative United States imaging systems by the end of 2025, carrying embedded uAI. | 高 | SU015, SU016 |
| CU010 | The uAI Clinical Portal unifies more than 60 applications in a single PACS-integrated workflow surface. | 中 | SU012, SU011 |
| CU011 | UII reports roughly 13 NMPA Class III approvals, 15 FDA clearances, and 31 CE certifications as of mid-2025, with 30 CE-certified applications featured at ECR 2026. | 中 | SU018, SU011 |
| CU012 | Third-party coverage of Chinese hospital AI describes radiology reporting workflow improvements on the order of 30 percent. | 低 | SU014, SU001 |
| CU013 | UII presented multicenter validation for its CE-certified applications across Europe, Asia, North America, and South America at ECR 2026. | 中 | SU011, SU008 |
| CU014 | UIH's broader installed base sits beneath the uAI institution count, and named publicly referenceable hospital deployments are a smaller subset still. | 中 | SU015, SU004 |
| CU015 | UII discloses no per-application utilization, active-use rate, repeat-purchase count, or denominator to convert 4,000 institutions into seats, studies read, or recurring revenue. | 中 | SU019, SU017 |
| CU016 | The newest AI agents shown at RSNA 2025 and ECR 2026 remain limited in independent, production-grade validation. | 中 | SU013, SU011 |
| CU017 | UII's adoption trajectory is rising and broad across 2025-2026, evidenced by expanding clearances and event showcases. | 中 | SU007, SU020 |
| CU018 | Native-AI products such as uMR Ultra and uCT SiriuX integrate real-time AI into clinical acquisition workflows. | 中 | SU020, SU006 |
| CU019 | At ECR 2026 UII featured 30 CE-certified medical AI applications validated in multicenter clinical settings. | 高 | SU011, SU012 |
| CU020 | United Imaging installations are named in Italy (including a uCT 960+), France, Germany, and Lithuania. | 中 | SU005, SU007 |
| CU021 | UC Davis Health lists a United Imaging 1.5T uMR 680 MRI system in its production radiology equipment fleet, an independent customer-side reference. | 高 | SU004, SU005 |
| CU022 | Flagship uMR Omega 3T ultra-wide-bore systems have been placed at US imaging providers such as X-Ray Associates of New Mexico. | 中 | SU010, SU006 |
| CU023 | Leading Chinese academic centers including Fudan University's Zhongshan Hospital in Shanghai are referenced as uAI users. | 低 | SU002, SU014 |
| CU024 | Cleared imaging applications rate high on production maturity and corroboration, while the agentic tier rates low on independent retention and outcome visibility. | 中 | SU011, SU013 |
| CU025 | Most named proof attaches to UIH hardware accounts and rarely isolates which sites pay for uAI software as a distinct line. | 中 | SU004, SU015 |
| CU026 | The newest autonomous agents are demonstrated and validated in research settings but lack independent production-grade outcome evidence at named sites. | 中 | SU013, SU001 |
| CU027 | ECR 2026 multicenter validation spanned research and clinical sites across four continents for the CE-certified application set. | 中 | SU011, SU008 |
| CU028 | No public disclosure reports any UII deployment that was blocked, churned, or publicly criticized on clinical grounds, though US-market trust skepticism is documented. | 中 | SU003, SU016 |
| CU029 | uAI is frequently embedded natively in UIH scanners and integrated into hospital PACS/RIS, raising switching costs because replacing the AI disrupts the imaging workflow or hardware relationship. | 中 | SU020, SU012 |
| CU030 | The uAI Clinical Portal's consolidation of 60-plus applications accrues training and integration lock-in for hospitals that standardize on it. | 中 | SU012, SU011 |
| CU031 | UII publishes no net revenue retention, gross retention, churn, renewal rate, contract length, or customer-satisfaction score, and no public cohort showing repeat usage over time. | 中 | SU017, SU019 |
| CU032 | Because most adoption is bundled with hardware, even the parent's filings do not expose AI-software renewal behavior. | 中 | SU022, SU015 |
| CU033 | Loss-making listed peers such as Airdoc show that broad clinical adoption does not guarantee durable, profitable revenue. | 中 | SU021, SU024 |
| CU034 | Every UII retention claim must currently be treated as a null with an explicit diligence ask for NRR/GRR, renewal cohorts, contract terms, and churn reasons. | 中 | SU019, SU017 |
| CU035 | UII runs a land-and-expand loop: a hospital buys a UIH scanner with embedded AI, adopts the uAI Clinical Portal for more cleared applications, then expands into agents and foundation models. | 中 | SU020, SU012 |
| CU036 | Overseas customer growth follows UIH's international build-out, with parent overseas revenue rising more than 50 percent in 2025. | 中 | SU015, SU025 |
| CU037 | UII's distribution, brand, and much of its installed base depend on parent UIH, so the customer relationship is only partly UII's own. | 中 | SU015, SU023 |
| CU038 | UII carries geographic and segment concentration in China's Tier-3 hospital market, where value-based procurement compresses prices and AI software lacks a national reimbursement line. | 中 | SU014, SU002 |
| CU039 | US and European expansion faces documented trust, supply-chain-security, data-privacy, and US-China trade-tension headwinds that can slow or block procurement at academic accounts. | 中 | SU003, SU016 |
| CU040 | No public disclosure quantifies UII's top-customer revenue concentration, leaving single-account and single-channel dependence a material open diligence item. | 中 | SU017, SU019 |
| CR001 | UII's risk surface is shaped by the policy and structural environment around a Chinese medical-AI company tethered to a listed imaging parent rather than by start-up fragility. | 中 | SR018, SR023 |
| CR002 | The top risk tier is external and largely outside management control - US national-security restrictions, NMPA approval complexity, and PIPL/DSL data rules. | 中 | SR006, SR001 |
| CR003 | The second risk tier is structural dependency on parent UIH for distribution, capital, brand, and installed base. | 中 | SR018, SR012 |
| CR004 | Severity here blends likelihood and impact with mitigation maturity, with regulatory and geopolitical risks scoring high impact and rising likelihood against low company-side mitigation. | 中 | SR006, SR002 |
| CR005 | Operational and technical risks - clinical-AI safety, model generalization, recalls, and cybersecurity - form the third tier with partial mitigation from clearances and quality systems. | 中 | SR021, SR020 |
| CR006 | Financial-model and execution risk is muted by parent strength but elevated by undisclosed standalone economics. | 中 | SR018, SR016 |
| CR007 | UII's principal threats are exogenous and slow-moving rather than acute, which makes them monitorable against measurable thesis-break triggers. | 中 | SR006, SR001 |
| CR008 | Unconfirmed items - litigation exposure, security incidents, and revenue concentration - are flagged with explicit diligence paths rather than assumed benign. | 中 | SR014, SR021 |
| CR009 | China's NMPA classifies high-risk diagnostic and decision-support AI as Class III, the highest-risk tier requiring mandatory clinical evaluation with locally representative data. | 高 | SR002, SR003 |
