{
 "updated": "2026-11-02",
 "about": "VC、研究机构画的 AI 视频 / 生成式媒体 / 短剧相关市场图谱。by = 谁画的，d = 日期，u = 原文，cats = 分了哪些类（中文转述），g = 对应 layers.json 的 code（all = 全景）。",
 "maps": [
  {
   "by": "a16z",
   "who": "Justine Moore",
   "title": "New market map: AI video generation",
   "date": "2025-06",
   "url": "https://x.com/venturetwins/status/1935367302834495832",
   "type": "vc",
   "topic": "视频模型",
   "layers": [
    "v1",
    "v2"
   ],
   "cats": "文生视频 / 图生视频模型公司，加上让角色开口说话的公司",
   "note": "X 帖子里的一张图（2025-06-18 发），不是 a16z.com 文章；2025-11 有更新版"
  },
  {
   "by": "a16z",
   "who": "Justine Moore",
   "title": "Updated market map on video generation models",
   "date": "2025-11",
   "url": "https://x.com/venturetwins/status/1990169320614453600",
   "type": "vc",
   "topic": "视频模型",
   "layers": [
    "v1",
    "v2"
   ],
   "cats": "只列视频生成模型厂商；作者指出几乎全在美国或中国，唯一例外是以色列的 Lightricks（LTX）",
   "note": "X 帖子（2025-11-16 发）；明确不含数字人模型（HeyGen、Synthesia、Tavus）"
  },
  {
   "by": "a16z",
   "who": "Justine Moore",
   "title": "There is no God Tier video model: But there is something better",
   "date": "2025-10",
   "url": "https://a16z.com/there-is-no-god-tier-video-model-but-there-is-something-better/",
   "type": "vc",
   "topic": "视频模型",
   "layers": [
    "v1",
    "v2",
    "v3",
    "v4"
   ],
   "cats": "论点文章：模型各有所长（Veo 3 物理与音画同步、Sora 2 多镜头叙事、Wan 开源、Grok 动漫等）；产品层机会在垂直模型与工作流封装",
   "note": "是文章不是图；a16z.com 页面发布 2025-10-22（substack 标 10-21）；提到 fal、Replicate、Krea 作为多模型入口"
  },
  {
   "by": "a16z × fal",
   "who": "Jennifer Li, Justine Moore",
   "title": "The State of Generative Media 2026",
   "date": "2026-02",
   "url": "https://a16z.com/the-state-of-generative-media-2026/",
   "type": "vc",
   "topic": "基础设施",
   "layers": [
    "v3",
    "v4"
   ],
   "cats": "a16z 对 fal 首份《生成式媒体现状》报告的五条解读：没有一个模型通吃、从推理走向编排、各类像素价值不同、广告电商游戏领先、开源追平",
   "note": "是报告解读不是地图；a16z 页面 2026-02-19 发布（fal 报告 02-18 发）；数据来自 fal 平台 600+ 模型"
  },
  {
   "by": "ummerr",
   "who": "ummerr（个人独立项目，GitHub: ummerr）",
   "title": "Generative Media Ecosystem Map — August 2026",
   "date": "2026-08",
   "url": "https://prompts.ummerr.com/market-map",
   "type": "研究机构",
   "topic": "全景",
   "layers": [
    "all"
   ],
   "cats": "九节：2023→2026 演变、市场地图、模型归属 vs 分发之争、创意 agent 与工作流层之争、基础设施与编排、基础模型版图、成本与采用信号、持久性与开源 / 地域、Momentum 25",
   "note": "数据截至 2026-08-19，约 150 个来源，估算值已标 est.；作者署名未见真名；页面无预览图"
  },
  {
   "by": "Machine Cinema",
   "who": "Minh Do",
   "title": "The AI Creative Market Map · Q2 2026（AI Creative Market Map - The Tools of the New Craft）",
   "date": "2026-05",
   "url": "https://text.machinecinema.ai/p/ai-creative-market-map-the-tools",
   "type": "媒体",
   "topic": "创作工具",
   "layers": [
    "all"
   ],
   "cats": "250+ 工具、19 类、5 个阶段：从制作协助到图像 / 视频 / 音频 / 3D 生成、聚合器、剪辑后期（放大、调色、Gen VFX）、AI 广告与社交、程序化视频，到版权 / 肖像保护",
   "note": "AI 电影社区 Machine Cinema 出品；substack 文章 2026-05-27，官网 PDF 版（需邮箱领取）标 Q2 2026：https://www.machinecinema.ai/ai-creative-market-map"
