{
 "sources": [
  "https://www.cbinsights.com/research/ai-drug-research-development-market-map/",
  "https://mmc.vc/research/ai-x-techbio-featuring-the-future-of-drug-discovery/",
  "https://x.com/FredaDuan/status/2106851949400469887",
  "https://www.nature.com/articles/d43747-023-00020-4"
 ],
 "about": "骨架：主轴照 CB Insights《The AI in drug R&D market map》（2025-05，discovery engines 卖药 / discovery platforms 卖软件 / clinical 临床），细分参照 MMC Ventures《AI x TechBio》（2026-06），湿实验与 CRO 参照 Altimeter / Freda Duan（2026-10），药企参照 Nature / McKinsey（2023-05）。outside_main_axis = CB Insights 主轴之外加的框（湿实验与 CRO、通用 AI 进场、药企）。",
 "layers": [
  {
   "code": "a1",
   "name": "小分子",
   "desc": "用自有 AI 发现并推进小分子管线",
   "box": "AI 原生药企",
   "box_desc": "卖药：用自己的 AI 发现、推进自有管线",
   "outside_main_axis": false
  },
  {
   "code": "a2",
   "name": "抗体与蛋白药",
   "desc": "AI 设计抗体、从头设计结合蛋白、蛋白药",
   "box": "AI 原生药企",
   "box_desc": "卖药：用自己的 AI 发现、推进自有管线",
   "outside_main_axis": false
  },
  {
   "code": "a3",
   "name": "新模态",
   "desc": "RNA、基因编辑、分子胶、天然产物、递送",
   "box": "AI 原生药企",
   "box_desc": "卖药：用自己的 AI 发现、推进自有管线",
   "outside_main_axis": false
  },
  {
   "code": "a4",
   "name": "资产型 / 临床开发型",
   "desc": "不从头发现分子，用 AI 买入、救援、加速已有资产",
   "box": "AI 原生药企",
   "box_desc": "卖药：用自己的 AI 发现、推进自有管线",
   "outside_main_axis": false
  },
  {
   "code": "m1",
   "name": "结构与共折叠",
   "desc": "结构预测、共折叠、结合亲和力",
   "box": "模型与软件平台",
   "box_desc": "卖软件、模型或 API，给药企和生物科技用",
   "outside_main_axis": false
  },
  {
   "code": "m2",
   "name": "蛋白生成模型",
   "desc": "生成式蛋白 / 抗体设计模型",
   "box": "模型与软件平台",
   "box_desc": "卖软件、模型或 API，给药企和生物科技用",
   "outside_main_axis": false
  },
  {
   "code": "m3",
   "name": "虚拟细胞与组学模型",
   "desc": "细胞、单细胞、病理、DNA 基础模型",
   "box": "模型与软件平台",
   "box_desc": "卖软件、模型或 API，给药企和生物科技用",
   "outside_main_axis": false
  },
  {
   "code": "m4",
   "name": "计算化学与模拟软件",
   "desc": "物理 + AI、计算化学、药物开发软件",
   "box": "模型与软件平台",
   "box_desc": "卖软件、模型或 API，给药企和生物科技用",
   "outside_main_axis": false
  },
  {
   "code": "t1",
   "name": "靶点发现",
   "desc": "多组学 + AI 找靶点与疾病生物学",
   "box": "靶点与数据",
   "box_desc": "先找对靶点，再攒别人没有的数据",
   "outside_main_axis": false
  },
  {
   "code": "t2",
   "name": "专有数据资产",
   "desc": "临床、基因组、表型、宏基因组等专有数据",
   "box": "靶点与数据",
   "box_desc": "先找对靶点，再攒别人没有的数据",
   "outside_main_axis": false
  },
  {
   "code": "c1",
   "name": "招募与匹配",
   "desc": "用病历匹配患者、招募受试者",
   "box": "临床试验 AI",
   "box_desc": "研发里最贵、最慢的一段",
   "outside_main_axis": false
  },
  {
   "code": "c2",
   "name": "试验设计与数字孪生",
   "desc": "方案优化、虚拟对照组、监管文件自动化",
   "box": "临床试验 AI",
   "box_desc": "研发里最贵、最慢的一段",
   "outside_main_axis": false
  },
  {
   "code": "w1",
   "name": "自动化 / 自驱动实验室",
   "desc": "机器人 + AI 闭环做实验、云实验室",
   "box": "湿实验与 CRO",
   "box_desc": "AI 产出更多候选药后，瓶颈转到真实实验",
   "outside_main_axis": true
  },
  {
   "code": "w2",
   "name": "AI 赋能的 CRO / CRDMO",
   "desc": "外包研发、生产企业的 AI 业务线",
   "box": "湿实验与 CRO",
   "box_desc": "AI 产出更多候选药后，瓶颈转到真实实验",
   "outside_main_axis": true
  },
  {
   "code": "l1",
   "name": "前沿实验室的生命科学线",
   "desc": "通用大模型用于查文献、提假设、调工具、设计分子",
   "box": "通用 AI 进场",
   "box_desc": "通用大模型公司和算力厂商的生命科学线",
   "outside_main_axis": true
  },
  {
   "code": "l2",
   "name": "AI 科学家与科研代理",
   "desc": "读文献、提假设、设计实验的 agent",
   "box": "通用 AI 进场",
   "box_desc": "通用大模型公司和算力厂商的生命科学线",
   "outside_main_axis": true
  },
  {
   "code": "l3",
   "name": "算力与模型托管",
   "desc": "GPU、云、生物模型托管与编排",
   "box": "通用 AI 进场",
   "box_desc": "通用大模型公司和算力厂商的生命科学线",
   "outside_main_axis": true
  },
  {
   "code": "p1",
   "name": "药企 AI 自建与合作",
   "desc": "每家药企一张卡：AI 合作对象和金额、自建平台、算力",
   "box": "药企",
   "box_desc": "买单的一方：自建 AI、和 AI 公司合作、收购",
   "outside_main_axis": true
  }
 ]
}