[
 {
  "t": "接近 IPO 的 AI 公司及其收入（英文原图标题：Emerging AI Companies Nearing IPO Scale）",
  "by": "Evelyn（小红书用户，转发/搬运；图上无署名，原图作者未查到）",
  "who": "个人",
  "d": "2026-02-08",
  "u": "https://www.xiaohongshu.com/explore/6988ab0b0000000028023eee",
  "img": "img/evelyn-xhs-ipo-scale.png",
  "cats": [
   "$1B+ ARR",
   "$200M+ ARR",
   "$100M+ ARR"
  ],
  "note": "三档共 23 家：$1B+ 为 OpenAI、Anthropic、Databricks、xAI、Cursor、Canva、Surge AI；$200M+ 为 Perplexity、Synthesia、Lovable、ElevenLabs、Glean、Suno、Gong、Fireworks AI、Dataiku；$100M+ 为 Clay、Gamma、Sierra、Harvey、Cognition、HeyGen、EliseAI（Databricks、Gong、Dataiku 旁有脚注 1，图中未显示脚注内容）。同一张图的英文版在 Followin 上被转为 '即将达到 IPO 规模的 AI 公司' 一帖，署名为 Upside GM 转述 'Chia'，也没有指向原制作者；图中没有来源水印。",
  "imgLocal": "C:/Projects/claude code/ai-revenue-map/research/img/evelyn-xhs-ipo-scale.png",
  "originNote": "原图出处未查到：小红书帖子 2026-02-08（点赞约 399、收藏约 454）只有图和一句话简介；图上无署名或水印。已搜英文标题全文、三档公司组合和 IPO 收入中位数（$381M→$878M）三种线索，均只找到转载，未找到原制作者。"
 },
 {
  "t": "Anthropic and OpenAI's Share of AI Startup Revenues Rises to 89%（Generative AI Database）",
  "by": "The Information",
  "who": "媒体",
  "d": "2026-05-17",
  "u": "https://feature.theinformation.com/articles/anthropic-openais-share-ai-startup-revenues-rises-89",
  "img": "",
  "cats": [
   "Anthropic + OpenAI（合占 89%）",
   "其余 32 家（含 Cursor、Cognition、ElevenLabs、Perplexity 等）"
  ],
  "note": "34 家头部 AI 初创年化收入合计近 $800 亿，半年 +112%；数据库追踪 700+ 家软件与服务公司；不含租 Nvidia 芯片的 AI 云和已上市公司；文中提示两家大头收入可能偏高（云伙伴按总额确认、微软分成）。付费墙。"
 },
 {
  "t": "Forbes AI 50 2026",
  "by": "Forbes（赞助：Mayfield；编辑 Rashi Shrivastava）",
  "who": "媒体",
  "d": "2026-04-16",
  "u": "https://www.forbes.com/lists/ai50/",
  "img": "https://www.forbes.com/dam/imageserve/69dfc661b968ebb871058b37/x.jpg",
  "cats": [
   "AI 50 主榜（50 家，20 家新上榜）",
   "Brink 早期榜（今年新增）"
  ],
  "note": "按技术实力、增长、市场影响评选，不是按收入排；榜内注明各家融资额与部分年化收入（如 Gamma 超 $1 亿、OpenAI 超 $250 亿、Anthropic 超 $300 亿）。OpenAI 与 Anthropic 合计融资 $2426 亿，约占全榜 $3056 亿的 80%。"
 },
 {
  "t": "CB Insights AI 100 2026（第 10 届）",
  "by": "CB Insights",
  "who": "研究机构",
  "d": "2026-05-05",
  "u": "https://www.cbinsights.com/research/report/artificial-intelligence-top-startups-2026/",
  "img": "https://research-assets.cbinsights.com/2026/05/04173014/AI00-FI-052026-1.png",
  "cats": [
   "企业应用",
   "行业应用",
   "基础设施与算力",
   "物理 AI"
  ],
  "note": "从 4 万多家公司里按交易活跃度、投资方实力、招聘势头、Mosaic 评分选出 100 家早期 AI 公司；无收入数据，是'最有前景'榜而非收入榜。"
 },
 {
