🚨 New market map: AI video in 2026 — 137 companies, 10 layers The most crowded layer isn't the mod...

AI Will(@FinanceYF5) · 人工智能

🚨 New market map: AI video in 2026 — 137 companies, 10 layers The most crowded layer isn't the models. It's the 34 creation platforms that wrap them: Higgsfield alone is at $700M annualized revenue, close to fal's $800M for serving the models underneath. Layer by layer: 1. Big-tech models (13) — Google Veo, ByteDance Seedance, Kling. Kling is at ~$500M ARR (Kuaishou, Q1 2026). OpenAI shut down Sora in March. 2. Independent models (12) — Runway (~$5.3B val.), Luma (~$4B), Vidu ($2B+). They survive on one thing done best: camera control, character consistency, live-action edits. 3. World models (4) — World Labs (~$5B), Decart (~$4B), Odyssey ($1.45B), Google Genie 3. Not a clip you watch, a world you walk into. 4. Inference APIs (9) — fal ($800M annualized), Replicate, Runware. One endpoint, hundreds of models, billed by the second. 5. Creation platforms (34) — Higgsfield ($700M), Artlist (~$300M ARR), Magnific/Freepik ($230M), OpenArt, InVideo ($70M each). Every model, one workspace. 6. Editing & post (21) — CapCut (736M mobile MAU), Descript, Opus Clip ($50M+ ARR). Timeline editing → auto-cut, captions, generated B-roll. 7. Ads & marketing (14) — Topview, Pencil, Waymark. Hundreds of ad variants the same day, A/B tested. 8–10. AI drama tools (30) — the layer most Western maps miss. Storyboard (Boords, Katalist), production studios (LibTV, SenseTime Seko, Tencent WorkRally), dubbing (ElevenLabs, Rask, YouTube auto-dubbing). A short-drama series that took a 30-person crew a month now takes 5–10 people a week. Data as of Nov 2026, public sources; company-reported numbers are labeled as such on the map. Bookmark this — we'll update it every quarter. Full list + sources: ai-video-drama-map.vercel.app Z 💬 1 🔄 0 ❤️ 1 👀 377 📊 1 ⚡ Powered by xgo.ing

查看原文