| CR010 | NMPA rules require predefined change-control for evolving or self-modifying algorithms, so each model update is a fresh, evidence-heavy regulatory project. | 中 | SR005, SR001 |
| CR011 | China's PIPL, Data Security Law, and Cybersecurity Law impose health-data localization and CAC security assessments for cross-border transfers of patient data. | 中 | SR001, SR020 |
| CR012 | The BIOSECURE Act became law on 18 December 2025 as Section 851 of the FY2026 NDAA, restricting federal agencies and federally linked contractors from procuring biotechnology equipment or services from designated companies of concern. | 高 | SR006, SR007 |
| CR013 | The OMB list of biotechnology companies of concern is due by December 2026 and DoD Section 1260H Chinese-military-company designations are already expanding in 2026. | 中 | SR010, SR009 |
| CR014 | Even before a specific UII or UIH designation, the BIOSECURE chilling effect plus FDA supply-chain and cybersecurity scrutiny can slow procurement at federally funded US hospitals. | 中 | SR008, SR011 |
| CR015 | BIOSECURE's definition of biotechnology equipment is broad and can encompass medical imaging devices, instruments, software, and data-analytics services. | 中 | SR007, SR008 |
| CR016 | No patent, trade-secret, or other litigation naming UII is found in the public record as of the 2026-06-23 run date. | 中 | SR013, SR015 |
| CR017 | Medtech is highly litigious and NPE-driven patent activity in radiology AI is rising, so the absence of disclosed UII cases is a diligence gap rather than a clean bill of health. | 中 | SR014, SR013 |
| CR018 | NMPA 2025-2026 standards programs and high-end-device guidance keep tightening and internationally aligning requirements while preserving China-specific clinical and cybersecurity demands. | 中 | SR004, SR002 |
| CR019 | Because uAI informs triage, detection, segmentation, and reporting, a false negative or failure to generalize to new scanners, protocols, or patient mix can cause clinical harm and regulatory action. | 中 | SR021, SR002 |
| CR020 | NMPA change-control rules mean fixing or updating a deployed model is a governed, evidence-heavy process rather than a quiet patch. | 中 | SR005, SR003 |
| CR021 | The newest autonomous agents shown at RSNA 2025 and ECR 2026 widen the gap between benchmark performance and accountable real-world behavior, with independent production-grade validation still limited. | 中 | SR028, SR021 |
| CR022 | Systems integrated into hospital PACS/RIS and native to UIH scanners create an attack surface and a PIPL data-protection obligation. | 中 | SR030, SR001 |
| CR023 | UII discloses little public detail on its security architecture, controls, or incident history, so resilience must be verified rather than assumed. | 中 | SR021, SR022 |
| CR024 | Supply-chain and manufacturing risk sits mostly with the parent UIH hardware business but transmits to UII through bundled deployments and any component or export restriction. | 中 | SR012, SR018 |
| CR025 | A confirmed clinical incident or recall tied to a uAI product would be a clear thesis-break trigger for the operational-risk thesis. | 中 | SR021, SR020 |
| CR026 | UII's quality systems and clearance breadth (13 NMPA Class III, 15 FDA, 31 CE) are real operational mitigants but do not substitute for undisclosed utilization-level reliability data. | 中 | SR021, SR022 |
| CR027 | UII depends on parent UIH for the hardware channel carrying embedded uAI, much of the brand and global service footprint, a large share of installed base, and capital. | 中 | SR018, SR012 |
| CR028 | Any shock to UIH - a 1260H listing, export restriction, or demand slump under value-based procurement - transmits directly to UII. | 中 | SR010, SR012 |
| CR029 | UII's market access is gated by regulators (NMPA, FDA, EU notified bodies), making each new application dependent on renewed approval. | 中 | SR002, SR022 |
| CR030 | Value is delivered through the single uAI Clinical Portal platform and a small set of foundation models, creating platform-displacement risk within accounts. | 中 | SR028, SR021 |
| CR031 | State-linked ownership and investors add political-alignment risk that cuts both ways - domestic preference in China and heightened suspicion in the US. | 中 | SR024, SR011 |
| CR032 | UII depends on capital providers for the next financing round to fund AI-agent build-out, partly cushioned by a 2026 China AI funding rebound. | 中 | SR029, SR024 |
| CR033 | Hospital procurement budgets pressured by China's value-based procurement are a customer-side dependency that compresses pricing. | 中 | SR020, SR016 |
| CR034 | UII discloses no top-account revenue split, so single-account or single-channel customer concentration is a material open item. | 中 | SR023, SR021 |
| CR035 | The June 2025 Series A of roughly USD 139 million at about a USD 1.4 billion valuation plus the listed parent make acute near-term solvency unlikely. | 中 | SR024, SR018 |
| CR036 | The financial risk is durability rather than survival - no standalone revenue or margin disclosure, no national AI-software reimbursement, and VBP price compression. | 中 | SR020, SR016 |
| CR037 | Loss-making listed peers such as Airdoc - net losses for five straight years - show that broad clinical adoption need not convert into profitable, durable revenue. | 中 | SR016, SR017 |
| CR038 | Execution and people risk centers on retaining scarce AI and clinical-regulatory talent, scaling pre-revenue AI agents, and managing the open-source-to-paid-product bridge. | 中 | SR021, SR028 |
| CR039 | Each top risk maps to a monitorable trigger and thesis-break, including a binding BIOSECURE/1260H designation, an NMPA enforcement action, a safety incident, parent deterioration, a down-round, or single-customer concentration. | 中 | SR006, SR010 |
| CR040 | Risks transmit into customers, revenue, margin, financing, and valuation, and where evidence is missing the chapter records exact diligence asks rather than assigning comfort by intuition. | 中 | SR018, SR020 |
| CV001 | United Imaging Intelligence's most recent public valuation anchor is its June 2025 Series A, which set a post-money valuation of roughly 1.4 billion US dollars. | 高 | SV002, SV003 |
| CV002 | The Series A valuation was set on a company about seven years old, consistent with a founding around 2017-2018. | 中 | SV002 |
| CV003 | United Imaging Intelligence raised approximately 139 million US dollars (about 1 billion yuan) in its June 2025 Series A, co-led by E Fund and Shanghai state-linked investors. | 高 | SV002, SV005, SV024 |
| CV004 | The Series A round was backed by E Fund Management, Shangguo / Shanghai Guotou investment, and Shanghai Alliance / United Investment alongside the listed parent. | 中 | SV005, SV024 |
| CV005 | United Imaging Intelligence reports deployment of AI tools across more than 4,000 institutions, supporting its platform-scale thesis. | 中 | SV004, SV026 |
| CV006 | United Imaging Intelligence holds roughly thirteen NMPA Class III clearances plus additional CE and FDA clearances across its imaging-AI portfolio. | 中 | SV006, SV026 |
| CV007 | United Imaging Intelligence positions itself as a platform spanning multiple imaging modalities and clinical applications rather than a single-product vendor. | 中 | SV007, SV029 |