  "t": "The Cloud 100 2025（Benchmarks Report）",
  "by": "Bessemer Venture Partners（与 Forbes、Salesforce Ventures 联合发布）",
  "who": "VC",
  "d": "2025-09-03",
  "u": "https://www.bvp.com/cloud100",
  "img": "https://www.forbes.com/lists/cloud100/",
  "cats": [
   "AI",
   "基础设施",
   "金融科技",
   "设计/生产力",
   "销售/营销/客户体验",
   "安全",
   "人力资源",
   "开发者工具",
   "垂直软件",
   "其他"
  ],
  "note": "私营云与 AI 公司榜，总估值首次超 $1 万亿；AI 类仅 22 家但占估值约 42%（据二手转述，十大领域和占比未在 BVP 官方页直接核到）。是估值榜，不按收入排。已被 AI 100 接替。",
  "dateNote": "发布日期取自二手来源"
 },
 {
  "t": "The AI 100（Bessemer × Newcomer）",
  "by": "Bessemer Venture Partners 与 Newcomer",
  "who": "VC",
  "d": "2026-11-12",
  "u": "https://www.bvp.com/collections/the-ai-100",
  "img": "https://cdn.sanity.io/images/2e9xvz8g/production/d9a2c71ab6dea44698618016d9c40b0b2e898e88-1920x1080.png?rect=0,36,1920,1008&w=1200&h=630&fit=crop",
  "cats": [
   "基础模型",
   "基础设施",
   "机器人",
   "其他（具体分类官方页未列）"
  ],
  "note": "Cloud 100 的接替者：全球最重要的 100 家私营 AI 公司年榜；提名 2026-10-15 截止，首份榜单定于 2026-11-12 在 Newcomer 与 Cerebral Valley AI Summit 发布，尚未发布（今天 11-02）。",
  "status": "未发布，需 11-12 后回查"
 },
 {
  "t": "ARR Club — Fastest Growing AI Startups",
  "by": "ARR Club（运营方官网未标明）",
  "who": "数据站（运营方未标明）",
  "d": "",
  "u": "https://www.arr.club/rankings/fastest-growing",
  "img": "https://www.arr.club/api/og",
  "cats": [
   "AI Core",
   "AI Coding",
   "AI Sales & Marketing",
   "AI Infrastructure",
   "Developer Tools",
   "Robotics",
   "AI Healthcare",
   "AI Consumer 等"
  ],
  "note": "按已验证 ARR 里程碑（最早到最新披露，至少间隔 100 天）算年化增速的前 100 家私营 SaaS 与 AI 公司，另有每周增速榜；增速榜会被小基数公司刷到前面（如 $1M→$120M）。"
 },
 {
  "t": "The Leonis AI 100",
  "by": "Leonis Capital",
  "who": "VC",
  "d": "2025-11-05",
  "u": "https://www.leoniscap.com/research/the-leonis-ai-100",
  "img": "https://www.leoniscap.com/_astro/drdXPlrHWn2psytEOACx07JxI4.B5wrOEiw.webp",
  "cats": [
   "按增长速度选出的 100 家 AI 应用公司（含创始人背景、团队规模、ARR 洞察）"
  ],
  "note": "索引 1 万多家 AI 初创，用资金、招聘、用量、GitHub、新闻和 Sacra、ARR Club 的 ARR 估算选出最快增长的 100 家；数据截至 2025-09，收入为估算。"
 },
 {
  "t": "Top 100 Gen AI Consumer Apps — 7th Edition",
  "by": "Andreessen Horowitz（Olivia Moore）",
  "who": "VC",
  "d": "2026-10-05",
  "u": "https://a16z.com/100-gen-ai-apps-7/",
  "img": "https://d1lamhf6l6yk6d.cloudfront.net/uploads/2026/10/Top-100-Gen-AI-&ndash;-Worlds-Fair-OG-Image.png",
  "cats": [
   "网页榜（Similarweb 月访问量）",
   "移动榜（Sensor Tower 月活）",
