TwelveLabs, the San Francisco-based video understanding startup, has raised a $100 million Series B co-led by NEA and NAVER Ventures, lifting its total funding above $207 million. Amazon, Radical Ventures, Korea Investment Partners, Index Ventures, Quadrille Capital and Red Bull Ventures also joined one of the largest pure-play video AI financings to date.
Teaching Machines to Watch
TwelveLabs' Video Cognition System ingests and indexes video across visuals, audio and on-screen text, turning archives into searchable, queryable data. The founders argue that video, not text, is the richest record of the real world, and that breakthroughs in multimodal perception have finally made frame-level reasoning tractable. Applications span media companies mining archives for clips, security teams flagging incidents in live feeds, and factories spotting defects on production lines.
Why Investors Piled In
Enterprise video data is growing roughly 40 percent per year, yet most of it remains opaque to software. Backers are betting TwelveLabs can do for video what search engines did for text, with Amazon's participation hinting at future cloud distribution. The raise echoes a broader shift of venture capital toward machine perception, from Apple's expanding computer vision research to Orbbec's AI-powered 3D vision systems for automation.
Competition and What Comes Next
TwelveLabs faces pressure from big-cloud research labs and generative video startups alike, but differentiates with an end-to-end system that embeds and archives video rather than just querying it live. The company plans to expand its multimodal foundation models and enterprise integrations, chasing markets where visual data drives decisions. Investor appetite for such training-data-rich plays is strong: General Intuition recently raised $320 million on the value of visual interaction data for robotics.
Reporting based on coverage from Tech Startups and company announcements.