| CV008 | The company is the dedicated AI subsidiary of United Imaging Healthcare, giving it distribution access to the parent's installed equipment base. | 中 | SV007, SV030 |
| CV009 | United Imaging Intelligence does not publicly disclose standalone revenue, gross margin, or customer-retention metrics. | 中 | SV003, SV006 |
| CV010 | Parent United Imaging Healthcare reported 2025 revenue of about 13.8 billion yuan, up roughly 34 percent, with net profit near 1.87 billion yuan. | 高 | SV001, SV021 |
| CV011 | HealthTech M&A valuations entering 2026 mostly cleared around mid-single-digit revenue multiples with low-teens EBITDA multiples. | 中 | SV009 |
| CV012 | Premium AI and data platforms can stretch above mid-single-digit revenue multiples, with the strongest names cited near ten-to-fifty-times forward revenue. | 中 | SV008, SV010 |
| CV013 | Recent Hong Kong Chapter 18A medical and biotech listings have priced near fifteen-to-thirty-times forward revenue when near-term revenue is projectable. | 中 | SV015, SV016 |
| CV014 | Hong Kong reclaimed the top global IPO ranking in 2025, signaling a deeper exit market for China-linked biotech and medical-AI issuers. | 中 | SV016 |
| CV015 | Because United Imaging Intelligence discloses no standalone revenue, the revenue multiple implied by its 1.4-billion-dollar mark cannot be calculated from public evidence. | 中 | SV003, SV006 |
| CV016 | US-listed Tempus AI traded near 7.1 times enterprise value to revenue in 2026 at roughly a 10-billion-dollar enterprise value. | 中 | SV011 |
| CV017 | Viz.ai was valued near 1.2 billion US dollars at its 2022 Series D and was reported to have reached profitability by 2025. | 中 | SV017 |
| CV018 | Listed Chinese imaging-AI peer Airdoc traded near 1.8 times enterprise value to revenue in 2026 while remaining loss-making. | 中 | SV012, SV019 |
| CV019 | Airdoc's reported market-capitalization-to-revenue multiple of about 5.4 times far exceeds its 1.8 times enterprise-value multiple, reflecting limited net cash and persistent losses. | 中 | SV012 |
| CV020 | Infervision is estimated near a 215-million-dollar valuation on about 12 million dollars of revenue, an outlier implied multiple likely reflecting data uncertainty. | 低 | SV012 |
| CV021 | Parent United Imaging Healthcare carried a market capitalization near 85 billion yuan in mid-2026 at roughly six times sales. | 中 | SV014, SV021, SV022 |
| CV022 | The parent's hardware-led valuation provides a scaled but imperfect anchor for the AI subsidiary, which lacks a pure-software comparable. | 中 | SV021, SV022 |
| CV023 | Broad Chinese AI peer Yitu Technology has been referenced near a 2.4-billion-dollar valuation, illustrating China platform-AI marks. | 低 | SV013 |
| CV024 | Healthcare AI is among the fastest-growing enterprise-software verticals, expanding from about 15 billion dollars in 2024 toward projections above 100 billion by 2030. | 中 | SV013 |
| CV025 | China's medical-AI market is supported by policy and hospital digitization, framing UII's domestic upside scenarios. | 中 | SV025, SV028 |
| CV026 | United Imaging Intelligence's export and international validation activity, including ECR 2026 presence, supports the bull-case export leg. | 中 | SV004, SV026 |
| CV027 | The Series A amount and lead investors are corroborated across multiple independent databases and news outlets. | 中 | SV027, SV002 |
| CV028 | Listed pure-play imaging-AI and imaging-software peers internationally tend to trade in roughly low single-digit to mid revenue multiples. | 中 | SV012, SV011 |
| CV029 | No single public comparable fully matches United Imaging Intelligence's parent-attached, multi-modality imaging-AI position. | 中 | SV011, SV012 |
| CV030 | Median AI software multiples vary widely by category, with infrastructure and platform names commanding the highest forward-revenue multiples. | 中 | SV008, SV010 |
| CV031 | At a 1.4-billion-dollar mark with an undisclosed denominator, the valuation already embeds successful AI monetization that Chinese imaging-AI peers have not demonstrated. | 中 | SV012, SV018 |
| CV032 | A disciplined recommendation on United Imaging Intelligence is research-more with medium confidence and a high risk rating. | 中 | SV003, SV012 |
| CV033 | UII could plausibly justify a double-digit revenue multiple if standalone revenue is modest but fast-growing with software margins, placing the mark inside the premium private band. | 低 | SV008, SV015 |
| CV034 | Persistent losses at listed Chinese imaging-AI peers signal weak hospital AI willingness-to-pay, a key downside transmission for UII's valuation. | 中 | SV018, SV019 |
| CV035 | Geopolitical export restrictions such as BIOSECURE-style measures and FDA scrutiny could remove the international leg of UII's bull case. | 中 | SV020 |
| CV036 | Heavy dependence on parent United Imaging Healthcare for distribution, IP, and demand reduces the durability of an independent standalone valuation. | 中 | SV008, SV030 |
| CV037 | Required diligence includes standalone revenue, margin, retention, related-party terms, and geopolitical export exposure before underwriting the mark. | 中 | SV003, SV006 |
| CV038 | On public evidence, the bear case implies equity value of roughly 0.15 to 0.55 billion dollars while the bull case reaches 2.0 to 4.2 billion, straddling the current mark. | 中 | SV011, SV012 |
| CV039 | Investment-KPI scoring is strong on backing and deployment but weak on standalone economic visibility and current price support. | 中 | SV004, SV012 |
| CV040 | The valuation call is highly sensitive to the standalone revenue figure, which the company has not disclosed. | 中 | SV003, SV011 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | United Imaging Intelligence | 联影智能官方网站 - UII official website | |
| SO002 | United Imaging Intelligence | UII Brochure - AI Solutions for Healthcare | Shanghai United Imaging Intelligence Co., Ltd. Address 701 Yunjin Road, Xuhui District, Shanghai 200030. |
| SO003 | G-MedTech | United Imaging Intelligence - Players | United Imaging Intelligence (UII), headquartered in Shanghai, is a subsidiary of United Imaging Healthcare Technology Group Co., Ltd. ... Since its founding in 2017. |
| SO004 | Health AI Register | United Imaging Intelligence - Health AI Register | |
| SO005 | CB Insights | United Imaging Intelligence - Products, Competitors, Financials | |
| SO006 | Datanyze | United Imaging Intelligence Company Profile - Management and Employees | |
| SO007 | 36Kr | United Imaging Intelligence completed a Series A financing round of 1 billion yuan | Shanghai United Imaging Intelligence Healthcare Co., Ltd. recently completed its Series A financing with a total scale of 1 billion yuan ... jointly led by E Fund ... and Shanghai Guotou Investment. |
| SO008 | 36Kr | United Imaging Intelligence Focuses on Multi-modal Medical Intelligent Agents After Raising 1 Billion Yuan | United Imaging Intelligence ... is planning to consider an independent listing. ... The company has launched 12 product platforms and over 100 AI applications. |
| SO009 | DealStreetAsia | China's United Imaging Intelligence snaps over $139m in Series A round | |
| SO010 | Insight (Fineline Cube) | United Imaging Intelligence Closes RMB 1B Series A Round to Boost AI in Healthcare | United Imaging Intelligence ... secured RMB 1 billion (USD 139 million) in a Series A financing round ... co-led by E Fund Management and SIG Property Management. |