   "美国消费者支出榜 Top 50（YipitData，新增）"
  ],
  "note": "第一次把流量与消费支出并排：50 家支出榜中 29 家不在任何流量榜上；三个榜都上的只有 7 家。支出榜是美国信用卡面板样本，不等于公司收入。"
 },
 {
  "t": "2025: The State of Generative AI in the Enterprise（企业 AI 支出拆分）",
  "by": "Menlo Ventures",
  "who": "VC",
  "d": "2025-12-09",
  "u": "https://menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise/",
  "img": "https://menlovc.com/wp-content/uploads/2025/12/enterprise_ai_growth-scaled.webp",
  "cats": [
   "应用 $19B（部门型、通用助手、垂直 AI $3.5B）",
   "基础设施 $18B",
   "编程 $4.0B（部门型 $7.3B 中约 55%）"
  ],
  "note": "不是公司榜而是市场拆分图：企业生成式 AI 支出 $37B；另有各模型厂商占企业 LLM 支出份额（二手转述 Anthropic 40%、OpenAI 27%，本次未在 Menlo 页面直接核到）。"
 },
 {
  "t": "Ramp AI Index",
  "by": "Ramp Economics Lab",
  "who": "研究机构",
  "d": "2026-10-14",
  "u": "https://ramp.com/data/ai-index",
  "img": "https://next.ramp.com/assets/images/data/EconomicsLabOG.webp",
  "cats": [
   "按厂商采用率（Anthropic、OpenAI）",
   "按企业规模（大、中型）",
   "按 token 支出份额",
   "支出分位（前 1%、前 10%、中位）"
  ],
  "note": "用 85,000+ 家 Ramp 客户的刷卡数据按月追踪：2026-09 采用率 55.9%，Anthropic 43.8%、OpenAI 40.6%；10 月周度 token 支出 Anthropic 占 60.1%。是买家支出，不是厂商收入；只覆盖 Ramp 客户。"
 },
 {
  "t": "Sacra — Companies（私营公司收入估算库）",
  "by": "Sacra",
  "who": "研究机构",
  "d": "",
  "u": "https://sacra.com/explore/companies/?sort=valuation",
  "img": "",
  "cats": [
   "按估值排序的公司库，含年化收入估算与增速",
   "各公司研究页（如 Cursor、Anthropic）"
  ],
  "note": "Sacra 对私营公司的收入估算库，页面持续更新（抓到时显示 Anthropic、OpenAI、xAI、Databricks、Anduril、Cursor 等）；是估算不是披露，引用时要标 'Sacra 估算'。页面无固定发布日期。"
 },
 {
  "t": "State of AI Report 2025 — 精选 16 家 AI-first 公司年化收入",
  "by": "Air Street Capital（Nathan Benaich）",
  "who": "VC",
  "d": "2025-10-09",
  "u": "https://www.stateof.ai/2025-explore/business",
  "img": "https://www.stateof.ai/assets/report-2025/report-pages/page-001.jpg",
  "cats": [
   "16 家 AI-first 公司合计 $185 亿（2025-08 年化）",
   "精益 AI 榜（二手转述：ARR >$500 万、员工 <50 人、成立 <5 年，44 家合计超 $40 亿）"
  ],
  "note": "官方强调这是精选 16 家的年化 run rate，不是全行业收入；'精益 AI 榜' 的细节来自 Founder Park 转述，未在官网页直接核到。"
 },
 {
  "t": "Foundation Model Estimated Revenue & Capital Raised（BOND 报告内图表）",
  "by": "BOND（Mary Meeker 等）",
  "who": "VC",
  "d": "2025-05-30",
  "u": "https://www.bondcap.com/reports/tai/",
  "img": "https://www.bondcap.com/report/png/tai/cover.png",
  "cats": [
   "OpenAI",
   "Anthropic",
   "xAI",
   "Perplexity（年化收入对比累计融资）"
  ],