  },
  {
   "by": "Machine Cinema",
   "who": "Machine Cinema",
   "title": "AI Storytelling Market Map Q3 2025",
   "date": "2025-06",
   "url": "https://text.machinecinema.ai/p/ai-storytelling-market-map-q3-2025",
   "type": "媒体",
   "topic": "创作工具",
   "layers": [
    "all"
   ],
   "cats": "覆盖「开发—制作—后期」整条 AI 叙事流水线的工具与工作流",
   "note": "该社区第一张地图，欢迎厂商自荐收录；是上面 Q2 2026 版的前身（约 100 个工具）"
  },
  {
   "by": "招商证券",
   "who": "顾佳, 谢笑妍, 刘晓珊",
   "title": "AI影视行业深度报告：AI影视 AI应用的扛旗者",
   "date": "2026-11",
   "url": "http://stock.finance.sina.com.cn/stock/go.php/vReport_Show/kind/search/rptid/846849258986/index.phtml",
   "type": "券商",
   "topic": "全景",
   "layers": [
    "all"
   ],
   "cats": "受益标的沿「IP 储备—AI 制作—分发变现」三条主线；覆盖 AI 视频模型（Seedance 2.5 等）、AI 短剧、AI 漫剧、AI 中长剧 / 电影",
   "note": "报告日期 2026-11-01；新浪研报页只给摘要，全文需下载；引 DataEye：2026 国内微短剧 1210 亿元，其中 AI 短剧与漫剧超 400 亿元"
  },
  {
   "by": "浙商证券",
   "who": "冯翠婷",
   "title": "浙商传媒互联网｜AI影视、漫剧：多模态将重塑内容生产与商业闭环",
   "date": "2026-09",
   "url": "http://stock.finance.sina.com.cn/stock/go.php/vReport_Show/kind/search/rptid/842641260204/index.phtml",
   "type": "券商",
   "topic": "短剧",
   "layers": [
    "d1",
    "d2",
    "d3"
   ],
   "cats": "AI 影视与漫剧的商业化、多模态模型如何改写内容生产和商业闭环",
   "note": "报告日期 2026-09-13，公众号推送 09-15；新浪研报页只给摘要；摘要里没有逐层分类，需看全文确认"
  },
  {
   "by": "艾媒网",
   "who": "艾媒网（页面未署个人作者）",
   "title": "（附名单）5分钟大数据读懂中国AI漫剧产业链：微短剧之后，下一个百亿级内容战场",
   "date": "2026-02",
   "url": "https://www.iimedia.cn/c1096/109825.html",
   "type": "研究机构",
   "topic": "短剧",
   "layers": [
    "d1",
    "d2",
    "d3"
   ],
   "cats": "AI 漫剧产业链分上游、中游（内容生产制作）、下游三段，并附企业名单",
   "note": "页面发布 2026-02-24；名单部分为文中列出，未逐家核对"
  },
  {
   "by": "亿欧",
   "who": "邱同春",
   "title": "盘点 | 2026年全球AI视频生成赛道的融资众生相",
   "date": "2026-08",
   "url": "https://www.iyiou.com/analysis/202608121137880",
   "type": "媒体",
   "topic": "全景",
   "layers": [
    "all"
   ],
   "cats": "全球 AI 视频生成工具生态图谱：底座层、基模层、工具层、场景层；并拆解北美、日韩、东南亚、拉美四个市场",
   "note": "2026-08-12 发布；以融资盘点为主（称 91 亿美元涌入赛道）；读者定位是出海的中国工具厂商；预览图是文章头图，未必是图谱本身"
  },
  {
   "by": "Sensor Tower",
   "who": "Donny Kristianto（博客署名）",
   "title": "State of Short Drama Apps 2026",
   "date": "2026-06",
   "url": "https://sensortower.com/blog/state-of-short-drama-apps-2026-report",
   "type": "研究机构",
   "topic": "短剧",
   "layers": [
    "d2",
    "d3"
   ],
   "cats": "短剧 App 的下载、收入、互动、广告：2026 Q1 全球下载超 8.5 亿次（同比 +140%）",
   "note": "只当短剧需求侧背景，不是工具地图；博客发布 2026-06-12，完整报告需表单下载；中文版 2026-08"
  },
  {
   "by": "NFX",
   "who": "NFX（James Currier 等）",
   "title": "Generative AI Market Map and 5-Layer Tech Stack",
   "date": "2022-12",
   "url": "https://www.nfx.com/post/generative-ai-tech-5-layers",
   "type": "vc",
   "topic": "全景",
   "layers": [
    "all"
   ],
   "cats": "开源的生成式科技地图（发布时 550+ 家公司）与五层技术栈",
   "note": "通用生成式 AI 地图，视频只是其中一类，不是视频专项；页面发布 2022-12-07，早于本图谱时间窗，只作历史参照；你线索里的「Generative Tech Market Map」是它的续篇，未单独核到"
  }
 ]
}