| SO011 | C114 | United Imaging Intelligence Secures A Round Financing of 1 Billion Yuan | |
| SO012 | MarketScreener | Shanghai United Imaging ... received CNY 1 billion in funding from a group of investors | |
| SO013 | Crunchbase | Series A - United Imaging Intelligence | |
| SO014 | Preqin | United Imaging Intelligence Asset Profile | United Imaging Intelligence - overview. Established 2017. |
| SO015 | PR Newswire | United Imaging Intelligence at RSNA 2025 - Reshaping Medicine with AI Agents | |
| SO016 | PR Newswire | United Imaging Intelligence at ECR 2026 - Validating, Expanding, and Applying Radiology AI at Scale | |
| SO017 | PR Newswire | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | United Imaging Healthcare (SSE 688271) achieved a record annual revenue of RMB 13.80 billion in 2025, marking a 33.98% year-on-year increase. |
| SO018 | PR Newswire Asia | United Imaging Intelligence at RSNA 2025 - Reshaping Medicine with AI Agents | |
| SO019 | Complete AI Training | RSNA 2025: United Imaging Intelligence Puts AI Agents to Work From Scan to Report | |
| SO020 | The Imaging Wire | FDA Updates AI List with New Clearances | Philips at 50 ... Canon at 45, United Imaging at 38, Aidoc at 31, and DeepHealth at 28. |
| SO021 | Baidu Baike | 上海联影智能科技股份有限公司 (Shanghai United Imaging Intelligence) | |
| SO022 | RocketReach | UII America, Inc. Information | United Imaging Intelligence (UII), founded in 2017, is a global medical artificial intelligence (AI) company with offices in Shanghai and Boston. |
| SO023 | Morningstar | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | |
| SO024 | MCRA | Empowering FDA Clearance of United Imaging Intelligence Technology | |
| SO025 | PR Newswire | World's first open-source LLM for medical video released (uAI NEXUS MedVLM) | United Imaging Intelligence (UII) presented uAI NEXUS MedVLM ... fully open source ... accepted by CVPR 2026. |
| SO026 | Medical Economics | Commercializing AI in medical technology: Challenges & opportunities | Reimbursement is a principal scaling bottleneck for medical imaging AI, and only a handful of companies globally have secured consistent payer coverage. |
| SM001 | Grand View Research | AI In Medical Imaging Market Size, Share Report, 2026-2033 | The global AI in medical imaging market size was estimated at USD 1.8 billion in 2025 and is projected to reach USD 20.2 billion by 2033, growing at a CAGR of 35.11% from 2026 to 2033. |
| SM002 | Grand View Research | China Medical Image Analysis Software Market Report, 2030 | The China medical image analysis software market size was valued at USD 422.8 million in 2022 and is expected to expand at a compound annual growth rate (CAGR) of 24.9% from 2023 to 2030. |
| SM003 | Grand View Research | China AI In Healthcare Market Size & Outlook, 2030 | |
| SM004 | Insights10 | China Artificial Intelligence (AI) in Medical Imaging Market Report 2022 to 2030 | |
| SM005 | The Business Research Company | AI In Medical Imaging Market Share, Size, Trends, Report 2026 | It will grow from $3.79 billion in 2025 to $5.29 billion in 2026 at a compound annual growth rate (CAGR) of 39.5%. |
| SM006 | Coherent Market Insights | AI in Medical Imaging Market Share & Opportunities 2026-2033 | Global AI in medical imaging market is estimated to be valued at USD 2.20 Bn in 2026 and is expected to reach USD 17.77 Bn by 2033. |
| SM007 | Fortune Business Insights | AI in Medical Imaging Market Size, Share | Growth Report | |
| SM008 | Precedence Research | AI in Medical Imaging Market Size, Share, and Trends 2026 to 2035 | By imaging modality, the CT segment led the market share of 37% in 2025. By clinical area, the oncology segment is growing at a notable CAGR of 30.20% between 2026 and 2035. |
| SM009 | Research and Markets | AI in Medical Imaging - Global Market Report | |
| SM010 | Business Research Insights | AI in Medical Imaging Market Size, Share | Report 2033 | |
| SM011 | DataM Intelligence | AI in Medical Imaging Market Size, Share, Report 2026-2033 | |
| SM012 | Intel Market Research | AI Medical Imaging Market Outlook 2026-2034 | |
| SM013 | IndexBox | Radiology AI Platforms Market in China - Report | The China Radiology AI Platforms market stands as a critical and rapidly evolving segment within the nation's broader healthcare technology landscape. |
| SM014 | Towards Healthcare | China AI in Healthcare Market Size, Statistics and Companies | The China AI in healthcare market is projected to grow from USD 2.67 billion by 2025 to reach USD 84.11 billion in 2035, at a CAGR of 41.20%. |
| SM015 | Made-in-China Insights | China's AI Healthcare Market - Growth Trends, Drivers and APAC Comparison | In 2023, the value of China's AI healthcare market reached approximately US$1.59 billion ... this market will expand to US$7.33 billion by 2028. |
| SM016 | China Med Access | China's Hospital AI Diagnostic Systems 2026: Pathology, Radiology and Deployment | By 2026, over 1,200 Tier-3A hospitals in China are reported to have adopted AI-assisted diagnostic systems for radiology, pathology, or both. |
| SM017 | China Med Access | China's Healthcare AI Regulation 2026 Updated: NMPA Class III Device | Most AI-driven diagnosis or therapy software falls under Class III due to their impact on patient care and safety. |
| SM018 | Cisema | China NMPA Issues 2026 Medical Device Industry Standards Development Plan | |
| SM019 | China Med Device | NMPA Roundup June 2026 | |
| SM020 | National Medical Products Administration (NMPA) | Announcement of the NMPA on Issuing Device Classification Guidance | |
| SM021 | NSF | China Regulatory Update - NMPA Launches New Standard Projects for 2026 | |
| SM022 | Asia Actual | New Guidance on Medical Device Classification Adjustments in China | |
| SM023 | Hawksford | China's Personal Information Protection Law (PIPL) - business guide | Medical and health data is explicitly categorized as sensitive personal information under PIPL. |
| SM024 | Crowell & Moring | How Businesses Can Navigate China's Data Regulations in 2025 | |
| SM025 | Chambers and Partners | The Final Piece of China's Cross-Border Personal Information Transfer Regulations | New Certification Measures take effect January 1, 2026. |
| SM026 | Fierce Biotech | Biotech Giants and Investors Gather to Solve the Commercialisation Puzzle | The real battle now is execution - converting promising AI into validated, reimbursable products woven into healthcare workflows. |
| SP001 | The Imaging Wire | Top 10 Radiology AI Vendors by Number of FDA Authorizations | GE HealthCare held its lead at 130, then Siemens Healthineers at 95, Philips at 58, Canon at 48, United Imaging at 40, Aidoc at 33. |
| SP002 | Tracxn | Infervision - 2026 Company Profile & Team | Infervision has raised a total funding of $213M over 8 rounds. |
| SP003 | Tracxn | DeepWise - 2026 Company Profile & Team | DeepWise has raised $115M in funding from investors like Legend Capital and CICC Capital. |
| SP004 | VCBeat Health | Deepwise Medical Unveils New AI-Powered Platforms at 2026 China Medical Equipment Exhibition | |