  "note": "报告内一张表：2025-04/05 估算 OpenAI 年化约 $92 亿、Anthropic $20 亿、Perplexity $1.2 亿，对照累计融资 $639 亿/$180 亿/$14 亿；注明来源 PitchBook、The Information、Bloomberg 等，全为估算。已过时（现数字大幅上升）。"
 },
 {
  "t": "Hypergrowth research — Comparison Tables",
  "by": "Seva Ustinov（个人研究站 etc.sevaustinov.me）",
  "who": "个人",
  "d": "",
  "u": "https://etc.sevaustinov.me/hypergrowth-research/data/comparison-tables.html",
  "img": "",
  "cats": [
   "(C) 一手来源确认",
   "(E) 推算",
   "(S) Sacra 等二手分析"
  ],
  "note": "跨公司表：成立时间、商业化时间、最新 ARR、估值、到 $10M/$100M ARR 用时（含 Sierra、Harvey、Decagon、Abridge、Glean、Writer 等）；每个数字带置信度标记，是难得标了口径的个人追踪表。页面无发布日期，作者同时是 The Recursive 文章作者，自己做 AI 创业。"
 },
 {
  "t": "Data on AI companies（AI Companies Data Hub）",
  "by": "Epoch AI",
  "who": "研究机构",
  "d": "2026-10-27",
  "u": "https://epoch.ai/data/ai-companies",
  "img": "https://epoch.ai/assets/images/datahub/thumbnails/ai-companies-thumbnail.png",
  "cats": [
   "前沿模型公司年化收入",
   "融资与估值",
   "人数",
   "算力开支",
   "用户量"
  ],
  "note": "Epoch 汇总公司披露与媒体报道的前沿 AI 公司数据库，可看收入图表和表格，页面标注 2026-10-27 更新；目前只覆盖 OpenAI、Anthropic 等纯模型公司，应用层公司不在内。"
 },
 {
  "t": "Anthropic could surpass OpenAI in annualized revenue by mid-2026（OpenAI vs Anthropic 年化收入趋势图）",
  "by": "Epoch AI（Luke Emberson、Yafah Edelman）",
  "who": "研究机构",
  "d": "2026-02-19",
  "u": "https://epoch.ai/data-insights/anthropic-openai-revenue",
  "img": "https://epoch.ai/assets/images/data-insights/anthropic-openai-revenue/t-anthropic-openai-revenue.png",
  "cats": [
   "OpenAI",
   "Anthropic"
  ],
  "note": "两家自各自首次报 $10 亿年化以来的收入曲线及交叉点外推（基线约 2026 年 8 月、约 $430 亿），只有两家公司，不是排行。"
 },
 {
  "t": "Anthropic and OpenAI earn more revenue per employee than major public tech companies（人均收入对比图）",
  "by": "Epoch AI（Luke Emberson）",
  "who": "研究机构",
  "d": "2026-05-08",
  "u": "https://epoch.ai/data-insights/revenue-per-employee-ai-companies",
  "img": "https://epoch.ai/assets/images/data-insights/revenue-per-employee-ai-companies/t-revenue-per-employee.png",
  "cats": [
   "Anthropic 约 $900 万/人",
   "OpenAI 约 $560 万/人",
   "Nvidia、Google、Apple、Meta"
  ],
  "note": "按人均年化收入把两家实验室和上市科技公司排在一张图里，两家均高于所有福布斯全球 2000 强科技公司；人数部分是外推值。"
 },
 {
  "t": "Charted: The Soaring Revenues of AI Companies (2023–2025)",
  "by": "Visual Capitalist（数据：Epoch AI）",
  "who": "媒体",
  "d": "2025-11-01",
  "u": "https://www.visualcapitalist.com/charted-the-soaring-revenues-of-ai-companies-2023-2025/",
  "img": "https://www.visualcapitalist.com/wp-content/uploads/2025/11/AI_Company_Revenue_Web.webp",