| SP005 | VCBeat Health | Airdoc, China's First AI Medical Imaging IPO, Lists on HKEX | Airdoc listed on the Hong Kong Stock Exchange ... this vision has given rise to a publicly listed company now valued at RMB 5.8 billion. |
| SP006 | Signify Research | Medical Imaging AI Market Update - Funding, Regulation & Generative AI | Aidoc raised $150 million in Series E funding led by Goldman Sachs ... a clear shift in capital allocation toward vendors positioning AI as enterprise infrastructure. |
| SP007 | Aidoc | Aidoc Raises $150 Million Series E Led by Goldman Sachs | The round brings total funding to over $500 million ... the company analyzes more than 60 million patient cases annually and is deployed across nearly 2,000 hospitals. |
| SP008 | Business 2.0 Channel | Aidoc $150M Series E 2026: Goldman Sachs Backs Clinical AI Before IPO | |
| SP009 | CB Insights | Airdoc - Products, Competitors, Financials, Employees | |
| SP010 | HKEXnews | Beijing Airdoc Technology Co., Ltd. - Listed Company Filing | |
| SP011 | pdpspectra | AI Radiology 2026: Vendors, FDA, PACS | |
| SP012 | 800 Oxygen | AI Radiology 2026: How Annalise, Aidoc & Rad AI Cut Missed Diagnoses | |
| SP013 | China Med Access | AI Healthcare Algorithms in China 2026 Updated | |
| SP014 | IndexBox | Medical Imaging AI Software Market in China - Report | |
| SP015 | PitchBook | Aidoc 2026 Company Profile: Valuation, Funding & Investors | |
| SP016 | Tencent Healthcare | Tencent Miying - Smart Medical Imaging Platform (AIMIS) | |
| SP017 | YITU Technology | YITU Healthcare - Explore the AI World | |
| SP018 | Ping An Group | Integrated Healthcare - A New Insurance Model in China | |
| SP019 | Sacra | Viz.ai revenue, valuation & funding | Viz.ai raised $252M over 7 rounds ... valuation $1.2B as of April 2022 ... deployed across nearly 2,000 U.S. hospitals. |
| SP020 | AI Health Apps | Annalise.ai Review 2026: Comprehensive AI Radiology Platform | |
| SP021 | Stabilarity Hub | Medical ML: China's Massive Medical AI Deployment | |
| SP022 | PitchBook | Shukun 2026 Company Profile: Valuation, Funding & Investors | |
| SP023 | AInvest | Ping An Healthcare's Strategic Breakthrough in Integrated Finance and AI Health Ecosystems | |
| SP024 | Viz.ai | Viz.ai Raises $100 Million in Series D Funding | |
| SP025 | CB Insights | Annalise.ai - Products, Competitors, Financials, Employees | |
| SP026 | AI Health Apps | Aidoc Review 2026: First Foundation Model AI for Comprehensive CT | |
| SI001 | Shanghai United Imaging Healthcare Co., Ltd. | 2025 Half-year Report Summary (Company Code 688271) | operating revenue reaching 6.02 billion CNY, a year-on-year increase of 12.79%. Net profit attributable to shareholders of the parent company was 998 million CNY. |
| SI002 | PR Newswire | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | In 2025, the company reported revenue of RMB 13.80 billion, representing a year-on-year increase of 33.98%, and net profit attributable to shareholders of RMB 1.87 billion, up 48.14%. |
| SI003 | PR Newswire | United Imaging Healthcare Reports 2025H1 Financial Results | revenue and net profit attributable to ordinary shareholders after deducting non-recurring profit or loss reached CNY 6.02 billion and 0.97 billion for 2025H1. |
| SI004 | Yahoo Finance | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | In 2025, the company reported revenue of RMB 13.80 billion, representing a year-on-year increase of 33.98%. |
| SI005 | 36Kr | "United Imaging Intelligence" Focuses on Multi-modal Medical Intelligent Agents After Raising 1 Billion Yuan in Series A | United Imaging Intelligence is planning to consider an independent listing. In June, it announced the completion of a 1 billion yuan Series A financing round. |
| SI006 | Minichart | Beijing Airdoc Technology Annual Report 2025: Financial Performance | the company's warning under Rule 18A.08(3) notes that ultimate commercialization risks remain, which investors should monitor closely. |
| SI007 | Simply Wall St | Market Sentiment Around Loss-Making Beijing Airdoc Technology Co., Ltd. (HKG:2251) | announced a latest loss of CN¥255m on 31 December 2024... They expect the company to post a final loss in 2025, before turning a profit of CN¥8.1m in 2026. |
| SI008 | Bamboo Works | Airdoc Tech enjoys healing properties of medical AI rally | The medical diagnostics company has been mired in net losses for five straight years, despite rising revenues. |
| SI009 | StockAnalysis | Beijing Airdoc Technology (HKG:2251) Company Profile & Description | Beijing Airdoc Technology provides AI-empowered retina-based early detection and diagnosis solutions. |
| SI010 | Crunchbase News | China Leads Asia's Startup Funding To Its Highest Level In More Than 3 Years | investors put $27.4 billion to work across seed- through growth-stage financings for Asian companies in Q1... with most going to China. |
| SI011 | DealStreetAsia | China's United Imaging Intelligence snaps over $139m in Series A round | United Imaging Intelligence raised over $139 million in a Series A round at a pre-money valuation of about 9 billion yuan. |
| SI012 | 36Kr | United Imaging Intelligence completed a Series A financing round of 1 billion yuan | United Imaging Intelligence completed a Series A financing round of 1 billion yuan. |
| SI013 | FLCube | United Imaging Intelligence Closes RMB 1B Series A Round to Boost AI | The funds will be used to accelerate R&D and the upgrade of large medical AI models and multi-modal intelligent medical agents. |
| SI014 | MarketScreener | Shanghai United Imaging Healthcare announced it received CNY 1 billion in funding | Shanghai United Imaging Intelligence announced that it has received CNY 1 billion in funding. |
| SI015 | United Imaging Intelligence | UII corporate site - AI solutions for healthcare | UII delivers AI solutions across imaging, diagnosis, treatment and research deployed in thousands of medical institutions. |
| SI016 | Health AI Register | Shanghai United Imaging Intelligence Co., Ltd. - company profile | UII holds multiple regulatory clearances across CE, FDA and NMPA pathways for its radiology AI portfolio. |
| SI017 | Crunchbase | United Imaging Intelligence - organization profile | United Imaging Intelligence is a medical AI subsidiary that raised a Series A in 2025. |
| SI018 | Crunchbase | United Imaging Intelligence Series A funding round | United Imaging Intelligence Series A round closed in June 2025. |
| SI019 | Preqin | United Imaging Intelligence - private markets profile | Preqin tracks United Imaging Intelligence as a venture-backed medical AI asset. |
| SI020 | CB Insights | United Imaging Intelligence - company profile | CB Insights profiles United Imaging Intelligence as a private medical AI company. |
| SI021 | VCBeat Health | China medical AI commercialization analysis | Chinese medical AI vendors face slow commercialization and the absence of national reimbursement for AI software. |
| SI022 | Morningstar | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | net profit attributable to shareholders of RMB 1.87 billion, up 48.14% year-on-year. |
| SI023 | The Imaging Wire | Numbers From the FDA Show Radiology Is Maintaining Its Lead | United Imaging ranks among the top radiology-AI vendors by number of FDA authorizations. |
| SI024 | Ping An Group | Integrated Healthcare: A New Insurance Model in China | Ping An's integrated finance-and-healthcare model monetizes health services through ecosystem cross-sell. |
| SI025 | ChinaMedDevice | NMPA AI Medical Device approval and market access | AI medical software in China faces approval and market-access hurdles, with no universal reimbursement pathway. |
| SI026 | AInvest | Ping An Healthcare 137% Profit Surge - Integrated Finance and AI Health Ecosystems | Ping An Healthcare posted a 137% profit surge driven by integrated finance and AI-driven health ecosystems. |
| SE001 | PR Newswire | The World's First Open-Source Medical Video LLM Released, Calling the Global Developer Community to Push It Further | United Imaging Intelligence has unveiled uAI NEXUS MedVLM... built on a monumental dataset comprising 531,850 video-instruction pairs across 8 clinical scenarios... accepted by CVPR 2026. |
| SE002 | GitHub (UII-AI) | UII-AI/MedGRPO-Code - reference inference pipeline for uAI-NEXUS-MedVLM | Reference inference pipeline for uAI-NEXUS-MedVLM-1.0a-7B-RL - a medical video-understanding model trained with SFT + MedGRPO on Qwen2.5-VL-7B. |
| SE003 | arXiv | MedGRPO: Multi-Task Reinforcement Learning for Heterogeneous Medical Video Understanding (arXiv:2512.06581) | we first introduce MedVidBench, a large-scale benchmark of 531,850 video-instruction pairs... we introduce MedGRPO, a novel RL framework for balanced multi-dataset training. |
| SE004 | Hugging Face (UII-AI, via Wayback Machine) | UII-AI/MedVidBench dataset | MedVidBench dataset of medical video-instruction pairs for benchmarking medical video understanding. |
| SE005 | UII-AI (GitHub Pages) | MedGRPO project page and leaderboard (CVPR 2026) | MedGRPO: Multi-Task Reinforcement Learning for Heterogeneous Medical Video Understanding - CVPR 2026. |
| SE006 | U.S. Food & Drug Administration | 510(k) Premarket Notification K242292 - uAI Easy Triage ICH | September 24, 2024 Shanghai United Imaging Intelligence Co., Ltd. - 510(k) clearance for uAI Easy Triage ICH. |
| SE007 | XrayInterpreter | Shanghai United Imaging Intelligence - Radiology AI company profile | uAI Easy Triage ICH... analyzes non-contrast head CT images to detect suspected intracranial hemorrhage... prioritize and triage cases. |
| SE008 | XrayInterpreter | uAI Portal | FDA Radiology AI Device (K240411) | uAI Portal... provides tools to view, manipulate, and evaluate CT angiography studies... uses AI-based segmentation algorithms. |
| SE009 | PR Newswire | United Imaging Intelligence at RSNA 2025 - Reshaping Medicine with AI Agents | a new generation of AI agents capable of understanding, reasoning, and acting across the entire care pathway. |
| SE010 | United Imaging (USA) | United Imaging Intelligence at RSNA 2024 - uAI PLUS foundation model | the debut of the uAI PLUS (Pretrained Large Models for Unified Solutions)... integrating Medical Imaging Foundation Models, Medical LLMs, and Medical Multimodal LLMs. |
| SE011 | MedTech News | United Imaging's AI Revolution at RSNA 2025 | They unveiled a truly ambitious suite of AI-driven technologies... to fundamentally revolutionize medical imaging and diagnostics. |
| SE012 | Nerds.xyz | United Imaging Intelligence releases open source medical video AI model | The company also released a massive dataset called MedVidBench with over 531,000 video instruction pairs, plus a benchmark and leaderboard so developers can actually test against it. |
| SE013 | PR Newswire (en.prnasia) | United Imaging Intelligence at ECR 2026 - Validating, Expanding, and Applying Radiology AI at Scale | the uAI Clinical Portal integrates with major hospital PACS, supporting over 60 AI applications... more than 30 CE-certified applications. |
| SE014 | United Imaging Intelligence | UII corporate site - AI solutions for healthcare | UII delivers AI solutions across imaging, diagnosis, treatment and research with 12 platforms and 100+ applications. |
| SE015 | MCRA | Empowering FDA Clearance - United Imaging Intelligence Technology (case study) | MCRA supported United Imaging Intelligence in achieving FDA clearance for its AI imaging technology. |
| SE016 | Health AI Register | Shanghai United Imaging Intelligence Co., Ltd. - company profile | UII holds multiple regulatory clearances across CE, FDA and NMPA pathways for its radiology AI portfolio. |
| SE017 | G-Medtech | United Imaging Intelligence - medtech player profile | United Imaging Intelligence develops AI applications for medical imaging across modalities. |
| SE018 | Complete AI Training | RSNA 2025: United Imaging Intelligence Puts AI Agents to Work From Scan to Report | The uAI Agent for Chest CT Reporting detects 73 thoracic diseases from a single scan; the Brain MRI agent identifies 47 neurological conditions. |
| SE019 | The Imaging Wire | Numbers From the FDA Show Radiology Is Maintaining Its Lead | United Imaging ranks among the top radiology-AI vendors by number of FDA authorizations. |
| SE020 | ChinaMedDevice | NMPA AI Medical Device approval and classification | NMPA Class III is the highest regulatory risk class, requiring rigorous clinical evaluation for AI medical software. |
| SE021 | NMPA (English) | National Medical Products Administration - regulatory updates | NMPA continues to advance the regulatory framework for medical devices including AI-enabled software. |
| SE022 | Cisema | NMPA roadmap for AI medical devices | China's NMPA has built a dedicated approval pathway and standards roadmap for AI medical devices. |
| SE023 | Stabilarity | Medical ML - China's massive medical AI deployment | China has deployed medical AI at large scale across hospitals, including radiology imaging analysis. |
| SE024 | China Med Access | China hospital AI diagnostic systems 2026 - pathology and radiology deployment | Chinese hospitals are deploying AI diagnostic systems across pathology and radiology workflows in 2026. |
| SE025 | PDP Spectra | AI radiology and medical imaging in 2026 | AI radiology tools in 2026 increasingly integrate triage, segmentation, and reporting into clinical workflow. |
| SE026 | United Imaging Intelligence | UII exhibition / product showcase | UII showcases its uAI product platforms and AI agents for medical imaging. |
| SU001 | Healthcare Intelligence Brief | United Imaging Intelligence Scales Radiology AI at ECR 2026 | United Imaging Intelligence showcased a portfolio of CE-certified AI applications validated in multicenter clinical settings, with workflow improvements for radiology reporting. |
| SU002 | Global Times | Chinese hospitals accelerate AI-powered health care transformation | Chinese hospitals are accelerating the integration of AI into clinical workflows, moving from pilot programs to scaled deployment across major hospitals. |
| SU003 | Martin Hunsinger | The Incursion of United Imaging into the U.S. Market: What It Means for Healthcare | US healthcare providers historically favor Western vendors, and there is recurring skepticism about supply-chain security, data privacy, and long-term support from a China-based firm. |
| SU004 | UC Davis Department of Radiology | Magnetic Resonance Imaging (MRI) Equipment | United Imaging 1.5T uMR 680 is listed among the MRI systems in the UC Davis Health imaging fleet. |