  "cats": [
   "OpenAI $130 亿（2025 年 8 月）",
   "Anthropic $70 亿（2025 年）",
   "xAI $5 亿（2025 年）"
  ],
  "note": "三家模型公司 2023–2025 年化收入折线图，数据来自 Epoch；只有三家且已过时，页面上的发布日期没抓到，11 月是图片上传月份推断。"
 },
 {
  "t": "Anthropic tops CNBC 2026 Disruptor 50（第 14 届榜单）",
  "by": "CNBC（Julia Boorstin）",
  "who": "媒体",
  "d": "2026-05-19",
  "u": "https://www.cnbc.com/2026/05/19/2026-cnbc-disruptor-50-rankings-anthropic-no-1.html",
  "img": "https://image.cnbcfm.com/api/v1/image/108289783-1775826656096-gettyimages-2264943542-20090101260308-99-764839.jpeg?v=1775826686&w=1920&h=1080",
  "cats": [
   "Anthropic #1",
   "OpenAI #2",
   "Databricks #3",
   "Anduril #4",
   "Ramp #5",
   "Cyera #9"
  ],
  "note": "私营创新公司 50 强，43 家以 AI 为核心；按规模化、用户和销售增长排名，不直接按收入；文章只列出前五和第 9 名，Anthropic 一季度收入同比约 80 倍（CEO 自述）。"
 },
 {
  "t": "AI Harness' ARR Multiples（估值对 ARR 对数图）",
  "by": "Tomasz Tunguz（Theory Ventures）",
  "who": "VC",
  "d": "2026-08-09",
  "u": "https://tomtunguz.com/ai-harness-arr-multiples",
  "img": "https://res.cloudinary.com/dzawgnnlr/image/upload/c_fill,g_auto,w_1200,h_630/uetsxogcas5xx8bpqlcf",
  "cats": [
   "Ramp ARR $14 亿",
   "Harvey / Legora / Sierra 各约 $1 亿+",
   "Decagon ARR $3500 万"
  ],
  "note": "把五家 AI 应用公司的历轮估值画在对数坐标的 ARR 上，并标 10x–200x 倍数线，再画 2024 年 12 月至 2026 年 6 月的估值/ARR 走势；只有五家公司，部分收入是估算。"
 },
 {
  "t": "State of AI Report 2026 — AI inference is becoming a vast business",
  "by": "Air Street Capital（Nathan Benaich）",
  "who": "VC",
  "d": "2026-10-08",
  "u": "https://www.stateof.ai/2026-explore/business",
  "img": "https://www.stateof.ai/assets/state-of-ai-2026-cover.png",
  "cats": [
   "OpenAI+Anthropic 合计约 $1050 亿",
   "Mercor 约 $20 亿",
   "Handshake AI 近 $10 亿",
   "micro1 $5 亿+",
   "Surge $12 亿（2024 财年）"
  ],
  "note": "第 9 版年度报告商业章，列出两大实验室合计年化收入和训练数据供应商的收入，页面说明各家口径和日期不同、不能加总；页内图表我没抓到，封面图作代替。"
 },
 {
  "t": "The state of the AI economy（生成式 AI 去重收入图）",
  "by": "Exponential View（Azeem Azhar）",
  "who": "研究机构",
  "d": "2026-06-25",
  "u": "https://www.exponentialview.co/p/the-state-of-the-ai-economy",
  "img": "https://substackcdn.com/image/fetch/$s_!3dcB!,w_1200,h_675,c_fill,f_jpg,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fec7d1c16-2ca0-49a5-beb8-63a11c40699b_1982x1112.png",
  "cats": [
   "全行业去重收入 $1100 亿（12 个月）",
   "年化 $1750 亿"
  ],
  "note": "自下而上建模 1000 多家公司，画出 2023 年 1 月至 2026 年 6 月的行业收入曲线和对折旧的比较；是全行业总量图，不是公司排行。"
 }
]