| SU005 | HealthManagement.org | Advancing Diagnostic Imaging with United Imaging's Global Expansion | United Imaging is expanding globally, installing diagnostic imaging systems in Italy, France and Lithuania, including the uCT 960+ and uMR 680. |
| SU006 | United Imaging (Europe) | uMR Omega: Ultra-wide Bore 3.0T MR | The uMR Omega features a 75 cm ultra-wide bore designed to expand access to MR and broaden clinical capabilities with uAiFI technology. |
| SU007 | United Imaging Healthcare | United Imaging Healthcare News Center | United Imaging brings more than 20 products and intelligent solutions across diagnosis, radiotherapy, and internet medical care to major industry events. |
| SU008 | La Ragione (PR Newswire distribution) | United Imaging Intelligence at ECR 2026: Validating, Expanding, and Applying Radiology AI at Scale | At ECR 2026 under the theme Rays of Knowledge, UII showcased AI solutions with 30 CE-certified medical AI applications clinically validated and applied in everyday care. |
| SU009 | Synapse Medical Systems | uMR Omega (3T) | The uMR Omega 75 cm ultra-wide bore 3T MRI is offered to expand clinical capabilities and patient access. |
| SU010 | DOTmed News | uMR Omega 3T by United Imaging | The uMR Omega is a 75 cm ultra-wide bore 3T MRI with advanced uAiFI technology used to expand access to MR in the community. |
| SU011 | PR Newswire | United Imaging Intelligence at ECR 2026: Validating, Expanding, and Applying Radiology AI at Scale | UII showcased 30 CE-certified medical AI applications validated in multicenter clinical environments across Europe, Asia, North America, and South America. |
| SU012 | PR Newswire Asia | United Imaging Intelligence at ECR 2026 (Asia release) | The uAI Clinical Portal unites more than 60 applications into a single workflow platform, with foundation models powering AI agents for Chest CT, Brain MRI, and Ultrasound. |
| SU013 | PR Newswire Asia | United Imaging Intelligence at RSNA 2025 - Reshaping Medicine with AI Agents | UII presented AI agents for radiology at RSNA 2025, demonstrating image-to-report and workflow agents integrated into clinical practice. |
| SU014 | ChinaMedAccess | China's Hospital AI Diagnostic Systems 2026: Pathology, Radiology and the Race to Deploy at Scale | By 2026 more than 1,200 Tier-3A hospitals in China have deployed AI-assisted diagnostic systems for radiology, pathology, or both, with reported radiology workflow improvements around 30 percent. |
| SU015 | PR Newswire | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | United Imaging products are in more than 100 countries and regions, with cumulative US installations exceeding 640 systems and overseas revenue up over 51% year-on-year. |
| SU016 | Yahoo Finance | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | United Imaging reported installations in over 90% of US states with more than 640 cumulative imaging systems by the end of 2025. |
| SU017 | Crunchbase | United Imaging Intelligence - Organization Profile | United Imaging Intelligence develops AI solutions for medical imaging, diagnostics, and clinical workflows deployed across medical institutions. |
| SU018 | Health AI Register | Shanghai United Imaging Intelligence Co., Ltd. - company profile | United Imaging Intelligence holds a portfolio of CE-marked and FDA-cleared radiology AI applications listed in the Health AI Register. |
| SU019 | United Imaging Intelligence | UII - AI Solutions for Healthcare (company website) | UII offers 12 platforms and more than 100 AI applications deployed across over 4,000 medical institutions. |
| SU020 | g-medtech | United Imaging Expands Multi-Modality Portfolio and Native AI Capabilities at RSNA 2025 | United Imaging expanded its multi-modality portfolio with native AI capabilities such as uMR Ultra and uCT SiriuX integrating real-time AI into clinical workflows. |
| SU021 | Bamboo Works | Airdoc Tech Enjoys Healing Properties of Medical AI Rally | Airdoc has posted net losses for five straight years despite rising revenue, underscoring how hard durable monetization is for Chinese medical-AI vendors. |
| SU022 | Shanghai United Imaging Healthcare Co., Ltd. | 2025 Half-year Report Summary (Company Code 688271) | Operating revenue reached 6.02 billion CNY with overseas revenue contributing a record share, reflecting accelerating international installation. |
| SU023 | 36Kr | United Imaging Intelligence completes large Series A financing | United Imaging Intelligence, the AI arm of United Imaging Healthcare, raised a roughly 1 billion yuan Series A round backed by E Fund and Shanghai state investors. |
| SU024 | Simply Wall St | Market sentiment around loss-making Beijing Airdoc Technology | Beijing Airdoc remains loss-making, illustrating the difficulty of converting broad medical-AI adoption into durable, profitable customer revenue. |
| SU025 | PR Newswire | United Imaging Healthcare Reports 2025H1 Financial Results | United Imaging Healthcare achieved record revenue in overseas markets in the first half of 2025, supporting its global installed-base growth. |
| SU026 | g-medtech | United Imaging Intelligence - medtech players database profile | United Imaging Intelligence is profiled as the AI subsidiary of United Imaging Healthcare with a portfolio of radiology AI applications. |
| SR001 | Hogan Lovells | China's digital therapeutics and AI in health care - where does regulation begin | AI-powered diagnostics and software-as-a-medical-device require stakeholders to navigate China's evolving NMPA, PIPL, and data-security regulatory landscape. |
| SR002 | NMPA (National Medical Products Administration) | Announcement on Promoting High-quality Development of High-end Medical Devices | AI-powered medical devices and high-end medical imaging equipment are key areas for fostering new quality productivity in the medical device industry. |
| SR003 | Medbot | How NMPA Class III Medical Device Approval Works in China | NMPA Class III is the highest-risk medical device category in China, and the registration certificate determines what can actually be sold to a hospital. |
| SR004 | Cisema | China NMPA Issues 2026 Medical Device Industry Standards Development Plan | The NMPA's 2026 plan outlines development and revision of more than 80 medical device industry standards with direct impact on manufacturers. |
| SR005 | Cisema | NMPA Roadmap for AI Medical Devices | China's regulatory roadmap for AI medical devices imposes clinical evaluation and change-control requirements specific to evolving algorithms. |
| SR006 | Arnold & Porter | The BIOSECURE Act Becomes Law in the United States | On December 18, 2025, the President signed a revised BIOSECURE Act as Sec. 851 of the FY26 NDAA, restricting use of certain foreign biotech equipment and service providers. |
| SR007 | Goodwin Procter | Final FY 2026 NDAA Includes BIOSECURE Act Provisions | The BIOSECURE Act restricts US executive agencies from contracting with biotechnology companies of concern and will significantly impact the biotech supply chain. |
| SR008 | Baker McKenzie | United States: The BIOSECURE Act Becomes Law | The BIOSECURE Act bars federal agencies from procuring biotechnology equipment or services from designated companies of concern with foreign-adversary ties. |
| SR009 | Ropes & Gray | BIOSECURE Act Enacted | The NDAA includes an updated BIOSECURE Act restricting federal agencies and contractors from procuring biotechnology equipment or services from companies of concern. |
| SR010 | Life Sciences Perspectives (Mintz) | BIOSECURE Update - 1260H List Released | The Department of Defense published its updated Section 1260H list of Chinese military companies, adding WuXi AppTec and 27 other new entities. |
| SR011 | Martin Hunsinger | The Incursion of United Imaging into the U.S. Market: What It Means for Healthcare | US providers historically favor Western vendors, with recurring skepticism about supply-chain security, data privacy, and long-term support from a China-based firm. |
| SR012 | Yahoo Finance | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | United Imaging reported installations in over 90% of US states with more than 640 cumulative imaging systems by the end of 2025. |
| SR013 | Radiology Business | Memorial Sloan Kettering sues GE HealthCare alleging willful patent infringement | GE HealthCare faces a lawsuit alleging it deployed and profited from imaging technology developed by an individual from another organization. |
| SR014 | UnitedLex | Medtech Patent Battles: Key Litigation Trends and How NPEs are Shaping the Landscape | The medtech industry has seen a surge in IP disputes, particularly driven by non-practicing entities, including in radiology AI. |
| SR015 | GreyB | US Patent Litigation Cases of 2025 (Jan-June) | The first half of 2025 saw US patent clashes over AI-powered technologies that could shape entire industries. |
| SR016 | Bamboo Works | Airdoc Tech Enjoys Healing Properties of Medical AI Rally | Airdoc has posted net losses for five straight years despite rising revenue, underscoring how hard durable medical-AI monetization is. |
| SR017 | Simply Wall St | Market sentiment around loss-making Beijing Airdoc Technology | Beijing Airdoc remains loss-making, illustrating the difficulty of converting broad medical-AI adoption into durable, profitable revenue. |
| SR018 | PR Newswire | United Imaging Healthcare Releases 2025 Annual Report and Q1 2026 Results | In 2025 United Imaging reported revenue of RMB 13.80 billion, up 33.98%, and net profit of RMB 1.87 billion, up 48.14%. |
| SR019 | Shanghai United Imaging Healthcare Co., Ltd. | 2025 Half-year Report Summary (Company Code 688271) | Operating revenue reached 6.02 billion CNY with net profit attributable to the parent of 998 million CNY in the first half of 2025. |
| SR020 | ChinaMedAccess | China's Hospital AI Diagnostic Systems 2026: Pathology, Radiology and the Race to Deploy at Scale | China's hospital AI systems moved from pilots to full-scale deployment in 2025-2026 under tightening regulatory and data-governance expectations. |
| SR021 | United Imaging Intelligence | UII - AI Solutions for Healthcare (company website) | UII holds 13 NMPA Class III certificates, 15 FDA clearances, and 31 CE certifications across its uAI platforms. |
| SR022 | Health AI Register | Shanghai United Imaging Intelligence Co., Ltd. - company profile | United Imaging Intelligence holds CE-marked and FDA-cleared radiology AI applications listed in the Health AI Register. |
| SR023 | Crunchbase | United Imaging Intelligence - Organization Profile | United Imaging Intelligence is the AI subsidiary of United Imaging Healthcare developing medical-imaging AI solutions. |
| SR024 | 36Kr | United Imaging Intelligence completes large Series A financing | United Imaging Intelligence raised a roughly 1 billion yuan Series A backed by E Fund and Shanghai state investors. |
| SR025 | g-medtech | United Imaging Intelligence - medtech players database profile | United Imaging Intelligence is profiled as the AI subsidiary of United Imaging Healthcare with a radiology AI portfolio. |
| SR026 | PR Newswire | United Imaging Healthcare Reports 2025H1 Financial Results | United Imaging Healthcare achieved record overseas-market revenue in the first half of 2025. |
| SR027 | Global Times | Chinese hospitals accelerate AI-powered health care transformation | Chinese hospitals are accelerating AI integration into clinical workflows under active state policy support. |
| SR028 | PR Newswire | United Imaging Intelligence at ECR 2026: Validating, Expanding, and Applying Radiology AI at Scale | UII showcased 30 CE-certified AI applications validated in multicenter clinical settings across four continents. |
| SR029 | News Crunchbase | China Leads Startup Funding as AI Seed and Growth Drive Asia Q1 2026 | First-quarter 2026 Asia funding of about USD 27.4 billion flowed heavily to China, led by AI, supporting follow-on capital availability. |
| SR030 | United Imaging (Europe) | uMR Omega: Ultra-wide Bore 3.0T MR | The uMR Omega integrates uAiFI technology, illustrating AI embedded natively in United Imaging scanners. |
| SV001 | United Imaging Healthcare | Shanghai United Imaging Healthcare 2025 Half-Year Report Summary | |
| SV002 | DealStreetAsia | United Imaging Intelligence raises Series A round | |
| SV003 | Crunchbase | United Imaging Intelligence - Company Profile | |
| SV004 | GetHealthBrief | United Imaging Intelligence at ECR 2026 - Radiology AI at Scale | |
| SV005 | Global Times | United Imaging Intelligence completes major AI funding round | |
| SV006 | Health AI Register | Shanghai United Imaging Intelligence Co., Ltd. - Company Record | |
| SV007 | United Imaging Intelligence | United Imaging Intelligence - AI Solutions for Healthcare | |
| SV008 | Qubit Capital | AI Startup Valuation Multiples | |
| SV009 | Nelson Advisors | HealthTech M&A Multiples January 2026 | |
| SV010 | Finro Financial Consulting | AI Valuation Multiples Q1 2026 | |
| SV011 | Multiples.vc | Tempus AI Valuation Multiples | |
| SV012 | Multiples.vc | Airdoc Valuation Multiples | |
| SV013 | AI Funding Tracker | Top AI Healthcare Startups | |
| SV014 | Yahoo Finance | Shanghai United Imaging Healthcare (688271.SS) Key Statistics | |
| SV015 | Law.asia | Chapter 18A listings of biotech companies on the HKEX | |
| SV016 | KPMG | Chinese Mainland and Hong Kong IPO Markets 2025 Review and 2026 Outlook | |
| SV017 | Sacra | Viz.ai Company Profile and Revenue | |
| SV018 | Bamboo Works | Airdoc Tech Enjoys Healing Properties of Medical AI Rally | |
| SV019 | Simply Wall St | Market Sentiment Around Loss-Making Beijing Airdoc Technology | |
| SV020 | Martin Hunsinger | The Incursion of United Imaging into the US Market | |
| SV021 | StockAnalysis | Shanghai United Imaging Healthcare (SHA:688271) Market Cap and Net Worth | |
| SV022 | MarketScreener | Shanghai United Imaging Healthcare - Valuation Ratios | |
| SV023 | Morningstar | 688271 Stock Price Quote | |
| SV024 | 36Kr | United Imaging Intelligence completes large AI financing round | |
| SV025 | China Med Access | AI Healthcare Algorithms China 2026 | |
| SV026 | PR Newswire Asia | United Imaging Intelligence at ECR 2026 - Validating, Expanding and Applying Radiology AI at Scale | |
| SV027 | Crunchbase | United Imaging Intelligence Series A Funding Round | |
| SV028 | Global Times | China medical AI sector developments | |
| SV029 | United Imaging Intelligence | UII Brochure - AI Solutions for Healthcare | |
| SV030 | United Imaging Healthcare | United Imaging Global News Center |